Point of View, Perspective and Focalization

Point of View, Perspective, and Focalization
Narratologia
Contributions to Narrative Theory
Edited by
Fotis Jannidis, Matas Martnez, John Pier
Wolf Schmid (executive editor)
Editorial Board
Catherine Emmott, Monika Fludernik
Jose Angel Garca Landa, Peter Hhn, Manfred Jahn

Andreas Kablitz, Uri Margolin, Jan Christoph Meister
Ansgar Nnning, Marie-Laure Ryan
Jean-Marie Schaeffer, Michael Scheffel
Sabine Schlickers, Jrg Schnert
17
Walter de Gruyter Berlin New York
Point of View, Perspective,

and Focalization
Modeling Mediation in Narrative
Edited by
Peter Hhn, Wolf Schmid
and Jrg Schnert
Walter de Gruyter Berlin New York
Printed on acid-free paper which falls within the guidelines
of the ANSI to ensure permanence and durability.
Library of Congress Cataloging-in-Publication Data

Point of view, perspective, and focalization : modeling mediation in
narrative / edited by Peter Hhn, Wolf Schmid, Jrg Schnert.
p. cm. (Narratologia, ISSN 1612-8427 ; 17) (Narratologia ; 17)
The majority of the papers collected in this volume are based on
talks given at the conference ... held at Hamburg University by the
Hamburg Research Group Narratology (Forschergruppe Narratologie) from October 13 to 15, 2006, titled Point of View, Perspective,
Focalization: Modeling Mediacy.
ISBN 978-3-11-021890-9 (alk. paper)
1. Point of view (Literature) 2. Mediation. 3. Narration (Rhetoric) I. Hhn, Peter, 1939 II. Schmid, Wolf. III. Schnert, Jrg.
PN3383.P64P65 2009
808dc22
2009012025
ISBN 978-3-11-021890-9
ISSN 1612-8427
Bibliographic information published by the Deutsche Nationalbibliothek
The Deutsche Nationalbibliothek lists this publication in the Deutsche
Nationalbibliografie; detailed bibliographic data are available in the Internet
at http://dnb.d-nb.de.
Copyright 2009 by Walter de Gruyter GmbH & Co. KG, D-10785 Berlin
All rights reserved, including those of translation into foreign languages. No part of
this book may be reproduced in any form or by any means, electronic or mechanical,
including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher.
Printed in Germany
Cover design: Christopher Schneider, Laufen
Contents
Introduction ...........................................................................
Part I: Re-Specifications of Perspective

JAN CHRISTOPH MEISTER, JRG SCHNERT
The DNS of Mediacy......................................................................
11
URI MARGOLIN
Focalization: Where Do We Go from Here? ..................................
41
TATJANA JESCH, MALTE STEIN

Perspectivization and Focalization: Two ConceptsOne
Meaning? An Attempt at Conceptual Differentiation .............
59
ALAIN RABATEL
A Brief Introduction to an Enunciative Approach to Point of
View................................................................................................
79
GUNTHER MARTENS
Narrative and Stylistic Agency: The Case of Overt Narration .......
99
DAVID HERMAN
Beyond Voice and Vision: Cognitive Grammar and Focalization
Theory.............................................................................................
119
BRIAN RICHARDSON
Plural Focalization, Singular Voices: Wandering Perspectives in
We-Narration...............................................................................
143
VI
Contents
Part II: Some Special Aspects of Mediation

VIOLETA SOTIROVA
A Comparative Analysis of Indices of Narrative Point of View
in Bulgarian and English ................................................................
163
TOM KUBEK
Focalization, the Subject and the Act of Shaping Perspective........
183
CHRISTIAN HUCK
Coming to Our Senses: Narratology and the Visual.......................
201
Part III: Transliterary Aspects of Mediation

ROLAND WEIDLE
Organizing the Perspectives: Focalization and the Superordinate
Narrative System in Drama and Theater.........................................
221
SABINE SCHLICKERS
Focalization, Ocularization and Auricularization in Film and
Literature ........................................................................................
243
MARKUS KUHN
Film Narratology: Who Tells? Who Shows? Who Focalizes?
Narrative Mediation in Self-Reflexive Fiction Films .....................
259
JAN-NOL THON
Perspective in Contemporary Computer Games.............................
279
Authors ...........................................................................................
301
PETER HHN
(Hamburg)
Introduction
The basic constellation constituting a narrative can be described as a
communicative act (narration) through which happeningsincluding existents such as characters, places, circumstances, etc., within the storyworld (fictional or factual)are represented and thus mediated through a
given verbal, visual or audio-visual sign system. This representation is inevitably shapedin the selection, combination, perspectivization, interpretation, evaluation of elementsby the agency producing it, ultimately
the author who, however, may delegate mediation, particularly in fictional
narration, to some intermediary agent or agents, typically a narrator (narrators voice) and, at a lower level, to one or more characters (characters
perspective) located within the happenings (in verbal texts) and,
according to some theorists, to the recording apparatus and/or voice-over
(in film). This process of transforming and transmitting the story in the
discourse, is what is meant by mediation in the broadest sense. One crucial problem concerning mediation in verbal texts as well as in other media is thus the extent and dimensions of its modeling effect, and more
particularly the precise relative status and constellation of the mediating
agents, i.e. the narrator or presenter and the character(s). The question,
then, is how are the structure and the meaning of the story conditioned by
these two different positions in relation to the mediated happenings perceived from outside and/or inside the storyworld?
The problem of mediation in narrative was the topic of a conference
held at Hamburg University by the Hamburg Research Group Narratology (Forschergruppe Narratologie) from October 13 to 15, 2006,
titled Point of View, Perspective, Focalization: Modeling Mediacy. The
majority of the papers collected in this volume are based on talks given at
the conference, supplemented by a few additional articles. In the conference title, mediacy was meant as an umbrella term covering all modes,
means, and instances of mediation, but since some (presumably above all,
Peter Hhn
German) readers might associate this word with Stanzels more limited
concept of Mittelbarkeit, it was replaced with the comprehensive term
mediation for the present collection. Like the conference, the book has
a twofold aim: to offer a fresh look and a systematic renewal of the notion
of mediation in narratology in its traditional focus on literary texts, and in
addition to apply this concept to narration in other media, including drama, film, and computer games. Mediation is intended to comprise all
possible aspects, forms, and means of constructing and communicating a
story in discourse: the selection, ordering, and segmentation of storyworld
elements, their transmission through a presenter (e.g. the narrators voice
or equivalent agents in film and drama), their presentation from a particular standpoint or perspective. In its range, this volume does not aim at
an exhaustive overview or neat differentiation and definition of the
various terms currently in use for this field. Rather, the individual articles
address some controversial aspects of the narratological conceptualization
and systematization of mediation in their application to both literary and
other media.
The first group of articles in this volumepart Icomprises contributions (Meister & Schnert, Margolin, Jesch & Stein, Rabatel, Martens,
Herman, and Richardson) whose aim, from various angles, is to re-define,
re-specify, or re-model perspective, especially Genettes concept of focalization, typically with regard to its distinction from narrative voice. Jan
Christoph Meister and Jrg Schnerts The DNS of Mediacy outlines a
comprehensive approach to the process of representation inherent in the
histoire/rcit or story/discourse distinction achieved by what they call the
Dynamic Narrative System, the instance to which the constitution and
communication of the narrative as a whole is attributed. The DNS comprises both voice and perspective (or narration and point of view), modeled as the integrated result of mental activities across the three (interconnected) dimensions of perception, reflection and mediation. These
dimensions differ as to the specific type of mental activity and the constraints exercised by that activity, whether determined, respectively, by
epistemological or sensory input: temporal and spatial proximity to the
object domain (perception); mental reaction to the input, i.e. the cognitive, emotive and evaluative relation to the object domain (reflection);
medial materialization of the output, i.e. the semiotic relation to the object
domain (mediation). All three dimensions are organized along the same
fundamental opposition of diegetic (narratorial) vs. mimetic (actorial),
such that mental activity can shift between the positions of narrator and
Introduction
character(s). The underlying premise is that the rigorous distinction

between speaking and perceiving as introduced by Genette, however
useful it may be to overcome a confusion between these notions, tends to
obscure their inherent interrelations with respect to the modes of mediating stories to readers, since variations in one dimension always entail
changes in the other two. By replacing the relatively rigid architecture of
structuralist concepts with the variable combination of three scalable parameters, this theoretical model also facilitates a more precise and flexible
description and analysis of mediation in practical terms.
Uri Margolin, in Focalization: Where Do We Go From Here?, rigorously restricts the term focalization to storyworld participants (characters) and their mental (and textual) representation of that world. For
him, focalization thus comprises perceptions (through all five senses) as
well as acts of simulation or empathy and recollection on the part of characters. It categorically excludes, however, both their acts of planning or
projection (because not focused on pre-existing storyworld elements) and
the content of an omniscient, impersonal narrators voice (because not
position-dependent and therefore not restricted). Narrators cannot be focalizers, in Margolins view, except for certain special cases in which the
narrator deliberately limits his perspective to that of a character or situates
himself in a concrete situation. Of the two alternative definitions of focalization proposed by Genetteperception and knowledge or informationMargolin thus opts for the former. Basically, for him focalization is
restricted to Genettes internal focalization, although he does concede that
internal focalization can also occur with homodiegetic or autodiegetic
narrators. Margolins conception of focalization has the merit of clearly
marking off mediation in the form of sensory or mental perception, but at
the same time it leaves other aspects of mediation unexamined, especially
narratorial perspective.
Taking issue with Genettes ambivalent conception of focalization as
knowledge (information content) or perception, Tatjana Jesch and Malte
Stein, in Perspectivization and Focalization: Two ConceptsOne
Meaning? An Attempt at Conceptual Differentiation, propose to redefine
the term by separating the two aspects and subdividing the complex of
mediation into two different dimensions. Focalization is taken as the
selection or withholding of information about the fictional world by the
author, whereas perspectivization refers to the subjective perception of the
world by a fictive entity, either one or several characters or the narrator
(thus partly resembling Margolins conception of focalization). Since
Peter Hhn
perspectivization is ultimately governed by the author, it can be used by

him as a means of focalizing or channeling information.
Based on a different critique of the delimitation and subdivision of
Genettes concept of perspective, as also addressed by the two preceding
articles, Alain Rabatels A Brief Introduction to an Enunciative Approach to Point of View offers a consistent clarification and simplifycation of the authors model of focalization on linguistic grounds. Rabatel
distinguishes between the first speaker or enunciator (who expresses the
whole of an utterance, produces the discourse) and the intratextual enunciator (who filters the object of discourse through his evaluations, quailfications, modalizations and judgments). Based on a carefully selected
corpus, he then goes on to identify and analyze the linguistic markers
which convey point of view as well as variations in perception, thought
and speech forming a continuum that passes from narrated through represented to asserted points of view. Speaker and enunciator may or may not
coincide, as the enunciator may be located on the same level or on different levels, i.e. that of the narrator or that of the character(s). Analysis
of external and internal markers within this framework opens the way to a
differentiated description of complex points of view.
Within this broad context of mediating devices, Gunther Martens
Narrative and Stylistic Agency: The Case of Overt Narration discusses
a specific point: the status and possible functions of overt narration that
come into view with the foregrounding of the narrators act of narration.
As a form of mediation, the act of narration must be considered a genuinely narrative device not to be neglected in the face of the widespread
privilege granted to presentation through characters internal focalization
since the beginning of the modernist period.
A new comprehensive approach to mediation is presented by David
Hermans Beyond Voice and Vision: Cognitive Grammar and Focalization Theory, drawing largely on the cognitive semiotics and linguistics of Talmy and Langacker. Herman reformulates focalization as a process of conceptualization and construal of a storyworld scene by an embodied mind, specifying a number of parameters which allow for a discriminating description and analysis of mediation as it occurs in a scene:
scanning the scene (static vs. dynamic, synoptic vs. sequential); the observers distance from the scene (distal, medial, proximal); scope (narrow
vs. wide); figure/ground alignment (foregrounding vs. backgrounding of
characters, elements, etc.); degree of granularity (how detailed the presentation is); spatial and temporal viewpoint (vantage point and orientation
Introduction
within a directional grid); degree of objectivity vs. subjectivity. This

system of categories offers a well integrated transmedial approach which
covers both narratorial and figural perspectives (voice and vision), cuts
across the distinction between information and perception and applies
equally well to mono-modal narratives (novels, short stories) and multimodal texts such as graphic novels.
Discussing we-narration and we-focalization as a means of presenting collective consciousness, Brian Richardson, in Plural Focalization, Singular Voices: Wandering Perspectives in We-Narration, analyzes an unusual literary technique which tends to transcend the boundary between homodiegetic and heterodiegetic narration as well as to blur
the distinction between narration (voice) and focalization, fluctuating ambivalently between narratorial and figural perspectives. As wandering focalization, this mediating device ultimately undermines the distinctions
construed by Genette or, in other ways, by both Jesch and Stein and
Rabatel.
The three contributions in part II (Sotirova, Kubek, and Huck) address relatively unexplored aspects of perspective. A linguistic aspect affecting mediation is the subject of Violeta Sotirovas A Comparative
Analysis of Indices of Narrative Point of View in Bulgarian and English.
Emphasizing verbal aspect, Sotirova investigates the influence of the divergent verb systems of Bulgarian and English on the rendering of internal focalization through free indirect discourse, and she identifies areas of
differences between the two languages, adding, however, that syntactic
analysis of verb forms must be supplemented by that of contextual and
semantic features.
A special topic of a different kind is discussed by Tom Kubek in
Focalization, the Subject and the Act of Shaping Perspective, notably
Mukaovsks concept of the subject of a literary work. The subject,
taken as a complex of value-based perspectives on the storyworld, constitutes the overall meaning of a work, comparable to what is termed in
other approaches the implied or abstract author. The comprehensive perspective, he argues, is not inherent in the text alone nor determined by it,
but ultimately depends on reception by the reader.
Christian Huck, in Coming to Our Senses: Narratology and the Visual, offers a critique of the sensory foundation of a central dimension of
mediation, drawing attention to the pervasive tendency in narratology to
literally or figuratively model perspective in analogy to the visual sense
(as revealed in the terms point of view, focalization, perspective), a
Peter Hhn
reflex of the dominant tradition of privileging vision in Western culture.

Huck argues that because of the differences among the senses (e.g. in the
degree of distance/proximity and detachment/involvement), the visual
cannot stand for all senses, even metaphorically, a point he illustrates with
reference to two factual 18th-century travelogues. The act of perception
and its specific sensory channel (slant of perception) as the basis of
what can be reported by the narrator needs to be studied more thoroughly,
with explicit reference to culturally defined default patterns and their
privileging of individual senses.
The final group of articles (part III) explores the application of narratological concepts to the analysis of mediation in the non-textual media:
drama (as performance), film, and computer games (Weidle, Schlickers,
Kuhn and Thon).
While previous narratological studies of drama have concentrated on
epic elements contained in the diegetic and hypodiegetic levels (including
stage directions), Roland Weidle, in Organizing the Perspectives: Focalization and the Superordinate Narrative System in Drama and Theater,
specifically turns to the extradiegetic level. The agency of extradiegetic
narration, which he calls the superordinate narrative system (cf. Meister
and Schnerts dynamic narrative system), is predominantly impersonal
and covert (close to what others call the implied or abstract author),
taking form indirectly by controlling selection, arrangement and focalization, but also through the compositional implication of internal connections (through motifs) or prolepses. Similarly, focalization in drama is
usually external, allowing the characters on stage to be apprehended only
from the outside (unless they themselves reveal their inner thoughts in
direct speech, as in soliloquies). Weidle further highlights these default
conditions by referring to rare (and extremely contrived) exceptions, such
as the overt appearance of a personal narrative agent (the author in Lauwers Isabellas Room) or the establishment of zero focalization (simultaneous presentation of two chronologically different scenes on stage, as
in Marbers Closer) or of internal focalization (the enactment of memory
on stage, as in Stoppards Travesties).
In Focalization, Ocularization and Auricularization in Film and Literature, Sabine Schlickers first distinguishes between the implied director
(equivalent to the abstract author in literature) and the so-called camera
as the narrative agent of the film that intermediates between visual and
acoustic information. Thus, what is distinguishable as voice and focalization in literary texts appears to be more closely intertwined in film,
Introduction
functioning as perspectivization in general. More precisely, perspectivization operates in the form of focalization (the narrators knowledge
about the characters) in its interplay with ocularization and auricularization (visual and acoustic information about the storyworld). These
three terms are then further subdivided into two modes: superior or broad
vs. restricted to characters perspective, resulting in zero vs. internal ocularization, auricularization, focalization (for the latter, an external mode is
added in which the spectators knowledge is more restricted than characters knowledge). Finally, comparing a few novels with their film adaptations, Schlickers demonstrates the analytic usefulness of these categories in complex examples of filmic mediation, e.g. the combination of
zero auricularization and zero ocularization with internal focalization.
Markus Kuhns article, Film Narratology: Who Tells? Who Shows?
Who Focalizes? Narrative Mediation in Self-Reflexive Fiction Films, is
based on a similar system of categories as Schlickers, although the terms
employed are somewhat different. Kuhn names the position of the narrator the filmic narrative agent, dividing it into a visual narrative instance and one or more (optional) verbal narrative instance(s), roughly
equivalent to the traditional distinction of showing and telling. Like
Schlickers, he defines focalization on the basis of knowledge (subdivided
into zero, external and internal), labeling the visual and auditory aspects
of perception ocularization and auricularization. The constellation of the
visual and verbal narrative instances is shown to be highly variable in
terms of both relative dominance and content, ranging from contradictory
through complementary to congruent tendencies. Again, no sharp distinction can be drawn between narration (voice) and perspective (image).
Kuhn then goes on to corroborate these findings through the analysis of
mediating techniques employed in several self-reflexive multilayered
films.
The special setup of the medium of computer games, due primarily to
the feature of interactionality, results in substantial differences in how
perspective is organized in this medium as compared text-based media. In
the concluding article, Perspective in Contemporary Computer Games,
Jan-Nol Thon differentiates three dimensions of perspective for the
player, i.e. three positions or points from which the game world is presented: the spatial perspective determined by the point of view; the actional perspective determined by the point of action; and the ideological
perspective determined by the point of evaluation. Of these, the first and
third are similar to the perceptual or spatial and the ideological facets of
Peter Hhn
focalization as they also occur in literary texts, although more restricted in

their variability on account of the rigidly agonistic nature of gameplots. However, the second dimension, actional perspective, is peculiar
to computer games, as it also involves the productive position of the author (or the narrator), the interactive player being called on to actively
participate in the plot-development.
While a number of the essays collected in this volume seek to clarify
and differentiate the terminology of mediation with regard particularly to
literary narratives, others represent attempts to apply narratological categories and terms originally designed for the analysis of literary texts to
other forms, aspects and modes of narrative communication, and in particular to the media in their various forms. The aim is to explore the great
variety of sensory, cognitive, ideological, semiotic and technical modalities of transmitting, representing and structuring happenings in narrative
communication. Special emphasis is given to the notion of mediation as a
basis for analysis and comparison of these various modalities with respect
to the constitution and constellation of mediating instances, but also to the
influence of social and cultural contexts and technological conditions on
mediation. The ultimate objective is to develop a system of categories to
account for mediation within the framework of a general and comparative
narratology. Such a system would make it possible to identify features
common to all forms of mediation as well as the features characteristic of
and peculiar to each specific medium and mode of narration.
***
Note on bibliographical references
In cases where the date of the original publication is important but a later
edition or a translation is quoted from the reference will combine the
original publication date with the page number(s) of the edition used.
Part I: Re-Specifications of Perspective
JAN CHRISTOPH MEISTER, JRG SCHNERT

(Hamburg)
The DNS of Mediacy

1 Who Sees? and Who Speaks?: All Questions Answered?
Whatever narrative might be, it is certainly not a simple 1:1 duplication
without loss or gain: like any representation it reduces the complexity of
its reference domain to the carrying capacity of its medium and to the
processing capacity of senders and receivers. By the same token it also
adds a specific type of semiotic and performative surplus value: a narratives content is always narrated content. Narratives do not present us
with information per se; they broker informationand mediacy is the signature of this brokering activity which combines quantitative reduction
with qualitative (semantic) enrichment.
In literary texts the effect of mediacy is triggered by a complex interaction of epistemological and rhetorical constraints. Some of these can be
detected by surface phenomena hard-wired into the text, like an explicit
narratorial intrusion, while others manifest themselves mainly in the form
of readerly inferences. Genettes work on the logic of discours, in its
double take on who sees (with the particular emphasis on focalization) and who speaks (voice), pays attention to both types of constraints. However, intuitive as it may be it is exactly the rather reductionist concept of focalization which counts among the most controversial
elements in Genettes structuralist heuristics1. While some theorists would
rather revert to a more holistic and less analytical model like Stanzels,
hard core structuralist narratologists tend to advocate an even more
1
In our opinion the methodological status of Genettes taxonomic system is indeed that
of a narratological heuristics, not that of a narrative theorysee Genettes own characterization of his approach in Narrative Discourse and Narrative Discourse Revisited
as a procedure of discovery, and a way of describing (Genette 1980: 265) or, in
short, as a method of analysis (23).The controversy surrounding Genettes concept
of focalization is summarized in Jahn (1996) and Jahn (2005); also see van Peer &
Chatman (2001).
12
Jan Christoph Meister, Jrg Schnert
differentiated analysis. The most advanced approach in this regard is that

in Schmid (2005), for here the phenomenon of perspective is no longer
treated as a distinct medium-level systematic aspect of mediacy. Rather,
against the background of his revised ideal-genetic model Schmid
conceptualizes perspective as a summary high-level effect of narratives to
which the constituent narrative transformations of selection, composition
and verbalization equally contribute2. In this model perspective integrates
aspects of perception and aspects of expression. This integration is an important step towards building a more dynamic model of mediacy.
Our present contribution wants to take the next logical step into this
direction. To date neither Genettes critics3 nor his followers seem to have
questioned the methodological premise that forms the shared basis of his
and most other approaches towards mediacy. Formalist, structuralist as
well as more traditional hermeneutic theories all tend to conceptualize
narrative as a given, that is, as an artefact that exists as a complete and
stable whole. This is particularly apparent in the ideal-genetic models (cf.
Schmid 1982; Genette 1980 and 1988; Rimmon-Kenan 1983) implicitly
guided by the question of how a narrative was made. A narrative, it is
implied here, is the outcome of a completed process: it is a product. In a
genetic perspective narrative is thus explained as the result of a sequence
of creative actions attributed to the empirical author in terms of real-world
ontology, or, in terms of systematic logic, to some abstract narrative
instancei.e., to a narrator. Reception theory based models do not really
differ in this regard; for they merely address the same question from the
opposite angle: how did the reader, on the basis of the narrative, construct
a logically (but not necessarily ontologically) preexisting narrated
world? Of course, such reconstruction is not thought of as a simplistic
re-enactment of the genetic process, but as one that is contingent on the
particular contexts and world knowledge of readers which no author and
narrator can fully anticipate. Cognitive narratology (e.g., Herman 2003),
2
See Schmid (2005: 12749); the ideal-genetic model presented here is an extended version of the one in Schmid (1982).
In the current volume this criticism is represented by Tatjana Jesch and Malte Stein:
Genettes concept of focalization is actually an amalgamation of two wholly independent elements for whichas the author himself might have anticipatedone actually needs two terms. The first element is the perception of the world invented by the
author through narrators and other agents also invented by the author; the second element is the regulation of narrative information within the communication between
author and reader (59).
The DNS of Mediacy
13
in paying due attention to the cognitive logic of narrative processing, has

eventually gone at least half the way: it conceptualizes narrative as a
product of cumulative cognitive processes. However, the constituent
micro-processes of narrative processing seem to remain below the
discriminative threshold, as the cognitivist focus on schematic knowledge
representation and synthetic high-level operators for process orientation
and control (scripts, frames, types, world knowledge etc.) indicates.
In a sense the quest for a narrative theory that can capture the differential between representation and representedan ambivalence inherent
in the very term representation, which in the European languages
generally denotes the symbolic artefact as well as the performative act of
symbolizationis often implicitly guided by two hidden assumptions.
The first (and fairly trivial) is the ontological post hoc, ergo propter hoc
fallacy which manifests itself in our (seemingly natural) reflex to misinterpret any instance of representation as conclusive proof for the temporal pre-existence of the represented. The second assumption is of a
methodological nature and hence more difficult to grasp. We will focus
on the latter.
In a given narrative the phenomena so far categorized under the headings of distance, focalization or perspective are certainly strong
indicators for a differential between representation and represented. Accordingly, it seems to make good sense to analyze them one by one on the
basis of distinct systematic typologies. Also, against the background of
these typologies, interdependency and interaction between the indicators
can then be explored by way of a combinatorial matrix: which type of focalization goes along with which type of mediacy, etc. The method is
compelling, for the how of discourse it seems, can now be concisely defined in terms of its position on the chess-board like tableau of focalization and mediacy.
The main problem with this approach is not one of terminological
overlap or inconsistency; rather it lies in the use of permutation logic
which combines statements concerning the relation among narrative instance and object domain on the one hand (e.g., attributions of the type
heterodiegetic vs. homodiegetic) with statements about the type of
information strategy (e.g., types of focalization) on the other. On closer
inspection these two types of attributions are in fact categorical apples
and pears: the former is ontological, the latter epistemological. Little
wonder then that the options for mapping the two typologies onto one
14
another are indeed rather limited4. To analyse the phenomenology of discourse in this wayparticularly across distinct and logically incompatible
categorical dimensionsis possible only if one posits the text as given
and read in its entirety, and as a stable object that can be dissected from
various angles. This, indeed, is the methodological assumption to which
we referred above. It has resulted in one of narratological theorys most
problematic blind spots: the process character of narrative.
There are many examples of structuralist descriptions which tend to
arrest processuality and re-interpret it by way of stratificatory models.
One is Genettes: while his taxonomy does of course admit the fact of
variable discourse organization in the constituent parts of a given narrative (one can place the narrator piece on the chess-board of discourse
differently within every segment, so to speak), it implicitly forces us to
model the logic of narrative representation in terms of independent a-temporal systematic layersa layer of focalization, then a layer of distance,
then a layer of order, etc.
The rigid systematic architecture of the analytical approach thus projects a-temporal systematicity onto its object. However, many existing
novels and novellas clearly defy this undertaking as our own reading experience shows. Narratives of 19th century realism may have tended to
present us with a discourse organization free of contradictory indicators,
but modern and particularly post modern literature clearly places more
complex demands on readers, and thus on narratological theory. Conscious profiling of the narratorour attempt to answer the questions of
where / who / how does s/he know / reflect / (dis-)informhas become
an increasingly difficult task. There is no one narrative instance; rather,
it is something that is in flux and can change throughout every reading:
it is a function, rather than a given.
1.1 Terminology Revisited
One might argue that this simply points us to the need for better definitions and more plausible attributions for the various types of focalization
For example, common sense shows that there is a strong affinity between an autodiegetic mediacy and an internal focalization, while an autodiegetic mediacy with (permanent) external focalization makes little prima facie sense and can thus be marked as
highly unlikely in the matrix of possible combinations.
The DNS of Mediacy
15
found in Genettes descriptive apparatus5. Some would indeed warrant at

least reformulation. For example, Genettes focalization zero is certainly counter intuitive in that what it refers to is not a strange nothingness, but rather a universal and non-restricted potency: the ability to
perceive, gain knowledge and pass on information about a reference domain ad libitum. Autonomous vision would therefore be a preferable
label for this epistemological position from which physical as well as
psychological phenomena (via introspection) can be perceived and represented in their absolute totality.
Moreover, the attribution of autonomy intuitively signals the ability
to change and restrict the representational potential at will: a position of
autonomous vision allows for the voluntary temporary adoption of a
more constrained epistemological vantage point which is either character
bound (Mitsicht or co-vision in Martinez & Scheffels terminology), or
reduced to the mere external aspect (Auen[an]sicht) that limits the representation to that of objective physical phenomena (cf. Rimmon-Kenan
1983: focalized from without).
Genettes internal focalization (restricted focalization) is equally
problematic. In addition to a straight forward single character bound epistemology it can also denote the mixed position of narratorial co-vision
with a defined reflector character which, however, goes along with a reductionist external perspective onto all other characters. Moreover, internal focalization can also be used to describe the case of an autobiographic narration where the mode of focalization is completely selfcentered. And this does not even exhaust the possibilities for the more
vexing constellations in first person narration listed by Stanzelchronicle, eye witness account, interior monologue, etc.
These examples show that individual definitions could of course be
improved onbut the architecture of the analytical model per se would
still be problematic. So how about a more rigid systematic approach? Perhaps we should set out with a principled distinction, like that between all
cases of an objectifying epistemology, of a looking-from-without, versus all cases of introspection, i.e. of an empathetic epistemology of
experiencing-from-within6. The scope of the objectifying continuum
5
See e.g. Niederhoff (2001); also see the contribution by Tatjana Jesch and Malte Stein
in the current volume.
It also calls for a clear-cut distinction among epistemology (how can we know what we
know?), psychology (how can we feel what we feel?) and ontology (how can we be
what we are?). The Genettian distinction of heterodiegetic vs. homodiegetic narrators
16
could then be marked by concepts such as autonomous visioncovisionself visionexternal vision7. All of these refer to cases where
the epistemological vantage point remains external to the object domain,
even if what one looks at is oneself (though the object remains located in
a spatial-temporal position distinct from the point of origin of the enunciative act). By contrast, the second continuum would then integrate the
various modes of introspection, of being focalized from within (sensu
Rimmon-Kenan 1983). This continuum could be delineated by the positions of indexical (subjective) visionpartial introspectionfull introspection.
In all of these the empathetic vantage point is centered within an observing and feeling instance. The analysis of a concrete narratorial statement would then amount to a definition of its narratorial position in terms
of an intersection of the epistemological and the empathetic. For example,
autonomous vision can integrate distinct positions of partial introspection, whereas co-vision grants access only to feelings and emotions attributable to the reflector character. External vision on the other
hand cancels all possibility for introspection.
The following table is an attempt to systematize the existing terminology. It is based on the binary model of external vs. internal focalization as it manifests itself with regard to the object domain of characters and their physical and mental states. n characters denotes all possible characters existing in a narrated world; p stands for a special set of
characters in a world that is accessible to one or more actors with perceptive abilities; character x (with its different internal states x1, x2, etc.)
stands for a particular actor who perceives.
mixes up epistemology and ontology, not to mention the implicit reversal of the Platonian (and Aristotelian) definition of mimesis (domain of the represented content) vs.
diegesis (domain of the acts of representation, but also the representation as suchalso
see footnote 12). With the exception of a first-person real-time report any act of telling
is, in a logical sense, a telling-from-without and thus heterodiegetic in an epistemological sense. The question whether or not a narrator exists within the narrated world
is thus an ontological one. However, we already have a term for a narrator who exists
within the narrated worldhe or she is, quite simply, a narrating character. In the
following we will disregard the heterodiegetic/homodiegetic distinction: it is simply
not needed.
Our definition of this continuum varies slightly from that of Martinez & Scheffel (cf.
1999: 64).
17
The DNS of Mediacy
external focalization
internal focalization
narratorial, unrestricted
perception (autonomous
vision / bersicht or
Allsicht resp.)
external vision (Auen[an]sicht) onto n characters
introspection of n characters (full introspection)
narratorial, partially restricted perception (co-vision /

Mitsicht with x)
external vision onto the p- introspection of x (partial

set of characters, excluding subjective introspection)
character x
actorial, partially restricted

perception (self vision of x
onto various states of x)
external vision onto the p- introspection of x at difset of characters, excluding ferent stages x1, x2 etc.
character x
(complex subjective introspection)
actorial, restricted perception

(self vision of x in simultaneity to the mediating
process: interior monologue)
external vision onto the pset of characters

(excluding character x),
dominated by subjective
introspection
introspection of x
(subjective introspection),
dominating mediation
actorial, restricted perception external vision onto the p- marginal to zero intro(external vision)
set of characters, excluding spection (of x)
character x
1.2 Before Terminology

This discussion of how narratologys conceptual tool-kit and terminology
might be improved with particular regard to the category of focalization
will be taken up later in this article. However, in the end all these terminological and pragmatic issues eventually point us back to the more principled problem mentioned above, that of the a-temporal systematicity embedded in structuralist narratologys methodological design. Perhaps the
most telling symptom of this orientation is the prevalence of visual metaphors in the narratological terminology dedicated to the analysis of perspective and focalization (sic!). These metaphors by their very nature imply that, at least in logical terms, narration is preceded by acts of perceptionand even Genettes choice of the more technical concept of focalization, which connotates a camera lens, cannot escape this association
either, although it does manage to overcome the anthropomorphism of the
viewing-metaphor. Generally speaking, the bias on perceptive and sen-
18
sual input significantly overshadows the mental activities by which we

embed the perceived content within the cognitive, emotional and evaluative frame works at our disposal. In fact, as fictional narratives prove
we do not even require any sensual or empirical input: perception may be
a sufficient trigger for narration, but it is certainly not a necessary prerequisite for it. Rather, the cognitive, emotional and evaluative frame
works at our disposal have endowed us with the unique ability to invent
the content that will fit them retrospectively and as if it had already
been perceived. And while such content-free fictional narratives might
be an extreme case even their factual counter parts are obviously based on
the complex procedures of cognitive pre-ordering and pre-processing
which characterize all acts of communication.
In the end the terminological predominance of visual, optical and spatial metaphors (perspective, focus, distance etc.) as well as the objectifying systematicity characterizing structuralist narratology both betray
its methodological disregard for the processual interdependency and dynamics of acts of perception (real or imagined), acts of cognition / emotion and acts of mediation (of expressing and passing on information).
Structuralist approaches have a tendency to freeze the narrative into a
single snap shot taken right at the end of everything. We are made to
believe that one can see the logic of discourse in a given scene ex post
like we see the compositional structure of a picture, or a landscape in
which everything seems to be present and presented simultaneously.
However, while the (real or hypothetical) objects referred to by a
painting or a narratives constituent symbols might as such indeed be
temporal and existential antecedents, their representation as a semiotic
construct necessarily develops and unfolds over timeand so does the
complementary cognitive and emotive activity of the recipient who transforms the signs into mental images. Before we can have a complete representation in the sense of a complex mental imagethe entire tale, the
complete picturewe encounter a multitude of representational and interpretive acts that interact with one another. And from time to time, this
interplay results in interference rather than coherence.
One way to overcome this restriction is to look at the process of perception in terms of its local and temporal constraints, such as the ideological and linguistic frames of reference within which it takes place.
However, relevant analytical approaches such as Rimmon-Kenan (1983)
and Schmid (2005) once again focus explicitly on the analysis of an ideal
type of how a narrative is narratorially produced. The ongoing mental ac-
The DNS of Mediacy
19
tivity required on the recipients part and coded into the medium in the
sense of processing instructions and controls remains beyond their scope.
2 The Model and its Terminology

In contrast to the terminological discussions which have kept narratology
busy over the past two decades we would now like to propose a more fundamental revision of the concept of narration. Our systematic point of departure is not the narrative as an artefact, but rather the processing of a
narrative; our goal is to arrive at a model architecture of the communicative process which drives narratives. In short, our goal is that of modelling mediacy, which is why our model is called the Dynamic Narrative
System (DNS)8. In order to develop this system architecture we will
have to ask questions that go beyond the traditional ones of who sees
and who speaks. For the time being the relevance of the DNS model of
mediacy is, however, restricted to the sub-set of literary narratives in the
narrower sense of the term. The examples given in section 4 (intended to
demonstrate the scope of our model and its terminology) will therefore
also be restricted to literary texts.
Irrespective of the significant methodological reorientation which we
just proposed there is certainly no need to re-invent all the terminological
wheels of narratology. The innovation lies in a different type of application of the established narratological concepts: rather than using them to
analyze narrative in the sense of a stable object and then sort terminologically defined states into the rigid slots of a taxonomy, we will try
and work with a scalar description of the variable processes which contribute to narration as a performative phenomenon. However, this approach is also not a licence to a pseudo-Heraclitean stance of everything
flows. In order to identify typical patterns of mediacy we will have to
arrest, from time to time, what in reality is an ongoing process. In the
continuum of processing time defined points of observation have to be
chosen in order to sample prototypical constellations of parameters and
8
Our concept of dynamic narrative system is only loosely related to the concept of
narrative system referred to by Roland Weidle (see his contribution to the current
volume: Organizing the Perspectives: Focalization and the Superordinate Narrative
System in Drama and Theater). Weidles approachwhich in turn is based on Jahns
concept of the dramatic superordinate narrative agent (Jahn 2001: 672)presents an
attempt to define a narrator concept specific to the case of theatric narration qua performance.
20
their values (cf. section 4). In defining these constellations we will try to
stick with existing terminology as far as possible, but also complement it
by additional concepts and parameters where necessary. Where new concepts are introduced we will try to make sure that they are intuitive, and
also translatable across languages without loss. Our aim is not a fully coherent rigid analytical apparatusrather, we want to try and keep our
model sufficiently flexible so that different historical (and perhaps not
even yet realized) constellations can be identified with it.
Even so, the model proposed in the following can only cover half of
what makes up narrative-in-operation: we will not be able to reflect sufficiently on the readerly aspect of narrative processing. However, it has to
be emphasized that our DNS model is not based on the idea of narrative
as an abstract and self sufficient semiotic machine which runs in and by
itself. In the reality of concrete narrative processing each and every
component and module of the DNS requires interaction with a human
mind in order to be activated. Where and how this mind engages with the
architecture and turns it into a live system remains to be explored. For the
time being we can only present the architecture as a blue print for the system as such. The model tries to explain how a narrative influences and
determines our profiling of its narrative instance, the inferential construct
commonly referred to as narrator. And finally, it is a modelnot a
theory, and not a taxonomy either: it simply tries to give us an idea of
how some of the crucial discourse phenomena are functionally interrelated, taking into account the dimension of time.
3 The DNS (Dynamic Narrative System)

The discourse/story distinction is perhaps the most fundamental contribution of narratology to literary theory. As we all know, its conceptual
ancestors are Saussures distinction of signifiant/signifi and, perhaps
more importantly, the linguistic distinction between expression plane and
content. All of these point back to a dichotomy inherent in the basic
notion of representation which is preserved in its etymology and ambivalent semantics. Representation has two functional dimensions: the symbolic dimension of being an image of as well as the pragmatic dimension of standing in for9. In both cases representation communicates
9
As narratologists we generally focus on the semiotic concept of representation, neglecting its wide spread usage in the political and legal sphere. Aestheticians and
The DNS of Mediacy
21
something which is not present itself: in the former case a sensory perception which our epistemological conditions do not allow us to make
ourselves; in the latter the intention of someone who is not personally
present. Narratives often merge these two dimensions of representation,
particularly on the level of discourse. Let us try to take them apart again.
3.1 Representation as Being an Image of
Narratives have a specific way of informing us about things that happen(ed), the things that they claim to be an image of. Particular to the
narrative representation of things that happened is its strong (though not
exclusive) focus on events and their temporal ordering. Events need not
be restricted to things that happen in the world (so-called object
events), but can also be mental events (processing events) that take
place in the mind of a character, or in that of the narrator, or, if nothing
else, in the readers own mind. Temporality is crucial to the narrative
mode of representation. Moreover, temporality is not just the principle
that allows for narratives sequential ordering of snap-shots of the world
into connected events, but also the principle by which we position ourselves vis--vis the flux of events, real or imagined. Narrative representation is, as Ricoeur has argued, therefore perhaps the privileged way
for humans to experience temporality. Narratives introduce physical
before-after time relations into what they are an image of; at the same
literary critics tendency to conceptualize the symbolic representation as an absolute,
self-motivating entity might be seen as a consequence of this neglect of the pragmatic
dimension of representation. By contrast, the political and legal concept of
representation features prominently in many contemporary dictionaries, the
Encyclopaedia Britannica being one. And yet in an etymological perspective the use of
the term in the former meaning has clearly preceded the latter significantly, as the
Oxford English Dictionary (2nd edition) entry on representation shows, which
differentiates among eight major variants. According to the OED, the fact of standing
for, or in place of, some other thing or person, esp. with a right or authority to act on
their account; substitution of one thing or person for another is first documented in
1624, while the fact of representing or being represented in a legislative or
deliberative assembly, spec. in Parliament; the position, principle, or system implied
by this only appears in 1769. As for the semiotic concept of representation, the
action of presenting to the mind or imagination; an image thus presented; a clearlyconceived idea or concept is first mentioned in 1647; however, the use of the term in
the fundamental sense of an image, likeness, or reproduction in some manner of a
thing is already found as early as 1425. This is the first documented occurrence of the
term representation in the English language and appears in a theological context.
22
time, they also activate subjective past-present-future positioning via indexical terms (now, then etc.) which transcend the realm of the represented and force us to engage ourselves mentally in the representational
game10. It is this interplay of time and tense, of the objectively perceived
(or imagined) time-line and our subjectively experienced position-in-time
which is so highly suggestive and lures us to immerse ourselves into a
fictional continuum of events.
3.2 Representation as Standing in for
But whose intention does a narrative stand in for? Obviously, it can
communicate the intentions of (real or fictional) agents that appear on the
content plane. More importantly, however, narrative also encourages us to
read it as a performative sequence made up partly of observations and
reflections, and partly of utterances, all of which we attribute to someone.
This someone is the product of a typical post hoc, ergo propter hoc reasoning: there is a thought, so somebody must have thought it; there is an
utterance, so somebody must have uttered it, etc. In everyday terminology
this instance is generally called the narrator; referred by structuralists,
however, as a narrative instance in order to avoid the anthropomorphism. While its ontological status is as problematic as its logical
genesis, the narrative instance is nevertheless a useful heuristic device in
acts of interpretation. However, if we want to understand how it works
then we will need to open the black box. And what we find in there is
what we propose to call the dynamic narrative system.
3.3 Constituents of the Model
In a process oriented perspective narrative representation is the function
of intellectual activities that run in parallel and across three dimensions:
10
McTaggart refers to these two fundamental principles of temporal ordering and positioning as that of the indexical (pastpresentfuture) A-line and the physical (beforeafter) B-line. Both, however, are logically dependent on the a-temporal C-line of
purely numerical or sequential ordering. By the same token one must regard the sequence of words that makes up the narratives text as a temporally neutral C-line. The
narrative texts often claimed temporality is in fact entirely induced by acts of interpretation. On the relevance of McTaggarts time philosophy for understanding narrative temporality see Meister (2005); on temporality and narrativity see Currie (2007).
23
The DNS of Mediacy
perception, reflection, mediation11. We can describe how these dimensions and the activities taking place in and across them are interrelated by way of what one could call the fundamental representational
formula:
representation = function of{ perception * reflection }
mediation
On the basis of this formula, we will start with an abstract overview of the
narrative systems components and then go into more detail later. In
diagrams 1, 5 and 6 we try to visualize the triangular relationship of mediation with perception and cognition in the form of a three dimensional,
dynamic intersection. The internal logic of each dimension is presented in
figures 2, 3 and 4. In section 4 the application of our model and taxonomy
to concrete literary examples will be demonstrated.
If representation is the output of the system, then perception, reflection and mediation are the three functional dimensions in which the system can (and must) perform in order to produce such output. In other
words, the system can only come alive and run if there is activity in all
three systematic dimensions. In order to make representation happen we
will therefore have to define the systems modus operandi in each of the
three dimensions: what are the constraints that govern perception, and
which goals have been set in this dimension? By the same token, what are
the constraints and goals set in the other two dimensions? This overall
mix of constraints and goals is what we call dimensional parameters.
11
Our term mediation refers to the process dimension of narrative representation,

whereas mediacy is a property of the product, i.e. of narratives. The complexity and
variety of narrative mediacy cannot be sufficiently captured in a two or three element
order pattern. Visualizations in terms of intersecting and mutually affecting dimensions
offer far better possibilities. The three dimensions in our modelperception, reflection
and mediationdiffer from tabular categorizations (cf. Genette), diagrams (cf. Rimmon-Kenan) or circle sectors (cf. Stanzel) in that they can display overlap to varying
degrees. This overlap can change gradually with regard to prototypical constellations
(see figure 5). In order to avoid the heterogeneous associations called up by terms such
as point of view, perspective and focalization our terminology consciously avoids any reference to these. Mediation, on the other hand, was chosen to in order to
capture processes of semiosis and representation without reference to specific media,
avoiding suggestive categories such as voice or verbalization. The closest resemblance
to our model can be found in Rimmon-Kenans parameter called facet (cf. RimmonKenan 1983: 7884) and in Schmids parameters of spatial-temporal positioning, i.e.
Standort, Zeitpunkt (for perception), Ideologie (for reflection), Verbalisierung (for
mediation) (cf. Schmid 2005: 13845).
24
What these dimensional parameters are, which of them are compulsory

and which optional, and what values they may take in principle is specific
to the type of representational system. In the system architecture of narrative representation all dimensional parameters are organized along one
and the same fundamental opposition of diegetic (narratorial) vs. mimetic (actorial)12. Each of the three DNS dimensions has at least one top
level parameter (cf. figures 2 to 4) which defines the dimension and is, as
such, not only logically indispensable, but also has the highest level of
impact.
Figure 1: DNS (Dynamic Narrative System) Model of Representation
12
Our identification of diegetic with narratorial and mimetic with actorial interprets
the Platonian (and Aristotelian) distinction in its narrower sense, i.e. as the two fundamentally opposed representational modes of telling vs. showing, or representational vs.
simulative. This is not to be confused with the second meaning of diegesis found in
Aristotle, where the concept denotes the narrators utterances in toto (i.e. in the modern
sense of narrative, Erzhlung or rcit).
The DNS of Mediacy
25
In a systems perspective, setting or changing one particular parameter is

always of more than just local consequence. Because the representational
system is dynamic, any parameterization can potentially affect not only
the remaining parameters within the same dimension; it will also result in
a systematic predisposition for the other two dimensions. This principle of
change one, change all is the key to the run-time logic that turns our
abstract DNS into a live system.
3.4 The DNSs Constituent Dimensions
Once again: the DNS as such is a model, not a taxonomy, and not even an
analytical or descriptive tool. Its intention is not to compete with existing
classifications of narrative phenomenology, in particular with those of
Genette or Schmid. Rather, its purpose is to provide such classifications
with a unifying theoretical frame of reference in order to prepare for the
next step (cf. section 4) where we will make a first suggestion towards a
typology consistent with our model.
3.4.1 The Dimension of Perception
In order to produce output a system requires some inputthe DNS is no
exception to this rule. In humans perception is one of three ways to provide us with relevant data, cognition and emotion being the other two.
Whether this data is sensory or mentally generated (i.e., thoughts), whether it is real or imagined, true or false is in the end irrelevant: if our aim is
to produce representational output, then all of it qualifies as valid input.
In Genettes narrative theory the specific conditions under which perception of narrated content takes place (or rather, is inferred to have taken
place) fall mainly under the category of focalization (zero vs. external
vs. internal). Focalization, it seems, regulates what a narrator or a
narrating character can know about the world, his epistemology, whereas
the particularities of voice (hetero- vs. homo- vs. autodiegetic narration)
determine the position in the fictional world from which he then utters his
communications13.
However, at least in the case of fictional narrative this distinction between knowing and communicating, between the epistemological and the
13
This reference to Genettes use of the term focalization is restricted to the aspect of
input. Of course, in Genettes own model focalization and voice at the same
time also account for the narratorial communicative strategy.
26
ontological constraints under which the representational system operates

seems methodologically problematic. Most of our inferences about the
narrating instances perceptive conditions will turn out to be based on the
very utterances which we ascribe to it. In other words, the narrating instances epistemological profile does not exist a priori, but is really just a
function of performative acts for which we hold it responsible: we
make the narrator by looking for one. The so-called autodiegetic I-narrator can best illustrate this dilemma: by necessity, its ontological position
must be in the world about which it informs usontology and
epistemology go hand in hand. Moreover, a narrative instance embodied
in the fictional world is in fact just a narrating character, and the quality
of information which he can relate to us is not just a matter of abstract
epistemological constraints, but also a question of his or her level of cognitive and emotive engagement. It is this mental closeness that matters
and profiles the narrating instance, not just temporal-spatial proximity.
Figure 2: Dimension of Perception
In the DNS model we define the dimension of perception as one which is

primarilybut not exclusively!characterized by epistemological input
The DNS of Mediacy
27
constraints and parameters14. The compulsory top level parameter defines the constraints of temporal and spatial proximity under which acts of
perception take place. When set to a diegetic value, such acts are
constrained by considerable distance between observer and the domain
observed: the typical epistemological position of an omniscient narrator
who plays the narrative game with an open deck.
On the other hand, when set to a mimetic value the constraints on
perception will generate close-ups from the contextually defined point of
view of a specific character. As our brief discussion of some of the problems in Genettes system has already indicated, we need to be aware that
this first dimensional take on the DNS introduces a systematic boundary where, in the reality of system performance, none exists. Perception
and processing are closely related, and any change in either dimension
will immediately have its effect in its counterpart. While the top-level
parameter insists on dimensional specificity, lower-level optional parameters create systematic overlap across dimensions.
3.4.2 The Dimension of Reflection
One of the tenets of structuralist narratology was the formalist conceptualization of narrated characters as mere surface layer representatives
for something that drives the narratives progress on the deep level of action logic: functions. Meanwhile, current narratological theory has begun
to rediscover the more traditional notion of character, demonstrating a
new interest in characters phenomenology and anthropomorphic qualities
(cf. Jannidis 2004). As a result, actants are extended into fictional
minds (Palmer 2004), a concept more apt to explain why and how readers engage with narratives across the full spectrum of mental activities.
The narrator, abstract as he or she might be, has never been at a similar
risk of being turned into a mere functional variable of representation and
14
By mapping identical parameters onto each of the three dimensions, yet in different sequence, we try to demonstrate their difference and interrelation at the same time. For
example, sensual perception will always go along with mental processes and rudimentary semiosis. The primacy assigned to the temporal and spatial parameter with
regard to perception links this dimension to the discussion of point of view/perspective. By contrast, the discussion on focalization is of relevance also to the dimension of reflection, as it is to the dimension of mediation (here with regard to information strategies).
28
action logic. One reason is that a narrator has a powerful real-life aid: the
author as whose alter ego he is often misread, if only covertly. One might
call this the narratorial fallacythe tendency to react to an inference
based construct, which is particular to the narrative type of representation,
as if it were real. This tendency is certainly not just found in cases
where readers confuse the categories of author and narrator. It has an even
more compelling motivation in what our DNS model integrates as
dimension of reflection.
Figure 3: Dimension of Reflection
What is contributed to representation in this dimension is a sort of mental

mark-up of the input that was derived in the dimension of perception and
then processed by what seems to function like a mind. The mark-up
which it generates defines the narrative systems cognitive, emotive and
normative relation to the object domain of the representation. These evaluative stances can be accentuated as diegetic and thus rendered as attributable to the narrator, or mimetic and therefore attributable to characters;
they can be contradictory or reinforce each other: in the end they all contribute to what one might call the Geisteshaltung, the mentality or,
literally translated, the mind position of the representation.
The DNS of Mediacy
29
A representation lacking this functional component is not narrative

or, to turn the argument around: our success in narrativizing a non-narrative representation like for example a photograph hinges to a large degree on our ability to add or identify this dimension in it post festum. Its
compulsory top level parameter is that of evaluative stance (cognitive,
emotive, normative). Optional lower-level parameters include semantic
profiling in terms of part-whole-relations, as well as spatial-temporal positioning. Again, these lower-level parameters overlap with the top-level
parameters of the other two dimensions.
3.4.3 Dimension of Mediation
In its dimension of mediation the DNS defines the semiotic constraints
that regulate the systems outputthe materialization of the representation in its double functionality of being an image of and standing in
for. This is where the conditions of possibility for concrete semiotic
realization are negotiated and stipulated, where paradigms, opposites and
isotopes are formed and highlighted, and where modes (e.g., tropes, metaphors, register) and media of articulation are selected and posited in relation to the object domain.
Figure 4: Dimension of Mediation
30
Must narrative representation meet specific criteria in this regard? Not

really, but its outer limits can be defined: a representation devoid of any
reference becomes an object in its own right; whereas a representation
devoid of any mark of non-identity vis--vis its referent cancels the referent and usurps its position in totality. Both are no longer representations,
but things present. Whether or not such an opaque mimetic representation
is indeed possible is hard to decide. One thing is obvious though: in our
daily lives the clear distinction between things that stand for other
things and things that are is increasingly hard to make.
We find ourselves directly affected by signs as if they were objects
even where these signs explicitly inform us about their semiotic status
(take the Dow Jones Index). Conversely, we regularly read empirical
objects and occurrences as signs for rather than experiencing them per
sewe have to do this in order to be able to learn or plan ahead. In short,
semiosis in interaction with human practice defies and subverts fixed
semiotic categorizations: in our existential practice, things are signs are
things are signs etc. Aesthetic semiotic practice, however, need not rely
on contextual markers in any event; it indicates its semiotic status inherently, by surplus structuring which indicates its poetic function (sensu
Roman Jakobson). Against this background the compulsory top-level
parameter in this dimension regulates the semiotic relation between signs
and their reference domain. Optional lower level parameters include those
which were top-level in the other two dimensions, spatial-temporal
relation (dimension of perception) and cognitive, emotive and normative
relation to the object domain (dimension of reflection).
3.5 DNS Run-Time Dynamics
The brief sketch of our DNS model presented thus far runs the risk of all
discursive prose: it can only present in sequence what in reality is a highly
recursive and dynamic process. Once the architecture has been activated
the live narrative system is constantly in flux: input that has been
processed in the dimension of reflection is passed back to the dimension
of perception; a new constraint in the dimension of mediation becomes
visible and forces the reader to re-run the system in his/her mind up to a
certain point, then jump back to the cut point, and so forth. Furthermore,
once actual reading takes place the system interfaces massively with
human mental processes which are way beyond the scope of our model.
The DNS of Mediacy
31
All we can try is to give at least a graphical indication of what the system at work would possible look like: three revolving planes that intersect
with one another in different ways on every rotation.
Figure 5: The DNS at Run-Time, i.e. Performing as a Dynamic Narrative System
4 From the DNS Model to an Analytical Heuristics

In the following we will present a typological table of modes of narrative
representation (see page 34). This seems to be in contradiction to our
initial criticism of combinatorial attempts to generate typologies. However, our typology is by no means intended to exhaust all valid combinations in parameter settings: it merely tries to project some of the typical
constellation that might occur along the three dimensional performative
continuum of perception, reflection, and mediation onto a two dimensional table. This is the first step; the second will be an attempt to demon-
32
strate how our abstract model might be applied in the practice of textual
analysis15.
This is how our table should be read: the constituent sub-processes of
narration are in the first (systematic, not real!) instance determined by
the different extensions of the narratives object domain (1st column).
Thereafter, we capture the synchronous processes of perception, reflection
and mediation in the form of three successive tabular dimensions (2nd to
4th column). In all three dimensions the standard qualification of a given
parameter is measured in terms of its relation to the object domain. The
parameters as such are (a) spatial / temporal proximity, (b) cognitive /
emotional / normative engagement, (c) semiotic disposition (which
increases step-by-step from an abstract disposition in the dimension of
perception to a realized manifestation in the right-most dimension of
mediation).
Within each of the three dimensions of narrative processing these parameters are graded along the continuum of lowmediumhigh impact. Every tabular dimension is continuously interacting with the other
two: the system is a fully dynamic one; in terms of computational programming approaches one might compare it to a recursive and highly interactive modular program architecture rather than a batch-mode first do
this, then do that algorithm. When we read a row in our table across its
three centre columns and their respective sub-columns we can see the
scope of variations in relations to object domain that fall under one particular representational type. In reality, the number of such types might
be huge; we have decided to limit ourselves to just six types which seem
to be best documented historically and can thus be cited as exemplary
cases. Finally, the two right-most columns compare the traditional
Genettian type-term with our suggested terminological replacement.
The measuring of a particular parameter in terms of its relation to object domain value is thus not a question of yes or no; it is a question of
attributing it a particular position within an array that extends from high
to low. If we want to measure the level of internal influence which the
initiating instance of the narrative process (or the textual instances that
represent it) can have, then we will differentiate along the axis of low
mediumhigh interest. If on the other hand our interest lies in measuring
the extent to which the process is constrained by text-external (historical
and cultural) factors then we will do so along the scale of fullymedi15
See Grabienski et al. (2006).
The DNS of Mediacy
33
umlow constrained. The values entered in our table are not of an absolute nature; rather they represent an ensemble of tendencies which in
their combination allow us to describe the dynamics of narrative processing. From a literary history perspective the few prototypical constellations represented in our table can only capture a glimpse of what has
beenor might still berealized empirically.
The typology of representations is based on the following premises:
(a) The qualification narratorial defines a position external to the narrated world. The narrating instance is by default completely autonomous
and unconstrained; however, it is marked as narratorial on a gradual scale
as soon as a level of limitation affecting its operations in the three dimensions becomes discernable.
(b) The qualification actorial defines a position within the narrated
world. Again, the narrating instance is marked on a gradual scale in terms
of its dimensional limitations: for example, by the spatially and temporally defined position from which the instance observes simultaneously
occurring events, as in the case of an eye witness account, or by the simultaneity of experience and narration, e.g. in a protagonists interior
monologue. The latter is in contrast with the so-called autobiographical
mode of narration. This mode allows for the narrating instances choice of
different spatial-temporal positions within the dimensions of perception
(which is, by definition, experience centered) and mediation (where the
focus is primarily on representation). In a typical autobiographical
narrative different situations in life are defined by different constellations
in the protagonists cognitive, emotional and normative engagement.
(c) Finally, a third type of mediacy is defined in terms of hybrid positions,
which we call mixed narratorial / actorial. Here the narrators acts of
evaluation and mediation take place from one position, but are combined
with acts of perception and reflection bound to a second position that
indicates an actorial stance. Actorial mediacy, in these cases, is graded on
a scale ranging from covert to overt. An example would be the
difference between a completely factual eye witness report, and an affected by-standers account displaying traces of personal engagement
with the ongoings16.
16
A term we deliberately avoid in our qualification of the six prototypes is extradiegetic. In our opinion the term is a tautology in that it merely captures the selfevident epistemological prerequisite of all narrative representations: as soon as we talk
about diegesis in any meaningful way, we have to associate the enunciative act with
an enunciator, and dissociate the product of enunciation (the narrative, the text) from it
34
As we will demonstrate, representational types can be attributed to texts

as a whole, but also to passages within texts as we will show below. In
some instances the distinction among types is not very clear cut, as in the
case of constrained narratorial representation vs. mixed narratorial /
actorial representation: the former type also subsumes phases where the
default autonomous vision of the narrating instance is temporarily restricted by constraining its powers of perception and reflection to those
attributable to one or more actors; in the latter type the narrative instance
is parameterized throughout in accordance with the epistemological and
at the same time. The logical opposite to extradiegetic would in fact not be intradiegetic, but simply mimetic. The current (Genettian) use of the qualifier intradiegetic, however, is not intended as a statement concerning the ontological status of
the representation as such: it merely tries to point out that the act of narration is, at the
same time, its own object; in other words: that diegesis is not organized as a two-level
affair of signifiant vs. signifi, but rather in the form of nested instances of narration.
The DNS of Mediacy
35
reflective position of a particular actor, but the overall semiotic disposition will nevertheless indicate a higher-level narratorial instance.
We will now analyze two textual examples in order to illustrate how
the dynamic narrative systems mode of operation might be measured in
terms of the continuously changing values which it assigns to the functional parameters in its three interrelated dimensions of perception, reflection, and mediation17.
Example 1: Charles Perrault Little Red Riding Hood
Once upon a time there lived in a certain village a little country girl, the prettiest creature
who was ever seen. Her mother was excessively fond of her; and her grandmother doted
on her still more. This good woman had a little red riding hood made for her. It suited the
girl so extremely well that everybody called her Little Red Riding Hood. One day her
mother, having made some cakes, said to her: Go, my dear, and see how your grandmother is doing, for I hear she has been very ill. Take her a cake, and this little pot of
butter. Little Red Riding Hood set out immediately to go to her grandmother, who lived
in another village.
As she was going through the wood, she met with a wolf, who had a very great mind
to eat her up, but he dared not, because of some woodcutters working nearby in the forest.
He asked her where she was going. The poor child, who did not know that it was dangerous to stay and talk to a wolf, said to him: I am going to see my grandmother and carry
her a cake and a little pot of butter from my mother. Does she live far off? said the
wolf. Oh I say, answered Little Red Riding Hood, it is beyond that mill you see there,
at the first house in the village. Well, said the wolf, and I'll go and see her too. I'll go
this way and go you that, and we shall see who will be there first.
The wolf ran as fast as he could, taking the shortest path, and the little girl took a
round-about way, entertaining herself by gathering nuts, running after butterflies, and
gathering bouquets of little flowers. It was not long before the wolf arrived at the old
womans house. He knocked at the door: tap, tap. []
Little Red Riding Hood took off her clothes and got into bed. She was greatly amazed
to see how her grandmother looked in her nightclothes, and said to her, Grandmother,
what big arms you have! All the better to hug you with, my dear. Grandmother, what
big legs you have! All the better to run with, my child. Grandmother, what big ears
you have! All the better to hear with, my child. Grandmother, what big eyes you
have! All the better to see with, my child. Grandmother, what big teeth you have
got! All the better to eat you up with. And, saying these words, this wicked wolf fell
upon Little Red Riding Hood, and ate her all up.
Moral: Children, especially attractive, well bred young ladies, should never talk to
strangers, for if they should do so, they may well provide dinner for a wolf. I say wolf,
but there are various kinds of wolves. There are also those who are charming, quiet, po-
17
The examples were taken from the web and have not been philologically verified. Yet,
for the purpose of a demonstration of our model in application they should suffice.
36
lite, unassuming, complacent, and sweet, who pursue young women at home and in the
streets. And unfortunately, it is these gentle wolves who are the most dangerous ones
of all.
These passages taken from a short text present an example for our type 1
(unconstrained narratorial representation), including a passage of quasimimetic scenic representation marked in italics. In this example the
dynamics of the three dimensions could be outlined as follows.
PERCEPTION: the narratorial instances perceptive abilities are generally not constrained by the spatial or temporal limitations of any single
actorial positionthe path of the wolf and the path of Little Red Riding
Hood are equally followed. Physical objects as well as the mental states
of the wolf and Little Red Riding Hood are being presented (viz. the short
sequence of introspection at the beginning of paragraph two). However,
the perception of narrated events is only marginally intersected by the
dimension of reflection. With a view to mediation, these formulations underline the narrators distanced and ironic position of cognitive superiority vis--vis the characters, as in this opening: The poor child, who did
not know that it was dangerous to stay and talk to a wolf [...]. This
superiority, however, will only be put to full effect in the concluding
moral of the story, where the focus of perception no longer lies on the
fantastic story of Little Red Riding Hood and the wolf, but rather on the
real constellations of social life.
REFLECTION: in the beginning, the amount of the narrating instances
cognitive, emotional and normative engagement during the act of perception is minimal, and the semiotic disposition is unmarked. However,
long before the final moral is explicated a first sign of reflection-processing is detectable in the narrating instances normative evaluations of
the actors (Little Red Riding Hood is being spoilt by her mother and
grandmother; the wolf is hungry and ravenous.) In the final moral of the
story the normative engagement of the narrative instance increases dramatically from medium to high interest.
MEDIATION: the initially gradual and then suddenly exponential increase in normative engagement is paralleled and supported by the curve
which the actualization of semiotic disposition along the mediating process follows. First a number of isolated semiotic determinants manifest
themselves (including introspection into the protagonists state of mind
and the representation of character-bound attitudes, even though the latter
are not explicitly marked as actorial in their verbalization) before the final
The DNS of Mediacy
37
moral is presented as an explicit marker of the narratorial activity of

diegesis (in the Platonian sense).
This first example demonstrated how a description in terms of traditional narratological categories can be formalized in terms of the DNS
model. Our second example will now demonstrate the added text analytical capacity which sets the DNS model apart from its precursors.
Example 2: Alfred Dblin The Murder of a Buttercup
[a] Yes, he had killed the flower, and it was no business of theirs, and he had the right,
which he would defend against all of them. He had the right to kill flowers, and he did not
feel obliged to justify it in any more detail. He could kill as many flowers as he pleased
for a thousand miles around, north, south, east, west, whether they scoffed at it or not.
And if they carried on laughing like that he would leap at their throats. [] [b1] Again he
runs hard against a low fir; it strikes down at him with raised hands. [b2] He breaks his
way through violently, the blood running in streams down his face. He spits, lashes out,
kicks the trees, yelling, slides down, sitting and rolling, finally runs headlong down the
last slope at the verge of the wood toward the lights of the village, his torn frock coat
thrown over his head, [c] while behind him the mountain rustles threateningly, shaking its
fists, and everywhere trees can be heard cracking and breaking as they run after him,
cursing.
This passage (an excerpt from a longer novella) presents a type 3 example
(mixed narratorial / actorial representation): in [a] free indirect speech
is used to represent the thoughts of the protagonist (he has beheaded a
butter cup with his walking stick). This is followed by the description of a
number of actions [b] in which actor centred perception [b1] merges into
the externally based perception of a narratorial instance [b2] and then [c]
reverts back to actorial perception. The dynamics of the three dimensions
could be outlined as follows.
PERCEPTION: as the historical present in paragraph 2 indicates, perception is bound to time and place of the fictional events, yet at the same
time it is intermingled to a high degree with actorial emotions and evaluationsand so is the semiotic disposition, which is determined by actor
centred patterns. But this is not a fixed constellation: the dimensions of
perception, reflection and mediation are being repeatedly and dynamically
repositioned against one another. The effect is such that perception, by
way of smooth transitions, is also characterized by the evaluations and
verbalizations of the narrator. One example is the second sentence in [b2],
where toward the lights of the village signals that the predominantly
narratorial perceptionby way of a mergingis momentarily juxtaposed
38
with a actorial orientation, thus subtly preparing the eventual shift to a

fully actorial perception in [c].
REFLECTION: the processing of perceptions mirrors the strong relationship with the actorial dimension: strong in [a], [b1] und [c]; weaker in
[b2]. However, semantics and syntax indicate the narrators intent to demonstrate the intense, pathological state of mind of the actor, without
resorting to explicit commentary.
MEDIATION: the predominant commitment to the spatial and temporal
position of the actor as well as to his cognitive, emotional and normative
position is amplified during mediation processing. However, in its actualization of the semiotic disposition the narratorial instance upholds its
claim to perform representational operations which transcend the actorial
disposition.
The following presents another example from Dblins text:
[a] He was paying, paying for his mysterious guilt. [b] He was performing divine service
with the buttercup, [c] and the calm businessman asserted now that each person had his
own religion; [d] one had to assume a personal position to an ineffable God. There were
things that not everybody could understand. [e] A trace of suffering had appeared in the
gravity of his monkeys face; his corpulence had also decreased, his eyes became deep
set. [f] The flower, like a conscience, watched over his actions, stringently, from the most
important to the smallest everyday deeds.
This is an example for a type 2 constellation (constrained narratorial representation). Passages of reflector bound co-vision in [a], [d] and [f]
alternate with passages of unconstrained narratorial representation [c]
and [e], which include the option of introspection and commentary. In between we find passages of gradual transformation from actorial to narratorial profiling of the three dimensions [b].
The summary effect is one of a medium-status which oscillates between unconstrained narratorial representation and mixed narratorial /
actorial representation with gradual transitions. The segments of overlap
of the three dimensions change from sentence to sentence. A precise definition of who sees? and who speaks? is only possible in a few prominently marked positions within this constant flow.
The DNS of Mediacy
39
Figure 6: DNS Oscillating Between the Extremes of Type A/B-Mediation
5 Outlook
At this stage our DNS model is a first draft which obviously requires refinement in terms of its design and the analytical categories derived there
from. In the current volume the contributions by Markus Kuhn and Sabine Schlickers demonstrate how a literature based Genettian descriptive
apparatus can be fruitfully applied to other media: it remains to be seen
whether the narratological DNS model and its typology, too, extend in
relevance beyond text based representations. However, we believe that
two particular characteristics might make our model a strong candidate
for such transmedial application: one, its constituent process dimensions
and functional parameters perceptionreflectionmediation do not
show the usual bias for a particular medium of representation, nor for visual metaphors. Two, the DNS model is designed to account for the generic as well as the historical dimension of narrative: it conceptualizes the
dynamics of narrative processing as one that governs all narrative specimen, yet it always remains susceptible to change and creative mutation
itself.
40
References
Currie, Mark (2007). About Time. Narrative, Fiction and the Philosophy of Time. Edinburgh: Edinburgh UP.
Genette, Grard (1980). Narrative Discourse. Tr. J. E. Lewin. Oxford: Blackwell.
(1988). Narrative Discourse Revisited. Tr. J. E. Lewin. Ithaca: Cornell UP.
Grabienski, Olaf et al. (2006). Stimmen-Wirrwarr? Zur Relation von Erzhlerin- und
Figuren-Stimmen. A. Blhdorn et al. (eds). Stimme(n) im Text. Narratologische
Positionsbestimmungen. Berlin: de Gruyter, 195232.
Herman, David (ed) (2003). Narrative Theory and the Cognitive Sciences. Stanford, CA:
CSLI Publications.
Jahn, Manfred (1996). Windows of Focalization: Deconstructing and Reconstructing a
Narratological Concept. Style 30:2, 24167.
(2001). Narrative Voice and Agency in Drama. New Literary History 32:3, 659
80.
(2005). Focalization. D. Herman et al. (eds). Routledge Encyclopedia of Narrative
Theory. London: Routledge, 17377.
Jannidis, Fotis (2004). Figur und Person. Beitrag zu einer historischen Narratologie.
Berlin: de Gruyter.
Martinez, Matas & Michael Scheffel (1999). Einfhrung in die Erzhltheorie. Munich:
Beck.
Meister, Jan Christoph (2005). Tagging Time in Prolog. The Temporality Effect
Project. Literary and Linguistic Computing, Vol. 20. Oxford, UK, 10724.
Electronic publication at <http://llc.oxfordjournals.org/cg>i/reprint/20/Suppl/107>
(last seen 01.07.2008).
Niederhoff, Burkhard (2001). Fokalisation und Perspektive. Ein Pldoyer fr eine
friedliche Koexistenz. Poetica 33:12, 121.
Palmer, Alan (2004). Fictional Minds. Lincoln: U of Nebraska P.
Rimmon-Kenan, Shlomith (1983). Narrative Fiction. Contemporary Poetics. London:
Routledge.
Schmid, Wolf (1982). Die narrativen Ebenen Geschehen, Geschichte, Erzhlung
und Prsentation der Erzhlung. Wiener Slawistischer Almanach 9, 83110.
(2005). Elemente der Narratologie. Berlin: de Gruyter.
Van Peer, Willie & Chatman, Seymour (eds) (2001). New Perspectives on Narrative Perspective. Albany, NY: State U of New York P.
URI MARGOLIN
(Edmonton, Canada)
Focalization: Where Do We Go from Here?

All the papers in this volume show a desire to revise current focalization
theory in the hopes of strengthening and improving it, and to try and resolve at least some of the many open issues in this field. The two wishes
are closely connected, since it is often the reformulation of assumptions
that leads to the resolution of issues. The desire for revision seems to take
one of three directions: (1) expansion of the domain of application of focalization theory to other media, with the necessary theoretical modifications (2) reconfiguration (add, delete, replace, rearrange) of the systems of categories and distinctions currently available and (3) a reconceptualization of the whole theory by placing it within a more fundamental theoretical framework, be it fictional world semantics or cognitive
linguistics, both of which are ultimately semantic theories. I will try to
contribute modestly to both reconfiguration and reconceptualization efforts, but let me start with the most basic question wozu? (to what end?)
or why do we need a theory of focalization to begin with?
Fortunately, several very good answers to this potentially devastating
question have been offered in recent years. In a major article, Mieke Bal
(1993) asserts that all cognitive activity is located in the projects and constructions of specifically positioned subjects, and that narrative has the unique capability to map differently positioned subjects in their relation to
knowledge and to each other. To this one might add that viewing knowledge as position-dependent enhances ones ability to imagine standpoints
different from ones own, with correspondingly different insights regarding the same data, and to accept these other standpoints as alternatives to
ones own. In a contribution to a collective volume on narrative perspective Ansgar Nnning (2001) asserts that literature is the only place where
the construction of world models is thematized, and that narrative
thematizes and structurally reflects the problems attached to the construction of world models. Furthermore, the cognitive processes through which
42
Uri Margolin
an individual or collectivity actively construct their subjective world

models are represented on the story level in the perspectives of the characters. Finally, the Canadian scholar Pierre Ouellet (1996) rightly claims
that works of fiction represent the cognitive and perceptual experiences of
subjects such as characters and narrators; that the novel in particular is
constituted much more by experiences in the phenomenological sense
than by objective states of affairs; and that in the modern novel great
importance is attached to the representation of perceptual and cognitive
processes and activities, to the exploration and dramatization of the
minds movement. But what is focalization?
1 Defining Focalization
Focalization is the general term (Ober- or Sammelbegriff) used to designate at least some of the mental activities just mentioned and their products. One can describe focalization informally as a view of a thing as it
presents itself from the personal subjective point of view of a character or
narrator. To be more precise (and this is my proposed definition): focalization in narrative involves the textual representation of specific
(pre)existing sensory elements of the texts story world as perceived and
registered (recorded, represented, encoded, modeled and stored) by some
mind or recording device which is a member of this world. In other
words, focalization involves at least the internal inscription of external
data. Conversely, any state or event mentioned in the text which can possibly be thought of as being perceived in any way can be considered to be
the product of an act of focalization, hence indexed to a particular individual, time and place.
1.1 Five Components of Focalization
Occams razor reminds us that entities should not be multiplied beyond
necessity. This lesson is not lost on me, yet I would claim that, as follows
from my definition, any adequate description of focalization involves essentially not less than five factors. These are (1) the story-world state or
event focalized; (2) the focalizing agent and its make-up; (3) the activity
of perceiving and processing this object-focalization as nomen actionis;
(4) the product of this activity, that is, the resultant take or vision and (5)
the textualization of all the above, which is the only thing directly accessible to the reader and not requiring his imaginative reconstruction. It is
43
obviously also the one from which all other components are (re)constructed in the reading process. Moreover, the very basic distinction between who sees and who says, one of the key elements of literary narratology, is ultimately grammar and lexicon based. Let me now expand a
little on each.
(1) The object focalized is an element or sector of the story world:
states, entities, actions, events and processes, some located in space and
time and some internal or mental such as memories of previous acts of
focalization. Such an element or sector is the object of the focalizers attention and subsequent inner processing. Whether everything in the story
world can be meaningfully considered as object of potential focalization
is a contentious issue to which we shall return (for more on objects of
focalization in literary contexts see 3.1 below).
(2) The focalizing agent is a human or human-like story world participant who concentrates or focuses selectively on a portion of the available
sensory information. At its core is a mind or recording device with its capabilities, faculties, structures and constraints. These would include embodiment, situatedness or space-time position (=vantage point), architecture (=mechanisms, categories, routines) and, for human minds, also
norms, values and epistemic attitudes. A focalizing agent may consequently be termed perspective and it is an agent that performs numerous acts of focalization in the course of the story, and is hence a narrative macro element. The inner structure of focalizercumperspective
has been the subject of detailed study in recent German narratology,
especially by Nnning (2000; 2001) and Carola Surkamp (2003).
(3) Modeling or processing is either a momentary act or an extended
activity consisting of perceiving, viewing, selecting, making discriminations, matching information to frames and scripts (=schemas and scenarios), categorizing, gestalt forming, making connections, interpreting,
evaluating and so on. These various operations of construal, and their
products, are studied in detail in cognitive psychology, especially psychology of perception, and in cognitive linguistics. The narratologist
could and probably should employ the distinctions and definitions already
available in these disciplines regarding cognitive modeling rather than
invent his own. Seen from this perspective, stream of consciousness for
example is largely a technique for representing in a non-mediated fashion
the process whereby the mind registers incoming instantaneous sensations
and tries to identify and relate them to other current or remembered
sensory experiences. Similarly, detective novels often highlight the dif-
44
Uri Margolin
ficulties involved in perceiving under adverse condition such as in the

dark or from a hiding place. Focalization becomes a difficult and drawn
out process, almost a struggle, which may end up in failure, partial success or full success (cf. Schubert 2005).
(4) Take or vision of the given, or mental scene, is the inner representation of information, the product of the perceptual and conceptual
processing activity with respect to the focalized, or, in other words, the
product of the construal operations brought to bear on the segment of reality serving as the focus of attention. David Herman, in his essay in this
volume, discusses in detail the dimensions or parameters involved in an
adequate description of such mental products. Informally, one should
mention at least the selection, degree of detail and pattern given to the incoming data, and the fore- and backgrounding of information items.
(5) Textual representation includes all the linguistic (tense, modality,
deixis) and stylistic devices employed by authors for portraying at least
the processing activity and its product, drawing our attention to the fact
that what we are facing is a particular character perspective through
which entities are being perceived and represented. In an insightful article, Laurel Brinton (1980) has proposed the term represented perception to designate some linguistic and stylistic elements that indicate
that we are watching the external world being transformed into an internal
one (369). Perceptions are thus rendered in an unmediated fashion (i. e.,
without verba sentiendi) as they occur in the characters minds; through
emotionally laden, subjective and evaluative adjectives or verbal forms,
such as darling little spots or chuckling, absurd sound, and the flashlike nature of some sensations is rendered through disjointed syntax or
nominal phrases (37581).
(6) Finally, both minds and recording devices may contain an optional
self-monitoring component referred to as consciousness or self-awareness when humans are concerned. This device may be active during the
focalization act, running in parallel to it and providing an internal feedback or commentary. Such a commentary would among others make the
perceiver aware that he is engaged in perceiving an object and forming an
internal representation of it, and may sometimes also cast doubt on the
scope or validity of the take even as it is being formed.
45
1.2 Terminology: What Are We Talking About?

A large number of terms exist in English language narratology for designating the focalizer, and they can be grouped according to the aspect they
highlight:
mirror, screen, reflector, filter, prism stress the mediating role;
angle of vision, point of view, origo, focus, vantage point, window and
perspective stress the specific situatedness of the agent: spatial, temporal
but also conceptual, cultural and epistemic;
viewer, perceiver, cognizer, and experiencer point to aspects of the
mental activity involved;
(finally) center of subjectivity, awareness or consciousness and mediating consciousness remind us that a human or human-like mind is behind
most focalizations in literature.
The resultant take or vision has also had its fair share of terms, but
here all of them stress the same point, namely, the difference between the
focalized object as such and its particular representation by a given mind
or device. Thus we have modeling of an object, perspectival version, projection, prismatic refraction, and mental picture. The most general term is
obviously mental representation, and situation model or mental model
where sensory perceptions of space-time objects are concerned.
2 Focalization and the Wider Epistemic Context

In classical narratology, the phenomenon of focalization has often been
treated as sui generis to narrative or even literary narrative, yet it is in fact
just one special case of a much larger and far more basic picture of the
mind in action, with its representational, semantic and epistemic components. Corresponding wider and more basic frameworks are available in
several disciplines, and each of them could actually serve as basis for a
reconceptualization of focalization theory. Let me now enumerate several
such frameworks in order of increasing abstraction:
Within narratology, focalization is one component of a general
theory of fictional minds, that is, of the literary representation of mental
activity in all its varieties.
In terms of fictional worlds semantics, each take is part of an agents
particular epistemic perspective on the story world. Modally seen, one
posits, following Marie-Laure Ryan (1991) a textual actual world consisting precisely of all the facts of this world. Now each agent in it, sub-
46
Uri Margolin
ject to his information access possibilities and inherent processing capabilities, forms in his mind a take on one or more items of this world, each
such take forming an element of his epistemic map of this world or his
personal belief world relative to the textual actual one. In text-world semantics, one can speak of each take as component of the textual subworld
set up by the focalizers mental activity.
For cognitive linguistics, each take on a given situation constitutes
one of several possible construals or conceptualizations of this situation,
one of several alternative conceptual structures the mind can impose on
the same external phenomenon. The specific lexical and grammatical
choices made by the author in portraying a given take are viewed as indicators of the perceptualcumcognitive operations (process) which gave
rise to this (fictional) take, as well as to its specific structure (product).
The basic kinds of objects that can be involved in such construal operations are scenes and events, entities and processes, motion and location, and force and causation (cf. Talmy 2006: 542). The basic structures
or schemas employed in organizing these objects include the configurational (objects in space and time and their relations), perspectival
(location or path of the point at which one places ones mental eye to
regard a scene), attentional (patterns in which different data are fore-and
backgrounded), and force dynamics (relations between entities such as
opposition, overcoming, helping and hindering, causing and preventing)
(54344).
Cruse (cf. 2004: 4673) provides a detailed discussion of the basic
construal operations occurring in the mental structuring of data. These include: (1) attention, which encompasses selection, focus, scope and degree of detail and its summary or sequential scanning; (2) comparison, organizing the incoming data into fore- and background elements; (3) perspective or situatedness, defining the vantage point and orientation of the
observation, as well as the location and path of attention; (4) the constitution from data of spatio-temporal objects and their interrelations, that
is, providing a structure for the experience; and finally, (5) the conceptualization of processes and events as involving different kinds of
forces acting in different ways upon the participants of the events.
For cognitive psychology, focalization as defined in this article
could be fully subsumed under perception in the wider sense. Thus, James
Pomerantz in his entry on perception (2003) defines perception as the
complex sequence of processes by which we take the information received from our senses and then organize and interpret it, which in turn
47
allows us to see and hear the world around us as meaningful, recognizable

objects and events, with clear boundaries in space and time. We are also
given some basic facts about perception: it is limited, selective, influenced
by context and not entirely veridical (=corresponding to facts), and it
requires time, memory, and internal representation.
In terms of Husserls phenomenology one distinguishes the object of
attention or intentional object, the ego that has this object in mind, the
acts of consciousness applied to it, and the resultant noema or intentional
content.
In cognitive science, one speaks of objects of attention, mind, cognitive processing and the resultant mental representation.
(Finally) in terms of information theory, one speaks of information
or data input, an information processor, a series of internal operations
based on computation in the wider sense and the resultant information or
data output.
The adoption of any such theoretical framework would entail at least
the translation of the terms of focalization theory into those of the higher,
more powerful theory, and a corresponding reformulation of focalization
theory claims in terms of the framework selected. If this operation is successful, focalization theory becomes a sub-theory of the higher one, but
quite possibly containing some claims specific to the literary domain, so it
cannot be derived from the higher one or reduced to it, and we still get to
keep our jobs. Let us further note that if this subsumption under a higher
theory succeeds, then many of its insights may also apply to the literary
domain, giving us extra knowledge for free, and also suggesting many
perspectives and issues that could not occur to us within the narratological
context in isolation. The prospects of any such interdisciplinary reconceptualization are daunting indeed, but I am going to evade this task
by invoking the scholars most trite excuse: ars longa, vita brevis. Instead, I propose to go back into more specifically narratological issues
concerning the revision of current distinctions.
Focalization in the narrow sense is an act or activity. One can thus
ask about its object (what is the object of this act); agent (who focalizes);
and product (or how, that is, kinds of focalization). One can also enquire
about combined phenomena, such as the intersection of who and how or
agent and manner, which underlies all typologies of focalization. Before
we get down to details let us remember that our goal is not to produce
some epistemic or psychological general truths or systematics, but rather
to provide fruitful heuristic tools for the description of artistic products
48
Uri Margolin
containing a significant mental component in their content level. And let

us further recall that fictional worlds do not necessarily correspond to the
actual one that science studies, and need not obey its rules or categories.
3 What, or Contents of Focalization

All focalization is a mental activity, but not all mental activity is focalization. Similarly, not all objects of mental activity are objects of focalization. Focalization concerns only specific kinds of mental activity and
limited kinds of content.
Everybody agrees that incoming data from all five senses as they are
being received and processed form an essential component of focalized
content. By the same token abstract class concepts, universal truths and so
on do not belong here, even though they are part of mental content in general. Analogously, while the registering in the mind of sense impressions
is central to focalization, activities of abstract reasoning and inferencing
do not belong here.
Acts of mental simulation or empathy whereby I try to put myself
mentally in the place of another and imagine what he is perceiving or perceived from his time space location and how he is or was perceiving it,
are objects of second order focalization. Here, both the others act and its
manner, as well as its contents, are objects of focalization.
Acts of recollection, where an agent activates his own episodic long
term memory in order to bring into mind what he experienced through the
senses at some earlier time space point and how he did so are parts of focalization.
Some theorists, like Jahn (1999) and Bal (1997), argue for including
under focalization the contents of acts of planning, projection of future
scenarios, dreams, delusions and hallucinations, that is, exclusively internally generated perceptions. I find this inclusion problematic. While these
contents are indeed specific and inner-sensory in nature (images and
sounds) and involve imagined external events and situations, they do not
fall under the heading of pre-existing or existing story world elements as
experienced and registered by a mind inside this world as earlier defined.
Moreover, issues of the scope and correctness of any take with respect to
the textual actual world, so crucial to our assessment of its focalizer, are
by definition inapplicable here. As these phenomena clearly fall outside
my initial definition of focalization, I must exclude them for better or
worse. On the other hand, when a character recalls the contents and
49
manner of his original past acts of projection or of dreaming, he is concerned with mental episodes that did take place in the actual story world,
so these could be considered objects of focalization, and the accuracy of
his recollection of them can be meaningfully discussed. To conclude: in
my view, it would be fruitful for our theoretical work to include under
contents of focalization the contents of mental episodes involving sense
data stemming from spatio-temporally determinate situations, events and
entities, whether these sense data are perceived directly or at a remove of
time or person.
3.1 Conventions Regarding Focalization
From contents of focalization we now move to some specific artistic assumptions underlying fictional focalization, since focalization as such is a
general discursive category. These assumptions concern access to the
contents of characters focalization acts, as well as to the mental activities
involved. In fact, these assumptions are a central part of the constitutive
conventions that establish the institution of narrative fiction, and are orthogonal to our default assumptions about embodied human experience in
the actual world.
First and foremost is the convention that mental representations in
characters minds can be accessible to narrator and reader even if not expressed by the characters through words, drawings, or any other public
means.
Secondly, it is assumed that all mental representations are verbal,
even though in reality some are propositional and others image based.
All human focalization is active and transformational, and contains
an element of interpretation. Any individual act of focalization is just one
particular perspective on the story world, and is always fallible and often
skewed, distorted or at least partial. As we know from Kant, human
beings can know the world only as a series of mental representations
whose shape is determined by the constitutive conditions of the human
mind. In narrative fiction, however, there is the assumption that one can
know states of affairs hors de toute focalisation, fully and with absolute
certainty, through the discourse of an impersonal anonymous narrating
voice, usually in the third person past tense. Each individual take can thus
be assessed relative to this full objective truth. And this in turn enables
the reader to evaluate different takes regarding the same data, and also in-
50
Uri Margolin
fer back from the nature of a take to the nature of the focalizer behind it
(+/-limited, reliable etc.)
4 Who, or the Focalizing Agent

What can be scenically presented or dramatized in focalization are the
contents of individual takes and their construction in the mind. The underlying perspective or focalizing mind itself with its architecture can
only be reported upon in summary mode, being a set of general features.
As we have seen before, it is up to the reader to make further inferences
on the nature of a given focalizing mind from what and how it focalizes.
Consequently, further questions about the nature of the focalizer as textually represented have to follow the structure of narrative discourse and
not psychology. Maximally, the following narrative instances can function as focalizers: a narrative agent or story world participant; some anonymous position in story space (Claude Simons O in La bataille de
Pharsale, Banfields [1987] empty deictic center); a hypothetical agent
inside story space (Hermans [1994] hypothetical focalizer); and, in my
view, some kinds of narrators all the time and all narrators on some occasions. With so many focalizing instances, is all information in narrative
focalized, as Ronen (1994) and Ouelette (1996) claim? Here opinions
vary. If focalization is defined as a relation between narrated domain elements and a mind inside this domain (Princes [2001] internal focalization), then precisely all information coming through minds inside this
domain is focalized, and thats it. If, however, focalization is defined as a
relation between narrated domain elements and a mind inside the narrated
or narrating domain, then information coming through narrators who are
or were members of the narrated domain they are reporting about may be
considered focalized, at least in the sense that it comes from a personalized instance or center of awareness. On the other hand, no one would
consider information coming from an omniscient impersonal voice or
speech position focalized, as this kind of narrating function is incompatible with any notion of a restricted experiencing mind or an embodied
and situated center of awareness. But even here there is an exception: information coming from such a voice can be considered focalized if the
voice pretends to look at events through the mind of one or more story
world participants, a phenomenon dubbed reflectorization by Monika
Fludernik (1996).
51
Two cases that have evoked lots of controversy, but for opposite reasons, are the anonymous focalizer and the personalized narrator. The first,
because it is not a textually inscribed speech position, and the other
because, according to scholars like Chatman (1986) it is nothing but a
speech position. Let us begin with the anonymous focalizer.
4.1 The Anonymous Focalizer
In many cases we encounter in fiction a passage that looks like a take on a
given spatio-temporal situation from some inside subject position, as indicated by scenic immediacy, deictics and so on. But no story world participant is textually indicated as the observer-experiencer or origin of this
take. So who sees then? In such cases we sometimes postulate a nameless
observer as the focalizer. Monika Fludernik (1996) suggests we call this
operation figuralization, as we are attributing to some anonymous
observer figure the information in question. Let us not forget, however,
that this observer is a mere interpretive Hilfskonstruktion, the product of
an operation of naturalization. But what does this anonymous observer or
witness position, which Herman (1994) has dubbed unspecified virtual
witness, mean? We are in fact claiming that the specific nature of the
given information can be realistically motivated by positing as its origin a
standard observer position on the scene whose location can sometimes be
pinpointed. Following Hermans notion of hypothetical focalization we
could also say: this is what would be seen by whoever, any human observer, including the reader, if they were located at this space time position. All the same, whether or not a given passage represents a take to
begin with, and, if so, whose exactly, are often interpretive and contextdependent decisions, and we may arrive at no clear answer or at several
alternate equally plausible ones. Also, the transition from one take to another, or from a take by one person to that by another, are often textually
unmarked and subject to interpretive debates. Such indeterminacy is the
constructive principle of Vargas Llosas novel Conversacin en la
Catedral, for example.
4.2 Narrators as Focalizers?
Quite probably, no issue in focalization theory has generated more controversy than whether or not narrators can be focalizers. I believe the
question is wrongly put. After all we are dealing with artificial artistic
52
Uri Margolin
constructs, not with facts of nature. I think the question should accordingly be reformulated as follows: in what cases is it meaningful or fruitful
to consider a narrator as a focalizer as well, in view of our initial definition of focalization. I believe it makes perfect sense in some cases, and
the categorical refusal to do so stems from a failure to distinguish between
role or function and individual and to realize that a focalizer or a narrator
are not flesh and blood monolithic entities which remain constant
throughout, but artistic constructs which can repeatedly change roles in
the course of a text according to the authors informational needs at each
juncture. The one thing everybody agrees on is that only a personalized or
individuated narrating instance with a clear I-here-now Ich-origo, self
reference, subjective semantics etc. can function as a potential focalizer.
Beyond this I think it is better to distinguish situational varieties rather
than jump to universal claims whether or not narrators can function as
focalizers. There are three varieties I can think of right now:
The first and most obvious case, ignored by most narratologists, is a
narrator, who is also observer or agent in the narrated sphere, reporting on
events and situations taking place in the narrated sphere simultaneously
with his act of narration. In this case, person, time and place of narrator
and narrative agent are clearly the same, and the individual cannot but
report events and entities, including himself, as he observes and experiences them at the moment of narration. Such a narrator fulfils two distinct
functions, saying and seeing, and must function as focalizer, focusing on
the setting, other agents, or himself qua agent, since focalization is his
only way to acquire any knowledge of the world around him as it unfolds.
An individuated narrator who is currently reporting on earlier events
or situations in the narrated domain in which he acted as observer or agent
is the standard case. Obviously, such narration involves current acts of recall whose content are earlier acts of witnessing or experiencing. As agent
or observer of the events as they occurred, this narrator qua story world
participant was clearly able to focalize. So our problem concerns not this,
but rather the status of his current acts of recollecting and reporting on his
own past acts of focalization. Are they too acts of focalization? I think the
answer depends on the kind of current mental activity. Recall can be a
distanced analytic retrospective summary I saw X, I experienced Y,
which is not focalization since it lacks the immediacy and experientiality
essential to focalization. But recall may also be more like an attempt to relive or re-experience the original act of focalization or sensory experience
and its resultant take, effecting a mental shift of deictic center. A clear in-
53
dicator of this kind of recall is the switch from past to present tense. In
Prousts famous madelaine scene, the narrator starts by one day in
winter, on my return home. He then describes dipping the cake in the tea
and sipping the tea: I raised to my lips a spoonful of the teaI drink a
second mouthful (Proust 1981: 48). Following Edminston (cf. 1989:
73942) one could say that the narrator now adopts the intradiegetic
vision of himself then, presenting his own mental activity and view of
others (and himself) at the moment of the event. The narrator restricts
himself accordingly to the experiencing self then with its deictic center, in
a word, not doing any further retrospective information processing. (On
this point see also Shen [2003].) I feel very strongly that it would be quite
sensible and useful to include this kind of recollection under focalization.
It goes without saying, though, that all acts of recollection of any kind are
fallible, since memory is an active faculty, not a passive slate.
Chatman, the great enemy of narrator as focalizer, remarks that a
narrator could look at events and existents in the discourse world or space
of narration he occupies, to the extent that this world is fleshed out (cf.
1990: 14344). The same observation has been made by James Phelan
(2001), a friend of narrator as focalizer. This rare agreement opens up a
third area of narrator as focalizer. Any individuated narrator, whether or
not he is a participant in the narrated domain, can always be considered a
focalizer when his object of attention is his current situation as narrator,
his activity of telling and so on. The specification and emphasis on the
narratorial sphere at the expense of the narrated, on the narrators immediate context and his writerly activities, has a long history going back at
least to Cervantes (cf. Alter 1975), and is a hallmark of postmodernism.
As Brian Richardson points out in his paper in the present volume, even
the anonymous teller at the beginning of Conrads Heart of Darkness perceives immediate sights and sounds, including the voice of Marlow, on
the boat on the Thames, and can therefore be considered a limit case of
focalizer.
5 What, or Typologies of Focalization

The traditional external vs. internal focalization is valuable, but the terms
are polysemous, designating three different binary oppositions: the identity of the focalizer and his location in the narrated and/or narrating system; the nature of the focalized, that is, public sense data or internal mental ones; and the possibility of access to the minds of others. Another
54
Uri Margolin
well-known typology deals with levels of focalization and its possible

hierarchical and embedded nature (cf. Bal 1997). More recent is Jahns
(1999) interesting scale of focalization, starting from strict individual
perception of an object, followed by ambient one, where two focalizers
observe simultaneously the same object from different points. The next
stage is weak focalization where there is still an object but no determinate point or center of awareness, and finally no object and no point of
observation.
Broader Horizons: Beyond the Limits of Space and Time
I would like to conclude by discussing briefly several additional cases
going beyond the singular focalizer and act of focalization, involving second order focalizations, and including embedded, transferred, and joint
ones. I think they can be systematized in terms of original and transferred
systems of personal coordinates, consisting of person, time and space.
The truth table method indicates eight possible combinations of these
three factors, ranging from I-here-now to not-I, not-here, and not-now.
The I-here-now is of course the prototypical or standard kind of focalization. Proceeding to I not-now and optionally also not-here we get
the case of recollection of past acts of focalization we spoke of before.
The I can also juxtapose here several past acts of focalization concerning the same object, as Prousts narrator does towards the end of A la
recherche, where he juxtaposes in an extended act of recollection several
mental snapshots of the princess de Guermante dating from different
times in his life. This is by definition embedded or second order focalization. How about I+now but not here? The focalizer in this case tries
to place himself imaginatively in a situation currently occurring elsewhere
and project what he would see, hear etc. if he were there himself now.
This is not far-fetched at all. Every Canadian who, in the dead of winter,
spoke on the phone to a friend basking in the Caribbean sun has experienced this kind of vicarious hypothetical focalization.
The not-I cases are even more interesting. In all of them, the
focalizer tries to bracket out his own mental makeup, and possibly time
space location as well, put himself in the spatial, temporal and perceptual
position of another mind, and simulate how this other mind would process
situational information and what the resultant take would be like. We
usually refer to this mental exercise as taking the point of view of another or putting oneself in someone elses position or shoes or seeing
55
things from someone elses perspective, sometimes quite literally. Characters can do it to one another, and of course a personalized narrator can
do it with respect to one or more characters. The act of simulation is most
poignant when one tries to simulate the others take on the simulator himself (what do I look like to her?). In this case the simulator literally looks
at himself as if he were another. I suggest we call all the foregoing varieties transferred focalization.
Brian Richardson in this volume and myself some years ago have
drawn attention to the possibility of trans-individual focalization. This too
could be viewed as a chain whose first link is provided by Jahns notion
of ambient focalization. Jahn describes ambient focalization as a case
where spatial deictics are relaxed and the vision is mobile, hence beyond
that of a single individual. One variety would be where the narrators
words convey the simultaneous takes of several individuals on the same
object. He cites an incident from Virginia Woolfs To the Lighthouse
where two focalizers, James and Cam, sitting at two opposite ends of a
boat are simultaneously observing the people seated in the center of the
boat. The singular textual representation of these people consequently involves two points of view from which they are observed concurrently.
Note that the individual ingredients of any simultaneous composite vision
or take may be overlapping or complementary, but they may also fail to
coincide, creating a discordant focalization, like a Cubist painting.
In Jahns example the individual acts of focalization were concurrent
but not coordinated. But cases also exist which tend towards a coordinated vision. One is that of focalizers negotiating a joint consensual
take, a clear example of the social mind in action. Two or more people
can seek to formulate a jointly held vision of a person, situation, object or
event either by comparing and adjusting their individual ones to yield one
unified homogeneous picture, or by each contributing a piece of the puzzle, a partial vision needing to be complemented by all others in a resultant composite picture. Either process can be seen when witnesses to the
same car accident for example discuss their visual and auditory impressions of it. In the case of the unified picture one starts with a plurality of
individual visions in dialogue, interacting and intersecting and, if a joint
one is attained, one ends with a plural we discourse conveying it, such
as we saw or we felt etc. This is the last variety, that of the uniform
communal vision where the individual experiencers feel themselves as
one and speak in the collective we (or even I in some choruses of
56
Uri Margolin
Greek tragedy) about their collective vision, a subject developed in detail

in Brian Richardsons paper.
References
Alter, Robert (1975). Partial Magic. Berkeley: U of California P.
Bal, Mieke (1993). First Person, Second Person, Same Person: Narrative as Epistemology. New Literary History 24, 293320.
Bal, Mieke (21997). Narratology. Toronto: U of Toronto P.
Banfield, Ann (1987). Describing the Unobserved: Events Grouped around an Empty
Centre. N. Fabb et al. (eds). The Linguistics of Writing. New York: Methuen,
26585.
Brinton, Laurel (1980). Represented Perception: A Study in Narrative Style. Poetics 9,
36381.
Chatman, Seymour (1986). Characters and Narrators: Filter, Center, Slant and InterestFocus. Poetics Today 9, 189204.
(1990). Coming to Terms. Ithaca: Cornell UP.
Cruse, Alan D. (2004). Cognitive Linguistics. West Nyack, NY: Cambridge UP.
Edminston, William (1989). Focalization and the First Person Narrator: A Revision of
the Theory. Poetics Today 10, 72944.
Fludernik, Monika (1996). Towards a Natural Narratology. London: Routledge.
Herman, David (1994). Hypothetical Focalization. Narrative 2, 23053.
Jahn, Manfred (1999). More Aspects of Focalisation: Refinements and Applications.
GRAAT 21, 85110.
Nnning, Ansgar (2001). On the Perspective Structure of Narrative Texts. W. van
Peer & S. Chatman (eds). New Perspectives on Narrative Perspective. Albany:
SUNY Press, 20723.
Nnning, Vera & Nnning, Ansgar (eds) (2000). Multiperspektivisches Erzhlen. Trier:
VWT.
Ouellet, Pierre (1996). The Perception of Fictional Worlds. C. A. Mihailescu &
W. Hamarneh (eds). Fiction Updated. Toronto: U of Toronto P, 7690.
Phelan, James (2001). Why Narrators can be FocalizersAnd why it Matters. W. van
SUNY Press, 5164.
Pomerantz, James R. (2003). Perception. L. Nadel (ed). Encyclopedia of Cognitive
Science, Vol. 3. London: Nature Publishing Group, 52737.
Prince, Gerald (2001). Point of View on Points of View or Refocusing Focalization.
W. van Peer & S. Chatman (eds). New Perspectives on Narrative Perspective.
Albany: SUNY Press, 4350.
Proust, Marcel (1981). Remembrance of Things Past, Vol. 1. Tr. C. K. Scott Moncrieff &
T. Kilmartin. New York: Random House.
Ronen, Ruth (1994). Possible Worlds in Literary Theory. Cambridge: Cambridge UP.
Ryan, Marie-Laure (1991). Possible Worlds, Artificial Intelligence, and Narrative Theory.
Bloomington: Indiana UP.
57
Shen, Dan (2003). Difference behind Similarity: Focalization in Third Person Center-OfConsciousness and First Person Retrospective Narration. C. Jacobs & H. Sussman
(eds). Acts of Narrative. Stanford: Stanford UP, 8192.
Schubert, Christoph (2005). Fallible Focalization. GRM 55, 20526.
Surkamp, Carola (2003). Die Perspektivenstruktur narrativer Texte. Trier: VWT.
Talmy, Leonard (22006). Cognitive Linguistics. Encyclopedia of Language and Linguistics, Vol. 2. Amsterdam: Elsevier, 54246.
TATJANA JESCH, MALTE STEIN

(Hamburg)
Perspectivization and Focalization:

Two ConceptsOne Meaning?
An Attempt at Conceptual Differentiation
The study of focalizations has caused much ink to flow in the estimation of the scholar who coined this narratological term a little too
much (Genette 1988: 65). To begin with, we would like to justify our
desire to invest a bit more ink in the discussion of focalization by stating
our initial question about the topic: on what grounds, we asked ourselves,
did the author of Narrative Discourse title one chapter of his Narrative
Discourse Revisited perspective and the other focalizations, especially given the fact that he had originally introduced the term focalization
in order to avoid terminology like vision or point of viewall terms
with too specifically visual connotations (Genette 1980: 189). As
Genette explains again in Narrative Discourse Revisited, his remarks on
focalization should not be seen as anything more than a reformulation
(Genette 1988: 65) of the classic descriptions of perspective. Butas an
avowed follower of Occamwhy had he not abandoned the term perspective altogether?!
In the following, we will demonstrate (section 1) that Genettes concept of focalization is actually an amalgamation of two wholly independent elements for whichas the author himself might have anticipated
one actually needs two terms. The first element is the perception of the
world invented by the author through narrators and other agents also invented by the author; the second element is the regulation of narrative information within the communication between author and reader. In the
latter, in our assessment, lies the innovative potential of the discourse
about focalization. This potential, however, is overlooked by narratologists, who have wholeheartedly adopted the new term as a mere substitute for the older one (section 2). We would like to suggest that the terms
60
Tatjana Jesch, Malte Stein
focalization and perspectivein clear differentiation from one another

be retained (section 3) and, using the example of Kafkas The Metamorphosis, we will attempt to demonstrate the relevance of this terminological differentiation to textual analysis (section 4).
1 Focalization in Genette: One ConceptTwo Meanings

Genette first introduced the term focalization in his Narrative Discourse
as a replacement for the term point of view: To avoid the too specifically visual connotations of the terms vision, field, and point of view, I
will take up here the slightly more abstract term focalization (Genette
1980: 189). As explained here, the goal of replacing these terms was to
make clear, on a terminological level, that an analysis of perspectival
structures should not be limited to the visual:
There would have been no point in taking great pains to replace point of view with focalization if I was only going to fall right back into the same old rut [the narrowing of
meaning to visual perception, T. J. & M. S.]; so obviously we must replace who sees?
with the broader question of who perceives? (Genette 1988: 64)
Genette criticizes the terms vision, field and point of view for implying the visual too strongly. Thus, he suggests the question who perceives? clearly as a means to include not only optical perception, but also the constructive, meaningful awareness that a character or a narrator
gains as a result of his or her capacities of knowledge (Genette 1980:
162).
Later, however, in his Narrative Discourse Revisited, Genette gave
another definition of the term focalization, this time based on wholly different criteria. In place of the question who perceives? one finds the
implicit question what can the reader know?:
So by focalization I certainly mean a restriction of fieldactually, that is, a selection of narrative information with respect to what was traditionally called omniscience. In pure fiction that term is, literally, absurd (the author has nothing to know,
since he invents everything), and we would be better off replacing it with completeness of informationwhich, when supplied to a reader, makes him omniscient. The
instrument of this possible selection is a situated focus, a sort of information-conveying pipe that allows passage only of information that is authorized by the situation.
(Genette 1988: 74)
With this definition, Genette no longer stresses the mode of perception of

a fictive entity, but rather the transfer of information between author and
reader. At the same time, he presupposes that the choice of a limited nar-
Perspectivization and Focalization: Two ConceptsOne Meaning?
61
ratorial or figural perspectivea restriction of fieldwill always

bring about an equivalent limitation or selection of narrative information available to the reader.
Although Genette continues to claim in the same volume that narratology has no need to go beyond the narrative situation (137), his (new)
conception of focalization requires him to return to the authors communication with the reader. He had already done this in his Narrative
Discourse, when he mentioned in the passage on alterations:
[...] the case of a novel [...], where the narrator, a trusting husband, is present at scenes
between his wife and a male friend that he recounts without thinking anything amiss
but whose meaning cannot escape the least subtle reader. This excess of implicit information over explicit information is the basis of the whole play of what Barthes
calls indices. (Genette 1980: 19798)
As we see here, Genettes narratological practice leads him to presuppose

an author (as a communicative entity) whose messages to the reader are to
be drawn out of the text (as an excess of implicit information) even
when they contradict the point of view of the fictional perceiving subject:
here it is as if the narrator [...] did not understand what he relates; this in
no way prevents the reader from interpreting it in conformity with the
authors intentions (198)1. In examples such as this one, it becomes evident that one must clearly differentiate between perspective (understood
as focus of perception [Genette 1988: 64]) and focalization (understood
as selection of narrative information [74] available to the reader). For it
is always possible (and in literature no rarity) that the fictional perceiving
subject does not see (realize, comprehend, understand, etc.) something of
which the reader is made fully aware. Contrary to what even Genette at
times presumessee the definition of focalization in Narrative Discourse
Revisiteda limited narratorial or figural perspective in no way rules out
the possibility for the author to communicate the completeness of
information (74) with the help of interspersed indices. The reader can
know more about the actual world (cf. Ryan 1991) than the fictional
perceiving subject, which leads to the conclusion that perspective and
focalization are independent of one another.
In a footnote to his Narrative Discourse Revisited, Genette refers to
Franois Josts works on the difference between focalization and ocularization (information and perception), which he judges to be the most
relevant contribution to the debate on focalization and to the necessary
1
See also Genette (1988: passim).
62
refining of that notion (Genette 1988: 7374). Regardless of this, it has

become commonplace in narratological discussions simply to substitute
the term perspective for focalization and, despite the shift in terminology, to continue thinking within the categories of perspective and
point of view (Niederhoff 2001: 16)2.
2 The Prevailing Reception of the Term Focalization

Instead of giving an overview here of the numerous positions within the
discussion of focalization, we choose two works out of the abundance of
publications on the topic. Of these two, Rimmon-Kenan (1983) has already achieved the status of a seminal work in the field and can thus serve
as representative of the prevailing use of the term. The second work, on
the other hand, is a more contemporary plea for a terminological differentiation between focalization and perspective.
Rimmon-Kenan begins her explanation of focalization with the demarcation, established by Genette, of the acts of perceiving and speaking. The
perspectival position underlying a narrative speech act (and influenced by
the perceivers knowledge and values) need not necessarily represent the
position of the narrator: speaking and seeing, narration and focalization,
may, but need not, be attributed to the same agent (72). However, in the
simple two-part breakdown of types of focalization that she subsequently
proposes, she departs from Genettes formulations. The (implied)
underlying criterion of Genettes typology of focalizations, at least in
Narrative Discourse Revisited, is the differentiation between a nonrestricted and a restricted regulation of information vis--vis the reader:
Information for the reader
non-restricted
zero focalization
through the perspective
of a narrator
restricted
through the perspective

of an agent
Figure 1: Types of Focalization According to Genette (1988)

2
Tr. of quotations from Niederhoff (2001): Tracy N. Graves & Katherine McNeill.
63
Rimmon-Kenan, on the other hand, consistently uses the term focalization

as a synonym for perception (i.e. perspective) and divides the types of
focalization according to the position relative to the story (74) that the
perceiving subject (center of consciousness) holds. Her opposition of
external versus internal is based upon this criterion, which she thenin
the course of differentiating perspective into perceptual, psychological and ideological facetsconnects with further oppositional
features:
Center of consciousness
(focalizer)
outside the represented events:

=
unrestricted knowledge
emotionally neutral
ideologically higher position
inside the represented events:

=
restricted knowledge
emotionally not neutral
ideologically lower position
Figure 2: Types of Focalization According to Rimmon-Kenan (1983)
It becomes clear that the limitation of information plays no role in her typology when Rimmon-Kenan writes about the external focalizer (or narrator-focalizer) that he knows in principle everything about the represented world (79)3. He can, however, limit his knowledgeshe
3
The postulate that an external focalizer (or narrator-focalizer) is in principle omniscient (emotionally neutral and ideologically superior) in relation to the represented
world may be applicable to many narrative texts, but cannot be maintained in theoretical terms. Omniscience in respect to the represented world can only be assumed in
fictional narrative for those who invented or composed this world. This term (as Genette [1988] makes clear) thus applies only to the author, who determines with what
perceptual, psychological, and ideological characteristics he provides the (also
invented) narrating entity. Just like any character in the story, a narrator can also serve
the author as a goulot dinformation, an information-conveying pipe (Genette 1988:
74). Rimmon-Kenan probably does not consider this possibility because she virtually
equates the narrator with the author and as such analyzes narrative fiction on the
basis of a communicative model that only encompasses the (fictional) communication
between character and narrator, but ignores that of the author with the reader.
64
probably means the communication of his knowledgeout of rhetorical

considerations (79). In Genettes model, this would be the transition
from zero focalization to external focalization. When Rimmon-Kenan
speaks of external focalization, she refers to both the restricted and nonrestricted communication of knowledge as long as the narrator remains
the center of consciousness (the focalizer) and his omniscience can be
assumed. Instead of emphasizing, as Genette (1983) does, that the possible knowledge of the reader is restricted through focalization, she merely returns to the (hypothetical) state of knowledge of the fictional perceiving subject. Therefore, her understanding of focalization remains
within the framework of the conventional description of perspective.
Burkhard Niederhoff recently expressed a legitimate criticism of such
a use of the term focalization, in which he objects that the advocates of
focalization often think in the categories of perspective (Niederhoff
2001: 1) and that they can only justify the change in terms with unconvincing arguments (5)4. He holds the view, similar to ours, that Genettes
neologism includes aspects of narratorial communication, which are
not covered by the related terms of perspective or point of view (5).
Thus, Niederhoff also sees a conclusive decision in favor of one of the
two terms as neither necessary nor desirable (12).
Based on definitions from the Oxford English Dictionary, what Niederhoff then seeks to establish are similarities between the two terms. According to him, both terms imply a limitation of the power of the narrator
to communicate information (7). In addition, their most important
similarity consists in that they describe a subjectification and restriction
of perspective (9). As has become apparent in the above statements,
Niederhoff, too, has a tendency to infer from the boundaries of perception
and knowledge of the fictive entities (narrator or character) an equivalent
restriction of the information communicated to the reader. From the outset, he does not factor in the possibility that an author uses clues to reveal
4
There are above all two reasons that critics, with an appeal to Genette, have repeatedly
introduced and occasionally played against each other, although neither of the arguments has any substance. The first argument reads that the term focalization is more
appropriate than perspective to differentiate between voice and perception in narratological analysisan argument that Genette did not formulate even once in his Narrative Discourse. The other argument claims that the use of the term focalization
makes it clearer that the question who perceives? does not refer solely to the visual
component of perception. As the term chosen by Genette comes from the field of
optics, this argument is equally unjustifiable.
65
more to the reader than the narrator or agents (can) perceive and know
in other words, the possibility of an excess of implicit information over
explicit information (Genette 1980: 198). Just like those whom he reproaches for having an understanding of focalization that is too perspectival he links Genettes term above all back to (figural and narratorial)
perception. The only difference between the two terms consists for him
in the element that creates the restriction of perception (Niederhoff
2001: 9). While with perspective this is the spatial position, the restriction of perception with focalization is a result of the choice of a particular kind of reality (9). He writes: If one places a camera at a particular
point, the perspective becomes fixed, but not the focus. One can focus on
the flowers in the foreground or the rock face in the background (9). We
will not consider here the feasibility of applying such a differentiation
taken from photo-optics to the field of narratology, nor the question as to
how much insight one actually gains from such a comparison. In any case,
it remains to be said that behind both of the aspects of analysis named by
Niederhoff lies, once again, only the question of the perceptual horizon
and scope of knowledge of the narrator and/or agent. However, the
question of the authors management of information is wholly disregarded.
3 Recommendations for a Conceptual Clarification

Taking up Genettes definition from Narrative Discourse Revisited, we
recommend that focalization be defined as the authors temporary or definitive withholding of information from the reader. Under the term perspectivization5, on the other hand, we understand the representation of
something from the subjective view of a fictive entity (narrator or character). At the same time, this mode of viewing is always a part of that
which the author depicts for the reader. A connection between focalization and perspectivization can exist to the extent that perspectivization
often serves to account for a restriction of information within the fictional
world. As such, perspectivization can become a way to achieve focalization. Once again, however, perspectivization is not focalization, for a
5
We are consciously not using the term perspective here, but rather perspectivization, in
order to indicate that the structure found in this casejust as with focalizationcan be
traced back to a constructive activity on the part of the author.
66
text can contain (implicit) information that transcends the figural and/or
narratorial capacities of knowledge.
In order to determine whether the communicated information in a particular passage of the text is complete (as regards the events that have
been narrated up to this point), one must have a standard for measuring
the completeness of information. One must specify under what conditions
one can say that the author enables the readers omniscience. This is the
case, according to our assessment, when the reader is placed in a position
in which (up to a certain point of the narrative discourse) he can:
(a) order the depicted incidents chronologically and spatially (coherence
level I),
(b) recognize the said incidents as to be expected within the represented
world (i.e., according to a stereotype or schema)6 or as eventful (i.e., diverging from the stereotype)7 (coherence level II),
(c) comprehend the incidents in their causal, final and consecutive relations (coherence level III).
Coherence
Level
I
Type of relation
Activities of the reader
Temporal and spatial

first x, then y, then z;
x here, y there, z there
Ordering of the presented incidents

according to their chronological sequence in the represented world.
Mapping of the presented incidents
onto spaces.
Ordering and expansion of the presented incidents according to intra-,
inter- and extratextually established
schemata.
Comprehension of the presented incidents with regard to causes, motivations and consequences.
II
Correlative
If x, then also y
III
Causal/ultimate/consecutive
y because of x
y in order to x
x so that y
Table 1: Levels of Coherence in the Reconstruction of Narrated Stories

6
Compare to the procedure of sequence formation as Barthes has described it in several

works (collected in Barthes [1985]).
Such breaks in schema on the level of the histoire are repeatedly described as a
structural characteristic essential for the narratability of events. Barthes speaks in this
context of narrative transgression (1985), Quasthoff of plot disruption (1980), Lotman (1977), Renner (1983) and Schmid (1992) of an event and Herman (2002) of
non-stereotypic actions and events.
67
We assume in traditional narratological fashion that a narrative is the representation of a story comprising at least one action. As Bremond has already demonstrated, every plot is sequentially structured insofar as it is
made up of multiple phases. The reconstruction of a plot can thus take
place on the basis of a universal sequencing schema, such as Bremonds
elementary sequence, which is composed of a possibility (ventualit),
a process of actualization (passage lacte), and an outcome (achievement) (Bremond 1973: 131). We offer another schema here in its place,
out of which one more clearly sees:
(a) that the so-called possibility is based upon a subjective perception of
situation and a subsequent (conscious or unconscious) formation of intention,
(b) that the process of actualization can have effects that the agent could
not foresee.
Action
Cause
Stimulus
Intention
Treatment
Actualization
Primary
Effects
Result
Secondary
Effects
Figure 3: Sequencing Schema Action
Complete knowledge with reference to an action is achieved as soon as

the reader comes to know its cause, its underlying intention(s), its actualization and its results. If the corresponding information (in relation to the
events presented up to this point) is not communicated to the reader, he is
dealing with focalization.
It is also necessary to further specify the mode of the communication
of knowledge: when can a piece of information be seen as communicated
68
or withheld by the author? The explicitness of a piece of information cannot be the criterion of differentiation in this regard, for even a text whose
semantics remain closely attached to a conventionally and invariably
fixed literality contains implicit information. In the domain of the Implicit I, which is connected to the wording and in which the recipient
moves without any lack of information, focalization can therefore not
exist (cf. Linke & Nussbaumer 2001: 437). The Implicit II, on the other
hand, is understood as non-literal and variable because it does not belong to the semantics of the wording and demands pragmatic inferences
on the part of the reader. Even so, it also conveys deducible propositions
to an adequately competent reader. Focalization does not occur in this
case either.
One can only speak of withheld information, then, in the case of a proposition:
(a) that the reader needs in order to reconstruct the occurrence in the
actual world (cf. Ryan 1991) according to coherence level III (see
table 1),
(b) that is not explicitly given where the reader would need it,
(c) that the reader cannot discover by combining given propositions with
relevant cultural knowledge8.
In the practice of narrative analysis, it is of course not necessary to
make explicit every element of the sequencing schema action for each
action that is mentioned in the text. However, the majority of actions (in
fictional narrative worlds as well as in everyday life) are generally selfexplanatory, as understanding a plot means being able to explain at any
moment its individual elements. Thus, ones attention should be drawn
above all to those elements of the sequence whose explicationas a result of focalizationseems difficult or disputable.
Whether cultural knowledge that is activated in the reception process holds factual relevance can be determined through its power of integration or its functionality. We can
apply here, for example, the rules formulated by Titzmann, which determine when a
potentially relevant piece of knowledge can be considered functionalized, and thus factually relevant in the text (cf. Titzmann 1977: 360). This is the case when a conclusion
which can be reached through a textual proposition with the help of this piece of
knowledge (a) is itself a textual proposition, or (b) is in its turn functionalized as
the implicit condition of another proposition.
69
4 Perspectivization and Focalization in Kafkas The Metamorphosis

If one conceives of perspectivization and focalization as we have suggested, then theoretically, there are four possible alternatives in fictional narration:
focalization through perspectivization,
perspectivization without focalization,
focalization without perspectivization,
neither focalization nor perspectivization.
To give an especially concise example of each of these possibilities would
have required referring to several texts in what follows. However, we
have decided to reference only one text in order to make it clear that
conclusions about focalizationconclusions about what the reader (oneself) can know about a depicted worldnecessitate a larger hermeneutic
intensity than the mere description of which fictive entity perceives whom
or what, from which position and by what means.
4.1 Focalization through Perspectivization
In Kafkas story The Metamorphosis (1915), the events are told primarily
from the perspective of the protagonist Gregor Samsa. This is apparent
from the outset in the strong limitations of the field of vision presented to
the reader (when Gregor is confined to his room, the accounts and descriptions of the other family members are restricted to acoustic impressions or to the speculations of the protagonist about their actions). Gregors perspective is also apparent in the absence of direct insight into the
mental processes of other characters: assertions about the unspoken
thoughts, feelings or intentions of the characters other than Gregor are
rendered only in his subjective interpretation of them. For the most part,
these interpretations turn out to be unfounded, a fact that the recipient will
sometimes only recognize retrospectively. In such cases, the perspectivization becomes associated with a temporary focalization.
After Gregor Samsas transformation, his parents and sister complain
above all that since they could hit on no way of moving Gregor, they
could not give up this apartment, which was much too large for their
present circumstances (172). The family members express an intention
they want to change apartmentsand complain at the same time about a
circumstance that would make the realization of that intention impossible.
70
However, the obstacle is invalidated by the depiction of incidents from

Gregors perspective:
Gregor, however, realized it was not just their consideration for him that held them
back, for they could have easily transported him in a suitable crate with a couple of air
holes in it. The main obstacle to the familys relocation was their utter despair and
their sense of being struck by a misfortune like no one else among their friends and
relatives. (172)
The protagonist recognizes that a problem with his transportation cannot

be what is hindering the Samsas in their move. The real reason that the
family stays in their old apartment, as he sees it, is the utter despair into
which they have fallen because they have had to take care of themselves
since he ceased to be the breadwinner. For this reason, he infers, they
have reached the limits of their strength (172). Towards the end of the
story, when Gregor dies, there is a change in the perspectivization to the
narrators perspective9, which allows for selective insight into the
thoughts of the figures (18792). With this shift, a completely contrary
picture emerges: as it turns out, Gregors surviving family members look
into their future full of confidence, as their jobs were all exceedingly
advantageous and also promising (192). After their former provider has
passed away, they make hopeful plans including the rapid realization of
the move, which they had put off until then:
Naturally, the greatest immediate improvement in their situation could easily be
brought about by their moving; they hoped to rent a smaller and cheaper apartment,
but with a better location and altogether more practical than their current place, which
had been found by Gregor. (192)
Only retrospectively can the reader discern that it is not perhaps utter
despair that hindered Gregors parents and Grete in their move. It seems,
rather, that it was out of consideration for their physically changed son
and brother, who they did not want to take with them to their new dwelling. Hence Gregors death appears as the liberation of his family members: Then all three of them left the apartment together, which they had
not done in months, and took the trolley out to the countryside beyond the
town (19192). At this moment, specifically the daughter of the family
9
A distinct indicator for this change in perspective is the change in the way characters
are referred to. The characters belonging to the family are no longer called father,
mother or sister, but rather simply Mr. and Mrs. Samsa, daughter and Grete
(cf. Kafka 1915: passim).
71
feels better than ever. Since her brothers transformation Grete has
changed gradually as wellshe had blossomed into a lovely, shapely
girl (192).
4.2 Perspectivization without Focalization
During his life Gregor shows a tendency to be possessive of his sister. Although it is not apparent to him, it is made recognizable to the reader
through the inclusion of certain details in the story:
He was determined to creep all the way over to the sister, tug at her skirt to suggest
that she take her violin and come into his room, for no one here would reward her
playing as he intended to reward it. He wanted to keep her there and never let her out,
at least not in his lifetime. For once, his terrifying shape would be useful to him; he
would be at all the doors of his room simultaneously, hissing at the attackers. His sister, however, should remain with him not by force, but of her own free will. (180;
emphasis added)
In the passage above, the figural perspectivization does not lead to focalization but, on the contrary, serves to inform the reader about the protagonists tendency toward denial. One notices instantly the discrepancy between Gregors desire to have his sister stay with him of her own free will
and the prescriptive nuance resonant in the modal verb should, which
becomes even stronger through his own deterrent behavior. Such
protective behavior seems unwarranted especially given the fact that the
text gives no indication of any outside aggression directed towards Grete.
Instead the reader receives clear information that Gregor, who wanted to
keep her there [in his room] and never let her out, longs for intimate
togetherness:
She should sit next to him on the settee, leaning down to him and listening to him
confide that he had been intent on sending her to the conservatory []. Gregor would
lift himself all the way up to her shoulder and kiss her throat, which she had been
keeping free of any ribbon or collar since she had first started working. (18081)
4.3 Focalization without Perspectivization

As is generally known, Kafkas story begins with a discovery that is surprising to both the reader and the protagonist alike: One morning, upon
awakening from agitated dreams, Gregor Samsa found himself, in his bed,
transformed into a monstrous vermin (119). The fact that a human being,
identified as such by virtue of his name and by mention of his
72
sleeping habits, has had his body transformed overnight into that of an
animal cannot simply be put down to the (extratextually anchored) world
view of Genettes potential reader (Genette 1988: 138).
Given the change in condition of the protagonist, it is a matter, rather, of a
break with expectations or, as Jurij Lotman would suggest, a matter of an
event in the emphatic sense (see also Renner [1983]). Every event of
this kind raises the question of its cause, for which, in this instance, the
reader does not find any information in the introductory passage cited
above. Readers of modern narratives are confronted again and again with
similar deficiencies of information at the beginning of texts. Often, as
with detective stories, readers are not introduced to the depicted world of
such stories step-by-step, but rather they are pushed abruptly into them.
They must negotiate their own way in these worlds, even if the necessary
information to do so is not provided.
If, as in the beginning of the story, such a case of focalization exists, it
is not necessarily motivated by the limited viewpoint of the protagonist. It
is not due to Gregors limited state of consciousness that the circumstances of his transformation at the beginning of the story are completely
unknown. In fact, only later in the text does the protagonist express his
perplexity about his changed condition: Whats happened to me? Gregor wonders (119), much like the other characters subsequently do. On
the level of the authors communication, this general wonderment is an
indication of the fact that the heros metamorphosis is a matter of an exceptional incident in the depicted world, one that, as such, demands an
explanation. And this explanation remains unrevealed to and unexpressed
by the fictive entities at the texts culminationa fact that says nothing
about the conclusions the reader might be able to draw in the end. To assume (as the debate over focalization does to a large extent) that the
readers state of awareness is limited by the awareness of the characters
and of the heterodiegetic narrator would mean to insinuate that the author
has transgressed the conversational maxim of adequate information (cf.
Grice 1975). Because this presumption can only be justifiedif at all
after reading the story in its entirety, readers will initially assume that the
missing information will be given gradually.
73
4.4 Neither Focalization nor Perspectivization

Immediately after Gregor has examined his changed body, his gaze turns
to an object above:
He lay on his hard, armorlike back, and when lifting his head slightly, he could view
his brown, vaulted belly, partitioned by arching ridges, while on top of it, the blanket,
about to slide off altogether, could barely hold. His many legs, wretchedly thin compared with his overall girth, danced helplessly before his eyes []. Above the table,
[] hung the picture that he had recently clipped from an illustrated magazine and inserted in a pretty gilt frame. The picture showed a lady sitting there upright, bedizened
in a fur hat and fur boa, with her entire forearm vanishing inside a heavy fur muff that
she held out toward the viewer. (119)
This change in the figural line of visionfrom his own covered body to
the body of the woman in the pictureis communicated by the narrator
only belatedly and, even then, only incidentally. Nevertheless, the turn of
the gaze is a (figural) action and the attentive reader will ask why Gregorunder these exceptional circumstanceslooks at the portrait of the
woman. On the level of literality10 and denotation11, the author does not
impart any information at all about the cause and intention of Gregors
eye movement. However, the omission of this information is neither perspectivally motivatedin other cases the narrator renders Gregors
thoughtsnor is it a sign of focalization. For, at the moment that the
reader becomes aware of the progression of Gregors gaze, there are also
indications of its motivation: through interpretative inference it is already
possible to explain the movement of Gregors gaze at the moment of its
narration.
If one compares the woman dressed in fur with Gregors insect body,
both figures turn out to be creatures belonging to the isotopy animal.
The vaulted belly (gewlbter Bauch) of the metamorphosed Gregor,
in turn, can be considered compatible with the isotopy feminine, which
is also to be attributed to the woman. Whereas the arching ridges (bogenfrmige Versteifungen) that he notices in his abdomen after waking
up conform to the opposite isotopy masculine.
10
11
See Linke & Nussbaumer (2001: 43637).

See Barthes (1964).
74
Taking into account cultural knowledge contemporary to the text12 allows for the hypothesis that, underneath the womans fur, which reminds
one of a growth of pubic hair covering the whole body, a kind of phallus
is hidden in the form of her entire forearm. This allows the isotopy
masculine to be attributed to the woman in the picture as well. Like
Gregor, the woman, therefore, also combines both of the classemes masculine and feminine. In this lies a noticeable commonality between the
two figuresin addition to their shared animal traits.
Moreover, the illustration of the beauty in fur hanging on the wall of
Gregors bedroom also brings to mind Sacher-Masochs novel Venus in
Furs (1869), in which the Gregor (!) of that text, alias Severin, also has
two pictures of Venus-figures in fur hanging on his walls: an original of a
dominatrix armed with a whip and a copy of Titians Venus, whose fur, as
the reader will learn, has become a symbol of feminine tyranny and
cruelty (Sacher-Masoch 1869: 910). Sacher-Masochs hero Gregor13 is
seduced by the Venus in furs, who with her erotic, over-encoded furfetish and the dominance she exercises over her lover, seems to be a
phallic female figure. She is symbolic of feminine dominance in a relationship based on subordination and satisfies the desires of the man in her
abuse of him.
As far as the motivation of the eye movement of Kafkas Gregor is
concerned, two hypotheses can, consequently, be formulated. First, it is
Gregors intention, in looking at the female form in the picture, to compare his strange new body to an ideal figure framed in gold (at which
point it is confirmed that there are certain similarities between himself and
the lady). Second, in depicting this comparative gaze (and intertextually
alluding to Sacher-Masochs Venus in Furs14), the author insinuates that
12
13
14
See Freuds thesis that the sexual fetish is an imaginary phallus (1905; 1910). Next to
feet and undergarments, fur, according to Freud, belongs to the most often chosen of
fetish-objects, which no doubt [] owes its origins to an association with the hair of
the mons Veneris (Freud 1905: 155).
Kafkas Gregor owes his first name to this figure; the last name Samsa could be
close to an anagram of the first two letters and first three letters, respectively, of the
authors name, Sacher-Masoch.
This reference will be continued throughout the text, e.g. when one of the moralizing
critiques issued by Sacher-Masochs VenusAnd if any of you ever has had the courage to kiss my red lips, he then goes on a pilgrimage to Rome, barefoot and in a
penitents shirt [...] (5)is transferred to Kafkas text. In order to alleviate the burning pains in his abdomen (Kafka 1915: 132), Kafkas Gregor crawls quickly up to
the picture of the woman clad in nothing but furs and presses his body against the
75
Gregors transformation is connected to a crisis in his psychosexual development. The protagonists gender identity is (has become) problematicas shortly thereafter it is pointed out that he is jealous of those
colleagues of his who live like harem women (Kafka 1915: 121).
These hypotheses also help to demonstrate the increasing neutralization of the introductory focalization, through which the author seeks to
ensure that the reader remains, at least initially, in the dark about the
cause of Gregors metamorphosis. Once one has perceived Gregors psychosexual problems, it appears obvious to look for an explanation about
the mysterious metamorphosis of the hero in his identity-forming family
relationships.
5 Conclusion
Through the textual examples given from Kafkas Metamorphosis, it
should have become clear that the question of Who (of the fictive entities) perceives (how much)? can be clearly separated from the question
of What can the reader know about portrayed world(s)?. The fictional
perspectivization, on the one hand, and the regulation of information
within the communication between the author and reader, on the other
hand, are two wholly independent phenomena and, therefore, must be
conceptually differentiated from one another. Whether temporary and definitive withholding of information can be readily denoted by the term focalizationor whether one ought not to refer to them with the term filtrationis, as a quarrel over terminology, not what we would like to discuss here15.
15
glass, which held him fast, soothing his hot belly (161). After his mother catches a
glimpse of him in this position and faints, crying out Oh God, oh God! (162), he
tears himself away from the lady and feels tortured by self-rebukes (163)in the
sense of the script previously sketched out by Sacher-Masochs Venus in furs.
As Chatman has already determined (cf. Chatman 1990: 14353) filtration (and also
filter) is a good term for capturing something of the mediating function of a characters consciousnessperception, cognition, emotion, reverieas events are experienced from a space within the story world (144, 149). Chatman sees the advantage
of the term filter in the fact that it catches the nuance of the choice made by the
implied author about [] which areas of the story world [he] wants to illuminate and
which to keep obscure (144). We could also agree with this argument. However, we
believe that the definition of the concept as provided is too broad. In suggesting, namely, that the term filter applies to every point of view of a figurethe [] range
of mental activity experienced by characters in the story-world (143)Chatman
76
What deserves further discussion, however, is how much cognitive inference is required on the part of the reader before a piece of information
can no longer be considered textually communicated. We have taken an
extreme position with our claim that there is no focalization whenever all
needed information can at least be inferred (even if it requires a great
effort to do so). Should it (on the basis of cognitive psychology) be possible in the future, to distinguish between different degrees of implicitness
within the Implicit II (cf. Linke & Nussbaumer 2001), one could also
speak of levels of focalizationdepending on how narrowly defined the
information-conveying pipe (Genette 1988: 74) is in each case.
Translated from German by Tracy N. Graves and Katherine McNeill.
References
Barthes, Roland (1964). lments de smiologie. Communications 4, 91141.
(1985). Laventure smiologique. Paris: Seuil.
Bremond, Claude (1973). Logique du rcit. Paris: Seuil.
(1964). Le Message narratif. Logique du rcit. Paris: Seuil, 1973, 1147.
Chatman, Seymour (1986). Characters and Narrators: Filter, Center, Slant and InterestFocus. Poetics Today 9, 189204.
(1990). Coming to Terms: The Rhetoric of Narrative in Fiction and Film. Ithaca:
Cornell UP.
Freud, Sigmund (1905). Three Essays on the Theory of Sexuality [Drei Abhandlungen zur
Sexualtheorie]. J. Strachey et al. (eds). The Standard Edition of the Complete Psychological Works of Sigmund Freud, Vol. VII. London: Hogarth Press and the
Institute of Psychoanalysis, 1953, 123245.
(1910). Leonardo da Vinci and a Memory of his Childhood [Eine Kindheitserinnerung des Leonardo da Vinci]. J. Strachey et al. (eds). The Standard
Edition of the Complete Psychological Works of Sigmund Freud, Vol. XI. London:
Hogarth Press and the Institute of Psychoanalysis, 1953, 57137.
Genette, Grard (1972). Discours du rcit. Figures III. Paris: Seuil, 65278.
(1980). Narrative Discourse. Tr. J. E. Lewin. Oxford: Blackwell.
(1983). Nouveau discours du rcit. Paris: Seuil.
Grice, H. Paul (1975). Logic and Conversation. Syntax and Semantics 3, 4158.
Herman, David (2002). Story Logic: Problems and Possibilities of Narrative. Lincoln:
U of Nebraska P.
assumes that figural perspectivization categorically accompanies a curtailing of the information available (149: distortion of story information). With this move, he also
allows for a jumbling of two terms, for whose clear and strict separation we have
argued here.
77
Kafka, Franz (1915). The Metamorphosis [Die Verwandlung]. The Metamorphosis

and Other Stories. Tr. J. Neugroschel. New York: Charles Scribners Sons, 1993,
11992.
Linke, Angelika & Nussbaumer, Markus (2001). Konzepte des Impliziten. Prsuppositionen und Implikaturen. K. Brinker et al. (eds). Text- und Gesprchslinguistik. Ein
internationales Handbuch zeitgenssischer Forschung. Berlin: de Gruyter, 43548.
Lotman, Jurij (1972). The Structure of the Artistic Text. Tr. R. Vroon. Ann Arbor: U of
Michigan P, 1977.
Niederhoff, Burkhard (2001). Fokalisation und Perspektive. Ein Pldoyer fr eine friedliche Koexistenz. Poetica 33:12, 121.
Quasthoff, Uta M. (1980). Erzhlen in Gesprchen. Linguistische Untersuchungen zu
Strukturen und Funktionen am Beispiel einer Kommunikationsform des Alltags. Tbingen: Narr.
Renner, Karl Nikolaus (1983). Der Findling. Eine Erzhlung von Heinrich von Kleist und
ein Film von George Moorse. Prinzipien einer adquaten Wiedergabe narrativer
Strukturen. Munich: Fink.
Rimmon-Kenan, Shlomith (1983). Narrative Fiction. Contemporary Poetics. London:
Routledge, 2002.
Sacher-Masoch, Leopold (1869). Venus in Furs [Venus im Pelz]. Tr. J. Neugroschel. New
York: Penguin, 2000.
Schmid, Wolf (1992). Cechovs problematische Ereignisse. Ornamentales Erzhlen in
der russischen Moderne. Cechov, Babel, Zamjatin. Frankfurt a. M.: Peter Lang,
10434.
Titzmann, Michael (1977). Strukturale Textanalyse. Theorie und Praxis der Interpretation. Munich: Fink, 1993.
ALAIN RABATEL
(Lyon)
A Brief Introduction to an Enunciative Approach to

Point of View
Even when narrowed down to the field of language, the concept of point
of view borrows from a variety of other fields ranging from vision (a
spectacular point of view) to the expression of an opinion which is more
or less justified, but which is distinct from a scientific truth (this is a
point of view which I share), and including the adoption of a central narrative perspective (referred to differently by Genette as focalization),
not to mention the linguistic operation of foregrounding important information, in particular through an emphatic operation (as in: The text by
Genette that I particularly have in mind is Narrative Discourse).
Point of view (POV) is defined, in an enunciative approach, in terms
of the linguistic means with which a subject 1 envisages an object 2 , and
encompasses all the meanings of the term envisage 3 , whether the subject be singular or collective and the object concrete or linguistic. The
subject, who is responsible for the referential values of the object, expresses his POV either directly, in explicit commentaries, or indirectly,
through the construction of referential values, in other words through
choices concerning the selection, combination, and realization of the linguistic material, and does so in all circumstances, ranging from the most
subjective choices to those which appear to be the most objectivizing, and
from the most explicit markers to the most implicit clues.
1
2
3
Or focalizer (Genette [1972; 1983]), enunciator (Ducrot [1984]), subject of consciousness (Banfield [1982]), modal subject (Bally [1965]), locus of empathization (Forest
[2003]), centre of perspective (Lintvelt [1981]; Rabatel [1997]), etc.
Or object of focalization (focalis) (cf. Bal 1977).
Going from perception to mental representation, as expressed in, and through, discourse.
80
Alain Rabatel
An examination of POV based on linguistic markers makes it possible

to advance the discussion of opposable arguments, but at the same time
has the disadvantage, at least in a first analysis, of appearing to be valid
only for a particular language, such as French (the language in which I
have conducted practically all my research), because of the specificities of
each linguistic system. In point of fact, this is not at all the case, given the
similitude between cognitive phenomena 4 and markers which play an
identical role in numerous languages. It can therefore reasonably be
claimed that the POV theory presented in the present study is to a certain
extent generalizable, as long as the greatest care is taken not to transpose
unchanged those analyses and markers which might not have an equivalent in another systemfor all languages, beyond their distinctive differences, are the vehicles of enunciative heterogeneity, in other words the
interweaving of the voices of others in ones own discourse, and this is a
phenomenon which is fundamental to POV theory.
Given the still dominant nature of Genettian theories, with which I am
in strong disagreement,hence the abandonment of the term focalizationI might have begun by giving a brief presentation of the specificity of my own position compared to that of Genette, but I prefer to commence by presenting my own framework for enunciative analysis 5 , examining the instances of point of view (1) and the various modalities of
POV (2), before presenting the external or internal markers of POV which
contribute to the more or less subjectivizing or objectifying expression of
these instances (3). In the final part of this study, I bring together the
points on which Genette and I differ, while clearly underlining the nature
of my enormous debt to him (4). In short, I would ask readers to bear with
me before I lay out my reasons, so they can better appreciate, on the basis
of the evidence presented, why I distance myself from the Genettian
model.
4
5
See Uri Margolins contribution in the current volume.

Especially as: (i) Genette does not claim that his is an enunciative approach, (ii) his
essentially structuralist conception of focalization is grounded, it is worth noting, on a
very limited number of linguistic clues, and (iii) the users of his model often disagree
over the analysis of focalizations see Rabatel (1997, ch. 12).
A Brief Introduction to an Enunciative Approach to Point of View
81
1 The Instances of Point of View

Enunciative linguistics, of which Beneveniste (1966) was the pioneering
proponent, seeks to analyse the way in which the enunciators choices of
exophoric linguistic reference influence the addressee. It is in this context
that the links between, or disconnection of, discourse and situation of utterance have been studied through the examination of person and spacetime markers. Rather than speak of the subjectivity of the enunciator
(cf. Kerbrat-Orecchioni 1980), however, it is preferable to favor the notion of point of view, since a point of view is not necessarily subjective
(cf. Latour 2006 6 ). Furthermore, traces of a point of view are not exclusively restricted to markers relating to Iherenow; rather, they are
to be found widely distributed in the way the enunciator constructs discourse objects. In my approach to POV (freely inspired by Ducrot
[1984]), the speaker is the instance that expresses an utterance which is
localized either deictically or anaphorically, while the enunciator 7 , who is
similar to Ballys modal subject, takes charge of the utterance, insofar as
the evaluations, qualifications, modalizations, and judgements on the
objects of discourse are filtered through his subjectivity 8 . It is, however,
important to distinguish the prime enunciator, the one who takes responsibility for the utterance made by a speaker, from secondary enunciators
who are the sources of a POV, where these POVs are not expressed in
words. Ducrot insists on the fact that the expression of POV is not
necessarily embodied in precise words:
6
7
8
What makes you think that adopting a point of view means being restricted? or being particularly subjective? [...] If it is possible for you to see a statue from different
points of view, it is because the statue itself is three-dimensional and allows you, yes,
allows you to walk round it. If something makes such a multiplicity of points of view
possible, it is because it is complex, intricate, well-organized and beautiful, yes, objectively beautiful. [...] Do not believe all the nonsense that is written about the fact of
being restricted to your own perspective. Each science has invented ways of shifting
from one point of view to another, from one frame of reference to another. [...] This is
what relativity is all about. [...] If I want to be a scientist and attain objectivity, I need
to be able to move from one frame of reference to another, from one point of view to
another. Without such shifts I really would be restricted to my own narrow point of
view. (Latour 2006: 21013)
See Charaudeau & Maingueneau (2002: 22024, 226).
It is useful for the theory to make a distinction between these two actualizations, even
if they often go together, see Rabatel (2005a) and Rabatel (2008b: ch.15).
82
Alain Rabatel
I call enunciators those entities which are considered as expressing themselves
through the act of enunciation, though precise words may not necessarily be attributed
to them; if they can be said to speak, it is only in the sense that enunciation can be
seen as expressing their point of view, their position or their attitude, but not, in the
concrete sense of the term, their actual words. (Ducrot 1984: 204)
This is why it is necessary to make a distinction between cases where the

speaker is superimposed over a single enunciator (himself) and those
where this superimposition is complicated by the co-presence of several
different enunciators. The former case corresponds to situations in which
utterances are built on a syncretism between speaker and enunciator, as on
each occasion the speaker thinks what he says and says what he thinks
or pretends to say what he thinks: this can be observed in pleas, threats,
oaths, or again in declarations of love, or orders (see the following
quotations in [1] and [2]):
(1) Behold, thou art fair, my love; behold, thou art fair; thou hast doves eyes. (KJV: 9
Song of Solomon 1:15)
(2) And the LORD said unto Moses, stretch out thine hand over the sea, that the
waters may come again upon the Egyptians, upon their chariots, and upon their horsemen. (KJV: Exodus 14:26)
However, the prime speaker/enunciator may also develop in his discourse

certain POVs 10 which he does not necessarily share, as in the case of
irony, hypotheses, free indirect discourse, or again in delocutive utterances expressing a point of view 11 , in reported heterodiegetic narratives in
the past, as in:
(3) Le Philistin regarda et, quand il aperut David, il le mprisa : ctait un gamin, au
teint clair et la jolie figure. (TOB: 12 Premier Livre de Samuel 17:41)
(And when the Philistine looked about, and saw David, he disdained him: for he was
but a youth, and ruddy, and of a fair countenance. [KJV: 1 Samuel 17:42])
Genettes model of focalization would be hard put to account for the

above example in so far as it contains few linguistic markers of internal
focalization, essentially free indirect discourse or internal monologue.
However (3) is not a case in point of such forms. While, admittedly, one
might be tempted to analyze (3) as an instance of free indirect discourse,
this can only be achieved by stretching the example somewhat since, in
9
10
11
12
The Holy Bible (Authorized King James Version).

Having as a reference point an intratextual enunciator.
Marked in italics in (3) and in subsequent examples.
Traduction cumnique de la Bible.
83
accordance with the generally admitted norm, free indirect discourse depends on the presence of a reporting verb or, more rarely, of a verb of
thought (cf. De Mattia 2001), neither of which is present here 13 . The text
positions Goliath as the perceptual subject, he looked about, and describes the precise nature of the intentional perception: he disdained
him. Here, in the French translation, quand (when) is equivalent to as
soon as, indicating that Goliath deliberately looked at David to see
whether the latter might be a formidable adversary. The text does not just
predicate the act of perception, in the plane of historical enunciation
(whose prototypic tense, in French, is the simple past), by giving an overall view of this event. With the use of the imperfect tense copula, tait
(was) 14 , by virtue of the secant view which it expresses, the reader finds
13
14
See the studies of free indirect discourse in English (De Mattia [2001]; Poncharal
[2003]) and on the value of demonstratives in the marking of point of view in Swedish
(Jonasson [2002]), et al. While, I agree with Authier-Revuz (1992 and 1993) on the
need to stress the close relationship between free indirect discourse and POV in cases
of free indirect discourse without a reporting verb or verb of thought, and while I also
share, with Rosier (1999) and Fludernik (1993), the idea of a continuum of forms, I
would not go so far as to place POV on the same level with free indirect discourse, or
to consider POV as one of the various forms of reported speech, contrary to the
position wrongly attributed to me by Marnette (cf. 2005: 61, 277). Admittedly, when
seen in the context of dialogism, perceptual reports are close neighbours to the reports
of speech and thought found in reported speech, since a prime speaker/ enunciator (the
narrator) envisages things from the point of view of a secondary enunciator (a
character), even when there is no explicit discourse, as will be seen infra with example
(3), but Baxtinian dialogism is much a broader phenomenon than the notion of reported
speech. See Rabatel (2008b, ch. 15 to 17).
The markers of POV are broadly the same in French and English apart from the tense
systems. The French imperfect is the prototypic tense of the second plane (cf. Combettes 1992). The use of the word prototypic here is to be understood as meaning that
it is the tense which is most often encountered, but it should be noted that this role can
also be played by other verb forms, such as the present participle, as in the Hebrew
text. The above analyses are valid for French, and cannot be applied unchanged to the
English verb system. It is clear that in example (3) the French imperfect has to be
translated by a simple preterite; however, it should in no way be concluded that the
English simple preterite is equivalent to the French imperfect, but simply that it shares
certain aspectual characteristics with the latter. Poncharal (personal communication)
observes, moreover, that the imperfect rarely corresponds to a form in be+-ing, and
that there is often more affinity between the simple preterite and the imperfect than between the simple preterite and the French simple past. As for the rest, the English
translation of example (3) denotes a POV, by virtue in particular of the aspectual
values of was, but also by virtue of the presence of for (cf. Danon-Boileau 1995:
26), not to mention other choices involved in exophoric reference.
84
Alain Rabatel
himself at the heart of the perception: at this point the text reveals details
or parts of this perception (general appearance, complexion, face). The
reader thus realizes, without the Philistine having to say a word, that the
term youth and the allusion to his fair countenance, in short his quasi
feminine grace, are more characteristic of women than of men, and
connote the disdain of the virile male of mature years for an upstart who
is not part of the world of virile men, and hence not a worthy adversary
for a man of his strength.
This explanation will, it is hoped, enable the reader to obtain a better
grasp of Ducrots definition, quoted above, of the (intratextual) enunciator. Thus, the utterance in (3), written by the narrator, who corresponds to
the prime speaker/enunciator, involves an intratextual enunciator, Goliath,
who is the enunciative origin of a POV, even though this POV in no way
corresponds to discourse uttered by Goliath, since the latter has said,
literally, nothing. To put it another way, the POV represented is a descriptive fragment which could, perhaps, be paraphrased by a sort of implicit
internal monologue along the lines of: Ill soon make short work of this
pretty young man! The prime speaker/enunciator conveys this POV
without endorsing its disdainful connotation 15 , even though he confirms
the denotation of the propositional content, the youth and beauty of
David, in the absence of any epistemic distancing 16 .
What, then, are the narratological conclusions that we can draw from
this enunciative analysis? If the origins of POV are enunciators, then cat15
16
Axiological distancing, though of a discreet nature, is nevertheless present in the contrast between the verb disdained and the description of David: the positively orientated attributives would not normally indicate disdain, unless they are seen through the
sadistic prism of a man who has full confidence in his strength, and reduces human relationships to a man-to-man fight to the death. This distancing indicates a dissonance
between the narrator and the character/perceiver. In the contrary case, we speak of consonance, see Cohn (1978); Rabatel (1998: ch. 4; 2001 and 2008b: ch. 19).
This is why I distance myself from Fludernik (1993) when she treats speech and
thought as a whole, even going so far as finding similarities between perceptions and
thoughts in the case of narrated perceptions. While I share the idea of scalarity in the
subjective expression of speech, thought and perceptions, I do not go so far as to consider the latter three as equivalent to each other. Furthermore, her conception of free
indirect speech (FIS) is based on the idea that the distinction between mode and voice
is unfounded. My analysis of (3) shows that the distinction remains pertinent as long as
the enunciators POV is expressed through exophoric linguistic reference, even if he
does not pronounce any words, since it is the voice of the narrator who envisages
things from the characters point of view.
85
egories of POV, linked to these enunciative sources, can only exist by

virtue of their relationship with a linguistic substrate. This enunciative reality explains why an authentic narrational POV is undoubtedly present
when the objects of discourse are referenced without being seen through
the perspective prism of one of the main characters. This would be the
case in (4) if David was described with the same terms, without reference
to Goliath:
(4) David appeared. He was but a youth, and ruddy, and of a fair countenance, an
adversary unworthy of respect.
The concept of zero focalization (which Genette glosses as an absence of

focalization, the point of view of the narrator or variable focalization resulting from all the other focalizationswhich are, for a scientific definition, incompatible, contradictory glosses) does not stand up to scrutiny.
Nor does external focalization, despite Rivaras claims 17 . Bal (1977),
some time ago, stressed the confusion between focalization by (a
focalizing subject/an instance) and focalization on (a focalized object)
and reassigned external focalization to objective description of the focalized entity. My proposals are paralleled by the Anglo-American approach, which only makes a distinction between external point of view
(that of the narrator) and internal point of view (that of the character),
even if the linguistic justification for the qualification as external is
questionable. This enunciative consideration is the nub of my difference
of opinion with Genette. It would take too much space to reiterate my
1997 demonstration (one which has never been refuted since): the supposed examples of external focalization are actually attributable either to
a characters POV or the narrators POV with an external view, in other
words limited to the description of some external aspect of an object 18 ,
such as the description of someones clothing or of an object and
expressed in objectifying utterances whose enunciation is historical, and
above all to a description from which manifest traces of subjectivity are
absent 19 .
17
18
19
See Rabatel (2009) for a detailed discussion of this point.

The physical (external) description of David, therefore, shows traces of the (internal) subjectivity of Goliath; this is why I felt it necessary to abandon this dichotomy (cf. Rabatel 1997), which is unfounded from the linguistic point of view.
See Rabatel (1997: ch. 3, 4 and 12, and 2003c).
86
Alain Rabatel
2 Different Modalities of Point of View: Dialogism of Represented,

Narrated and Asserted Points of View
POV, then, corresponds to those elements which, in the exophoric linguistic reference to the objects (of discourse), reveal, from a cognitive and
an axiological point of view, a particular enunciative origin and indicate,
explicitly or implicitly, the latters representations and, where relevant,
his judgements on the referents. This definition makes it possible to give
an explanation for the close relationships between POV and reported discourse, on the one hand, and POV and assertion, on the other, without
limiting POV to perceptions or to the narrative genre alone: a POV exists
when reference to the object also entails the representation of an enunciator, even in the absence of explicit judgements, whether the object of
discourse be an opinion or a perception, and whether the latter appear in a
description, a narrative, a news item, an explanation or an argumentation.
This means that POV is not limited to the expression of represented
perceptions, as analyzed by myself in Rabatel (1998), even though I do
not disown my analyses; quite simply, these analyses are not, and never
claimed to be, the final word on POV. In this sense, reported discourse
(cf. Rosier 1999) and POV are subsets of the general problematic of dialogism. Speech, thought and perceptions may be reported/represented
using identical syntactic and enunciative patterns by means of direct, indirect, free indirect and free direct reporting. The fragment in italics in (5)
is a direct account of perception, which is largely analogous, at the
syntactic level, to its paraphrases containing direct discourse, as proposed
in (6), or containing indirect discourse in (7), free direct discourse in (8),
free indirect discourse in (9), and psycho-narration in (10):
(5) And the angel of the LORD appeared unto him in a flame of fire out of the midst
of a bush: and he looked, and, behold, the bush burned with fire, and the bush was not
consumed. (KJV: Exodus 3:2)
(6) And the angel of the LORD appeared unto him in a flame of fire out of the midst
of a bush: Moses looked and said to himself, the bush is burning with fire, and the
bush is not consumed.
(7) So Joshua sent messengers, and they ran unto the tent. They cried out that it was
there, hid in his tent, and the silver underneath. And they took them out of the midst
87
of the tent, and brought them unto Joshua, and unto all the children of Israel; and laid
them out before the LORD 20 .
(8) So Joshua sent messengers, and they ran unto the tent; and there, indeed, it is, hid
in his tent, and the silver underneath. And they took them out of the midst of the tent,
and brought them unto Joshua, and unto all the children of Israel; and laid them out
before the LORD.
(9) And when Pharaoh drew nigh, the children of Israel lifted up their eyes, and,
behold, the Egyptians marched after them; and they were sore afraid: and the children
of Israel cried out unto the LORD. (KJV: Exodus 14:10)
(10) So Joshua sent messengers, and they ran unto the tent and saw that what Achan
had said was true. And they took [it and the silver] out of the midst of the tent, and
brought them unto Joshua, and unto all the children of Israel; and laid them out before
the LORD.
These variations objectify the cognitive and linguistic continuums between perception, thought and speech, and explain, at the semiotic level,
the similarities of their textual values in the construction of reality effects
(mimesis), in the provision of information (mathesis) and the handling of
textual structuring (semiosis), according to Adam and Petitjean (1989).
All these various accounts of perception can be grouped together in a
continuum, labelled as follows, according to their greater or lesser degree
of visibility and their greater or lesser aptitude to express the enunciators
interiority, subjectivity and reflexivity:
Embryonic or narrated POV (cf. Rabatel 2001; 2004), corresponding
to perceptual points of view limited to traces on the first plane, as in (10);
Represented POV (cf. Rabatel 1998: 54), expressing the accounts of
perception (possibly associated with speech or thought) developed in the
second plane, as in the passages in italics in (3), (5);
Asserted POV (cf. Rabatel 2003b; c; Rabatel 2008b: ch. 15 to 17), corresponding to POVs expressed in word or thought, as exemplified by
conventional forms of reported discourse (cf. [6] to [9]) or in assertions,
outside of the context of reported discourse, as in the examples below.
Beyond the question of labelling, one should not lose sight of the underlying affinity between these forms which are capable of expressing the
POV of a character or that of the narrator: thus, the embryonic POV is not
an absence of POV, but is rather a minimal, minor POV, one which, while
less reflexive and subjectifying than an asserted POV, is nevertheless
already a POV. One should never lose sight of their affinity or their
20
Examples (7), (8) and (10) are based on the Biblical text: So Joshua sent messengers,
and they ran unto the tent; and, behold, it was hid in his tent, and the silver under it
(KJV: Joshua 7:22).
88
Alain Rabatel
complementarity in the expression of mimetism and reflexivity (cf. Rabatel 2003e), without forgetting that the dialogism of POVs goes beyond
the framework of reported discourse, since every assertion, or even a
word, expresses, in one way or another, a POV, as can be seen in the enumerations of the names of countries in which the pleasure of the Jews in
naming the different parts of their promised land is clearly perceptible or,
again, can be found in the genealogies which, through the enumeration of
filiations which have ensured the survival of Israel (Joshua 1521, or
1 Chronicles 2), give a glimpse of the hard-earned pleasure of enduring.
But, besides these forms of dialogism (cf. Bakhtin 1929; Rabatel
2008b: ch. 13), all propositional content, even that which is not concerned
with perception, thought or speech, expresses by default the POV of the
prime speaker/enunciator, or that of an intratextual enunciator. Thus, the
selection of information in the construction of propositional content is
highly significant. This is the case in the First Book of Chronicles 11:1:
the narrative, which states that Then all Israel gathered themselves to
David unto Hebron, elides seven and a half years of the reign of David
over Hebron (while Ishbosheth, one of the sons of Saul, ruled over the
rest of Israel), so as to give it to be understood that the reign of David
concerned all Israel, and that nothing having to do with Saul was of any
importance. This is confirmed by the ensuing direct discourse, in which
all Israel declares to David: Behold, we are thy bone and thy flesh.
And moreover in time past, even when Saul was king, thou wast he that
leddest out and broughtest in Israel: (KJV: 1 Chronicles 11:12).
3 Markers of Point of View

Parallels are often drawn between the linguistic nature of POV and the
idea contained in a proposition. All propositions, being centered on propositional content (=PC) express a POV. Does this, then, mean that there
can be no POV above the level of the proposition, for instance at the text
level, or below the level of the proposition, for instance at the word level?
It is relevant to consider that the PC is the heart of any POV, for it relates to a predication which, in an assertion, always already expresses the
point of view of the speaker by virtue of the choice of words. But that
does not imply that the lower threshold of POV is predication, as a single
word can at times suffice to express a worldview, relating back to a POV,
as long as the word links back to an enunciator and a POV which can both
be clearly identified for a given linguistic community. Above the pro-
89
positional level, it is desirable 21 to group PCs together according to

referential content (all the PCs concerning the same referent), enunciative
origin (all the PCs having the same enunciative origin) and argumentative
orientation (all the PCs sharing the same orientation to which, if necessary, could be added PCs with a contrary orientation integrated into the
argumentative drift of the principal enunciator)in short, according to
the similarity of the effects of the POV upon the reader.
The external markers are the opening or closing markers indicating the
beginning and the end of a POV. Such markers are fragile, once one
moves outside the clearly marked-out framework of written direct discourse. With an asserted POV, the external markers are those of reported
(or represented) discourse, and only direct discourse has clear external
markers (and even then, this is only true of written discourse), for with
indirect discourse, while there is unquestionably an opening marker, the
closing marker is often far from being present in every case. The difficulties with external delimitation increase with free indirect discourse and
narrativized discourse. For the represented POV, the opening limit is indicated by a perceptual predicatormost often a verb, but also conceivably a nounand a related perceiving subject; in the absence of such
markers, the perceptual process may be inferred from a verb of motion;
the opposition between global tense forms (such as the preterite or the
historical present) and secant forms (such as the French imperfect or pluperfect) which play an opening role (passage from the first to the second
plane) or closing role (passage from the second to the first plane) comparable to what happens with free indirect discourse (cf. Vuillaume 2000) 22 .
With an embryonic POV, we move on from external markers to internal
markers.
Generally speaking, the more tenuous the external markers, the more
internal markers of modal actualization act as signals of enunciative alterity and have the merit of embodying the instances behind the POVs.
Numerous lexical elements are capable of playing this role such as, at the
level of the cohesion between nominals, lexical designations containing
21
22
In order to optimize the cognitive processing of information, the multiplication of myriads of enunciators should be avoided.
Tense alternation such as, in French, preterite>perfect>preterite, does not necessarily signify the end of a particular point of view and a return to narrative text. As was
shown in Rabatel (2003a), it also indicates the transition from a represented to an embryonic point of view (or vice versa), in other words different degrees in the reflexive
apprehension of percepts by an intratextual enunciator.
90
Alain Rabatel
positive or negative value judgements (11) or phenomena of actualization

of nouns (12). As examples (11) to (23) cover, with some variations, the
same ground as example (3), it will not be necessary to restate the original
analysis. Suffice it to say that the POV remains both the same and, in
terms of expressivity, not quite the same. We would remind the reader
that the part in italics corresponds to the expression of the POV as analyzed in example (3), and that the term(s) underlined correspond(s) to the
internal markers which intensify the original POV and generally give it a
specific semantic colouring, which can be more or less expressive, depending on their particular value. It is particularly important, therefore, to
stress that each of these markers contributes to the location of the POV,
and more particularly, to its expressivity (underlined); hence the more
numerous these markers are, the more strongly marked is the POV:
(11) And when the Philistine looked about, and saw David, he disdained him: for he
was but a youth, an ephebe, and ruddy, and of a fair countenance.
(12) And when the Philistine looked about, and saw David, he disdained him: this
youth, ruddy, and of a fair countenance.
At the level of logical cohesion, this function is also fulfilled by presentatives (cf. Rabatel 2008a, ch. 3), connectors, spatio-temporal markers
and intensifiers, by virtue of their enunciative-argumentative value (cf.
Rabatel 2008a: ch. 4) (13):
was barely a youth, but his ruddy complexion and fair countenance could not hide his
shifty look.
At the verbal level, temporal-aspectual distinctions, the semantic content

of verbs (cf. Rabatel 2003a), etc., also contribute to modal actualization in
the representation of events.
At the syntactic 23 level, most of the markers indicating dialogism act
as internal markers, skewing the account of perception towards one of
speech or thought, and confirm that the POV is that of Goliath: thus, the
dialogue markers which go to make up the dictum (propositional content)
reinforce the reflexive content of the perceptions, whether the case be one
of an interrogative (14), a rhetorical question (15), presupposition or negation (16), a cleft sentence (focus, foregrounding) (17), a concessive
(18), opposition (19), confirmation (20), rectification (21) or intensification (22):
23
Their syntactic basis is, of course, inseparable from their semantic dimension.
91
(14) And when the Philistine looked about, and saw David, he disdained him: who
was to be the enemys champion? For he was but a youth, and ruddy, and of a fair
countenance.
(15) And when the Philistine looked about, and saw David, he disdained him: had
anyone ever seen such a ridiculous enemy? He was but a youth, and ruddy, and of a
fair countenance.
(16) And when the Philistine looked about, and saw David, he disdained him: the
enemy really was not to be feared, for he was but a youth, and ruddy, and of a fair
countenance.
(17) And when the Philistine looked about, and saw David, he disdained him: the
enemys champion, he was but a youth, and ruddy, and of a fair countenance.
(18) And when the Philistine looked about, and saw David, he disdained him: for even
if the enemy looked bold, he was but a youth, and ruddy, and of a fair countenance.
was not a battle-hardened fighter, he was but a youth, and ruddy, and of a fair
countenance.
(20) And when the Philistine looked about, and saw David, he disdained him: yes, he
really was but a youth, and ruddy, and of a fair countenance.
(21) And when the Philistine looked about, and saw David, he disdained him: he was
a young man, in fact, but a youth, and ruddy, and of a fair countenance.
was but a youth, an effeminate weakling in fact, and ruddy, and of a fair countenance.
These dialogue markers make the potential underlying responsive dimension in (3) explicit: for instance, the dialogic perceptual account of (15) is
a response to a presupposed implicit objection; (16) and (18) are answers
to a prior objection (or at any rate anticipate such an objection), and so
on. All these markers (the list of which is not exhaustive; cf. echoic repetition, travesty, irony, the hypothetic, etc.) participate in the construction
of intratextual modal subjects.
Inversely, the POV can delete these markers, or even erase verbs of
perception (in some contexts, it is the perceiving subject who is implicit),
as long as there are enough clues in the exophoric reference to the object
for it to be understood as the source of the perception:
(23) The Philistine advanced towards his enemy: he was but a youth, and ruddy, and
of a fair countenance.
Thus, in function of the dialogism of the linguistic markers 24 involved in

the exophoric reference of perceptions, the latter denote perceptual processes which may be more or less intentional and more or less associated
24
By dialogism, we mean the fact that an utterance allows several voices, which answer
each other, to be heard. This concept is often confused with that of polyphony, but it is
desirable to make a distinction between the two, see Rabatel (2008b: ch. 13).
92
Alain Rabatel
with epistemic and axiological dimensions, to the point that the saturation
of dialogue markers blurs the distinction between, on the one hand, perception and, on the other, thought and speech 25 .
4 Marked Differences With the Genettian Model of Focalization

Besides the fundamental disagreement with Genette over the number and
nature of instances of POV, there are differences as to the status of firstor third-person narrative utterances, a question which intersects with that
of the status of utterances with a heterodiegetic narrator and those mediated by a character, on the plane of the expression of subjectivity and
knowledge.
As the previous markers may be present with an I- or he-POV, it follows that a POV expressed in the first person is not necessarily subjectifying by virtue of its expression, no more than a third-person point of
view necessarily implies an objectifying utterance. This scale of subjectivity in the expression of POV is due to the fact that what is perceived
is expressed, be it with an I or a he, through lexical or syntactic
markersmuch in the same way as were those mentioned in connection
with examples (11) to (23); and these markers, by virtue of their presence,
indicate the reactions of a subject toward an object. Thus, the I-POV in
(24) is totally subjectifying, through the comparison of the loved one with
a gazelle or a young stag. The I-POV in (25), describing Ezekiels
vision of the construction of the new temple, is objectifying, whereas the
same prophets vision of glory in (26) is a combination of various
objectifying data from which there emerge, despite his desire to describe
faithfully what he had seen in a dream, a certain number of subjective
reactions:
(24) The voice of my beloved! behold, he cometh leaping upon the mountains,
skipping upon the hills. My beloved is like a roe or a young hart: behold, he standeth
behind our wall, he looketh forth at the windows, shewing himself through the lattice.
(KJV: Song of Solomon 2:89)
(25) So he measured the court, an hundred cubits long, and an hundred cubits broad,
foursquare; and the altar that was before the house. And he brought me to the porch of
the house, and measured each post of the porch, five cubits on this side, and five
cubits on that side: and the breadth of the gate was three cubits on this side, and three
cubits on that side. (KJV: Ezekiel 40:4748)
25
See also Fludernik (1993) and supra, note 16.
93
(26) And I looked, and, behold, a whirlwind came out of the north, a great cloud, and
a fire infolding itself, and a brightness was about it, and out of the midst thereof as the
colour of amber, out of the midst of the fire. Also out of the midst thereof came the
likeness of four living creatures. And this was their appearance; they had the likeness
of a man. And every one had four faces, and every one had four wings. And their feet
were straight feet; and the sole of their feet was like the sole of a calfs foot: and they
sparkled like the colour of burnished brass. And they had the hands of a man under
their wings on their four sides; and they four had their faces and their wings. (KJV:
Ezekiel 1:48)
The presence and the combination of lexical and syntactic markers produce the same effects of subjectifying or objectifying expression with a
he-POV: the heterodiegetic POV of Solomon, describing the temple, is
objectifying in (28), the POV taking the form of a description of actions:
(27) And in the most holy house he [Solomon] made two cherubims of image work,
and overlaid them with gold. And the wings of the cherubims were twenty cubits long:
one wing of the one cherub was five cubits, reaching to the wall of the house: and the
other wing was likewise five cubits, reaching to the wing of the other cherub. And one
wing of the other cherub was five cubits, reaching to the wall of the house: and the
other wing was five cubits also, joining to the wing of the other cherub. The wings of
these cherubims spread themselves forth twenty cubits: and they stood on their feet,
and their faces were inward. And he made the vail of blue, and purple, and crimson,
and fine linen, and wrought cherubims thereon. (KJV: 2 Chronicles 3:1014)
The heterodiegetic POV in (3), on the other hand, includes numerous subjectivemes, as was seen above. For its part, the following extract from
Genesis occupies an intermediary position:
(28) In the beginning God created the heaven and the earth. And the earth was without
form, and void; and darkness was upon the face of the deep. And the Spirit of God
moved upon the face of the waters. And God said, Let there be light: and there was
light. And God saw the light, that it was good: and God divided the light from the
darkness. And God called the light Day, and the darkness he called Night. And the
evening and the morning were the first day. (KJV: Genesis 1:15)
The same reasoning denies the pertinence of the erroneous simplifications

which portray the POV of a heterodiegetic narrator as, by definition,
objective, and a characters POV as, by definition, subjective 26 . This type
of confusion between the origin of and the linguistic expression of subjectivity relies in an overly naive fashion on the mistaken idea that narration is so objectifying that no one is speaking here, the events seem
to narrate themselves, as Benveniste (cf. 1966: 241) claims in an as26
See supra the quotation from Latour, note 6.
94
Alain Rabatel
sertion which is contradicted by the presence of all sorts of subjective

markers (cf. Kerbrat-Orecchioni 1981), even in written texts whose enunciation is historical, because anaphoric localization still does not prevent
traces of the modal subject from coming to the surface (cf. Rabatel 2005c:
11720), as can be verified if one reflects on the presence or absence of
such-and-such a linguistic marker, as in examples (11) to (23), opening
the way to the possibility for a narrators or a characters POV to be more
or less objectifying or subjectifying.
The last point of disagreement concerns the attribution of a volume of
untouchable knowledge contiguous with each perspective, ranging from
narratorial omniscience to the maximum retention of information in external focalization, whereas this is only a theoretical possibility. This
eventuality is accepted as far as the narrator is concerned, even if, in
practice, he is confined to omniscience 27 , while on the other hand characters are denied variability in the volume of their knowledge. Now, omniscience is a datum which is not always verified in texts, depending on the
genre, the type of narrator, expositional strategy, etc.; nor, moreover, is it
reserved exclusively to narrators, insofar as it is manifest, since there exist
knowledgeable characters and since, generally speaking, the thesis
according to which the point of view of characters is a limited one (restricted to external vision, according to Vitoux [1982]), because they are
supposedly unable to have access to the thoughts of other characters,
quite simply does not stand up to close linguistic inspection, as is shown
by the examples analyzed in chapter 12 of my 1997 book. The fact that a
character can indeed evoke the thoughts of others, particularly in reported
discourse, is the surest indicator that characters, as the centers of narrative perspective, can have access to other characters interiority or, at
least, represent this interiority, as the narrator does, with the same margins
of certainty and error. What is certain is that the existence of this actorial
knowledge about other characters is not guaranteed: in this sense, there is
a clear difference between the authorial and the actorial instance 28 , but it
27
28
See Rabatel (2009) for a more detailed analysis of straightforward examples of omniscience or of an equally obvious absence of omniscience, and for the development of a
bridge between the enunciative approach to point of view and an interactional conception of narration which gives all due importance to the reader/co-enunciator. See
also Pier (2004) and Coste (2006).
It could be objected that the characters knowledge depends on their status as narrator-characters, who are the authors of embedded narratives. This objection, however, backfires on those who voice it: the fact that a character can act as a second-level
95
is one that has to do with the fiduciary relationship. In short, it is possible

for a character to have access to the introspection of someone else,
contrary to what J. Lintvelt writes: adopting the perspective of an actor,
the narrator is limited to the extrospection of this actor-perceiver, with the
result that he will only be able to give an external presentation of the
other actors (Lintvelt 1981: 44).
But beyond the differences, what remainsand this is one of Genettes unsurpassed (and unsurpassable) achievementsis the distinction
between mode and voice, in other words the possibility for the narrator to
tell a story with his own voice while allowing other enunciative sources to
be heard, even when they do not take the form of discourse. In my own
work, up to now, I have constantly endeavoured to identify the linguistic
markers which allow one to hear these POVs, with their profoundly dialogic nature, within the framework of a continuum, by paying particular
attention to the least obvious forms of subjectivity, in contexts of enunciative effacement, in which traces of a modal subject are nevertheless
perceptible 29 .
These different forms function conjointly (cf. Rabatel 2001; 2005b),
allowing the prime speaker/enunciator tomore or less explicitly
express his own point of view or adopt that of the characters, or even to
superimpose or oppose various narrative perspectives, inviting the reader,
on the basis of the effects produced by the POV, to get ever closer,
through empathy, to the characters reasons, as to those of the narrator,
since the former are expressed through the voice of the latter. This is why
the necessarily schematic presentation of forms and markers must be confronted with texts and their interpretations, which it has not been possible
for me to do here, having chosen to demonstrate and substantiate the deep
underlying enunciative unity which brings together such a diversity of
linguistic forms. But the reader may wish to refer to the numerous publications 30 in which critics have been good enough to point out the thor-
29
30
narrator exposes the vacuousness of the arguments that relegate characters to a role
which only allows them limited knowledge. This in no way reduces differences in
function and status: the cognitive superiority of the character-narrator, which is higher
than that of all other characters, remains lower than that of the first narrator.
For a more complete approach, see my re-reading of Genette, in Rabatel (1997 and
2008b: ch. 2).
See, in particular, my analyses of the Bible (cf. Rabatel 2008a, ch. 6 to 8), Maupassant
(cf. Rabatel 2008a: ch. 9 and 10), Pinget (in Bouchard et al. [2002]), Ernaux, Renaud
Camus or Semprun (cf. Rabatel 2008b: ch. 8 to 10).
96
Alain Rabatel
oughness of the linguistic descriptions and the resulting interpretative advances.

Translated from French by Rodney Coward (University of Tours), including the quotations from Ducrot (1984), Latour (2006) and Lintvelt
(1981).
References
Adam, Jean-Michel & Petitjean, Andr (1989). Le texte descriptif. Paris: Nathan.
Authier-Revuz, Jacqueline (1992). Repres dans le champ du discours rapport (1).
Linformation grammaticale 55, 3842.
(1993). Repres dans le champ du discours rapport (2). Linformation grammaticale 56, 1015.
Baxtin, Mixail (Bakhtin, Mikhal) [Voloinov (Volochinov), V. N.] (1929). Marxisme et
philosophie du langage. Paris: Editions de Minuit, 1977.
Bal, Mieke (1977). Narratologie. Paris: Klincksieck.
Bally, Charles (1932). Linguistique gnrale et linguistique franaise. Berne: Francke,
1965.
Banfield, Ann (1982). Unspeakable sentences: Narrative and representation in the language of fiction. London: Routledge & Kegan Paul.
Benveniste, Emile (1966) Problmes de linguistique gnrale 1. Paris: Gallimard.
Bouchard, Robert et al. (2002). Dclencher le mcanisme de la construction / dconstruction du texte romanesque. E. Roulet & M. Burger (eds). Les modles du
discours au dfi dun dialogue romanesque: lincipit du roman de R. Pinget, Le
Libera. Nancy: PU de Nancy, 153211.
Charaudeau, Patrick & Maingueneau, Dominique (2002). Dictionnaire danalyse du discours. Paris: Seuil.
Cohn, Dorrit (1978). Transparent Minds. Narrative Modes for Presenting Consciousness
in Fiction. Princeton: Princeton UP.
Combettes, Bernard (1992). Lorganisation du texte. Metz: Centre dAnalyse Syntaxique.
Coste, Didier (2006). Le rcit comme forme-mouvement. Acta Fabula 7:5,
<http://fabula.org/revue/document1641.php>.
Danon-Boileau, Laurent (1995). Du texte littraire lacte de fiction : lectures linguistiques et rflexions psychanalytiques. Paris: Ophrys.
De Mattia, Monique (2001). Mrs Dalloway de Virginia Woolf ou linstabilit du discours
rapport. M. De Mattia & A. Joly (eds). De la syntaxe la narratologie
nonciative. Paris: Ophrys, 22764.
Ducrot, Oswald (1984). Le dire et le dit. Paris: Minuit
Fludernik, Monika (1993). The Fiction of Language and Languages of Fiction. London:
Routledge.
Forest, Robert (2003). Empathie linguistique et point de vue. Cahiers de praxmatique
41, 85104.
97

Jonasson, Kerstin (2002). Rfrences dictiques dans un texte narratif. Comparaison
entre le franais et le sudois. M. Kesik (ed). Rfrences discursives sur les
langues romanes et slaves. Lublin: Wydawnictwo Uniwersytetu Marii-CurieSklodowskiej, 10721.
Kerbrat-Orecchioni, Catherine (1981). Lnonciation. La subjectivit dans le langage.
Paris: Colin.
Kleiber, Georges (2003). Adjectifs dmonstratifs et point de vue. Cahiers de praxmatique 41, 3354.
Latour, Bruno (2006). Changer de socitRefaire de la sociologie. Paris: ditions de la
Dcouverte.
Lintvelt, Jaap (1981). Essai de typologie narrative. Paris: Corti.
The Holy Bible (Authorized King James Version) [KJV] (1992). Oxford: Oxford UP.
Marnette, Sophie (2005). Speech and Thought Presentation in French. Amsterdam: John
Benjamins Publishing Company.
Pier, John (ed) (2004). The Dynamics of Narrative Form: Studies in Anglo-American Narratology. Berlin: de Gruyter.
Poncharal, Bruno (2003). Divergences nonciatives et stylistiques au DIL en anglais et
en franais. G. Mathis et al. (eds). Stylistique et nonciation: le cas du discours indirect libre. Nanterre: Publications de luniversit de Paris 10, 24966.
Rabatel, Alain (1997). Une histoire du point de vue. Paris: Klincksieck/CELTED.
(1998). La Construction textuelle du point de vue. Paris: Delachaux et Niestl.
(2001). Fondus enchans nonciatifs. Scnographie nonciative et points de vue.
Potique 126, 15173.
(2003a). Une lecture nonciative des hypothses aspectuo-temporelles et commentatives dans les suites PS + IMP : point de vue du locuteur ou de lnonciateur ? Journal of French Language Studies 13:3, 36379.
(2003b). Le point de vue, entre langue et discours, description et interprtation :
tat de lart et perspectives. Le point de vue. Cahiers de Praxmatique 41, 724.
(2003c). Entre usage et mention : la notion de re-prsentation dans les discours
reprsents. R. Amossy & D. Maingueneau (eds). Lanalyse du discours dans les
tudes littraires. Toulouse: Presses universitaires du Mirail, 11121.
(2004). Argumenter en racontant. Brussels: De Boeck.
(2005a). La part de lnonciateur dans la construction interactionnelle des points de
vue. Marges linguistiques 9, 11536, <http://www.marges-linguistiques.com>.
(2005b). Une catgorie transversale, le point de vue. Le franais aujourdhui 151,
5768.
(2008a). Homo narrans. Pour une analyse nonciative et interactionnelle du rcit.

Tome 1. Les points de vue et la logique de la narration. Limoges: ditions LambertLucas.
(2008b). Homo narrans. Pour une analyse nonciative et interactionnelle du rcit.

Tome 2. Dialogisme et polyphonie dans le rcit. Limoges: ditions Lambert-Lucas.
98
Alain Rabatel
(2009). Pour une narratologie nonciative ou pour une approche nonciative des
phnomnes narratifs ? J.-M. Schaeffer et al. (eds). Narratologies contemporaines.
Paris: CNRS ditions (forthcoming).
Rivara, Ren (2000). La langue du rcit. Introduction la narratologie nonciative. Paris:
LHarmattan.
Rosier, Laurence (1999). Le discours rapport. Brussels: Duculot.
Traduction cumnique de la Bible [TOB] (1975). Paris: Cerf.
Vitoux, Pierre (1982). Le jeu de la focalisation. Potique 51, 35968.
Vuillaume, Marcel (2000). Le style indirect libre et ses contextes. Cahiers Chronos 5,
10730.
GUNTHER MARTENS
(Ghent; Brussels; Antwerp)
Narrative and Stylistic Agency: The Case of Overt Narration

1 Introduction
This article mainly discusses aspects of style (rhetorical figures, imagery)
in relation to the question of agency. Narrative theories distinguish various textual agents such as narrators, focalizers, focal characters or
reflector-characters and invest them with differing degrees of autonomy, power and causation. This article first reviews some recent cognitive
attempts to turn formerly central stylistic elements of narratorial agency
into triggers and even objects of internal focalization. On the basis of this
inquiry, I argue that any account which treats narratorial discourse as a
kind of default stylistic norm is reductive in terms of stylistics. Its reductiveness becomes evident when marked stylistic and rhetorical features of the narratorial discourse are seen as intrusive interventions and
authoritative points of view of the narrator and thus infractions on the
carefully perspectivized text. This conception of narratorial agency,
which I call institutionalist, can be remedied by approaching overt aspects
of narration from a rhetorical and stylistic point of view. In order to
illustrate this approach, I address a number of cases in which stylistic features complicate the establishment of perspective, features that can then
be described as part of a lesser-known, more encompassing register of
overtness.
2 On Triggers of Internal Focalization

Existing narratological approaches to the establishment of (internal) perspective encounter problems when attempting to link narratorial agency
with stylistic elements. This can be illustrated by referring to the remarkable discussion surrounding the initial sentence of Joyces Eveline in
Dubliners: She sat at the window watching the evening invade the avenue (Joyce 1961: 34). According to Chatman, the metaphor invade
100
Gunther Martens
cannot belong to the diction of the protagonist. Because the metaphor is

said to reflect an inventive, literary mind, it is considered to be the narrators or even the authors infraction on the characters perspective. In
other instances, the text conveys how the character would phrase or verbalize that perception in Evelines true idiom, which is characterized by
colloquialisms (the man out of the last house) and strong deictic
anchoring (such as exclamations) (cf. Chatman 1978: 20406). Attridge,
however, claims that the rhythms of the patterned sounds (evening invade, avenue) are in fact graceless, the metaphors dead and the diction
commonplace. (Attridge 1990: 5) Echoing Genettes typically structuralist argument for internal focalization (the substitution test, cf. Genette
1972: 210; 1988: 194), Attridge argues that the passage can be translated
into first-person present tense with no difficulty, thus ruling that the
narrators style has given way to one that mimics the speech and thought
patterns of the character (Attridge 1990: 5). Other theorists (cf. e.g.
Fger 1993: 51; Jahn 1997: 46164; Rimmon-Kenan 2002: 51) have
continued to refine the account of why this passage is internally focalized
from the very beginning, linking the discussion about what one can assume the character to know or command with detailed attention to shifts
in the anchoring of the deictic center. Increasingly, however, aspects of
free indirect discourse that indicate the lingering indirectness of the
framing narration are interpreted as a key to the internality of a focalization, taking familiarizing articles and perception indicators (such as
watching, she heard) to be marking transitions to figurally focalized
passages (Jahn 1997: 463), while imagery is delegated to narratorial intervention 1 . From a cognitive point of view, Jahn even makes a case for
the view that the frame of internal focalization, once cognitively established, easily overrules the remnants of the narrators perceptibility. In
this register, Jahns most adventurous claim is that even conspicuous elements of the narratorial discourse pointing to indirect speech, such as the
inquit and percepit formulae, will not only act as triggers (and hence, as
an identification of objects) of internal focalization, but can also be integrated within that internal focalization. I aim to question Jahns sugges1
Narratorial localization typically uses descriptive imagery while reflector-mode focalization is usually cast in a mind style comprising referentless pronouns, the familiarizing article, minimized narratorial perceptibility, in actu presentation. (Jahn 1996:
257)
101
tion that these incipits are standardized and have lost their framing trigger
function. I will do so by illustrating that this level of indirectness can expand into a stylistic register with relevance to the agency of narration. By
arguing that even direct discourse and inquit formulae are amenable to internal perspective, Jahn tries to evacuate the site of the narrators agency
by giving a receptive twist to it: thus Mrs. Dalloways mind is [] only
minimally involved in the conversation, passively registering her own
automatic questions (Jahn 1997: 456). In a later article, Jahn (1999: 104)
specifies the much-needed excuse for treating third-person, past-tense
passages such as [a similar, typical Hemingway-passage of covert figuralized narration] as fully internally focalized segments: it is provided
by backgrounding deictic residue. Specific (modernist) texts featuring
stream of consciousness (such as Mrs. Dalloway) may strive towards that
evacuation, but this does not prevent other (not only older, even
modernist and postmodernist) texts and genres from experimenting with
types of narration with a former particular historical sedimentation 2 .
From a rhetorical point of view, it is striking to note that the metaphoric element invade continues to be discussed in terms of diegetic
appurtenance (cf. Genette 1972: 48) or rooting (cf. Stanzel 1979: 297).
As such, its discussion remains restricted to the circumscription of idioms
as guidelines for attributing individual voices and visions: Jahn proposes
to consider invade as a piece of the puzzle of ones mind (Jahn 1997:
462) in tune with the characters feeling of being beleaguered (462) by
impressions. By extending the isotopy as the characters frame of reference, Jahn gives a receptive twist to metaphor as an experiential
repertoire in line with Lakoff and Johnsons conceptual metaphor. Nevertheless, Jahns suggestion testifies to a rather resultative approach that
tends to disregard the narrative function of indirect, stylistic aspects such
as figurative patterns.
To conclude this section, a number of ideal-typical positions can be
distilled from this debate. On the one hand, the substitution test for internal focalization (cf. Genette 1972: 210, with reference to Barthes 1966:
40), which points to the structuralist legacy of narratology, remains inconclusive with regard to stylistic agency, precisely because structuralism
saw the radical formalization and construction of the object as one of its
2
In fact, Jahn (1999: 105) allows for such more active interferences between narratorial
and reflectorial conceptualizations to be grouped under Stanzels label colouring.
102
Gunther Martens
strengths. Aczel rightly criticized that in classical narratology the central

criterion for overtness remains the speakers relationship to the discourse
spoken (i.e. the position, level and time from which the story is told), and
not the quality of the discourse itself. Distinctive idiomatic traits are not
seen as indicators of voice (Aczel 2005: 635) 3 . According to Fludernik,
most approaches to perspective or focalization unfortunately (Fludernik
1993: 75) continue to stress the transformational nature of free indirect
discourse as a kind of linguistic clue from which to infer a clear-cut substitution of agency or an effacement of narrative command. On the other
hand, the cognitive models which continue to advance refinements and
nuances in the register of internal focalization tend to diminish the agency
of the narrator as such in favour of a discussion of narrative in terms of
embedded narratives: Real readers conceive of canonical fictional
language (that is, narrative without a narrator) as self-constituting rather
than emanating from a fictional teller (Galbraith 1995: 32). Jahns argument is somewhat akin to Galbraiths position, which he quotes approvingly (cf. Jahn 1999: 10304). Jahn is of course right to criticize what
he considers to be the attributive fallacy and what shall be termed here
an institutionalist-logical understanding of the attributive frame.
However, the fact that Jahn resorts to the term conative solicitude (borrowed from Bonheim; cf. Jahn 1997: 456) is particularly telling. This
terms at least suggest that texts that are highly aware and expressive of
their attributing activity and of their orientation towards the addressee (in
Jakobsons term) are solicitous, i.e. overly and redundantly meticulous
in a register which in fictional texts goes without saying on institutional
and conventional grounds 4 .
In the following, I aim to group the extended range of stylistic expressivity under a less evaluative version of the term overtness. In doing so,
I assume that the conditions for covert narration have sufficiently been
detailed in the many models which add refinements to internal focalization. What is left out of Jahns cognitive account is precisely that some
types of narration do in fact foreground and expand the narratorial frame
3
In his rhetorical approach to fictionality, Walsh discusses the concept of voice in three
senses, as instance, as idiom, and as interpellation (cf. Walsh 2006: 89-101).
Although Bonheim needs to be credited for his awareness that authors like Henry
James are real acrobats of the inquit (Bonheim 1982: 78), Collier comments that in
fact Bonheim displays a covert endorsement of unobtrusive or minimal tagging
(Collier 1991/1992: 48).
103
and the concomitant indirectness, and that this expansion fulfils the added
narrative function of reflecting the act of delegation itself. Both
Chatmans and Jahns interpretive moves are variations of the similar assumption that the metaphor invade must presuppose either a narrator or
a character as agent or deictic centre. Jahns conception of ambient
focalisation (Jahn 1999: 98), which may be that of an overt communal
narrator who adopts or substitutes the perception of one or more characters, comes close to the descriptive-summary textual phenomena that will
be studied in section 5 of this article. Yet, despite the sophistication of
Jahns account, he continues to define in negative terms as vagueness and
lack of spatio-temporal situatedness what is in factin terms of stylea
highly profiled, interpellative dimension of narrativity. In what follows, I
aim to consider how reading in terms of overtness may thwart the very effort of ascribing such stylistic traits to either narrator or character, inviting
us to consider the concept of narrative agency in relation to that of
stylistic agency.
3 Concepts of Overtness: A Critical Round-Up

Before I go on to theorize perspective in relation to agency, some further
terminological clarifications are in place. First, I need to clarify what exactly is meant by overt narration. Overt narration (from the French ouvert,
in turn derived from Latin aperire) can comment both on the content of
the narration (story world) and on the narrating function itself; the address
to a narratee is a part of this meta-narrative performance (Fludernik
1993: 443). Especially heterodiegetic overt narration continues to be
associated with older texts 5 . Within the scope of this article, it is not
possible to discuss the conceptual history of overt narration. It needs to be
noted that in Stanzels version of overt narratorial mediacy, the speaker is
5
Prince defines overtness as a narrator presenting situations and events with more than
a minimum of narratorial mediation, an intrusive narrator. He only refers to classical
texts as evidence: Eugnie Grandet, Barchester Towers, Tom Jones, Tristram
Shandy (Prince 1988: 69). In a passionate attack on the neo-modernist leanings of
Genettes system, Sternberg argues that assumptions that omnisciences time has
passed are counterfactual and stated in blissful disregard for the evidence
(Sternberg 2008: 716). According to Sternberg, Genette and even more his followers
fashion zero-focalisation as marginal and uninteresting, if at all discussible (715),
whereas this licensed excess of knowledge (715) is rather common and not limited to
mythological or biblical, law-making narrative with its typical epithetic overtness.
104
Gunther Martens
necessarily personified. In Stanzels opinion, the authorial narrator leads

us to infer the role of a companiable, talkative, intrusive speaker very
similar to Wayne Booths dramatized narrator. Stanzel exempted from
what I would call the overt type of narrative mediacy summaries, gnomic
comments and synoptic titles (as typical of satirical literature), which he
regarded as a kind of degree zero of narrativity or even non-narrative
direct information. Stanzels theory features a second, covert type of
mediacy as the dominant trend of modern(ist) literature, in which the
reflector-character functions as a kind of autonomous medium, escaping
the narrators control as mediating agent. It is not so much the way in
which Stanzel juxtaposed these two types of mediacy, but rather the
organicist overtones of his terminology and its visualization that met with
critique 6 . Genette is less outspoken about overtness. While Genette
maintains that the structuralist-linguistic inspiration of his model as well
the specific case of Proust (with special reference to metaphor) cut across
and foreclose in principle any (backwards-)compatibility with the tellingshowing distinction (cf. Genette 1972: 188), a kind of ironic dialogue
with the standard definition of overt narration is carried out in the
footnotes. Although Genette playfully denegates a penchant for Flaubert,
James, and Hemingway at the expense of Fielding, Sterne, and Thomas
Mann (Genette 1988: 54), he distances himself ironically from the
authorial style of Fielding in a footnote (cf. Genette 1988: 95) and finally
leaves the business of stylistics (which is still relevant to Stanzels
definition of colouring) to linguistic predecessors (52). Both Stanzel
andto a lesser extentGenette take into consideration overtness,
accompanied either by aperspectivism or zero-focalization. But in both
cases, the aforementioned telling-style utterances are mainly held to
affect and thematize the status of fictionality, they are judged to display a
typified, standardized capacity whose actual textual surface is only
relevant in quantitative terms of presence and visibility (see also Martens
[2008]). Apart from being embedded in a different interpretation of the
rhetorical tradition, the apotropaic gestures towards extreme summary
mentioned above may have a common ground: In non-fictional narrative,
telling-style phenomena like summary, abstract and metanarrativity are
very much linked to their institutional speech act functions: they are
6
In this respect, see also the thoughtful contribution to this volume by Tatjana Jesch and
Malte Stein, pages 5977.
105
meant to signal that instead of turn-based interaction, a longer stretch of

narrative report is coming up next (cf. Zipfel 2001: 130). There is no explicit need to signal this in fiction, although several authors (from Sterne
to Robert Walser) see stylistic opportunities in questioning this institutional self-evidence.
Although it could be argued that neutral functional descriptions of
narrators as instances of mediation do exist (cf. e.g. Nnning 1997), especially theories that conceptualize the role of the narrator in terms of a
more neutral logical necessity (cf. Ryan 1981), a set of narratorial functions (cf. Ryan 2001) or a deictic field of orientation, tend to assume that
the narratorial level of transmission or attributive quotationality functions
as a kind of stylistic default. Basically, they tend to discuss narratorial
discourse as a neutral and merely functional ground on which the peculiarities of characters diction in passages of free indirect discourse are set
into relief (Fludernik 2001b: 627). This conception entails the institutionalist-logical account of narratorial agency as exemplified in the discussion of Eveline. It may explain the basically substitutional quest for
either plain production sentences as expressive of the narrators idiom
or plain action sentences as embedded focalization. This conception of
agency inevitably pitches the local and particularized agent of focalization
against the logical but (ideally) reduced agency of the overall narration.
This conception ultimately rests on the assumption that vision is a more
flexible, individualized and diverse (multi-facetted) quality (cf. Warhol
1996; Horstkotte 2005), whereas the prominence and indirectness of the
overt narratorial framing discourse is deemed potentially detrimental to
the readers immersive activity: On account of its proximity to a voice of
a collective subject, [] offered in a mode of persuasion (Culler 2004:
31), Culler assumes it to be more prone to become the mouthpiece of
generalization, typification and standardization 7 . In fact, Chatmans own
usage of the term overtness, as illustrated in the Eveline example above,
and his explanatory distinction between the narrators slant and the
characters filter promptly testify to the institutionalist conception of
agency illustrated above: slant casts the narrators attitudes and other
mental nuances as a kind of ontological attitude and bias which the
poor narrator cannot really help. Slant captures the psychological,
sociological, and ideological ramifications of the narrators attitudes,
7
See, however, Nnnings concept of secondary mimesis (cf. Nnning 2001: 21).
106
Gunther Martens
which may range from neutral to highly charged (Chatman 1990: 143).
Filter, on the other hand, is intended to give a more positive and active
twist to the much wider range of mental activity experienced by
characters in the story worldperceptions, cognitions, attitudes, emotions, memories, fantasies, and the like (Chatman 1990: 143). Shifting
away from the polemical overtones used to treat overtness in terms of
(omni)-presence, 8 Fludernik rightly stressed the local nature and
functional flexibility of the signals responsible for the establishment of
internal perspective through the deictic centre (Fludernik 2001a: 107).
The same focus on small-scale linguistic features that trigger readers
establishment of narrators, reflectors, and observers (10203) needs to
be applied to the markers of overtness. Unlike first-person overt narration,
which is currently widely discussed in the instantiation of unreliable
homodiegetic narration, anonymous overt narration often refuses or fails
to pay off in terms of extracting or divinating both the real persona of
the narrator or in extracting the real past course of events and causalities
obfuscated in the overly phatic communication itself. The overt comment
on the performance may take the form of an axiological overdetermination (Schmid 2005: 79) without growing into a personified
presence. It is strictly this type of non-personified overtness I will be
concerned with in this article.
While personified, psychologically motivated overtness has been
widely discussed with regard to homodiegetic or unreliable narration, its
anonymous counterpart, which is prominent but nevertheless lacking in
clear-cut spatio-temporal situatedness and individuating features (Nnning 2001: 212) cannot adequately be discussed in terms of experience or
individual perception. What threatens to get obfuscated by this stress on
the mental and the receptive is precisely a neutral assessment of the interactive nature and function of overt narration, which can be said to occur
no less frequently 9 . In conversational story-telling, a narrator, instead of
8
In Chatmans view, an overt narrator is necessarily a distinct, personified agent. The

overt narrator writes of the burdens of authorship, modestly disavows artistic competence, speaks freely of the need to push this narrative button, tip that lever, and apply a
brake now and then []. This squeaky machinery may annoy us if we are overly
committed to the smoothly purring Jamesian style, and it hardly inspires a profound
contemplation of the nature of narrative artifice (Chatman 1978: 248).
In this respect, see also Nnnings foregrounding of the narrator function (cf. Nnning
2001: 39).
107
setting out to embody or quote the actual words and intonation of a

speaker (letting other people in on the actual experience), frequently
makes short shrift of this speech or thought representation by abstracting
the conversation, using epithets, metaphors and proverbs to characterize
the general speech habitus of the represented person rather than the actual
event or content. Such narrative performances are characterized by
dissociation, distancing, disengagement rather than by immersion. In this
respect, we enter the domain of distancing and irony, not devoid of
conative (but by no means imperative) and phatic aspects of bonding
(confiding) with the addressee at the expense of the reported experience.
In this type of narration, one can foreground the experiental clues it offers
in relation to an inferred speaker, but one can also highlight the extent to
which the level of ongoing communication (with its addressee-oriented
rhetoric) outweighs the past experience reduced to summary. By exempting from default narrativity the more resultative or evaluative summary
and thought report typical of overt and authorial narration, one risks to
ignore a dimension of style that is bound to fail the deictic substitution
test because it is established in a more dynamic, interpellative way. Both
Herman (cf. 2002: 365) and Palmer (cf. 2004: 57, 59) already warned that
this stress on internality may run the risk of implying a certain degree of
egocentricity in narratological terminology. Herman argues in favour of
resituating deixis in a wider, more sociocentric network of (more or less)
virtualized speech positions (366).
Ultimately, Hermans aim to discuss focalization in terms of experiential repertoires and competences differs from the present articles
concern to discern types of stylistic agency that may run counter to storyworld agency. In what follows, I will contrast the standard approach to
overtness (as intrusive, opinionated comment, hinging on an a priori personified speech position) with a possible alternative: according to this
alternative conception, readers may infer, on the basis of figurative constellations and other stylistic phenomena, marked shifts in agency which
supplement or even rival the more common (grammatical, deictic) indexes of narratorial agency. To that purpose, I shall discern a free-floating
dimension of stylistic overtness, which is more likely to obtain in the case
of (but not limited to) an overt profiling of the narratorial discourse. Such
a more flexible and dynamic understanding of narrative agency has
already been reached by focusing on areas in which the agency of narration is less dependent on prototypical historical instantiations, as in the
thorough theorization of hybrid you-narration by Monika Fludernik and
108
Gunther Martens
David Herman. In contrast to first or third-person narration, you-narration, by virtue of its double appeal to both the virtual-imaginable and
the particular, turns out to be rather indeterminate in terms of attributing
internality or externality. One of its striking features is that it challenges
readers to negotiate and compare the suggested experiential repertoires
(Herman 2002: 344) wavering between actual and virtual experience.
Below, I will not focus on you-narration nor on metaphor, which would
require a more sustained text-related approach. Instead, I will focus on
what I think poses a similar challenge to the narratological enterprise,
namely on less straightforwardly deictic elements of anchoring, as evidenced in rhetorical and stylistic aspects of narrative texts such as
summary characterization and praeteritio.
4 A Rhetorical Approach to Stylistic Agency

As one of the most straightforward and palpable elements of the establishment of internal perspective, one tends to rely on familiarizing elements such as etic (Stanzel 1979: 216) descriptions and the familiarizing article. That such familiarizing strategies may cease to function as
seemingly automatic cataleptic references and intuitive figural triggers
when set within texts of overt narration can be illustrated with reference
to the Swiss author Robert Walser. In many cases, Walsers narrators feel
the need to explicate straightforward deictic references by means of
parentheses.
the hot-headed lout of a Pole, that is, none other than our Master Steward happened to
surprise me at my quiet, attentive reading. (Walser 1990: 88)
Concerning a keg of the finest rye whiskey that, to the delight of the steward and that
of a certain additional personnamely to my very own, grinning, hand-rubbing
delightshowed up only to be closely inspected and quite thoroughly investigated
and examined by the two abovementioned important or insignificant personages, I
shall take care not to waste another word. (Walser 1990: 88)
While going in many, even contradictive evaluative directions, these utterances do not really allow for the inference of a troubled, yet consistent
individual mind. In fact, they point to an interface for which the interaction with the reader (our) is a constant task, sublimated in a quasi-automated display of codes of politeness, confession and self-referentiality.
A variation of this technique is particularly exploited in Walsers highly
overtly narrated Der Ruber (1925, published posthumously 1972), in
109
which the inescapably thetic and artificial quality of the figural trigger is
foregrounded:
The Robber now came to a house that was no longer present, or, to say it better, to an
old house that had been demolished on account of its age and now no longer stood
there, inasmuch as it had ceased to make itself noticed. He came, then, in short, to a
place where, in former days, a house had stood. These detours Im making serve the
end of filling time, for I really must pull off a book of considerable length, otherwise
Ill be even more deeply despised than I am now. (Walser 2000: 7475)
Walsers narrator stumbles over the fact that one assumes focalization to
be silently maintained: the accumulative utterances concerning an old
house that is no longer there markedly refuse to lend experiential credibility to this perception, which normally would familiarize the reader with
a real or virtual observer within the textual world 10 . The narrator goes to
great lengths to clarify that probably the house that was no longer there
had been demolished due to old age. The statement that the house had
ceased to make itself noticed is of course, by the time it occurs, a
performative self-contradiction from the readers point of view. The digressive legitimations lead up to the mock (metacompositional) motivation that the accumulation of verbal material is there for the book to be
long enough. Here the narrator slips into the present tense, with clear-cut
metaleptic overtones. The stylistic detours of the narratorial discourse,
motivated by the need to produce a book of considerable length,
continue to resonate with an agent covering a lot of distance in the storyworld. Throughout this novel, the narrator keeps on linking his (or, given
Walsers unsettling narration, one could also say her) own spatio-temporal position and motion as intradiegetic observer with the quantitative
aspects of the operation of both narrating and writingand ultimately the
readers progression through the novel. This is a classic and playful
case of overt narration: the reference to material writing acts as a comment on the performance of the narration. In addition, the quote illustrates
in a strikingthough ambiguousway that overt heterodiegeticextradiegetic narration can no longer simply be equated with performative authoritativeness (Culler 2004: 26). Although seemingly redun10
In order to make the stylistic contrast clear: Joyces Eveline unequivocally integrates a
similar reference to a disappeared object within the (however distal) deictic spatio-temporal frame of the character: One time there used to be a field there in which they
used to play every evening with other peoples children (Joyce 1961: 34).
110
Gunther Martens
dant, the contrived stylistic variation points to an underlying logic of

amplificatio which at times suggests that the quantity of the verbal and its
materiality playfully reflects and iconically mimics its content 11 . In
referring to the first Walser-passage, one may be reminded of Booths
attempts to demonstrate the impossibility of a narrative that remains fully
within the compass of a single character (cf. Booth 1989). However, the
latters hypothetical formulations were somewhat contrived and precisely
geared towards displacing the modernist canon. Instead, I argue that this
passage points towards lesser-known conative and otherwise rhetorical
aspects of narrative communication, which may obtain both in homo- and
in heterodiegetic narration.
The purpose of the analysis of the next example in this section is to illustrate that stylistic overtness does not necessarily signal a narration fully
in control of things. In fact, stylistic overtness may attach to narrators
whose stated ambition it is to remain covert. This can briefly be illustrated
by means of Gnter Grass novella Im Krebsgang. Grass novella installs
a homodiegetic narrator who aims to remain at a distance from the events
to be related: the narrator, a journalist by profession, survived the sinking
of the Wilhelm Gustloff in his mothers womb. In its multi-tiered frame
narrative, the novella displays an awareness of its attributive indirectness,
which is meant to caution against self-victimization and the retrospective
attribution of causality. Grass narrator does so by means of the master
trope of the crabs movement (seitlich ausscherend, scuttling sideways) as
well as with numerous caveats against the self-propelling and self-serving
nature of non-reflexive narrativization, summed up by the mechanical
activity of abspulen (cf. Grass 2002: 8, 54; unreeling). Nevertheless,
his attempt at distance is thwarted on the thematic level by the fact that
the narrators son Konrad turns out to have committed a crime inspired by
the historical events. Based on his apologetic stance towards antiSemitism and Nazi ideology, Konny killed an internet friend whom he
thought to be Jewish and who had defended the sinking of the ship as a
legitimate act of retribution. In this context, it is all the more striking to
see that the narrator momentarily lapses from his self-awareness and his
11
Snow does not fall lickety-split, but slowly, i.e. bit by bit, which means flake by
flake, down to the earth. (Walser 2002: 135) Originally in German: Schnee fllt
nicht Knall auf Knall, sondern langsam, d.h. nach und nach, will sagen flockenweise
zur Erde (Walser 1978: 369).
111
stylistic control of the narration. At a crucial point in the novel, the narrator summarizes his sons speech of defence in court: After that Konrad
offered a fairly vivid account of the state funeral rites in Schwerin. 12
Konrad refers to the historical event in Schwerin commemorating the
Nazi Wilhelm Gustloff, who had been murdered by a Jew. Although the
narrator says that Konnys evocation of this event was both ample as well
as bildhaft (colorful, vivid), the narrator carefully aims to bracket and
evacuate its ideological bias by means of the passing summary. In
addition, the meta-narrative utterance ablaufen continues to stress the
mechanical and foreseeable aspects of Konrads act of narrating 13 , and the
causative use of lassen strategically mixes quasi-metaleptic implications
of causation with connotations of allowing for, indulging. All the while
the narrators summary aims to convey that Konrads evocation of the
event is biased. The narrations conundrum at this point can be
summarized as follows: the narrators summary is there in order to avoid
giving the floor to an ideological type of discourse which it frames as
highly infectious and conative. Yet, in order not to do so, it has to make
use of the same linguistic strategies of partiality. While this might be
considered to be a momentary lapse, still setting off the narrators selfconsciousness from the tabloid-style sensationalism (Dye 2004: 481)
employed by Konny, it is hard to ignore the fact that, by similar means,
the novel frames not only the ideology but also the youths preferred
medium (internet communication) in its entirety as suspect, conatively
solicitous and liable to abuse.
Despite the narrators declared intention to remain unobtrusive, the
stylistic option discussed constitutes a kind of stylistic overtness which
leads the reader to reflect on the very mechanism of delegation and attribution. The strategic decision to reduce reflectorial delegation is rooted in
either frequent situational constraints (in conversational story-telling) or
in deliberate decisions of narrators not to spell out the dialogue or the in
actu presentation of thought and perception designated by Jahn as
mind-style (Jahn 1996: 257). One could interpret the ensuing texture as
another sign of the I, as a sign of the narrators judgmental or mental dis12
13
Danach lie Konrad den in Schwerin abgefeierten Staatstrauerakt ziemlich bildhaft

ablaufen (Grass 2002: 190). Here, as in the following, English translations of German
texts are mine, if not indicated otherwise.
Cf. [D]och setzte sich die Rede meines Sohnes wie selbstttig fort (Grass 2002: 191);
But my sons speech sped on, as if under its own steam (Grass 2004: 206).
112
Gunther Martens
tance, but that would ignore its dynamic nature as an interferential, compound phenomenon.
In literary studies, such strategic bracketing of access to a characters
mindset continues be associated with ironic and satirical purposes, although in fact recent research has shown that the narrative function of
stylistic agency extends well beyond this particular usage (cf. Biebuyck
2007). In the case of the acrobats of the inquit (cf. Bonheim) I will
briefly deal with in the following, it is quite obvious that the information
is not presented as a result of action or as we may have speculatively
come to learn (Culler 2004: 31), but by way of parentheses and appositions, highlighting iterative and typical features of characters that often
anticipate further developments. Nevertheless, such drastic parentheses
are even to be found in modernist novels normally considered to be devoted to the expansion of figural viewpoint(s):
Mamsell Jungmann, who was now already 35 years old and who could pride herself
on having withered away at the service of affluent circles []. (Mann, Buddenbrooks,
1901: 159)
Clarisses governessa family heir-loom, pensioned off in the honourable guise of
serving as an assistant mother []. (Musil, The Man Without Qualities, 1930: 902)
In its more moderate form, the iterative and incriminating description may
match the common, socially codified disregard of servants of that period.
In its extreme form of thematized narratorial command either enhancing
or disrupting fictionality, it climaxes in the topos of the narrator losing
sight of characters or forcefully throwing characters out of the novel.
And what about his brother Fritz? We make no secret of it that he does not interest us.
Not a single word from his mouth has been handed down to us. Even if it had been
handed down, it would not interest us. (Schneider, Schlafes Bruder, 1992: 51) 14
This kind of abbreviation mainly affects personae minores. Whereas the

appositions could be explained as a placeholder while a crucial phase in
the biography of a more important character is recounted, iterative characterization can also occur during center-stage events. This is the case in
14
For more violent ways of eliminating potential perspective-bearers with metaleptic

overtones, see Deupmann (2001) on Doderer. In most cases mentioned so far, the iterative description of flat characters is accompanied by hypothetical focalizations: see
also the exit of the midwife in Schneider (1992: 1920).
113
Musils The Man without Qualities, when the attempt to find a single
unifying idea for the modern age is abandoned:
a Frau Weghuber, a manufacturers wife with an impressive record of charitable
works and quite impervious to any idea that there might be something more pressing
than the objects of her concern, rose promptly to her feet. She advanced a proposal for
a Greater Austrian Franz Josef Soup Kitchen to the meeting, which listened politely.
[] Had they [=those present] been asked on their way to this meeting whether
they knew what historical events or great events of that sort were, they would
certainly have replied in the affirmative; but confronted with the weighty imperative
of making up such an event on the spot, they slowly began to feel faint, and something
like rumblings of a very natural nature stirred inside them. (Musil 1930: 18384;
italics added)
In this case, the apposition is underscored by the ensuing hypothetical focalization which, through its stylistic dissonance, highlights the impossibility for the characters to articulate or even allow for the recognition of
this mundane feeling (i.e. of hunger) themselves. This summary characterization may seem outrageously biased against the character(s) and
may strike many readers as parody. However, such appositional shortcuts
possess a jarring, satirizing effect of paralepsis only in texts that adhere to
presenting events in a more or less realistic way and that abide by the
unity and primacy of the storyworld as the focus of narrativity. In the case
of Musils meta-novel, however, this narrative short-circuiting even
affects the protagonist. Although his mindset is rendered more consonantly, the protagonist is not exempted from similar abbreviations, such
as the lopsided and quasi-tautological reference that two weeks later he
had had a lover for fourteen days (Musil 1930: 26). While expressive of
the modernist suspicion of the retrospective establishment of narrative
causality, the overtness goes to signal that there is in fact no unpredicated
version of a reality that one could get to know otherwise 15 .
One could of course argue that diegetic summary, focusing exclusively on the commonality of the topic and ignoring individual variations
of manner and inner verbalization, with its inherent tendency towards
typification, schematization of recognizable shared stances, perspectives,
views or common opinions (Margolin 2000: 605) is straining the nature
of narrativity as such. In fact, it has been argued, that its occurrence is in
fact more widespread in (hybrid) essayistic writing, sociological and
15
Since I deal with this at length elsewhere, the reader is referred to the examples and
markers of overtness I quote and discuss in Martens (2006).
114
Gunther Martens
historical discourses (Margolin 2000: 606) or even more appropriate to

journalism or lyrical writing (cf. Margolin 2001: 253). Ultimately, this
touches upon culturally determined preferences, i.e. the question whether
we see the narrative representation of reality as emanating from individual
mental activity rather than vested in a collective interaction, ridden with
argumentative and imaginative moves as a kind of contest of perspectives.
Stylistic overt agency comprises more than the conventionalized illusion-shattering display of personified narratorial agency. There is a wide
variety of less spectacular forms of authorial mediation that obtain more
than just a metafictional or metacompositional valency: the use of parenthesis and apposition (cf. Martens 2006), distancing appellations (Cohn
2000: 137), alienating comments (143) and various degrees of nominalization and appropriation (cf. Fludernik 1993) need to be taken into
account in order to systematize the effects of narratorial overtness. Diegetic summary, content paraphrase (Rimmon-Kenan 2002: 109), and we
could add: gnomic we-utterances can be a narratively productive undertaking too 16 . More than often in those cases representation is schematic
in content and projected rather than quoted in its verbalization (Margolin
2000: 606), yet, as illustrated in the passage by Gnter Grass, such
rhetorical and stylistic features may encode an alternative version of
agency, thus acting as a kind of prosody of narrative performance or a
diacritical surplus adding focus to the information distribution. This allows for the conclusion that the scope of overtness is finally realized to
be greater than the occasional slippage into the first person (Aczel 1998:
472). Elements of stylistic agency may reinforce the profile of an individual narrator, but they may also refuse to reinforce deictic localization; in
such case, they rather supplement or even contradict the construction of
well-defined focalizers or narrators (e.g. when imagery fails to be classified as either diegetic or non-diegetic). If one accepts the idea that stylistic agency may reshape the profile of narrative agents (cf. Martens &
Biebuyck 2007: 355-356), this phenomenon can be treated as the paranarrative dimension of narrative communication, which in itself does not
16
Whilst only gnomic utterances have made it into Stanzels theory of narrative, De
Temmerman (2007) sets out to reinsert the broader rhetorical techniques of
characterization (such as chreia) as outlined in the progymnasmata into narrative
studies.
115
result in the construction of a distinctive agent next to focalizers or narrators: figurativeness as second order agency
does not give rise to the construction of a second order agent, a paranarrator or a paracomposer. For it is clearly the recipient who performs or carries out the actions, even
though he or she is nothing more than the executor. [...] there is not one single narrating voice to be detected in the paranarrative; the figurative forms always entail a
multiplicity of voices. (Biebuyck 2007)
The paranarrative is similar, but not entirely assimilable to related concepts such as hypothetical focalization (cf. Herman) and the disnarrated
(cf. Prince). The paranarrative differs from the widely known aspects of
metanarrativity in that the former does not give rise to the inference of a
personified textual agent. It should have become sufficiently clear by now
that considering narratorial discourse as either an intrusive blockage or a
neutral stylistic default from which to distinguish character perspectives
poses problems when dealing with the stylistic aspects of overt narration.
When analyzing the agency of narration as a kind of interactional relay of
information, even stylistic elements acting precisely as an impediment to
the attribution of individualized perspective can be considered as functional parts of that interaction.
Conclusion
The outline presented here proposes a redefinition of the scope of overt
narration. This redefinition has taken as its point of departure the observation that the techniques and strategies of narrators going covert are
well-documented and currently even being expanded, whereas the conception of narratorial agency as either a stylistic default or an opinionated intrusion is itself difficult to reconcile with the objectives of stylistics
and rhetorical narratology. Instead, the criteria for stylistic overtness,
often unilaterally considered to be non-narrative schematic descriptions or
illusion-shattering disruptions of fictionality, contribute to alternative
versions of stylistic agency. Hence, this article has documented a number
of stylistic, rhetorical and narrative elements (especially in the domain of
summary and epithet-like descriptions) that are not straightforwardly deictic and yet relevant to narrative agency. These elements can be said to
primarily invite a reflexive reconsideration of the agency involved in the
act of narration. Further studies will need to systematize in what ways the
agency of narration relates to rhetorical and argumentative patterns pres-
116
Gunther Martens
ent in texts. As such, the hypothesis of stylistic agency is a topic in need

of further investigation with relation to a broader corpus 17 .
References
Aczel, Richard (1998). Hearing Voices in Narrative Texts. New Literary History 29: 3,
467500.
(2005). Voice. D. Herman et al. (eds). The Routledge Encyclopedia of Narrative
Theory. London: Routledge, 63336.
Attridge, David (1990): Reading Joyce. D. Attridge (ed). The Cambridge Companion to
James Joyce. Cambridge: Cambridge UP, 130.
Barthes, Roland (1966). Introduction lanalyse structurale des rcits. R. Barthes et al.:
Potique du rcit. Paris: Seuil, 1977, 157.
Biebuyck, Benjamin (2007). Figurativeness figuring as a Condenser between Event and
Action. How Tropes Generate Additional Dimensions of Narrativity. Amsterdam
International Electronic Journal for Cultural Narratology 4. Electronic publication
at <http://cf.hum.uva.nl/narratology/a07_biebuyck.htm>.
Bonheim, Helmut (1982). The Narrative Modes: Techniques of the Short-Story. Cambridge: Brewer.
Booth, Wayne (1989). Are Narrative Choices Subject to Ethical Criticism? J. Phelan
(ed). Reading Narrative: Form, Ethics, Ideology. Columbus: Ohio State UP, 5778.
Chatman, Seymour (1978). Story and Discourse: Narrative Structure in Fiction and Film.
Ithaca: Cornell UP.
(1990). Coming to Terms. The Rhetoric of Narrative in Fiction and Film. Ithaca:
Cornell UP.
Cohn, Dorrit (2000). The Distinction of Fiction. Baltimore: The Johns Hopkins UP.
Collier, Gordon (1991/1992). Style at the interface: Speech-tags. REAL 8, 33107.
Culler, Jonathan (2004). Omniscience. Narrative 12:1, 2234.
De Temmerman, Koen (2007). A Narrator of Wisdom. Characterization through gnomai
in Achilles Tatius. <http://www.princeton.edu/~pswpc/pdfs/detemmerman/030701.
pdf> (17.07.2008).
Deupmann, Christoph (2001). Ein fragwrdiges Kapitel. Erzhlte Gewalt und gewalthaftes Erzhlen bei Heimito von Doderer. Text+Kritik. Zeitschrift fr Literatur,
Vol. 150: Heimito von Doderer, 4856.
Dye, Elizabeth (2004). Weil die Geschichte nicht aufhrt: Gnter Grasss Im Krebsgang. German Life and Letters 57:4, 47287.
Fludernik, Monika (1993). The Fictions of Language and the Languages of Fiction: the
Linguistic Representation of Speech and Consciousness. London: Routledge.
(1996). Towards a Natural Narratology. London: Routledge.
17
The author is Postdoctoral Research Fellow of the Fund for Scientific ResearchFlanders. I wish to thank the anonymous reviewers and the editors for their suggestions.
117
(2001a). The Establishment of Internal Focalization in Odd Pronominal Contexts.

(2001b). New Wine in Old Bottles? Voice, Focalization, and New Writing. New
Literary History 32:3, 61938.
Fger, Wilhelm (1993). Stimmbrche: Varianten und Spielrume narrativer Fokalisation. H. Foltinek et al. (eds). Tales and their telling difference: Zur Theorie und
Geschichte der Narrativik. Festschrift fr Franz K. Stanzel. Heidelberg: Winter, 43
60.
Galbraith, Mary (1995). Deictic shift theory and the poetics of involvement in narrative.
J. F. Duchan et al. (eds). Deixis in Narrative: a cognitive science perspective.
Hissdale, NJ: Lawrence Erlbaum, 1960.
Genette, Grard (1972) : Discours du rcit. Figures III. Paris: Seuil, 65282.
(1988). Narrative Discourse Revisited. Tr. J. E. Lewin. Ithaca: Cornell UP
Grass, Gnter (2002). Im Krebsgang. Eine Novelle. Munich: DTV.
(2004). Crabwalk. Tr. K. Winston. New York: Harvest Books.
Herman, David (2001). Towards a Socionarratology. New Ways of Analyzing NaturalLanguage Narratives. W. van Peer & S. Chatman (eds). New Perspectives on
Narrative Perspective. Albany: SUNY Press, 21846.
(2002). Story Logic: Problems and Possibilities of Narrative. Lincoln: U of
Nebraska P.
Horstkotte, Silke (2005). The Double Dynamics of Focalization in W. G. Sebalds The
Rings of Saturn. J. C. Meister (ed). Narratology beyond Literary Criticism.
Mediality, Disciplinarity. Berlin: de Gruyter, 2544.
Jahn, Manfred (1996). Windows of focalization: deconstructing and reconstructing a narratological concept. Style 30, 24167.
(1997). Frames, Preferences, and the Reading of Third-Person Narratives: Towards
a Cognitive Narratology. Poetics Today 18:4, 44168.
(1999). More Aspects of Focalization: Refinements and Applications. GRAAT 21,
85110.
Joyce, James (1961). Dubliners. Harmondsworth: Penguin.
Mann, Thomas (1901). Buddenbrooks. Verfall einer Familie. Frankfurt a. M.: Fischer,
1975.
Margolin, Uri (2000). Telling in the Plural: From Grammar to Ideology. Poetics Today
21:3, 591618.
(2001). Collective Perspective, Individual Perspective, and the Speaker in Between: On We Literary Narratives. W. van Peer & S. Chatman (eds). New
Perspectives on Narrative Perspective. Albany: SUNY Press, 24155.
Martens, Gunther (2006). Beobachtungen der Moderne in Hermann Brochs Die Schlafwandler und Robert Musils Der Mann ohne Eigenschaften. Rhetorische und
narratologische Aspekte von Interdiskursivitt. Munich: Fink.
(2008). Revising and extending the scope of the rhetorical approach to unreliable
narration. E. Dhoker & G. Martens (eds). Narrative Unreliability in the TwentiethCentury First-Person Novel. Berlin: de Gruyter, 77105.
118
Gunther Martens
Martens, Gunther & Biebuyck, Benjamin (2007). On the Narrative Function of Metonymy in Chapter XIV of Heines Ideen. Das Buch Le Grand. Style 41:3, 34265.
Musil, Robert (1930). The Man without Qualities. Tr. S. Wilkins. New York: Picador,
1997.
Nnning, Ansgar (1997). Die Funktionen von Erzhlinstanzen: Analysekategorien und
Modelle zur Beschreibung des Erzhlerverhaltens. Literatur in Wissenschaft und
Unterricht 30:4, 32349.
(2001). On the Perspective Structure of Narrative Texts: Steps toward a Constructivist Narratology. W. van Peer & S. Chatman (eds). New Perspectives on
Prince, Gerald (1988). A Dictionary of Narratology. Aldershot: Scolar P.
Rimmon-Kenan, Shlomith (2002). Narrative Fiction: Contemporary Poetics. London:
Routledge.
Ryan, Marie-Laure (1981). The Pragmatics of Personal and Impersonal Fiction. Poetics
10:6, 51739.
(2001). The Narratorial Functions: Breaking Down a Theoretical Primitive. Narrative 9:2, 14652.
Schmid, Wolf (2005). Elemente der Narratologie. Berlin: de Gruyter.
Schneider, Robert (1992). Schlafes Bruder. Roman. Leipzig: Reclam.
Stanzel, Franz K. (1979). Theorie des Erzhlens. Gttingen: Vandenhoeck & Ruprecht.
Sternberg, Meir (2007). Omniscience in Narrative Construction: Challenges Old and
New. Poetics Today 28:4, 683794.
Walsh, Richard (2007): The Rhetoric of Fictionality. Narrative Theory and the Idea of
Fiction. Columbus: The Ohio State UP.
Warhol, Robyn R. (1996). The Look, the Body, and the Heroine of Persuasion: A Feminist-Narratological View of Jane Austen. K. Mazei (ed). Ambiguous Discourse:
Feminist Narratology and British Women Writers. Chapel Hill: U of North Carolina
P, 2139.
Walser, Robert (1972). Der Ruber. Roman. Frankfurt a. M.: Suhrkamp, 1986.
(1978). Winter. Das Gesamtwerk VIII. Verstreute Prosa I. (19071919). Frankfurt
a. M.: Suhrkamp.
(1990). Tobold II. Masquerade and Other Stories. Tr. S. Bernofsky. Baltimore:
The Johns Hopkins UP, 80100.
(2000). The Robber. Tr. S. Bernofsky. Lincoln: U of Nebraska P.
(2002). Selected Stories. Tr. C. Middleton. New York: NYRB Classics.
Zipfel, Frank (2001). Fiktion, Fiktivitt, Fiktionalitt. Analysen zur Fiktion in der Literatur und zum Fiktionsbegriff in der Literaturwissenschaft. Berlin: Schmidt.
DAVID HERMAN
(Columbus, OH)
Beyond Voice and Vision: Cognitive Grammar and

Focalization Theory
1 Introduction
1.1 Perspective, Construal, and Cognitive Narratology
This essay draws on ideas from cognitive linguistics to explore the structure and dynamics of narrative perspective1. More specifically, I suggest
the advantages of supplementing narratological accounts of focalization
with cognitive-grammatical research on construal or conceptualization
factoring in related ideas developed by Leonard Talmy (2000) under the
rubric of cognitive semantics. In the approach outlined here, perspective
takes its place among a wider array of construal operationsways of organizing and making sense of domains of experiencethat are anchored
in humans embodied existence and that may be more or less fully exploited by a given narrative. Besides attempting to recontextualize previous
frameworks for research on narrative perspective, my essay also has a
broader goal: namely, to suggest that by incorporating into the domain of
narrative analysis research exploring the nexus of language, mind, and
world, theorists can help promote the development of cognitive narratology as one of a number of postclassical approaches to narrative inquiry.
At issue are frameworks for narrative study that build on the work of
classical, structuralist narratologists but enrich that work with concepts
and methods that were unavailable to story analysts such as Roland
Barthes, Grard Genette, Algirdas J. Greimas, and Tzvetan Todorov du1
Herman (forthcoming a) sketches a less developed version of part of the analysis presented more fully in section 2 of this essay. I am grateful to Jeroen Vandaele for insightful comments and criticisms that led to revisions of the earlier version of the analysis and that also inform the expanded treatment given here.
120
David Herman
ring the heyday of the structuralist revolution2. Cognitive narratology constitutes one such framework, or rather cluster of frameworks, and in the
present essay I outline a cognitively grounded approach to the problem of
perspective to suggest directions for research in this wider, emergent area
of narrative inquiry3.
Further, my essay tests the descriptive and explanatory power of a cognitive approach to perspective by using as case studies both monomodal
print narratives (namely, stories from Joyces 1914 collection Dubliners)
and a multimodal text, namely, Daniel Clowess 1997 graphic novel
Ghost World4. By focusing on cognitive dimensions of focalization in different kinds of narrative texts, I suggest, story analysts can overcome limitations arising from the restricted corpora on which scholars working in
separate traditions of research have based their concepts and methods. In
particular, by cross-comparing how perspectives are represented and interpreted in different narrative media, theorists can explore the scope and relevance of ideas developed by cognitive linguists for stories not solely dependent on verbal language. My research hypothesis is that ideas from
cognitive grammar and cognitive semantics help illuminate perspectivetaking processes in narratives that exploit multiple kinds of semiotic resourcese.g., graphic novels that involve a coordinated interplay of verbal and visual information tracks. Hence, in the model proposed here, cognitive narratology should be viewed as a subdomain of the broader enterprise of cognitive semiotics (cf. Brandt 2004; Fastrez 2003)to which
cognitive linguistics also belongs. Cognitive semiotics is the super-category containing frameworks for studying how the use and interpretation of
sign-systems of all sorts are grounded in the structure, capacities, and dispositions of embodied minds. Cognitive narratology, for its part, studies
the design principles for narratively organized sign-systems, drawing on
2
For a fuller account of classical versus postclassical approaches to narrative theory, see
Herman (1999). For accounts of the structuralist revolution and of the way it shaped
structuralist theories of narrative in particular, see, respectively, Dosse (1997) and Herman (2005).
See Jahn (2005) for a synoptic account of developments in cognitive narratology; see
also Herman (2003 and forthcoming b).
Multimodal narratives can be defined as narratives that exploit more than one semiotic
channel to evoke a storyworld (for a fuller account, see Herman [forthcoming c]). On
methods for studying multimodality in textual artifacts and in face-to-face communication interaction, see Kress & van Leeuwen (2001) and Norris (2004), respectively.
Cognitive Grammar and Focalization Theory
121
tools from (cognitive) linguistics, ethnography, the philosophy of mind,

social and cognitive psychology, and other disciplines to explore the interfaces among narrative structure, semiotic media, and humans fundamental cognitive abilities5. I return to these distinctions and relationships in
my concluding section, using the analysis outlined in sections 13 as the
basis for a reassessment of the place of cognitive narratology within the
architecture of inquiry.
1.2 The Illustrative Narratives
To illustrate the advantages of moving from talk of focalization to talk of
conceptualization, and to suggest the productiveness of this approach for
the analysis of various kinds of narratively organized sign-systems, multimodal as well as monomodal, I focus here on two case studies. My first
case study, discussed in detail in section 2, involves a literary (i.e., monomodal print) narrativemore precisely, three stories from Joyces Dubliners (1914): Araby, Ivy Day in the Committee Room, and The Dead.
All of them completed between 1904 and 1906, the three stories form
something of a comparison set; I use them to redescribe contrasting modes
of focalization as, instead, alternative patterns of construal or conceptualization. The three stories respectively exemplify the first-person, authorial, and figural narrative situations described by Stanzel (1979); the stories
thus enable me to test out the implications of the model presented in this
paper for print texts using a range of strategies for encoding perspectives
on narrated situations and events. Araby is told retrospectively by a homodiegetic narrator revisiting his thwarted attempt to buy at a local bazaar
a gift for a girl with whom he had become infatuated. Ivy Day, which
takes place on election day as well as the anniversary of the death of the
Irish political leader Charles Stuart Parnell, is told almost exclusively
through dialogue and set in a meeting room where a group of men canvassing for the local candidates discuss the issues of the day. Finally, The
Dead is told in the third person but refracted through the vantage-point of
Gabriel Conroy, who undergoes an adventure of consciousness and a rethinking of his own attitudes and values when he learns about a former
5
See Herman (2007) for a fuller discussion of the challenges and opportunities of integrating cognitive narratology into the domain of interdisciplinary research on the mindbrainand vice versa. See also section 4 below.
122
David Herman
lover of his wife who (at least in Gretta Conroys interpretation of events)
died for her sake.
Section 3 of my essay then turns to a second case studyspecifically,
a single page from Daniel Clowess graphic novel Ghost World. In Clowess text, the coordinated interplay of two semiotic channels or information tracks, words and images, marks shifts in the vantage-point onmore
broadly, the construal ofrepresented situations and events. The narrative
focuses on two teenage girls trying to navigate the transition from high
school to post-high-school life, standing out contrastively against the
backdrop afforded by the tradition of superhero comics. Far from possessing superhuman powers, Enid Coleslaw6 and Rebecca Doppelmeyer struggle with familial and romantic relationships, resist the stereotypes their
peers try to impose on them, and are bought face to face, on more than
one occasion, with the fragility and tenuousness of their own friendship.
In this way, closer in spirit to the female Bildungsroman than action-adventure narratives, Ghost World, which was originally published as installments in the underground comics tradition and subsequently assembled into a novel, overlays a graphic format on content matter that helped
extend the scope and range of comics storytelling generally. My discussion of the illustrative page from Clowess graphic novel focuses on how
constellations of verbal and visual signs encode processes of construal that
are fundamentally isomorphic with those structuring monomodal print
texts. Analysis of word-image combinations in Ghost World thus reinforces the central claim of this essay: namely, that narrative perspective is
best understood as a reflex of the mind or minds conceptualizing scenes
within storyworlds. Accordingly, construal constitutes the common root of
voice and visionthe common denominator shared by types of narrative
mediation, no matter how many semiotic channels (or what specific channels) may be involved in the mediational process.
2 Beyond Voice and Vision: Rethinking Focalization Theory

In essence, the study of narrative perspective concerns how vantage-points
on situations and events in the storyworld are encoded in narrative discourse and interpreted as such during narrative processing. In this section,
after discussing foundational research on narrative perspective (2.1) and
6
As reported by Taylor (2001), Enid Coleslaw is an anagram for Daniel Clowes.
123
reviewing Jahns (1996 and 1999) own proposals about how to reconceptualize that earlier work (2.2.1), I explore how ideas from cognitive
grammar might enable narrative scholars to circumvent impasses created
by classical narratological theories of focalization (2.2.2 and 2.3). Narrative perspective, as I have suggested, can be interpreted as a reflex of
the mind or minds conceptualizing scenes represented in narrative texts,
such that construal becomes the common root of voice and vision. This
approach has wide-ranging consequences for previous accounts of perspective in stories7. For one thing, the focus of analysis shifts from taxonomy building, or the classification of types of focalization, to a functionalist account of perspective as sense-making strategy.
Joyces three stories constitute my main test cases in this section. Although I refer to the stories in their entirety, my discussion will use the
following three passages as touchstones or specific illustrative examples:
(a) Mr Hynes sat down again on the table. When he had finished his recitation [of The
Death of Parnell] there was a silence and then a burst of clapping: even Mr Lyons
clapped. The applause continued for a little time. When it had ceased all the auditors
drank from their bottles in silence. (Ivy Day, Joyce 1914: 135)
(b) I watched my masters face pass from amiability to sternness; he hoped I was not
beginning to idle. (Araby, Joyce 1914: 32)
(c) The piano was playing a waltz tune and he [Gabriel Conroy] could hear the skirts
sweeping against the drawing-room door. People, perhaps, were standing in the snow
on the quay outside, gazing up at the lighted windows and listening to the waltz music.
The air was pure there. In the distance lay the park where the trees were weighted with
snow. (The Dead, Joyce 1914: 202)
Then, in section 2.3, I draw on another passage from The Dead, represented as (d), to draw together the strands of my discussion and underscore the advantages of moving from classical narratological theories of
7
Likewise, Grishakova (2002 and 2006) richly synthesizes semiotic, narratological, and
cognitive-linguistic research to argue that Genettes voice and vision (perception)
are the two sides of the same process of sense-generation (Grishakova 2002: 529)
that perception is the common root of different modes of sense-production (verbal,
visual and others) (529). As I do in the present study, Grishakova draws on
Langackers ideas to underscore the parallelism of perception and conception and to
challenge Genettes understanding of focalization as pure perception, on the one
hand, and the existence of [...] non-focalized narration, on the other (Grishakova
2006: 153). See Broman (2004) for a comparable critique of Genettes attempt to drive
a wedge between narration and focalization.
124
David Herman
focalization to a postclassical account informed by Langackers cognitive

grammar and Talmys cognitive semantics:
(d) When he saw Freddy Malins coming across the room to visit his mother Gabriel
left the chair free for him and retired into the embrasure of the window. The room had
already cleared and from the back room came the clatter of plates and knives. Those
who still remained in the drawing-room seemed tired of dancing and were conversing
quietly in little groups. Gabriels warm trembling fingers tapped the cold pane of the
window. How cool it must be outside! How pleasant it would be to walk out alone,
first along by the river and then through the park! The snow would be lying on the
branches of the trees and forming a bright cap on the top of the Wellington Monument.
How much more pleasant it would be there than at the supper-table! (The Dead,
Joyce 1914: 192)
2.1 Narrative Perspective: Classical Accounts

In the narratological literature, the concept of focalization, originally proposed by Genette as a way to distinguish between who sees and who
speaks in a narrative, has generated considerable debate. In the Genettean
tradition, focalization is a way of talking about perceptual and conceptual
frames, more or less inclusive or restricted, through which participants,
situations, and events are presented in a narrative (cf. Prince 2003: 3132;
Herman 2002: 30130). Thus, in what Genette (1980) calls internal focalization, the viewpoint is restricted to a particular observer or reflector whereas in what he calls zero focalization (which Bal [1997] and
Rimmon-Kenan [1983] term external focalization) the viewpoint is not
anchored in a localized position. Also, internal focalization can be fixed,
variable, or multiple. Hence the focalization in Araby and The Dead
is, in Genettes terms, internal: as suggested by passages (b) on the one
hand and (c) and (d) on the other hand, the younger, Experiencing-I is the
focalizer in Araby whereas in The Dead Gabriel Conroy provides the
vantage-point on situations and events in the storyworld. Meanwhile, Ivy
Day (a) relies mainly on externalized views of the group of election
workers commemorating the anniversary of Parnells death. Hence,
whereas the focalization is fixed and internal in Araby and The Dead,
Ivy Day uses what Genette (as opposed to Bal and Rimmon-Kenan)
would term external focalization, in which what is presented [is] limited
to the characters external behavior [words and actions but not thoughts or
feelings], their appearance, and the setting against which they come to the
fore (Prince 2003: 32). There is, however, a departure from (what
Genette might call a paraleptic infraction against) this dominant code of
125
focalization when the narration dips briefly into the contents of Mr Croftons mind and reveals that he refrains from speaking because he considered his companions beneath him (Joyce 1914: 142).
So far, so good: the structuralist approach to focalization yields important insights into the contrasts and commonalities among texts like
Joycesand, in principle, among all texts categorizable as narratives.
Yet the classical picture of narrative perspective is complicated both by
(1) tensions between different approaches within the Genettean framework and by (2) a separate tradition of research stemming from the work
of Franz K. Stanzel (1979) on narrative situations, which is inconsistent
with or at the very least orthogonal to Genettes approach. In the first
place, Shlomith Rimmon-Kenan (1983) and Mieke Bal (1997) are among
the narratologists who argue that processes of focalization involve both a
focalizer, or agent doing the focalizing, and focalized objects (which can
in turn be focalized both from without and from within). Yet Genette
(1983) himself disputes these elaborations of his original account. Invoking Occams razor, Genette maintains that only the gestalt concept of focalization is needed to capture the modalities of narrative perspective8.
Stanzel for his part assimilates narrative perspective to the more general process of narratorial mediation, which he characterizes in terms of
three clines or continua: internal vs. external perspective on events, identity vs. non-identity between narrator and narrated world, and narrating
agent (or teller) vs. perceptual agent (or reflector). For example, the figural narrative situation, exemplified by The Dead globally and also locally in passages (c) and (d) above, obtains when a given stretch of narrative
discourse is marked by an internal perspective on events, a position toward the reflector end of the teller-reflector continuum, and non-identity
between narrator and storyworld. Authorial (=distanced third-person) narration, exemplified by passage (a), obtains when the discourse is marked
by an external perspective, a position toward the teller end of the tellerreflector continuum, and, again, non-identity between narrator and storyworld. More generally, whereas Genette and those influenced by him
8
Broman (2004) notes a further division among researchers working within the
Genettean tradition: namely, between those who follow Genette himself in developing a
global, typological-classificatory approach, whereby differences among modes of focalizations provide a basis for categorizing novels and short stories, and those who follow
Bal in developing the minute analysis of shifts in points of view between text passages
and sentences, and in certain cases even within the same sentence (71).
126
David Herman
strictly demarcate who speaks and who sees, voice and vision, narration
and focalization, the Stanzelian model suggests that the voice and vision
aspects of narratorial mediation cluster together in different ways to comprise the different narrative situations. Furthermore, for Stanzel, these aspects are matters of degree rather than binarized features. As the gradable
contrast between the authorial and figural narrative situations suggests, the
agent responsible for the narration can in some instances, and to a greater
or lesser degree, fuse with the agent responsible for perceptionyielding
not an absolute gap but a variable, manipulable distance between the roles
of teller and reflector, vocalizer and visualizer (cf. Shaw 1995; Nieragden
2002; Phelan 2001). Contrast Kate Chopins The Awakening, which
shuttles back and forth between the authorial and figural modes in order to
extrapolate general truths from internal views of Edna Pontelliers situation, with Franz Kafkas The Trial, which suggests the impossibility of
any such extrapolation by remaining scrupulously close to Josef K.s
position as reflector.
As even this cursory overview suggests, the lack of consensus or even
convergence among researchers after several decades of research in this
area, as well as the problematic incommensurability of the Genettean and
Stanzelian paradigms, points up the need to rethink foundational terms
and concepts of focalization theory itself. After providing in section 2.2.1
a brief overview of Manfred Jahns (1996 and 1999) innovative proposals
along these lines, in sections 2.2.2 and 2.3 I use ideas from cognitive linguistics to outline another strategy for reconceptualizing the study of narrative perspective.
2.2 Reframing the Classical Accounts
2.2.1 Jahns Model
Jahn (1999) has developed a powerful model of focalization based on folk
understandings of the structure of vision as well as the cognitive science
of seeing. Figure 1 reproduces what Jahn characterizes as a mental model
of vision (Jahn 1999: 87; cf. Jahn 1996: 242)a model grounded in how
we think we see things, as opposed to a precise mapping of the physiology
of vision. In this model, focus-1 corresponds to the burning point of an
eyes lens (Jahn 1999: 87) and also suggests an origo or vantage-point on
perceived scenes within a larger storyworld, i.e., a point at which all perceptual stimuli come together, a zero point from which all spatio-temporal
127
and experiential coordinates start (Jahn 1996: 243). Focus-2, meanwhile,

corresponds to the focused-upon object or scene within a field of vision
that is in turn nested within the surrounding environment or world.
Figure 1: Jahns Mental Model of Vision (Jahn 1999: 87)
Jahn (1999) builds on this basic model to suggest a scale of focalization possibilities, ranging from zero focalization (where no particularized center of consciousness filters the focused-upon events) to strict focalization of the sort found in first-person narration or figural narration
such as that used by Kafka. Figure 2 reproduces the scale at issue.
Figure 2: A Scale of Focalization Possibilities (Jahn 1999: 96)
The passages from Joyce quoted above would occupy different positions along this scale. For example, passage (b) would be located at the
rightmost position along the scale, it being strictly focalized through the
128
David Herman
vantage point of the Experiencing-I of Araby. Here focus-2, the schoolmasters face, is perceived from (or by [focus-1, i.e., the Experiencing-I])
under conditions of precise and restricted spatio-temporal coordinates
(Jahn 1999: 97). By contrast, passage (c) from The Dead can be located
at a position to the left of passage (b), since in addition to Gabriel Conroys perceptions the imagined perceptions of outside observers serve
briefly as a deictic center or vantage-point on the scene. Passage (a) from
Ivy Day, for its part, would need to be positioned to the left of passage
(c), somewhere in the vicinity of weak focalization, where all focalizing
agents, and with them all spatio-temporal ties, disappear, leaving only a
focused-upon object (Jahn 1999: 97).
In my next section I draw on ideas from cognitive linguistics to suggest another strategy for model-building in this contexta strategy likewise motivated by the dilemma of conflicting approaches to focalization
theory in its classical form (Jahn 1996: 241). Whereas Jahn rethinks
earlier accounts via mental models of vision, the heuristic framework outlined next emphasizes the inextricable interconnection between narrating
and perspective-taking. In other words, all storytelling acts are grounded
in the perceptual-conceptual abilities of embodied human minds.
2.2.2 From Focalization to Conceptualization
Building on studies by Langacker (1987) and Talmy (2000), among
others, the present section suggests how narrative analysts can move from
classical theories of narrative perspective toward a unified account of construal or conceptualization processes and their reflexes in narrative. Such
construal operations, which underlie the organization of narrative discourse, are shaped not just by factors bearing on perspective or viewpoint,
but also by temporal, spatial, affective, and other factors associated with
embodied human experience.
The basic idea behind conceptualization or construal is that one and the
same situation or event can be linguistically encoded in different ways, by
means of locutions that are truth-conditionally equivalent despite more or
less noticeably different formats (for a detailed overview, see Croft &
Cruse [2004: 4073]). Langacker (1987) suggests that a range of cognitive
abilities, including comparison, the deployment of imagery, the transformation of one construal into another or others, and focal adjustment, support the processes of conceptualization that surface as dimensions of semantic structure. In other words, these cognitive abilities are also design
129
parameters for language. A subset of the parameters at issuenamely,

those associated specifically with focal adjustmentderives from the enabling and constraining condition of having an embodied, spatio-temporally situated perspective on events.
A brief thought-experiment can illustrate the general processes at issue.
Assuming that (i)(vi) all truly apply to the same spatiotemporal configuration of participants and circumstances, differences among them reflect
humans ability to mentally construe one and the same situation in alternative ways:
(i) The family of raccoons stared at the goldfish in the pond.
(ii) The goldfish in the pond were stared at by the family of raccoons.
(iii) A family of raccoons stared at some goldfish in a pond.
(iv) The family of raccoons stared at the goldfish in the pond over there.
(v) That damned family of raccoons stared at the goldfish in the pond.
(vi) The family of raccoons stared at those damned goldfish in the pond.
(i) and (ii) show how alternate figure-ground relationships afford contrasting conceptualizations; (i) and (iii) how different locutionary formats
can represent different construals of hearer knowledge; (i) and (iv) how
conceptualizations can be more or less subjective in Langackers sense of
that term, i.e., include the situation of utterance more or less prominently
in the scene being construed; and (i), (v), and (vi) how different affective
registers can surface in alternative construals. Although cognitive grammarians tend to study such construal operations at the clause and sentence
level, my claim is that the operations themselves are scalable and can be
mapped onto discourse-level structures in narrative.
To return to the parameter of focal adjustment in particular, Langacker
(1987) identifies a number of sub-parameters relevant for the study of
how perspective shapes the construal of events. Combined with Talmys
(2000) account of perspective as a conceptual structuring system, Langackers account yields a rich framework for studying perspective-taking
processes in narrative contexts. Langacker decomposes focal adjustment
into the following sub-parameters (and sub-sub-parameters):
(1) selection, which concerns the scope of a predication, i.e. how much of
the scene that one is construing is included in the conceptualization;
(2) perspective, which includes
(2.1) figure-ground alignment, i.e., foreground-background relations (see
also Talmy [2000, 1: 31144]),
(2.2) viewpoint (=vantage point + orientation within a directional grid
consisting of vertical and horizontal axes);
130
David Herman
(2.3) deixis (deictic expressions include some reference to the ground or

situation of utterance in their immediate scope of predication),
(2.4) subjectivity/objectivity (for Langacker, the degree of subjectivity of
a construal varies inversely with the degree to which the ground is included in the immediate scope of a predication: the more the ground is included, the more objectivized the construal); and
(3) abstraction, which pertains to the level of specificity of a construal,
i.e., its degree of granularity (how much detail is included).
Meanwhile, in Talmys (2000) cognitive semantics, perspective constitutes a schematic system. On the basis of this system, languages establish
a conceptual perspective point from which [a referent entity can be] cognitively regarded (1: 68). In parallel with Langackers model, Talmys
account of the perspective system encompasses several categories or parameters that find reflexes in the semantic system of a given language
(1: 68-76), including
() the location of a perspective point within a referent scene;
() the distance of a perspective point from the regarded scene (distal, medial, proximal);
() perspectival mode, including motility, i.e. whether the perspective
point is stationary or moving, and mode proper, i.e., synoptic versus sequential viewing;
() direction of viewing, i.e., sighting in a particular direction (spatially
or temporally) from an established perspective point.
My larger point here is that classical theories of focalization, deriving
from the work of Genette and Stanzel, capture only part of this system of
perspective-related parameters for construal.
By shifting from theories of focalization to an account of the processes
and sub-processes involved in conceptualization, story analysts can explore how narratives may represent relatively statically (synoptically) or
dynamically (sequentially) scanned scenes (or event-structures). Scenes
will have a relatively wide or narrow scope, focal participants and backgrounded elements, an orientation within a horizontal/vertical dimensional
grid, and a more or less objective profile (i.e., encompass the ground of
predication to a greater or lesser extent). Scenes are also sighted from
particular temporal and spatial directions, and viewpoints on scenes can
be distal, medial, or proximal, that is, range from being far away to being
up close. Each such distance increment, further, may carry a default expectation about the degree of granularity (or level of detail) of the construal. Closer perspectives on scenes generally yield finer-grained (=more
131
granular, more detailed) representations; more distant perspectives generally yield coarser-grained (=less granular, less detailed) representations.
Analysts can investigate how these parameters for construal are realized textually (or, more broadly, semiotically)and in turn how particular
kinds of textual or semiotic cues guide readers efforts to parse narrative
representations into scenes that are variably structured, paced, and distributed over the course of a given story. Passage (a), for example, can be
redescribed as an instance of narrative discourse in which the conceptual
perspective point is static rather than dynamic and situated at a medial distance from the regarded scene, yielding a medium-scope construal of the
characters and their environment. Yet, despite the constant distance between the vantage-point on the scene and the scene itself, there is a shift in
the level of granularity of the representation: over the course of the passage, the focal participants move from particularized individuals (Mr
Hynes, Mr Lyons) to the characters viewed as a group (all the auditors).
Conversely, passage (c) (and also passage d, discussed in my next subsection) is remarkable for the way fluctuations in perspectival distance do not
affect the degree of granularity of the construal. Gabriel is at a proximal
distance from the drawing room, but as the sentential adverb perhaps9
indicates, his vantage-point is distally located vis--vis the scenes he imagines to be outside: namely, the quay and, still farther away, the park.
Yet there is no appreciable difference in the granularity of the construals
afforded by shifts along this chain of vantage-points. Working against default expectations about how much granularity is available from what perspectival distance, Joyces text evokes the power of the imagination to
transcend the constraints of space and timeboth here and again at the
end of story, when Gabriel imagines how the snow is general all over Ireland. The conceptualization processes portrayed in the story thus emulate
the spatio-temporal transpositions accomplished by Joyces own fictional
discourse; the concern in both contexts is the process by which one set of
space-time parameters can be laminated within another, to use Goffmans (1974) term. In other words, the scene outside the party becomes
proximate to Gabriels minds eye through the same process of transposi9
In Fauconniers (1994) terms, perhaps functions here as a space-builder, opening

within the storyworld an embedded mental space constructed by Gabriels imagination.
This space could also be characterized, in Paul Werths (1999) terms, as a subworld
within the text world evoked by Joyces story as a whole.
132
David Herman
tion that allows readers to relocate, or deictically shift (cf. Zubin & Hewitt
1995), to the spatial and temporal coordinates occupied by Gabriel as the
reflector through whom perceptions of the fictional party are filtered.
In passage (b), meanwhile, what is noteworthy are the cross-cutting directions of temporal sighting: the older, Narrating-I looks back on the
younger, Experiencing-I, whose observation of the increasingly dissatisfied expression on his schoolmasters face is in turn forward-oriented.
This bidirectional temporal sighting, the signature of first-person retrospective narratives (whether fictional or nonfictional), is complemented
by a combination of synoptic and sequential scanning. The passage reveals a construal of the masters face as undergoing change over time, but
the construal itself is summative, compressing into a single clause an alteration that one can assume unfolded over a more or less extended temporal duration.
2.3 Underscoring the Advantages of a Cognitively Grounded Approach
Drawing on the enriched analytic framework outlined in my previous subsection, theorists can ask questions about narrative perspective that could
not even be formulated within the classical models, while still preserving
the (important) insights afforded by Genettean and Stanzelian focalization
theory. The approach thus affords a more unified, systematic treatment of
perspective-related aspects of narrative structure that previous narratological research treats in a more piecemeal or atomistic way. These gains
can be underscored by a somewhat more extended analysis of one textual
segment, namely, the portion of The Dead excerpted as passage (d)
above.
In this passage, the Genettean analyst would speak of internally focalized narration; the Stanzelian, of narration in the figural mode. As he does
throughout the story, Gabriel functions in the quoted passage as the reflector. Accordingly, although the narrator remains distinct from Gabriel
(hence the use of the third person pronoun he), the narration is filtered
through Gabriels vantage-point on the scenes he encounters over the
course of the story. Further, drawing on the speech-category approach
to consciousness representation (cf. Cohn 1978; Palmer 2004: 5386), the
classical narratologist interested in tracing moment-by-moment shifts in
the perspective structure of the passage would be able to note the movement from actual to imaginary perceptions in the second half of the passage. In particular, the last four sentences of the passage feature imagina-
133
tive projections by Gabrielhypothetical forays into the way it is or

would be like outside the house where the party is taking place. For one
thing, the exclamation marks suggest sentiments or thoughts that have forcibly struck Gabriel, and that are therefore linked to his subjectivity rather
than the neutral, non-exclamatory discourse of the narrator. In addition,
the sequence of clauses containing verbs with modal auxiliaries (How
cool it must be..., The snow would be lying..., How pleasant it would
be..., etc.) exemplifies a process that linguists have termed the irrealis
modality. This modality encompasses all the semantic resources that enable language users to signal that they that are not fully committed to the
truth of a proposition about the world (cf. Frawley 1992: 38790). In this
case the main resource is the subjunctive mood signaled by the auxiliary
verbs. Coupled with the exclamation marks that express Gabriels subjectivity, the subjunctive indicates that Gabriel is again framing inferences
about the storyworld, but in this case inferences based on probabilistic
reasoning rather than on evidence to which he has direct, perceptual
access.
Beyond the study of such expressivity markers, however, recruiting
from Langackers and Talmys frameworks allows the analyst to capture
how the factor of perspective bears on a wider range of textual details, and
to uncover systematic interconnections among those details that remain
hidden when classical narratological approaches are used. In Talmys
terms, passage (d) reveals how Gabriels perspective constitutes a conceptual structuring system, in which Freddy Malins and his mother are, initially, the focal participants in a sequentially scanned scene. The past-tense
indicative verbs indicate that the scene is sighted from a temporal viewpoint located later on the time-line than the point occupied by the represented events. Spatially, the scene is sighted from a viewpoint situated on
the same plane as the represented action: Gabriel is not observing the
scene from below, for example, as is the case when he construes Gretta as
a symbol of something at the top of the stairs (Joyce 1914: 210)10.
Further, Gabriels initial medium-distance viewpoint on the scene (from
10
Likewise the factors of orientation and (spatial) sighting come into play in passage (c).
Gabriel first imagines others looking up at the lighted windows and listening to the
music in the house; then, mentally shifting to the deictic coordinates occupied by those
hypothetical outside observers, he imaginatively takes up their vantage-point and sights
the imagined scene in the park along a horizontal rather than vertical axis.
134
David Herman
the chair next to Mrs Malins) affords a medium-scope representation with

a corresponding, mid-level degree of granularity or detail. Then, when
Gabriel takes up his new position in the embrasure of the window, his distance from the scene increases, producing a wider-scope conceptualization
of the scene that has a correspondingly lower degree of granularity:
Gabriel construes the scene in terms of groups rather than individuals. As
they did in passage (c), then, the factors of distance, scope, and granularity
of construal co-vary systematically: as you get farther away from something, you see more of the context that surrounds it but with less overall
detail, and these perspectival constraints on peoples mental lives also
shape how they use languagefor example, how they produce and interpret narratives. Meanwhile, Gabriel has now moved much closer to the
window, his position affording a proximal, narrow-scope, and highly
granular, detailed representation of his own fingers tapping the cold pane.
The shift to free indirect thought in How cool it must be outside! marks
the onset of a new conceptualizationthis time of an imagined scene outside. As the new construal gets underway, distance, scope, and granularity
again co-vary: the hypothetical scene is farther away than the window, encompasses the whole area by the river and through the park, and is not envisioned in any detailed way. But then Gabriel imagines specific features
of the scene, the degree of granularity increasing dramatically to the point
where the snow on the branches of trees and on the top of the Wellington
monument comes into focus. Working against default expectations about
how much granularity is available from what perspectival distance,
Joyces text once again evokes, structurally as well as thematically, the
power of the imagination to transcend the constraints of space and time.
In short, in contrast with earlier focalization theory, a cognitive-grammatical approach points the way toward a more unified, integrative account of perspective and its bearing on other aspects of narrative production and processing, including stylistic texture (e.g., verb tenses and
moods), the spatio-temporal configuration of storyworlds, the representation of consciousness, and narrative thematics. A task for future research
is to consider other ways in which the idea of construal might afford new
foundations for narrative inquiryfor the study of how strategies for
telling are inextricably interlinked with strategies for conceptualizing the
world told about. In my next section, rather than venturing to explore
these further domains, I sketch the relevance of the approach outlined here
for transmedial narratology. Though cognitive grammar and cognitive semantics were of course designed for linguistic analysis, my argument is
135
that these frameworks are extensible for the purposes of cognitivesemiotic and also cognitive-narratological research, insofar as they point
to general capacities and constraints associated with embodied human
cognition. These capacities and constraints can be assumed to shape the
use and interpretation of all sign-systems, nonverbal as well as verbal,
whether narratively organized or not. I now turn to the perspective-indexing functions of word-image combinations in graphic narratives to test
this research hypothesis.
3 Perspective and Construal in Multimodal Narratives
My case study in this section is the page from Ghost World represented as
figure 3 (see next page). The visual-verbal organization of this page or sequence of panels encodes information about how the scene is being construed, and by whom, across the corresponding sequence of time-slices in
the storyworld. At issue is a temporally structured representation consisting of shifting figure-ground alignments, changes in the vantage-point or
location of the perspective point within the referent scene, and alterations
in perspectival mode and direction of viewing. Again, classical theories of
focalization capture only part of this system of perspective-related parameters for construal, and furthermore tend to be geared toward perspective-marking features of print texts. The model proposed here, by contrast,
allows story analysts to study how the logic of narrative perspective intersects with the constraining and enabling properties of particular modes
(=semiotic channels viewed as a means for the construction or design of a
representation) and media (=semiotic channels viewed as a means for the
dissemination or production of a given representation)11. In this way,
study of perspective-marking resources of different storytelling environments constitutes a key aspect of transmedial narratology (cf. Herman
2004 and forthcoming c). Theorists can hold constant the underlying, cognitively grounded system of capacities that supports narrative perspectivetaking, while comparing and contrasting how different storytelling environments (print texts, films, graphic narratives, plays, etc.) promote or inhibit the reliance on various elements of that system to encode perspective-based information in a given instance.
11
The distinction between mode and medium articulated here is drawn from Kress & van
Leeuwen (2001) and Jewitt (2006).
136
David Herman
Figure 3: Page from Daniel Clowes Graphic Novel Ghost World (1997: 26).
Copyright 2008 Daniel Clowes; courtesy of Fantagraphics Books (Fantagraphics.com).
137
In the opening panels of the page reproduced as figure 3, the texts wordimage combinations encode variable vantage-points on the scene, indexing construal processes operative at multiple levels. In the first panel, the
perspective point is encoded via the inclusion of Rebeccas and Enids
bodies within the first panel, putting the reader in a position of looking
over the characters shoulders in order to construe their own construal
processes; these processes form the common root of the verbal and visual
mediations of narrated content and find reflexes in the direction of the two
main characters gazes, the orientation of their torsos, and the structure as
well as the topical content of their utterances12. To reiterate: the perspective structure of both the first and the second panel situates the reader
at a point from which he or she can construe Rebeccas and Enids own
joint or at least coordinated acts of construal, which in turn center on the
non-focal characters second-order acts of mutual observation. Here processes of construal involve a telescopic chain of observational acts.
Note, however, that the two information tracks of the text feature different focal participants: the verbal track foregounds the visually backgrounded male characters, whose smaller size suggests their distal position vis--vis the orienting perspective point; by contrast, the visual track
represents Rebecca and Enid as the focal participants, thanks to their
larger size and implied proximity to the orienting viewpoint. In this instance readers are not likely to have any difficulty reconciling these information tracks within the larger perspective structure of the text. Other
multimodal narratives, however, might create more jarring discordances
as interpreters attempt to integrate reflexes of construal manifest in different information tracks. For example, the affective dimension of construal might be thematized (and thus de-automatized) by disjunctive information presented simultaneously through different semiotic channels,
as when a film soundtrack overlays on distressing, horrific images ebullient extradiegetic music, or vice versa.
To return to Clowes graphic narrative, as readers move to the third
panel on the page, they can use the context established by the visual
design of first two panels, together with the patterning of speech attributions in the form of word balloons, to draw an inference concerning the
12
To adapt Langackers terms: the ground of the characters discourse is placed within
the immediate scope of their predications, thanks to demonstrative pronouns and deictics in expressions such as see that guy and look behind him.
138
David Herman
status of the image represented in this third panel. Specifically, interpreters are likely to infer that this image of the former bass player is
mediated through the perceptions of one of the two main characters
most probably Rebeccas, given her physical location and the orientation
of her torso and gaze in the preceding panel13. That inference is reinforced
by the absence of a speech balloon in the third panel, even though the bass
player is shown talking on the phoneand even though the narrow-scope
or proximal representation makes the male character the focal participant
in the visual track. Here readers can assume that, because of the representation of the male characters location at the far side of the restaurant
in the first two panels, Rebecca cannot hear what he is saying on the
phonealthough he has acquired focal status in the domain of visual perception. By contrast, in the second panel on the page, in the case of the utterance represented by means of the leftmost speech balloon, readers can
assume that this remark was made within Rebeccas and Enids perceptual
range and is therefore included in the visual report of their perceptions at
this point in the unfolding action.
In panels 4 and following, the interplay of words and images prompts
readers to pull back from the internalized view of the ex-bass player in
panel 3 and adopt shifting perspectives during Rebeccas and Enids debate concerning what Rebecca characterizes as Enids impossibly high
standards for men. In a manner reminiscent of the shot/reverse-shot technique in cinematic narratives, the text first provides, in panel 4, an overthe-shoulder view of Rebecca from Enids perspective, followed in panel
5 by an over-the-shoulder view of Enid from Rebeccas perspective. Then
in panel 6 the perspective shifts again, to a more externalized view that
captures Enids angry expression as she defends her preference for the
cartoonist over the guitar plunkin moron (=ex-bass player) whom
Rebecca had alluded to favorably. Such shifts between perspective points
more or less proximally positioned vis--vis elements of the storyworld
are structurally homologous with (and arguably derive from the same cognitive capacities as) third-person or heterodiegetic narration that moves
13
Although it is arguable that this panel shifts away from the characters acts of construal
to a straightforward narratorial report of a moment of storyworld time, Rebeccas use of
the demonstrative pronoun in I just hate all these obnoxious, extroverted, pseudobohemian art-school losers might also be interpreted as a reference to a feature of the
storyworld that falls within the domain of the characters current perceptions.
139
along a spectrum from relatively more external to relatively more internal

viewsthat is, from external focalization (in Stanzels terms: the authorial narrative situation), where the vantage-point on events is not associated with a character in the storyworld, to internal focalization (in
Stanzels terms: figural narration), where the vantage-point is in fact a
characters.
Thus, in graphic narratives like Clowess both the design of individual
panels and sequential links between panels align readers with particular
vantage-points on the storyworld. More precisely, constellations of verbal
and visual cues encode shifts in the foreground-background relations
and fluctuating distances from focused-upon elements of scenesthat
structure the processes of construal represented within the text. Integral to
the interface between mind and world, and thus shaping the design and
use of all semiotic systems, these processes underlie narrative across media. However, as Clowess text also suggests, in multimodal narratives exploiting more than one information track, cognitive capacities and constraints associated with what Langacker terms the parameter of focal
adjustment (and what Talmy calls a perspective-based conceptual structuring system) can be distributed across more than one semiotic channel,
requiring interpreters to integrate these reflexes of perspective-taking
processes into a more or less coherent mental representation of a given
time-slice of the storyworld. A task for future research on narrative
perspective is to explore the processing mechanisms supporting integration of this sortand the relation of those processing mechanisms to the
ones brought to bear on manifestations of construal in monomodal print
texts such as Joyces.
4 Conclusion: Narrative Perspective, Cognitive Narratology,

and the Architecture of Inquiry
In the model sketched in this paper, cognitive semiotics is the broadest domain in which study of cognitive bases for perspective-takingthat is,
processes of construalcan be situated. In this domain, the question is
how the perspectival structure of embodied human experience finds reflexes in sign systems of all kinds, and how perspective-based information
is parsed out in monomodal versus multimodal texts (or communicative
practices). To restate the question: how do relevant cognitive capacities
and constraints affect the organization and sequencing of constellations of
signsin sign systems that may exploit either a single semiotic channel or
140
David Herman
else multiple channels to encode perspectivally enabled and constrained

construals of a storyworld? Treatments of perspective in cognitive linguistics are narrower in scope, focusing on how underlying cognitive abilities related to perspective are reflected in the organization of verbal language in particular. Meanwhile, cognitive-narratological research on perspectivethat is, research on the cognitive bases for perspective-taking
processes in narratively organized sign systems of all sorts, verbal as well
as nonverbal, monomodal as well as multimodalis situated at an intermediate level of generality. Compared with cognitive semiotics, cognitive
narratology takes a more targeted approach: it explores the perspectival
grounding not of sign systems as such but rather of sign systems that
exhibit a narrative profile. Compared with cognitive-linguistic research on
perspective, the cognitive-narratological approach is at once more general
and more specific: it is not restricted to the study of construal processes in
narratives conveyed through verbal language, but by the same token it
limits itself to how construal operates in narratively organized discourse
as opposed to language use more broadly.
Although the present paper explores ideas pertinent for each of these
domains of inquiry, its chief concern has been to begin characterizing
what a specifically cognitive-narratological approach to perspective might
involveto start detailing the distinctive scope, methods, and aims of
such an approach. As I hope to have demonstrated, further work in this
area will require closer scrutiny of the areas of intersection among processes of construal, dimensions of narrative structure, and the representational properties and capacities of the particular semiotic environments through which stories are told. In turn, integrative research of this
kind will have the added benefit of helping to establish the place of cognitive narratology within the architecture of inquiry. For the purposes of
further theory building, including but not limited to future work on cognitive aspects of narrative perspective, the following working definition
may suffice: cognitive narratology is the domain of study whose essential
concern is the nexus of narrative and mind, that is, the mind-relevant aspects of storytelling practices, whereverand by whatever meansthose
practices occur.
141
References
Bal, Mieke (21997). Narratology: Introduction to the Theory of Narrative. Toronto: U of
Toronto P.
Brandt, Per Aage (2004). Spaces, Domains, and Meaning: Essays in Cognitive Semiotics.
Frankfurt a. M.: Peter Lang.
Broman, Eva (2004). Narratological Focalization Models: A Critical Survey.
G. Rosshold (ed). Essays on Fiction and Perspective. Frankfurt a. M.: Peter Lang,
5789.
Clowes, Daniel (1997). Ghost World. Seattle, WA: Fantagraphics Books.
Cohn, Dorrit (1978). Transparent Minds: Narrative Modes for Presenting Consciousness
in Fiction. Princeton: Princeton UP.
Croft, William & Cruse, D. A. (2004). Cognitive Linguistics. Cambridge: Cambridge UP.
Dosse, Franois (1997). History of Structuralism, Vol. 1. Tr. D. Glassman. Minneapolis:
The U of Minnesota P.
Fastrez, Pierre (ed) (2003). Smiotique CognitiveCognitive Semiotics. Special issue of
Recherches en communication/Research in Communication 19.
Fauconnier, Gilles (1994). Mental Spaces: Aspects of Meaning Construction in Natural
Language. Cambridge: Cambridge UP.
Frawley, William (1992). Linguistic Semantics. Hillsdale, NJ: Lawrence Erlbaum.
(1980). Narrative Discourse. Tr. J. E. Lewin. Ithaca: Cornell UP.
Goffman, Erving (1974). Frame Analysis: An Essay on the Organization of Experience.
New York: Harper.
Grishakova, Marina (2002). The Acts of Presence Negotiated: Towards the Semiotics of
the Observer. Sign Systems Studies 30:2, 52953.
(2006). The Models of Space, Time and Vision in V. Nabokovs Fiction: Narrative
Strategies and Cultural Frames. Tartu: Tartu UP.
Herman, David (1999). Introduction. D. Herman (ed). Narratologies: New Perspectives
on Narrative Analysis. Columbus: Ohio State UP, 130.
(2002). Story Logic: Problems and Possibilities of Narrative. Lincoln: U of
Nebraska P.
(2004). Toward a Transmedial Narratology. M.-L. Ryan (ed). Narrative across
Media: The Languages of Storytelling. Lincoln: U of Nebraska P, 4775.
(2005). Histories of Narrative Theory (I): A Genealogy of Early Developments.
J. Phelan et al. (eds). The Blackwell Companion to Narrative Theory. Malden:
Blackwell, 1935.
(2007). Storytelling and the Sciences of Mind: Cognitive Narratology, Discursive
Psychology, and Narratives in Face-to-Face Interaction. Narrative 15:3, 30634.
(forthcoming a). Cognitive Approaches to Narrative Analysis. G. Brne et al.
(eds). Foundations for Cognitive Poetics. Berlin: de Gruyter.
(forthcoming b). Cognitive Narratology. J. Pier et al. (eds). The Living Handbook
of Narratology. Berlin: de Gruyter.
142
David Herman
(forthcoming c). Word-Image/Utterance-Gesture: Case Studies in Multimodal

Storytelling. R. Page (ed). New Perspectives on Narrative and Multimodality.
London: Routledge.
(ed) (2003). Narrative Theory and the Cognitive Sciences. Stanford, CA: Publications of the Center for the Study of Language and Information.
Jahn, Manfred (1996). Windows of Focalization: Deconstructing and Reconstructing a
Narratological Concept. Style 30:3, 24167.
(1999). More Aspects of Focalization: Refinements and Applications. GRAAT 21
(Issue Topic: Recent Trends in Narratological Research), 85110.
(2005). Cognitive Narratology. D. Herman et al. (eds). Routledge Encyclopedia of
Narrative Theory. London: Routledge, 6771.
Jewitt, Carey (2006). Technology, Literacy, and Learning: A Multimodal Approach.
London: Routledge.
Joyce, James (1914). Dubliners. New York: Penguin, 1967.
Kress, Gunther & van Leeuwen, Theo (2001). Multimodal Discourse: The Modes and
Media of Contemporary Communication. London: Hodder Arnold.
Langacker, Ronald W. (1987). Foundations of Cognitive Grammar, Vol. 1. Stanford:
Stanford UP.
Nieragden, Gran (2002). Focalization and Narration: Theoretical and Terminological
Refinements. Poetics Today 23:4, 68597.
Norris, Sigrid (2004). Analyzing Multimodal Interaction: A Methodological Framework.
London: Routledge.
Phelan, James (2001). Why Narrators Can Be Focalizersand Why It Matters. W. van
SUNY Press, 5164.
Prince, Gerald (22003). A Dictionary of Narratology. Lincoln: U of Nebraska P.
Methuen.
Shaw, Harry (1995). Loose Narrators. Narrative 3, 95116.
Stanzel, Franz K. (1979). A Theory of Narrative. Tr. C. Goedsche. Cambridge: Cambridge
UP, 1984.
Talmy, Leonard (2000). Toward a Cognitive Semantics, Vols. 1 and 2. Cambridge, MA:
MIT Press.
Taylor, Craig (2001). Girls World. The Guardian, November 3, 2001. <http://www.
guardian.co.uk/Archive/Article/0,4273,4290067,00.html>.
Van Peer, Willie & Chatman, Seymour (eds) (2001). New Perspectives on Narrative
Perspective. Albany: SUNY Press.
Werth, Paul (1999). Text Worlds: Representing Conceptual Space in Discourse. London:
Longman.
Zubin, David A. & Hewitt, Lynne E. (1995). The Deictic Center: A Theory of Deixis in
Narrative. J. F. Duchan et al. (eds). Deixis in Narrative: A Cognitive Science Perspective. Hillsdale, NJ: Lawrence Erlbaum, 12955.
BRIAN RICHARDSON
(College Park, MD)
Plural Focalization, Singular Voices:

Wandering Perspectives in We-Narration
We-narration and focalization is little studied and poorly known; apart
from the recent studies of Adelaide Morris, Uri Margolin, Celia Britton,
Joel Woller, and Amit Marcus, it has been largely unexplored in the
history and theory of narration. Nevertheless, I believe it is a compelling
type of perspective that has a substantial (if largely unknown) history, and
poses very powerful questions concerning basic assumptions of narrative
theory as currently understood. Though this is a somewhat new and fairly
uncommon practice in the West, in other cultures it is widespread, even
standard. Hertha D. Sweet Wong has analyzed collective speaking
subjects in traditional and recent Native American womens autobiography. She notes that a Native autobiographer, whether a speaking or a
writing subject, often implies, if not announces, the first person plural
weeven when speaking in the first person singular. We often invokes
a (sometimes the) Native community (Wong 1998: 171) 1 . She also notes
that contemporary Native American writers use the technique in a less literal manner, gesturing toward a larger community that is invoked rather
than depicted by the pronoun. As we will see, some indigenous South African speakers also employ first person plural narration.
It goes without saying that a first person narrator who frequently uses
the plural pronoun to denote the action of a group is not unknown in fiction. In a recent story by Eilis Ni Dhuibhne, for example, the narrator,
referring to herself and her sisters perspective, states, The only tolerable
part of the expedition occurred immediately after this, when we bought
1
Thus, one Yukon Native begins the story of her life with a history of her nation, the
histories of her mother and other close relatives, and the origin myth of her people.
She does not even get to her own birth until page 52 (and then it is buried in a long
list of her brother and sisters arranged in birth order). (174)
144
Brian Richardson
souvenirs at a stall [...]. There we stood and scrutinized the wares on

display: beads, statuettes, medals, snowstorms. Reverting to our consumerist role, we [...] do I mean I? I assume my sister felt the same about
it all [...] (Wong 1988: 117). This is an unproblematic case of first person perspective, with the appropriate epistemic boundaries of conventional fiction firmly in place. Not even a sister always knows what is passing
through the mind of her sibling. The purpose of this essay is not to examine the many kinds of usages of we in ordinary discourse, an analysis
that has been admirably performed by Margolin. Instead, the narratives I
am concerned with differ from natural narratives insofar as they produce a
tension concerning the identity, speech situation, or knowledge claimed
by the we voice. This occurs when, for example, an I pronoun is never used in the text, despite its appropriateness at many points, when the
focalization is unrealistically broad, or when the we-speaker possesses
knowledge that they cannot normally have acquired. These are distinctively literary uses and are not normally found in natural narratives.
The first sustained example of first-person plural narration seems to be
Joseph Conrads The Nigger of the Narcissus (1899). The work opens
in the conventional manner of standard, omniscient third person narrative.
Soon, a collective focalization is established through third person plural
reports of the crews perceptions. We-focalization does not appear until
a common bond develops among the seamen and we becomes the
primary perspective: We hesitated between pity and mistrust (36).
Conrad uses different types of narration in counterpoint to the
consciousness of the men he is depicting: the greater their cohesion, the
more insistent the use of we. At a few points later in the voyage, as
sailors retreat back into their own individual selves, the classic third
person form returns. It might be noted that collective focalization seems
more transgressive when it appears in we-narration as opposed to
they-narration, no doubt because a character narrator cannot know the
thoughts of other, under the conventions of mimetic narration.
We in this text refers to most of the crew, not the officers, the resentful Donkin, the West Indian James Wait, or old Singleton, the boatswain. The seamens perceptions of Donkin are represented in terms that
express difference: He stood on the bad eminence of a general dislike.
[...] Our sea-boots, our oilskin coats, our well-filled sea-chests, were to
him so many causes for bitter meditation: he had none of these things, and
he felt instinctively that no man, when the need arose, would offer to
share them with him (40). Intriguingly, in this passage the we-nar-
145
ration is focalized through the mind of Donkin, something of course a

first person narrator, singular or plural, is not supposed to be able to do.
As the narration continues, similar or shared thoughts are depicted as
if they were part of a single mind. This practice, however, grows more
odd (and less susceptible to a realistic recuperation) the longer it continues, as can be seen in the progression in the following passages: We
were appalled. We perceived that after all Singletons answer meant nothing. We began to hate him for making fun of us. [...] We suspected
Jimmy, one another, and even our very selves (43). Especially curious is
the claim that we suspected each other, since each individual is said to
be entirely united in distrusting the others. For the most part, we-narration and its peculiar kind of collective focalization appear to be entirely
reliable, a modest approximation of the third person narration elsewhere
in the narrative. At times, however, what might be called group unreliability strikes: the crew is convinced that Waits desire aided by
Wamibos presumed magic spells delayed the ship in the open seas (142).
At other times, the unreliability becomes self-conscious and openly acknowledged: we [...] sympathised with all [Waits] repulsions, shrinkings, evasions, delusions (139). That is to say, the we-perspective affirms what it wishes to believe even when it knows it is mistaken.
The we-perspective appears for the final time as the Narcissus
comes into its port (166). Once on land, the collective perspective in a
different manner as they-narration is used to depict the men in their last
moments together in the shipping office where they are paid off. And in
the last paragraphs of the novel, Conrad introduces his final, concluding
transformation, as an I-narrator suddenly intrudes into the text and continues the story seamlessly: Charley and Belfast wandered off alone. As I
came up I saw a [...] woman [...] fall on Charleys neck (170). Collective
narration and focalization are over; the narrator goes off by himself, with
nothing but his money, his memories, and an isolated consciousness.
Conrads use of we-narration transcends the strictures of realism in
the alternation of the we and they segments, since within a mimetic
framework, the we-voice cannot know the private thoughts of many of
the seamen. More audaciously, the two collective perspectives are not
clearly separated and often glide into each other. This juxtaposition of
mutually exclusive narrative stances can be seen prominently in passages
where a sentence of we-narration is followed in the next line by a third
person plural account: Our little world went on its curved and unswerving path carrying a discontented and aspiring population. They found
146
Brian Richardson
comfort of a gloomy kind in an interminable and conscientious analysis of

their unappreciated worth (103). From the perspective of a mimetic
theory of narration, the speaker either is or is not part of the group and
therefore one of the pronouns is entirely misleading. Other examples are
even more resistant to a realistic recuperation. As many Conrad critics
have pointed out, the entry into the consciousness of Singleton, Wait, and
others is incompatible with the narrators claims of having performed
physical action onboard: Groaning, we dug our fingers in, and very
much hurt, shook our hands, scattering nails and drops of blood (68). If
the narrator is a character on the ship, he cannot enter the minds of others
or report conversations he has not observed; if he is omniscient, he cant
break fingernails onboard. The narrator, that is, is simultaneously homodiegetic and heterodiegetic.
In the history of this narrative technique after Conrad, we find a number of repeated issues and concerns. The precise nature and identity of the
we is often varied and modulated as groups change or themes alter. The
we-form raises interesting issues concerning reliability: insofar as it is a
subjective form, it is enmeshed in issues of reliability and discordance,
but these are issues that are potentially different from those in first person
singular narratives since they may involve more accurate intersubjective
beliefs as well as communal misprisions or even mass delusion. In
addition, the we-perspective is used to present the collective consciousness of a tightly joined group who can be expected to share closely
aligned ideas and emotions, and the novels focus on groups like merchant
seamen, soldiers, revolutionaries, peasants, and even mice in the case of
Kafkas Die Sngerin Josephine. Henri Barbuses Feu, a 1916 novel
about World War I, is mostly narrated in the we-form. It explicitly discusses the unusual nature of the individuals that have come to form its
collective subject: Despite all the variations in age, origin, education and
status, and everything that used to be, despite the gulfs that used to divide
us, broadly speaking were the same. Behind the same crude shape we
conceal and exhibit the same manners, the same habits, the same simplified character of men who have reverted to their primal state (18).
And, we might add, the same set of perceptions. We find other early examples in the work of Victor Serge, Ignazio Silone, William Faulkner,
and Raja Rao.
A striking and sustained use of we-narration and focalization appears in a nonfictional text, Richard Wrights 12 Million Black Voices
(1941). As Joel Woller (1999) has remarked, the first chapter is narrated
147
in a cross-generational voice. [...] It functions as a kind of prologue, articulating a collective memory of the middle passage and slavery (348). In
a paragraph describing the horrors of the middle passage, the iterative
narration includes the experiences of both the living and the dead: In the
summer, down in the suffocating depths of those ships, on an eight- or
ten-week voyage, we would go crazed for lack of air and water, and in the
morning the crew of the ship would discover many of us dead (14). At
the time of its appearance, the work was castigated by some reviewers for
its authors temerity in presuming to speak for many other subjectivities
from within the we-perspective. This strategy would be greatly extended by Ayi Kwei Armah in his novel, Two Thousand Seasons (1973).
His we are black Africans, and the term is a tool of resistance; it is routinely opposed to the discourse of a colonizing they, Arab or European.
The we here is especially inclusive, stretching for a thousand years.
This leads to some particularly daring effects of voice, focalization, and
temporality, as when we remember a prophet speak to a we that existed ten centuries ago to inform it of events that will soon ensue, even
though we have also seen its tragic results centuries later (1213). As
such, this practice is a most interesting embedding of chronological relations within a largely iterative framework. This voice produces other
compelling features, including a collective memory which is set forth as
authoritative and a denunciation of the notion of an individual consciousness:
Of unconnected consciousness is there more to say beyond the clear recognition this is
destructions keenest tool against the soul? [...] That the passion and the thinking and
the action of any one of us should be cut off from our communal consciousness by
mere physical things, walls of wood or walls of stonethat would indeed be the
manic celebration of deaths white empire. (12829)
Uri Margolin has correctly noted that the we may shift in identity,
scope, size, and temporal location in the course of the narration (Margolin 2001: 245), and observes that, in Faulkners A Rose for Emily,
most we references are to the townspeople, a community that encompasses three different generations, not all of whom could be alive
together at any given moment (245). Wright and Armah show just how
extensive this we can be 2 .
2
As Margolin points out, in Armahs novel no less than half a dozen reference groups
with complex relations of inclusion or partial overlap can be distinguished, including
all Africans of the past one thousand years (Margolin 2001: 245).
148
Brian Richardson
We may now identify the main kinds of we-narration and -focalization. An inventory of the most salient varieties of we-narration, differentiated according to the degree to which they diverge from the poetics
of realism, could be aligned as follows:
(1) Standard: Largely realistic narration that nevertheless stretches
verisimilitude at key points, especially when the narrator discloses the inner thoughts, perceptions, or feelings of a group. For instance, in Joan
Chases novel, the we-voice of shared experience and the third person
accounts of each girls individual actions cannot be realistically squared.
During the Reign of the Queen of Persia (1983) is a narrative centered on
a group of cousins: There were four of usCelia and Jenny, who were
sisters, Anne and Kate, sisters too, like our mothers, who were sisters
(48). In the course of the narrative, each girl is described in the third person with zero focalization. This means that the writer of the narrative
must at some points be (misleadingly) referring to herself and her actions
in the third person, something not normally done in realistic representation. In the few cases where this occurs, such as Henry Adams referring
to himself in the third person in The Education of Henry Adams, we all
know very well that it is a first person narration that is told in the third
person form. In Chases novel, however, we dont know and cannot determine who the narrator is. The we-sections, however, are internally
focalized and present a single perception or emotion shared by three or
four individuals.
Another example can be adduced to show how standard we narration regularly examines its own practices self-reflexively within the
work itself. In Jeffrey Eugenidess The Virgin Suicides (1993), the wenarrator is a partially indeterminate collection of neighborhood boys who,
despite years of investigation and speculation, never begin to understand
the motives of the girls who commit suicide. This novel includes a number of subtle, self-reflexive allusions to the idea of a multiple, protean
subject, including the depiction of a shared consciousness with a single
focalization: while reading one of the girls diary together, we learned
about their lives, came to hold collective experiences of times we hadnt
experienced, harbored private images of Lux leaning over the side of a
ship to stroke her first whale (4243). Here Eugenides describes a realistic experience that is an analogue of his narrative practice that strays beyond the boundaries of realistic representation.
(2) Nonrealistic: In the texts by Conrad and Wright noted above we
have more flagrant violations of the parameters of realistic representa-
149
tion. Conrads are done solemnly without remark; Wrights narrator (like
that of Armah) discloses sentiments that stretch over centuries and range
across continents. Even more interesting in this context is a recent novel
by Zakes Mda, Ways of Dying (1995). Mda provides the most playful and
sustained interrogation of the curious epistemology of the we-narrator;
an early passage in Ways of Dying reads as if it were intended to answer
critics of the practice of Conrad, Wright, and many of their successors
concerning what should be mimetically impossible kinds of knowledge
and acts of focalization:
We know everything about everybody. We even know things that happen when we
are not there; things that happen behind peoples closed doors deep in the middle of
the night. We are the all-seeing eye of the village gossip. When in our orature the
story-teller begins the story, They say it once happened ..., we are the they. (12)
In an attempt to ground the impossible knowledge of the contents of other

minds in a first person form, the speaker playfully locates the source of
such knowledge in a more unreliable (yet widely believed) source, village
gossip, and then goes on to associate this narration with the well-spring of
traditional oral literature. The speaker continues with some salient reflections on the control and selection of narratives in this kind of speech
community and a direct address to a potentially skeptical audience: No
individual owns any story. The community is the owner of the story, and
can tell it the way it deems fit. We would not be needing to justify the
communal voice that tells this story if you had not wondered how we
became so omniscient (12). Here we have a slyly ironic and politically
charged explanation (or pseudo-explanation) for information not available
under a rigorously mimetic framework as well as a defense of traditional
practices of narration and focalization from the vantage point of a first
person plural voice.
As in Native American we-narration, there is also a socio-historical
component to this choice of pronoun. As Mda states:
The communal voice we is quite common in both the Nguni and the Sotho groups
of languagesespecially in the day-to-day speech. In Sepedi, for instance, it is
respectful to address a person in the plural form. In the storytelling traditions of folk
tales the communal voice is not that common since most stories are told in the third
person. But in the narration of legends, myths and history (and often the boundaries
are blurred here) we do find the communal voice sometimes, depending on the
storyteller and his/her distance to the events. 3
3
Private communication to the author, 18 December 2006.
150
Brian Richardson
(3) Our final category is the Antimimetic: Sarrautes Tu ne taime pas

(1989) eschews realism altogether, and functions instead as experimental
constructions of multiple discourses that can inhabit a we. This novel is
a representation of a collection of contiguous voices, some of them contradictory, that seem to form a single, polymorphous decentered consciousness. Here, the instabilities that flavor nearly all we-narration are
blended into the voices of a deconstructed self. As one of the voices complains at the out-set of the book: We all do that all the time. What else
can we do? Every time one of us shows himself to the outside world he
designates himself as I, as me, [...] as if he were the only one, as if you
didnt exist (Sarraute 1990: 2). This continues: Of course, we were a
little restless, a little ill at ease, embarrassed ... Not all of us, though ... We
never turn out in full force ... there are always some of us who are
dozing, lazing, relaxing, wandering ... this we can only refer to the ones
who were there when you came out with that remark (2, Sarrautes
ellipses). The work is not so much a psychological study as a philosophical allegory of the multiple subjectivities and voices bound together in a
self; its unstable, shifting, and always incomplete we-voices provide an
apposite image of this polydirectional entity. Intriguingly, this practice
yields a multiple subject with multiple focalizers. I refer to this as an antimimetic type of we-narration since it traduces the conventions of nineteenth century realismeven though the author (and many others) might
well claim it is a more accurate representation of mental events than that
found in realist works; Sarraute views the self as composed of many
selves, not unlike a flock of birds in flight. Hazard Adams novel Many
Pretty Toys (1999) also contains extended anti-mimetic narration of a
fragmented collectivity (e.g., All of usor rather most, for there are
always believed to be dissentersinsist finally that how we reach agreement is the major part of the story, in fact is the story, of its form, or
frame, or the sanction, in short, for our existenceif we existeverything else being a filter, as it were [16]).
In his 1996 article, Uri Margolin suggested that we-narratives are
rare for three related reasons: because the exact scope of the we may
remain ambiguous and may contain different members at different points
in the narrative, because the question of the narrators mental access of
others minds remains inherently unresolved, and because the sense of a
collective subject is more easily conveyed in lyric or meditative texts. The
examples above, I would argue, suggest a different conclusion on all three
points. As recent critical studies are making apparent, we-narration is
151
increasingly shown to be a supple technique with a continuous history of

over a century that continues to be deployed in a considerable number of
stories and novels, particularly those that emphasize the construction and
maintenance of a powerful collective identity. It has been utilized by a
considerable number of major twentieth century authors as well as
significant figures prominent in oppositional literatures. Substantial
examples of its sustained use appear in every decade of the twentieth and
twenty-first centuries and its practitioners include Conrad, Kafka,
Faulkner, Ignazio Silone, Raja Rao, Richard Wright, Robbe-Grillet,
Butor, Sarraute, Vargas Llosa, Cortzar, Patrick Chamoiseau, Zakes Mda,
Jeffrey Eugenides and Joyce Carol Oates 4 .
It is certainly the case that it is an excellent vehicle for expressing a
collective consciousness; the relative rarity of its use enhances its ability
to highlight traditional formulas and foreground its difference from the
autonomous individual consciousness associated with the rise of the novel
in England and the development of modernist techniques of representing
minds. For socialists, feminists, and Third World intellectuals who
denounce the extremes of bourgeois egoism and the poverty of an isolated
subjectivity, we-narration must seem a prefiguration of the new, more
communal, and more egalitarian society they are working to promote. The
form is also singularly adept in expressing the shared perspectives of a
number of different groups, including Conrads seamen; the isolated rural
communities of Silone, Rao, and Faulkner and the preindustrial
Gemeinschaft they share; the revolutionaries of Armah; the segregated
urban poor of Mda; the soldiers of Barbusse; the childrens sensibility
depicted by Faulkner and Vargas Llosa; the crass cliques portrayed by
Joyce Carol Oates; and even the society of mice depicted by Kafka.
Margolins second objection has been partially answered by Margolin
himself in a later study of narration in the plural (2000). In this article he
clarifies that collective representations of mental activity
are best interpreted as typification, schematization, or contraction of recognizable
shared stances, perspectives, views, or common opinions held by numerous members
of the group. The specific words employed are supposed to echo what most or all
group members may have thought on a given occasion, but rather than being a verbatim quotation, they are in fact a condensation of numerous expressions, an image of
collective inner speech, projected or invented by the narrator. (60506)
4
I discuss many of these examples at greater length in my book, Unnatural Voices.
152
Brian Richardson
These comments indicate something of what is at stake in the use of such

narration and focalization. I feel, however, that it is most useful to see the
we-narrator as a different kind of figure from the realistic type of first
person narrator and more like a postmodern first person narrator who refuses to be bound by the epistemological rules of realism; the same is true
of we-focalization 5 . I argue that we is an essentially dialectical perspective that typically (and most successfully) plays with its own boundaries. As Celia Britton observes,
its extreme elasticity provides a point of view that is not limited to any one character
or period of time but moves around from one to another. [...] As such it creates a different representation of intersubjective relations between the individual characters,
suggesting that peoples most intimate feelings are known to the community. (Britton
1999: 142)
Rather than an inherently flawed, problematic, or scandalous technique,

we-focalization is instead an extremely flexible strategy that works precisely because of its variable referents. The drama created for the reader is
thus to determine how literally and how figuratively to take each such
expression of shared mental events. The we glides between the individual subjectivity and collective omniscience, between a strict and a more
lax denotation, and between mental experiences that are entirely, partially,
or minimally shared. Its focalization remains fixed even as those whose
perspective it describes shift positions, change, and grow larger or
smaller. Much of the drama of reading such a work comes from observing
the fluctuations in the group that constitutes the we, assessing its explicit epistemological statements concerning the origin and veracity of its
beliefs, attending to moves away from realism and toward a more paradoxical discourse, and noting fundamental changes in the general reliability of the we-narrator.
Indeed, one of the great challenges of reading this kind of fiction is to
establish the relative objectivity or subjectivity of the we. One may
even discern a general, if intermittent, historical trajectory that moves
from more reliably intersubjective narrators earlier in the century to ever
more unreliable ones and then back again to a playful, postmodern communal omniscience. The we-narrators of Silone, Rao, Wright, and
(most of the time) Conrad, when confined within their own spheres of ex5
Indeed, the most recent criticism and theory of we-narrations often explicitly rejects
the parameters of realism: see Britton (1999: 136); Woller (1999: 34648); Fulton
(2003: 1113, note 3).
153
perience, are utterly reliable. Those of Eugenides are fallible, while Mda
self-consciously provides his narrator with the authoritative knowledge he
should not normally be able to possess. The we of Wright can even
include the perspectives and voices of the dead.
Margolins latest treatment of the subject still attempts to provide a
mimetic framework that can explain these phenomena. But there is no
need to insist on such a framework: if an author ignores these parameters,
as Conrad does, or gives them a postmodern wink, as Mda prefers, then
the problem dissolves. If Conrads depictions of his crews sensibilities
are inherently unresolvable given the existing models based on realist
conventions, then we should not limit ourselves to realist conventions
when grounding our theories.
The larger theoretical problem foregrounded by more extreme forms
of we-narration is starkly present in Mario Vargas Llosas The Cubs
(Los cachorros [1967]). In this text, we and they forms alternate, not
merely in successive sections or passages, but within the same sentence:
They were still wearing short pants that year, we werent smoking yet, of
all the sports they liked football best, we were learning to ride the waves
[...] (Vargas Llosa 1989: 1). As Jean OBryan-Knight (1997: 340)
comments, in a single sentence [...] we observe the group [of four boys]
subjectively and objectively. Vargas Llosa has thus compressed the
epistemological antinomy devised by Conrad into a starkly unnatural
form, thereby foregrounding the transgression that we-narration always
threatens to enact: the collapsing of the boundary between the first and
the third persons and thereby minimizing the foundational difference
between the implicit fallibility of all first person narration and the inherent infallibility of third person fiction.
Such epistemic slippage has appeared before in the history of the novel: many classic examples of first person narration and focalization are by
no means innocent of the transgressions apparent in we-narratives. As
Peter Rabinowitz points out,
Anton Lavrentevich, the narrator of Dostoyevskys Possessed, offers a limited perspective on events at the beginning of the novel. But while he remains the nominal
narrator throughout the text, his persona and limitations fade away for long passages
in the middle, where we receive a great deal of information to which he could have no
possible access. (Rabinowitz 1987: 12627)
Genette discusses a number of such examples, which he terms paralepses, in Proust (cf. Genette 1980: 20711), including
154
Brian Richardson
the last thoughts of Bergotte on his deathbed, which, as has often been noted, cannot
have been reported to Marcel since no onefor very good reasoncould have
knowledge of them. That is one paralepsis to end all paralepses; it is irreducible by
any hypothesis to the narrators information, and one we must indeed attribute to the
omniscient novelistand one that would be enough to prove Proust capable of
transgressing the limits of his own narrative system. (208)
We have, Genette explains, a paradoxicaland to some people shamefulsituation of a first-person narrating that is nevertheless occasionally omniscient (252). Genette offers a way out of this impasse by concluding that the decisive criterion is not so much material possibility or
even psychological plausibility as it is textual coherence and narrative
tonality (208)typical features, it will be noted, of we-narration.
Through this analysis, Genette provides us with the tools by which we can
best comprehend we-focalization (though he himself unfortunately
dismisses the collective witness as narrator as an unremarkable variant
of homodiegetic narration [245]).
Later work by Dan Shen and by Henrik Skov Nielsen has further explored and foregrounded this elusive kind of oscillating narrative perspective. Shen writes that paralepses draw attention not only to the
limitations of the violated modes of focalization, but also to the fact that
the barriers between modes of focalization are very much conventional
(Shen 2001: 172). Nielsen takes his analysis further, arguing for two separate theoretical entities in the case of such paraleptic texts: the narratingI and what he calls the impersonal voice of the narrative. The latter
can say what a narrating-I cannot say, produce details that no person could remember,
render the thoughts of other characters, speak when the character remains silent, etc. It
speaks, however, in the first person, both when the possibilities of the person referred
to by the first person are abandoned and when it says what this person cannot say.
(Nielsen 2004: 13940)
We can now further pinpoint the distinctive theoretical differences of

we-narration. Genette has stated that the novelist must choose between
two narrative postures, either to have the story told by one of its characters, or to have it told by a narrator outside the story (Genette 1980:
244). Whenever a text uses a first person plural narrator to depict the
thoughts of others, it tends to straddle the line between first and third
person fiction, as a homodiegetic character narrator discloses that which
can only be known by an external heterodiegetic intelligence. These narrations are thus simultaneously first and third person discourses, and thus
transcend either subtly or flagrantly the foundational oppositions set forth
155
in different ways by Stanzel and Genette (though, as we have noted,

Genette is able to see these fused in Proust). Whereas most second person
narration oscillates between these two poles, we-narration curiously occupies both at once. Genette mentions two instances of sudden shifts in
narration where the hero moves from I to he or he to I (cf.
Genette 1980: 246); he calls these cases narrative pathology but goes on
to admit that the contemporary novel has passed that limit [...] and does
not hesitate to establish between narrator and character(s) a variable or
floating relationship, a pronominal vertigo in tune with a freer logic and
more complex conception of personality (246). It is, I suggest, this
pathology that needs to be investigated fully, and this floating relationship
that needs to be theorized by narratologists.
We-narration, as I describe it above, is an artificial perspective that
shares the characteristic features of the two other, more common, perspectives. But they too, of course, are themselves artificial conventions.
The distinguishing feature of third person narration, as Kte Hamburger
and others have insisted, is the disclosure of the thoughts and mental
events of persons other that the narrator. This of course is not possible in
life by human beings, and is an ability granted by convention to such a
narrator. The first person is equally, if differently, a construct. In addition
to the slippages just noted in the work of Proust and Dostoevskij, violations which could be added to indefinitely, there is the curious situation
that a governess, writing in the first person, may not know what she has
not perceived, but she may write in a highly (and unlikely) compelling
narrative style, and one that reads just as if it were written by Henry
James. Further, as Alan Palmer has recently stressed, interior monologue
is not an entirely accurate representation of the jumble of thoughts, impressions, and subvocalized speech that goes on inside our heads, but
rather a stylized and often misleading technique of representing consciousness. Finally, in fiction the borders between first and third person
narration are often more porous than many theorists would allow. Nikolai
Gogol in his 1842 story, The Overcoat, draws attention to both of these
conventions. As his protagonist is strolling in the streets of St. Petersburg
at night in his new overcoat, he sees a salacious figurine in a shop window and smiles. The narrator asks a few rhetorical questions concerning
the cause of this smile, before going on to aver that there is no creeping
into a mans soul and finding out what he thinks. Throughout the text,
however, he has been doing just that, revealing private thoughts, disclosing scenes that were unobserved, and generally assuming the prero-
156
Brian Richardson
gatives of an omniscient third person narratoreven as he complains that

his memory is growing dim and he cannot recall all the details of the
events he narrates. Gogol is obviously mocking these conventions and
refusing to be bound by them; he is clearly telling his reader that when it
comes to narration, he can do whatever he pleasesa sentiment many
other novelists have obviously shared.
Concerning a theory of narration, we need both the traditional opposition of first and third person or homo- and heterodiegetic forms, as well
an additional category that generally exists only in fiction which deliberately conflates the other two perspectives. Concerning focalization, I
suggest that in addition to the standard conception of who sees, in
Genettes shorthand, which is limited to the narrator in first person texts
and to character focalizers in third person texts, we should add an additional term, wandering focalization, which partakes of both of the others to describe the peculiar features of we-narratives. We might add this
to the growing list of concepts to depict unusual or deviant forms, including Genettes paraleptic focalization, David Hermans hypothetical
focalization (cf. Herman 2002: 30923), and what I have elsewhere
called pseudo-focalization (Richardson 2006: 89) to designate apparent
heterodiegetic depictions of a characters consciousness that turn out to be
projections by a homodiegetic narrator, as found in most of the depictions
of thoughts and emotions in Ian McEwans Atonement, a novel that seems
to be an omniscient, third-person work until it is revealed at the end to
have been written by one of its characters.
References
Adams, Hazard (1999). Many Pretty Toys. Buffalo: SUNY Press.
Adams, Henry (1907). Democracy, Esther, Mont saint Michel and Chartres, The Education of Henry Adams. New York: Library of America, 1983, 7151192.
Armah, Ayi Kwei (1973). Two Thousand Seasons. London: Heinemann.
Barbusse, Henri (2003). Under Fire. Tr. R. Buss. New York: Penguin.
Britton, Celia (1999). Collective Narrative Voice in Three Novels by Edouard Glissant.
S. Haigh (ed). An Introduction to Caribbean and Francophone Writing. Oxford:
Berg, 13547.
Chase, Joan (1983). During the Reign of the Queen of Persia. New York: Harper and
Row.
Conrad, Joseph (1899). The Nigger of the Narcissus. Complete Works. London:
Doubleday, 1921.
Eugenides, Geoffrey (1993). The Virgin Suicides. New York: Farrar, Straus, and Giroux.
157
Fulton, Dawn (2003). Romans des Nous: The First Person Plural and Collective Identity
in Martinique. The French Review 76:6, 1104114.
(1980). Narrative Discourse. Tr. J. E. Lewin. Ithaca: Cornell UP.
Herman, David (2002). Story Logic: Problems and Possibilities of Narrative. Lincoln:
U of Nebraska P.
Marcus, Amit (2001). A Contextual View of Narrative Fiction in the First Person Plural.
Narrative 16:1, 4664.
Margolin, Uri (1996). Telling Our Story: On We Literary Narratives. Language and
Literature 5:2, 11533.
(2000). Telling in the Plural: From Grammar to Ideology. Poetics Today 21, 591
618.
Mda, Zakes (1995). Ways of Dying. New York: Farrar, Straus, and Giroux.
Morris, Adalaide (1992). First Persons Plural in Contemporary Feminist Fiction. Tulsa
Studies in Womens Literature 11, 1129.
N Dhuibhne, ils (1988). Blood and Water. Dublin: Attic Press.
Nielsen, Henrik Skov (2004). The Impersonal Voice in First-Person Narrative Fiction.
Narrative 12, 13350.
OBryan-Knight, Jean (1997). From Spinster to Eunuch: William Faulkners A Rose for
Emily and Mario Vargas Llosas Los cachorros. Comparative Literature Studies
34, 32847.
Rabinowitz, Peter (1987). Before Reading: Narrative Conventions and the Politics of
Interpretation. Ithaca: Cornell UP.
Richardson, Brian (2006). Unnatural Voices: Extreme Narration in Modern and Contemporary Fiction. Columbus: Ohio State UP.
Sarraute, Nathalie (1989). Tu ne t'aimes pas. Paris: Gallimard.
(1990). You Don't Love Yourself. Tr. B. Wright. New York: Brazilier.
Shen, Dan (2001). Breaking Conventional Barriers: Transgressions of Modes of Focalization. W. van Peer & S. Chatman (eds). New Perspectives on Narrative Perspective. Albany: SUNY Press, 15972.
Silone, Ignazio (2000). The Abruzzo Trilogy: Fontamara, Bread and Wine, The Seed
Beneath the Snow. Tr. E. Mosbacher. South Royalton, VT: Steerforth Italia.
Vargas Llosa, Mario (1989). The Cubs and Other Stories. Tr. G. Kolovakos & R. Christ.
New York: Noonday.
Vargas Llosa, Mario (1967). Los jefes; Los cachorros. Barcelona: Seia Barral, 1982.
Woller, Joel (1999). First-person Plural: The Voice of the Masses in Farm Security
Administration Documentary. Journal of Narrative Theory 29:3, 34066.
Wong, Hertha D. Sweet (1998). First Person Plural: Subjectivity and Community in Native American Womens Autobiography. S. Smith & J. Watson (eds). Women,
Autobiography, Theory: A Reader. Madison: U of Wisconsin P, 16878.
Wright, Richard (1941). 12 Million Black Voices. New York: Thunders Mouth, 1988.
158
Brian Richardson
Appendix:
Bibliography of We-Narratives
Narratives Entirely or Substantially Composed in the We-Form
Joseph Conrad (1899). The Nigger of the Narcissus.
Henri Barbusse (1916). Feu.
Franz Kafka (1924). Die Sngerin Josephine, oder das Volk der Muse.
Thomas Mann (1929). Mario und der Zauberer.
Ignazio Silone (1930). Fontemara.
William Faulkner (1930). A Rose for Emily.
(1931). That Evening Sun.
(1931). A Justice.
(1931). Divorce in Naples.
(1932). Death Drag.
(1935). That Will Be Fine.
(1943). Shingles for the Lord.
(1948). A Courtship .
Raja Rao (1938). Kanthapura.
Richard Wright (1941). 12 Million Black Voices.
Alain Robbe-Grillet (1954). The Way Back.
Hans Erich Nossack (1963). Das Mal.
Mauro Senesi (1963). The Giraffe.
Michel Butor (1964). La Gare Sainte-Lazare.
Amos Oz (1966). Elsewhere, Perhaps.
Mario Vargas Llosa (1967). Los Cachorros.
Gabriele Wohmann (1971). Stories in Gegenangriff.
(1975). Stories in Lndliches Fest.
Pierre Silvain (1971). Les Eoliennes.
Ayi Kwei Armah (1973). Two Thousand Seasons.
Arlette and Robert Brechon (1974). Les noces dor.
Edouard Glissant (1975). Malemort.
(1981). La Case du commandeur.
Donald Barthelme (1978). We dropped in at the Stanhope ....
Julio Cortzar (1981). Queremos Tanto a Glenda.
Mark Helprin (1981). North Lights .
John Barth (1982). Sabbatical.
Joan Chase (1983). During the Reign of the Queen of Persia.
T. C. Boyle (1985). Greasy Lake.
Jim Crace (1988). The Gift of Stones.
Louise Erdrich (1988). Tracks.
Elfriede Jelinek (1989). Lust.
Nathalie Sarraute (1989). Tu ne taimes pas.
Jeffrey Eugenides (1993). The Virgin Suicides.
Zakes Mda (1995). Ways of Dying.
159
Patrick Chamoiseau (1997). LEsclave vieil homme et le molosse.

Joyce Carol Oates (1999). Broke Heart Blues.
Hazard Adams (1999). Many Pretty Toys.
Alice Elliott Dark (2001). Watch the Animals.
Jill McCorkle (2002). Billy Goats.
Nadine Gordimer (2003). Visiting George, Look-Alikes, Karma. Loot and Other
Stories.
Yiyun Li (2005). Immortality, Persimmons. A Thousand Years of Good Prayers.
Narratives with Significant Sections in the We-Form

Gustave Flaubert (1857). Beginning of Madame Bovary.
Yevgeny Zamyatin (1924). We.
Victor Serge (1931). Naissance de notre force.
Gertrude Stein (1937). Everybodys Autobiography, chapter 4, America.
Albert Camus (1947). La Peste.
Samuel Beckett (1950). LInnommable.
Vladimir Nabokov (1951). Speak Memory, chapter 15.
Carlos Fuentes (1964). Alma Pura.
Maurice Roche (1966). Compact.
Ngugi wa Thiongo (1967). A Grain of Wheat.
Toni Morrison (1970). The Bluest Eye.
Julia Alvarez (1991). How the Garcia Girls Lost Their Accents.
Jean Echenois (1992). Nous trois.
Patrick Chamoiseau (1992). Texaco.
Part II: Some Special Aspects of Mediation
VIOLETA SOTIROVA
(Nottingham)
A Comparative Analysis of Indices of Narrative Point of View

in Bulgarian and English
1 Aspectual Differences Between Bulgarian and English
In her 1993 book, The Fictions of Language and the Languages of Fiction, Monika Fludernik notes with regret that we know little as yet about
how narrative viewpoint is expressed in free indirect discourse in languages such as Russian and Japanese, or other non-indo-European languages. She calls for new studies of free indirect discourse in underinvestigated languages, suspecting that knowledge of their linguistic features would have an impact on the theories narratologists construct to account for the presentation of viewpoint and consciousness. Fludernik herself outlines some of the features of free indirect discourse in Russian and
she compares these with the linguistic cues she identifies for free indirect
discourse in English, German and French (cf. 10002). Although Fluderniks comments cannot be accepted as entirely justified because work
on Russian and Japanese has been carried out (cf. Kuroda 1973, 1987;
Pascal 1977; Schmid 2003), her monumental study of the linguistic technique of free indirect discourse hints at the possibility of using crosslinguistic evidence when attempting to theorize the form and function of
free indirect discourse.
My aim in this paper will be to examine features of free indirect discourse across two languages, Bulgarian and English, and to analyze these
features in relation to some unresolved issues in narrative theory. I shall
focus in particular on the use of tense and aspecttwo verbal categories
where the differences between the two languages are most prominent and
which also bear significantly on the narratological issues of voice and
perspective.
Although similar to Russian in many respects, Bulgarian offers a
richer system of aspectual, tense and mood categories which present complex possibilities for rendering past events and states. Similarly to verbs
164
Violeta Sotirova
in other Slavonic languages, Bulgarian verbs are marked on the stem for
aspect. So, pairs of verbs with the same meaning exist in the language, the
only difference between them being their finished/perfective or
unfinished/imperfective aspect. This aspectual distinction would roughly correspond to the English simple and progressive aspect, although one
has to bear in mind that aspect in English is grammaticalized, while in
Slavonic languages it is considered to be lexical. Unlike English, Bulgarian only allows the imperfective verb to be used in the indicative mood in
present tense. The perfective stem is reserved for certain modal constructions and can also be used with the future marker. Where the differences
between the aspectual systems of the two languages are most apparent is
with the so called class of stative verbs (cf. Quirk & Greenbaum 1977:
47). Bulgarian, unlike English, maintains the aspectual distinction even
on stative verbs which in English cannot take the progressive aspect.
Stative verbs which according to Leech fall into four main semantic types,
can in most cases be expressed in Bulgarian with either of a pair of
verbs 1 :
verbs of inert perception (feel, hear, see, smell, taste)
verbs of inert cognition (believe, forget, hope, imagine, know, suppose,
understand)
state verbs of having and being (be, belong to, contain, consist of, cost,
depend on, deserve, have, matter, own, resemble)
verbs of bodily sensation (ache, feel, hurt, itch, tingle) (cf. Leech 1971)
Verbs of inert
perception
Perfective stem
Imperfective stem
I feel:
I hear:
I see:
I smell:
I taste:
Verbs of inert
cognition
Perfective stem
Imperfective stem
I believe:
I forget:
I hope:
1
I base my analyses of Bulgarian verbs on Andreichin et al. (1998).
Indices of Narrative Point of View in Bulgarian and English

I imagine:
I know:
I suppose:
I understand:
State verbs of
having and being
Perfective stem
Imperfective stem
I am:
I belong to:
I contain:
I consist of:
I cost:
I depend on:
I deserve:
165
I have:
I matter:
I own:
I resemble:
Verbs of bodily
sensation
Perfective stem
Imperfective stem
it aches/hurts me:
it itches me:
it tingles me:
The semantic difference between the two aspectual types in Bulgarian

finished and unfinishedlies in the way in which they present the
event or situation. Similarly to other languages, the finished aspect is
synthetic, it views the event as a whole in its completion. The unfinished
aspect is analytical and views the event in progress, in the process of
unfolding. All verb stems in Bulgarian belong to one of the aspectual
classes: perfective/finished or imperfective/unfinished, with most of them
having an aspectual counterpart. Some verb stems only exist in one of the
classes and their corresponding aspectual partner is formed through affixation. The use of a prefix (-; -) in the case of feel and ache is
thus normal in the transformation of an imperfective stem into a perfective verb. But while in some cases the difference between such pairs of
166
Violeta Sotirova
verbs is purely aspectual, in others there is also a more perceptible change

in lexical meaning (e.g. the verb becoming inchoative or semelfactive).
However, the aspectual distinction is present, even if the lexical meaning
also undergoes a change when transforming these imperfective verbs into
perfective. Bulgarian morphologists explain the existence of these so
called defective verbs with the fact that their lexical meaning does not
readily allow them to take the perfective. Such would be the case with
most of the verbs in Leechs class of state verbs of having and being.
However, the other semantic classes of stative verbs can take either of a
pair of verbs in Bulgarian. Stative verbs, then, are of special interest
because in one of my languages for the purposes of this comparison we
have two possibilities, while in the other, in English, only the simple
aspect is allowed by the grammar. The Bulgarian pairs of verbs I have
given in the tables are both in the present tense. Further, more complex
distinctions would be present in the rendering of past states.
What emerges so far from the brief sketch of aspectual differences between Bulgarian and English is that dynamic verbs in most cases would
be translatable with equivalent forms across the two languages, but the
rendering of stative verbs would be more intricate because of the two
aspectual possibilities in Bulgarian. This difference between the two languages would be interesting to pursue, not only for a purely linguistic
comparison, but also because it bears on some important issues in narrative theory.
2 Implications for Narrative Theory

It has already been noted by narratologists that stative verbs in the past
tense make sentences of free indirect discourse indistinguishable from
narration because they cannot take the progressive aspect. This class of
stative verbs also poses problems for another reason. They usually denote
some kind of mental state of the character and as such hover on the brink
between narration and character point of view. There is disagreement as
to whether such sentences should be classed as free indirect discourse or
whether they stem from the narrators point of view. All such sentences,
containing mental verbs, such as believe, know, feel, suppose, suspect,
expect, etc. fall under Short et al.s (1996) category of internal narration which the authors add to Leech & Shorts (1981) cline of modes
of speech and thought presentation after extensive work on a large corpus
of narrative texts. The need for new categories is suggested by the numer-
167
ous examples drawn from these texts, or as the authors assert: In

fictional texts, narrators may provide reports of characters cognitive
activities and emotional states which do not fall under any of Leech and
Shorts categories for thought presentation (Short et al. 1996: 124).
Short et al.s examples of Internal Narration include sentences, such
as: Her approval filled the military young man with happiness (Aldous
Huxley, Point Counter Point); For a moment he was rendered motionless by surprise, a kind of respect (Rupert Thomson, The Five Gates to
Hell). The reason why the presentation of these states does not fit into the
Leech & Short (1981) model is that all of the categories associated closely
with the characters point of view in this initial taxonomy are verbal
categories. Most importantly, free indirect thought has to be translatable
into direct thought in order to qualify as such. What we have in these examples, on the other hand, are states that are not necessarily verbalized or
articulated in inner speech. When Short et al. (1996) introduce this new
category of Internal Narration, they position it at the leftmost end of their
cline, next to pure narration. This means that the narration of internal
states is viewed by them as almost fully in the narrators control and not
tinged with the characters voice or perspective, or as they say: we are
given insights into a characters internal states or changes, but no representation of specific thoughts of the character. (125) The reason for their
interpretation of sentences of internal narration as akin to narratorial discourse, I think, lies not least in the fact that these states are not readily
verbalized into inner speech. But no more elaboration on these decisions
is given at this stage than simply to state: clearly, NI lies at the interface
between narration and thought presentation (125). The speech and
thought presentation clines then look like this:
Narrator
in control
N
N
NV
NI
NRSA
NRTA
NNarration
NVNarrators report of voice
NRSANarrators representation of
speech act
ISIndirect speech
FISFree indirect speech
DSDirect speech
FDSFree direct speech
IS
FIS
IT
FIT
Norm
Norm
DS
FDS
DT FDT
Character
in control
NNarration
NINarration of internal states
NRTANarrators representation of
thought act
ITIndirect thought
FITFree indirect thought
DTDirect thought
FDTFree direct thought
168
Violeta Sotirova
Sentences of internal narration are thus less closely associated with the
characters internal point of view than sentences of indirect thought or
even narrative reports of thought acts, examples of which would be: Jed
thought he understood (Rupert Thomson, The Five Gates to Hell); As
she walked down the Charing Cross Road, she put to herself a series of
questions (Virginia Woolf, Night and Day). More surprisingly, Short et
al. draw a further distinction between cognitive or emotional experiences which fall under internal narration and reports of characters perceptions, whether the stimuli are internal (She felt a pain in her stomach) or external (She felt the softness of his hair) which they say
would be coded as narration (Short et al. 1996: 125).
This type of sentence, rendering the mental states of characters, is
classed as psychonarration by Dorrit Cohn (1978). Cohn explains that
such sentences can give us a glimpse into the characters almost unconscious states and as such allow for non-articulated thoughts and feelings
to be presented to the reader. Fludernik later identifies psychonarration
with the narratives external description of figural consciousness
(Fludernik 1993: 136). Although Cohn considers the mode of psychonarration important, she like Short et al. privileges what Palmer calls the
speech category account of the presentation of fictional minds (Palmer
2002: 28). Palmer takes issue with this account because he thinks that all
of the modes for the presentation of thoughts and states of the mind,
identified by narratologists and stylisticians, tend towards viewing the
content of consciousness as internalized speech and because these concepts do not add up to a complete and coherent study of all aspects of the
minds of characters in novels (Palmer 2002: 28). What he identifies as
missing from existing accounts of thought presentation is the role of
thought report in describing emotions and the role of behavior descriptions in conveying motivation and intention (28), or all of these sentences that Short et al. would class under their category of Narration of
Internal States.
Palmer is right in arguing that analyzing consciousness as consisting
of articulated, verbalized speech would result in viewing it as highly selfreflective. This tendency is apparent in a number of stylistic accounts of
free indirect discourse where features of direct discourse, such as direct
questions, exclamations, imperatives, are seen as some of its central indices. Only perception among the non-verbal processes of consciousness
has been recognized as part of free indirect discourse and only by some
theoreticians (Brinton [1980]; Banfield [1982]). But perception as a valid
169
component of free indirect discourse only encompasses characters perceptions of the external narrative world that surrounds them. It is usually
identified by the use of deixis: progressive aspect or proximal deictic adverbs and its content refers to the characters outside world. The kind of
states of the mind that Palmer is arguing have been excluded from studies
of narrative or simply relegated to a less important position in the presentation of fictional minds include, in his words, mental phenomena as
mood, desires, emotions, sensations, visual images, attention, and memory (Palmer 2002: 31). These can be exemplified by one of the extracts
that he quotes from Austens Emma (1816):
Not that Emma was gay and thoughtless from any real felicity; it was rather because
she felt less happy than she had expected. She laughed because she was disappointed;
and though she liked him for his attentions, and thought them all, whether in friendship, admiration, or playfulness, extremely judicious, they were not winning back her
heart. She still intended him for her friend. (362; cited in Palmer 2002: 35)
Here Emma is reflecting on Frank Churchills behavior towards her. Her

emotional response to his attentions, which we as readers know were
misjudged, is captured in a series of sentences that contain either a verb,
or an adjective, or an adverb, or a noun that corresponds to a mental state
of the character: gay, thoughtless, felicity, felt, happy, expected, laughed, disappointed, liked, thought, heart, intended. Palmer captures all such
narrative sentences under a category which he argues is equally, if not
more, important than what has traditionally been defined as free indirect
discourse: what he calls thought report (Palmer 2002: 30). One of the
reasons behind the neglect of this category by narratologists and stylisticians he sees in the fact that the presentation of consciousness here is
done by the narrator, whereas narratologists have tended to privilege
modes of presentation that stem more directly from the character or are
perceived as more mimetic (cf. 31).
Palmers argument then clashes with the other stylistic and narratological accounts reviewed here in respect of the importance of the different categories used for the presentation of minds in fiction. He argues
for a more inclusive and more comprehensive analysis of the different
ways in which readers can have a glimpse of characters fictional minds.
But he concurs with them on one important point: that the report of internal states which corresponds to Short et al.s category of internal narration is under the narrators control. I would agree with Palmer that
bracketing out the internal states of characters from our analysis would
mean to miss out a very important dimension of characterization in nar-
170
Violeta Sotirova
rative. But I would question the understanding of internal states as closely

related to narration and as stemming entirely from the narrator.
3 Case-studies
If Bulgarian offers two alternatives for most stative verbs denoting mental
states of characters, then this should complicate the position adopted by
Short et al. that such sentences are entirely in the narrators control. I will
begin my comparison of the use of aspect as a cue of narrative view-point
across two passages from Lawrences Sons and Lovers (1913) and their
Bulgarian translations:
(A) One day in March he lay on the bank of Nethermere, with Miriam sitting beside
him. It was a glistening, white-and-blue day. Big clouds, so brilliant, went by overhead, while shadows stole along on the water. The clear spaces in the sky were of
clean, cold blue. Paul lay on his back in the old grass, looking up. He could not bear
to look at Miriam. She seemed to want him, and he resisted. He resisted all the time.
He wanted now to give her passion and tenderness, and he could not. He felt that she
wanted the soul out of his body, and not him. All his strength and energy she drew
into herself through some channel which united them. She did not want to meet him,
so that there were two of them, man and woman together. She wanted to draw all of
him into her. It urged him to intensity like madness, which fascinated him as drugtaking might. (239; italics added)
Most of the sentences in this passage would probably be classed not as

free indirect discourse, but as internal narration or even narration on Short
et al.s model. They contain verbs of cognition and emotion which on
their analysis presuppose an external reporter of the state of the character,
i.e. the narrator. Because Bulgarian offers at least two alternatives for
most of these verbs, it would be interesting to follow through their
Bulgarian translations.
One of the complications arising at this point would be that apart from
the aspectual distinction on the stem of the verb, in the past tense Bulgarian also offers two alternative endings. These are regarded as tenses by
Bulgarian linguists: the Past Complete Tense, or the Aorist, and the Past
Incomplete Tense, or the Imperfect 2 . Most commonly, verbs with
perfective or finished stems would take Aorist endings and verbs with imperfective or unfinished stems would take Past Incomplete endings. But
crossovers are also possible and new shades of meaning are created. Thus,
imperfective stems can take Aorist endings in which case most typically
2
Strictly speaking, this distinction in the past tense is also aspectual.
171
they usually mean that the action is completed in the past but it has had a
certain duration while being carried out. For example, in a sentence like:
For nine years he wandered about, homeless, sleepless, restless 3 , the
imperfective verb would be combined with the Aorist to express this
meaning of duration and completion at the same time. Or, it is also possible to combine imperfective verbs with the Aorist in order to express
iterative events within a limited period of time: e.g. Several times during
the night I was awoken by the loud barking of the dog.
The Past Incomplete Tense, or the Imperfect, denotes events in progress, concurrent with another past orientational moment, which may not
have finished before the moment of speaking. Most typically, the Past Incomplete Tense combines with imperfective or unfinished verb stems. But
perfective stems can also take the Past Incomplete Tense in some special
circumstances. Usually, this combination of perfective verb stem with
Past Unfinished Tense, or Imperfect, results in modal meanings
conditional, optative etc.: e.g. If this happened, then the trip would be
most delightful; If only the damned telegram would arrive. And finally,
perfective stems in the Past Incomplete could express habitual events:
He would get up in the morning, lay the table and begin to wait for the
others. Thus, a four-way distinction of past meanings is possible in Bulgarian, a feature unique to Bulgarian and Macedonian, which does not
occur in other Slavonic languages. Most typically, however, perfective
verbs would take the Aorist and imperfective verbs the Imperfect past
tense endings. This is the case in the first Lawrence passage that I quote
above. All of the verbs I have underlined are rendered by the Bulgarian
translator as imperfective verbs in the Past Incomplete Tense. In the table
below I have given both the imperfective form in the Past Incomplete and
the perfective counterpart in the Aorist in order to highlight the possibility
of an alternative choice.
he lay:
(perfective Aorist) (imperfective Imperfect)
was:
(imperfective Imperfect)
they went by:
(perfective Aorist) (imperfective

Imperfect)
they stole along:
Examples are directly translated from Bulgarian examples in Andreichin et al. (1998).
172
Violeta Sotirova
they were:
(perfective Aorist) (imperfective Imperfect) [verb derived from blue]
he could not bear:
(imperfective Imperfect) [had no strength]
she seemed to want:

[desired]
he resisted:

Imperfect)
he wanted:
he could not:
he felt:
she drew:
(perfective Aorist) (imperfective Imperfect) [suck out]
she did not want:
it urged:
(perfective Aorist) (imperfective Imperfect) [filled throughout]
it fascinated:

[fascinated, intoxicated]
In all of the sentences that focus on Pauls internal states the translator has
chosen the imperfective verb with Past Incomplete endings. This renders
the experience as being in the process of unfolding. Bulgarian linguists
point out that the Past Incomplete corresponds in all of its meanings to the
present tense and as such, when used in quasi-direct discourse, it denotes
experience which is current and immediate for the character. These
semantic properties of the Past Incomplete and of imperfective verbs give
a different shade of meaning as opposed to perfective verbs in the Past
Complete. Each and every sentence of this passage could have been
translated using perfective verbs in the Past Complete. However, their
typical value of denoting punctual events, completed in the past, and
arranged chronologically might not have been entirely adequate stylistically. I think that given the semantic properties of all the verbs in the
passage, and given the meaning of the whole episode, the translators decision to choose imperfective verbs in the Past Incomplete has been sensitive and justified.
173
But apart from rendering the episode successfully in a foreign language, the translators choices also signal the strong semantic connotations of these sentences: they are closely associated with the characters
experiences and as such stem from his point of view. Translating all of
these events and states in a tense and aspect that foreground the progress
of the experience makes this experience more immediate.
An immediate conclusion that can be reached at this point is that mental verbs which denote states of the character are always translated as imperfective verbs in the Past Incomplete. But it is not the case, as the following passage from the same novel demonstrates:
(B) Morel watched her shyly. He saw again the passion she had had for him. It blazed
upon her for a moment. He was shy, rather scared, and humble. Yet again he felt his
old glow. And then immediately he felt the ruin he had made during these years. He
wanted to bustle about, to run away from it. (243; italics added)
Interestingly, the verbs in this passage are mostly rendered as perfective
Aorists. The list of pairs of verbs, where available, is included below:
he watched:
he saw:

[=felt]
it blazed:
(perfective Aorist)/ (imperfective Imperfect)
he was shy:

Imperfect) [=got anxious]
he felt:
he felt:
he had made:
(Past indefinite) [=had transformed himself into]
he wanted:

Imperfect)
There is only one imperfective form in the Imperfect in the whole passage
and that is the form of the first verb watched. In all other instances
where the English verb is in the Past Simple the Bulgarian translator has
chosen to render it with a perfective verb in the Aorist. Although some
verbs in the passage could have been rendered in the Imperfect with imperfective verb stems, the choice has in this case fallen on the perfective
174
Violeta Sotirova
Aorist. I use this passage to demonstrate that both alternatives are available in the language. If we are only guided by the type of verb, stative
verbs can be translated as imperfective Imperfects as we saw in passage
(A) and they can also be translated as perfective Aorists as we see here in
passage (B). The question now would be: is this a random decision on the
part of the translator?
This passage, as opposed to the one quoted in (A), displays certain
signals of temporal ordering that might perhaps account for the translators choice of perfective Aorists. In sentence two where the translator
switches to perfective Aorists, we have the adverb again which suggests
a new occurrence of an event and as such punctuates the series of events.
The adverb features in the Bulgarian translation and might be taken as a
signal of a particular moment in the chronological development of the
narrative. Sentence three also displays an adverbial phrase which denotes
instantaneousness: for a moment. The verb that the translator uses here
begins with a prefix - that is a common perfective prefix on verb
stems. There is a possibility to derive an imperfective verb from this
perfective stem, but its meaning in this case would be of an intermit-tent
event, e.g. it can be used with the verb to light as in
(=there are lightnings). Without the prefix the imperfective form of the
verb could readily take the Imperfect and mean simply it was blazing.
Since here English also offers a choice between simple and progressive,
the translator has adhered closely to the writers choice of verb, but also
has observed the semantic restrictions imposed by the adverbial phrase.
The verb in the next sentence he felt is a stative verb, so its form in
English is limited to the Past Simple, but the Bulgarian form chosen is
again of a perfective Aorist. It seems to me that the presence of the adverbs yet again is once more taken as a contextual clue for the punctuality and chronological ordering of the events described.
A similar reasoning might have resulted in the choice of a perfective
Aorist for the verb he felt in the next sentence. Here, the explicit chronological adverb then and the punctuality denoted by the other adverb
immediately have probably triggered the choice of verb form made by
the translator. Bulgarian would not permit the combination of either of
these adverbs with an imperfective verb in the Imperfect. Although here
the translator omits then, immediately on its own imposes the same
restriction. The final sentence would permit the use of an imperfective
Imperfect verb, but what seems to have happened here is that the constraints on some of the verbs in the passage influence the rest of the verb
175
choices. Perhaps switches from one aspectual class to another within the
boundaries of a short paragraph like this would have resulted in incoherence, or as Jacob Mey describes this phenomenon, here we have the principle of interpretative obstination or syntactic inertia (Mey 1999: 33).
In other words, unless strongly prompted to reshape an interpretation, a
reader will keep an established interpretation within sentence boundaries,
and even across sentences and within paragraphs. Once the pattern of
using perfective Aorists is established in sentence two, the translator adheres to it throughout the paragraph, thus suggesting an interpretation of
these sentences as stemming from the narrators point of view. A shift to
the other aspectual class would have resulted in a shift to the characters
point of view which if chosen in the final sentence alone would have
required too big an interpretative leap.
The translators choices of verb forms in (A) would strongly suggest
that character states are not entirely in the narrators control. Even though
they cannot, and probably are not, consciously articulated by the character, it is semantically implausible to position them under the narrators
control on Short et al.s cline of modes of thought presentation if a narrative internal viewpoint is suggested through the use of imperfective
verbs in the Imperfect past tense. On the other hand, the verb forms used
in the translation of (B) would seem to support the position adopted by
Short et al. since perfective Aorists would imply that these states are
viewed holistically as punctual, discrete and completed events in the past.
Perhaps this would support the hypothesis that these states, precisely because they cannot be verbalized by the character are more likely to stem
from the narrators viewpoint. Another passage from Sons and Lovers,
quoted in (C), would address this issue further:
(C) Miriam was astonished and hurt for him. It had cost him an effort. She left him,
wanting to spare him any further humiliation. A fine rain blew in her face as she
walked along the road. She was hurt deep down; and she despised him for being
blown about by any wind of authority. And in her heart of hearts, unconsciously, she
felt that he was trying to get away from her. This she would never have acknowledged.
She pitied him. (241; italics added)
Only in the first sentence the translator has chosen to render the English
past tense with two perfective verbs in the Aorist. All of the other English
verbs in the past simple appear in the Bulgarian translation as imperfective Imperfects:
176
Violeta Sotirova
she was astonished:
she was hurt:

Imperfect)
had cost him:
she left him:

Imperfect)
she walked:
() (perfective Aorist) (imperfective Imperfect)
it blew:
she was hurt:
she despised him:

Imperfect)
she felt:
she would never have

acknowledged:
What is of particular interest in this passage is the sentence: And in her

heart of hearts, unconsciously, she felt that he was trying to get away
from her. The explicit lexical signals that the feelings of the character are
buried deep beneath the level of consciousness and that they are indeed
unconscious would perhaps prompt some analysts to attribute this
sentence to the narrator. After all, Miriam is not consciously aware of
these feelings and what is more she would never have acknowledged
them. But interestingly, this transcription of what is supposedly hidden
from the character herself is rendered once again in Bulgarian with verbs
that very much implicate her own experience of these states in the text.
Do we, therefore, need to posit an external observer for the presentation
of these states; is the voice of the narrator necessary in sentences of character internal states for the reasons that Short et al. list?
On the evidence of the examples so far I would argue that the element
of unconscious half-sensing of certain states on the part of the character is
not a valid reason to attribute sentences of this kind to the narrator. A
language like Bulgarian that allows aspectual distinctions for these kinds
of verbs can be used as a nice test-case of the two possibilities of at-
177
tributing such sentences to the narrator or to the character. My passage

(A) offers evidence that the internal states of the character are more
strongly linked with that characters viewpoint than with the narrators
voice. Passage (B), on the other hand, shows that imperfective Imperfects
are not the only available option for rendering past tense stative verbs into
Bulgarian and that contextual signals impose semantic restrictions on the
interpretation of viewpoint in the paragraph as a whole. The most
instructive example, perhaps, is the passage quoted in (C). It shows that
Bulgarian allows the use of imperfective verbs in the Imperfect, even
where we are explicitly told that a character is unconscious of a certain
state. In this case it seems that the semantic properties of the verbs which
denote mental states are so strongly linked with character subjectivity that
their imperfective forms coupled with the Imperfect are considered to be
the stylistically appropriate option. The interpretation of the passage, thus,
is once again as arising from Miriams point of view, rather than as a
report delivered by a controlling narrator.
Further confirmation of the strong semantic links between stative and
mental verbs and character point of view is also witnessed in the comparison of a passage from Woolfs To the Lighthouse (1927) with its Bulgarian counterpart:
(D) But his son hated him. He hated him for coming up to them, for stopping and
looking down on them; he hated him for interrupting them; he hated him for the exaltation and sublimity of his gestures; for the magnificence of his head; for his exactingness and egotism (for there he stood, commanding them to attend to him); but
most of all, he hated the twang and twitter of his fathers emotion which, vibrating
round them, disturbed the perfect simplicity and good sense of his relations with his
mother. By looking fixedly at the page, he hoped to make him move on; by pointing
his finger at a word, he hoped to recall his mothers attention, which, he knew angrily,
wavered instantly his father stopped. But no. Nothing would make Mr Ramsey move
on. There he stood, demanding sympathy. (44; italics added)
he hated:
he stood:
(perfective Aorist) (present)
vibrating:
(perfective Aorist) [did not stop (to flutter)]
he hoped:
(imperfective Imperfect)
he knew:

wavered:
(perfective Aorist) (imperfective Imperfect) [got distracted]
178
Violeta Sotirova
would make:
(imperfective Imperfect) [nothing was in the

ability to]
he stood:
The pattern of choosing imperfective Imperfect verbs is repeated here as

well. In some cases, as with the verb hope, the translator has had no alternative; in others, as with the non-stative verb stood, the alternative
perfective Aorist could have been chosen. But all of the mental verbs that
denote states of the character are once again rendered with imperfective
verbs in the Imperfect. What is perhaps most revealing about this extract
is that in one place the translator spontaneously chooses the present tense
verb: appropriately, this occurs in the clause inserted in parentheses that
interrupts the coherence of the series of parallel clauses beginning with
he hated him. There are no strict rules about the sequence of tenses in
Bulgarian which allows the use of the present in subordinate clauses for
example, so these transitions from past to present are perceived as far less
ungrammatical than in English. In this case, the digression in the characters mind which is inserted in parentheses, is felt by the translator to be
an experience so immediate that she renders the verb in an aspect and
tense that foreground the duration of the event and its presentness, thus
placing us amidst its unfolding.
On the evidence of the Bulgarian translations of (A), (C) and (D), then,
I would like to suggest that the narrative internal observation point from
which the events and states described in these passages are viewed is the
characters. The imperfective aspect, coupled with the Past Incomplete,
creates this sense of experientiality. In the three passages from Lawrence
and Woolf(A), (C) and (D)we have witnessed a consistent pattern of
choosing imperfective verbs in the past Imperfect which strongly suggests
that the semantic cues of the three passages foreground the characters
own experience of the situation. Even when these experiences are not
verbalized and even when they are explicitly stated to be unconscious
states of the character, the aspectual rendering of the verbs in Bulgarian
provides evidence that we are still within the characters consciousness.
Example (B), on the other hand, which makes consistent use of perfective
verbs in the Aorist, brings to the fore the role played by other contextual
signals, such as adverbial expressions. In both cases, where imperfective
Imperfect verbs are chosen and where perfective Aorist verbs are chosen
179
by the translator, the semantics of the whole passage has obviously played
the strongest part in these decisions.
This analysis does not mean that for English speakers such sentences
sound more removed from the characters viewpoint than for Bulgarian
speakers. The semantic properties of these verbs suggest strongly enough
the characters internal experience of events and states, regardless of
whether the language system grammaticalizes or lexicalizes these semantic meanings in aspectual distinctions. In English, grammar and meaning
part in that the grammar does not allow stative verbs to take the progresssive aspect of experientiality. In Bulgarian, these distinctions are available
and it is the context that determines whether one chooses the imperfective
or the perfective. Once this choice is made, the interpretation of viewpoint
is very strongly suggested by the verbal aspect and past tense ending,
with the imperfective Imperfect placing us inside the characters
consciousness and the perfective Aorist denoting an external report of the
states of consciousness experienced by the character.
4 Conclusions
From a narratological and stylistic standpoint, then, sentences denoting
the internal states of characters merit a semantic analysis. Simply assigning them to narration, or placing them on the borderline between narration
and the other modes of thought presentation does not capture adequately
their effects on readers. The fact that such sentences cannot readily be
transformed into direct speech should not be taken as proof that they are
not expressive of the characters point of view, but rather should prompt
us to question the transformational account of free indirect discourse. If
we revisit some of our examples so far, we will find that there are enough
many other markers of subjectivity in these sentences to warrant an
interpretation of character internal point of view: in the Emma passage, all
of the words denoting emotion from different classes (gay,
thoughtless, felicity, happy, disappointed, etc.), the intensifier
extremely; in passage (A) from Sons and Lovers, the modal verb
could, the proximal deictic now, and again a series of nouns subjectively referring to the inner life of the protagonist (soul, strength,
energy, intensity, madness); in passage (C) from Sons and Lovers,
the past perfect of characters past experience (it had cost him), the
progressive of he was trying, the proximal deictic this; and finally, in
passage (D) from To the Lighthouse, the evaluative exaltation, sublim-
180
Violeta Sotirova
ity, egotism, twang and twitter of his fathers emotion, perfect simplicity, angrily. These contextual signals, along with the verbs which
denote mental and emotional states of the character are the semantic
guarantors for reading these passages as stemming from the characters
point of view and perhaps should be included in a broader definition of
the free indirect mode, not as discourse, but as a particular style of writing
consciousness. On the other hand, passage (B), with its contextual signals
of punctuality of the states and events and of their chronological ordering
would perhaps invite also from English readers a more narrator-orientated
interpretation. Ultimately, what cross-linguistic comparisons of this kind
bring to light is the importance of semantic and contextual analyses rather
than purely syntactic transformational accounts of the different modes of
consciousness presentation.
As McHale (1978: 263) points out, already Voloinov (1973) had
shown that free indirect discourse, or his quasi-direct discourse, cannot be
theorized in purely syntactic terms as a fusion of two possible modes of
report, direct and indirect. Rather, he, and later Baxtin (1975), see it as the
collision or sounding in harmony of two voices, of two angles of vision,
of two points of view. This argument in favor of a semantic analysis of
free indirect discourse does not mean that the syntactic properties of this
mode should be ignored, but syntax should not be allowed to determine
an interpretation; it should only be an explanatory tool in the process of
unpacking interpretations. The semantic argument that Voloinov and
Baxtin advance is fully justified linguistically by Adamson (1994) who
finds the semantic roots of the free indirect mode in the everyday
practices of empathetic deixis and echoic, or quotative, modalized
utterances. On this analysis, the style of writing character point of view
emerges as an independent form which is not derived from a transformation of a pre-existing construction (direct or indirect) or as a blend
of two constructions, but is based on linguistic practices that also have
their psychological counterpart: empathy and echolalia. This semantic account of the free indirect style opens up the possibility of broadening out
its parameters. It does not have to be the product of a syntactic transformation, but is governed by the semantics of experience. If we set it
apart from its alleged base forms: direct and indirect discourse, then the
connection that many have seen between sentences of free indirect discourse and inner speech also becomes more tenuous and the name discourse itself becomes questionable.
181
References
Adamson, Sylvia (1994). Subjectivity in Narration: Empathy and Echo. M. Yaguello
(ed). Subjecthood and Subjectivity. Paris: Ophrys, 18398.
Andrejin, Ljubomir et al. (1998). Bulgarska gramatika na suvremennija literaturen ezik.
Tom 2: Morfologija [Bulgarian Grammar of the Contemporary Literary Language,
Vol. 2: Morphology]. Sofia: Abagar.
Austen, Jane (1816). Emma. London: Penguin, 1966.
Baxtin (Bakhtin), Mixail (1975). Slovo v romane [Discourse in the novel]. Voprosy
literatury I stetiki [Questions of literature and aesthetics]. Moscow:
Xudoestvennaja literatura, 72233.
Banfield, Ann (1982). Unspeakable Sentences: Narration and Representation in the Language of Fiction. Boston: Routledge & Kegan Paul.
Brinton, Laurel (1980). Represented Perception: A Study in Narrative Style. Poetics 9,
36381.
Cohn, Dorrit (1978). Transparent Minds: Narrative Modes for Presenting Consciousness
in Fiction. Princeton, NJ: Princeton UP.
Fludernik, Monika (1993). The Fictions of Language and the Languages of Fiction.
London: Routledge.
Kuroda, Shigeyuki (1973). Where Epistemology, Style and Grammar Meet: A Case
Study from the Japanese. P. Kiparsky & S. Anderson (eds). A Festschrift for
Morris Halle. New York: Winston, 37791.
(1987). A Study of the So-Called Topic wa in Passages from Tolstoi, Lawrence,
and Faulkner (of course, in Japanese translation). J. Hinds et al. (eds). Perspectives
on Topicalisation. The Case of Japanese wa. Amsterdam: Benjamins, 143 61.
Lawrence, David Herbert (1913). Sons and Lovers. London: Penguin, 1973.
(1990). Sinove i ljubovnitsi [Sons and Lovers]. Tr. L. Aleksandrova. Sofia:
Profizdat.
Leech, Geoffrey (1971). Meaning and the English Verb. London: Longman.
Leech, Geoffrey & Short, Mick (1981). Style in Fiction: Linguistic Introduction to
English Fictional Prose. London: Longman.
McHale, Brian (1978). Free Indirect Discourse: A Survey of Recent Accounts. Poetics
and Theory of Literature 3, 24987.
Mey, Jacob (1999). When Voices Clash. A Study in Literary Pragmatics. Berlin:
de Gruyter.
Palmer, Alan (2002). The Construction of Fictional Minds. Narrative 10:1, 2846.
Pascal, Roy (1977). The Dual Voice. Free Indirect Speech and its Functioning in the
Nineteenth Century European Novel. Manchester: Manchester UP.
Quirk, Randolph & Greenbaum, Sidney (1977). A University Grammar of English. London: Longman.
Schmid, Wolf (2003). Narratologiia. Moscow: Iazyki slavianskoi kul'tury.
Short, Mick et al. (1996). Using a Corpus for Stylistics Research: Speech and Thought
Presentation. J. Thomas & M. Short (eds). Using Corpora for Language Research.
London: Longman, 11031.
182
Violeta Sotirova
Voloinov (Volochinov), Valentin (1973). Marxism and the Philosophy of Language. Tr.
L. Matejka & I. R. Titunik. New York: Seminar Press.
Woolf, Virginia (1927). To the Lighthouse. London: Grafton Books, 1977.
(2000). Kum fara [To the Lighthouse]. Tr. I. Vasileva. Pleven: EA.
TOM KUBEK
(Prague)
Focalization, the Subject and the Act of Shaping Perspective

A study of the history of the notion of focalization shows that the two distinct structural levels upon which it is considered 1 have, time and again,
given rise to an endemic conflict, and that the ensuing disputes between
two opposing camps of theorists have taken place not within the framework of either level, but across them. This is why they are constantly latent, and always ready to erupt. It would be possible to say that this occurs in this way because their subject (the subject of their investigation),
in connection with the definition of focalization, is not the same. For
Genette and his disciples this subject is the narrative (which itself is focal1
Genette speaks about focalization at the narrative level, and for him this corresponds
with selection of information (Genette uses the term selection even though he is
aware of the danger this brings owing to an implied concept of mimesis). Focalization
thus provides the basis for some kind of complex view of narration. Against this, Bal
situates the entire theory of focalization on the level at which the characters of a narrative focalize the world of narrative events most simply: the story. She states: In a
story, elements of the fabula are presented in a certain way. We are confronted with a
vision of the fabula (Bal 1985: 100). The sentence, Elizabeth saw him lie there, pale
and lost in thought, is therefore for her an example of focalization, in which, for Bal,
Elizabeth is the focalizer who does the focalizing, and she does not enquire whether
there is also someone else, focalizing both him and Elizabeth. Thus the focus on
narrative (or, as Genette says, the narratoror, beyond the convention that takes the
fiction into account, the author) is a strategy that brings about this type of focalization.
Moreover, Genette speaks also of focalization of the narrator, which, according to him,
is logically implied in the case of first-person narration. Genette thus regards it as inapposite to personify focalization at the level of focalizers by whose mediation the elements of the story would be focalized: according to him only the narrative itself may
be focalized, and as a complex entity. And Bal also says: Focalization is the
relationship between the vision, the agent that sees, and that which is seen. This
relationship is a component of the story part, of the content of the narrative text: A
says that B sees what C is doing (Bal 1985: 104; italics added). Thus she provides a
foundation for her theory of a mutually conditioned relationship between subject and
object, in which she characterizes the seen as that which sees.
184
Tom Kubek
ized); for Mieke Bal and her disciples it is the story (which is focalized by
means of the focalizers). The difference between their approaches is
reflected also in Genettes reply to Bal in the book Narrative Discourse
Revisited:
The rest of the Balian theory of focalizations develops according to its own logic,
based on her innovation (establishment of an instance of focalization composed of
focalizer, a focalized, and even, page 251, recipients of the focalizing), like the idea
of a focalization in the second degree. (Genette 1988: 76)
Bal thus establishes the instance (agent) of focalization, whereas for Genette it is rather the situation of focalization that is established. On the one
hand, Gerard Genette refuses to connect focalization with any of the
elements of narrative that we designate as the narrator or as characters, considering it a higher category than these 2 ; on the other, Mieke
Bal claims it is possible to delimit the term focalization on the level of
those elements, and in consequence to personify it, in the form of a focalizer. For Genette, focalization is connected with a limitation of the amount of information that the reader obtains through the text about the
fictional world. In Bal, focalization is not connected with this phenomenon. It is a matter of an activity (of relationships) that produces information that has already been selected; this is why Bal does not speak of a
zero-value focalization, and why she regards it as necessary to study both
object and subject of the relationship, separately, in other words, what is
focalized and what does the focalizing.
Therefore these two concepts can never be reconciled, despite the fact
that they both use the common term focalization. Both consider the question that they answer with the term focalization quite legitimately
however, as mentioned above, each on its own structural level. One
should be aware of these ambiguities in the term, and perceive them as a
productive area to which the attention of narratology must be directed.
2
Genette defines focalization in terms of the characters knowing more or less than
the narrator; however, in specific examples his narrator comes close to the concept
of the author, and often overlaps with the concept of the implied author, which Genette
refuses to admit to his theory. The peculiarity of the relationship between narrator and
focalization is attested also by Genettes dictum: For me, there is no focalizing or focalized character: focalized can be applied only to the narrative itself, and if focalizer
applied to anyone, it could only be the person who focalizes the narrativethat is, the
narrator, or, if one wanted to go outside the conventions of fiction, the author himself,
who delegates (or does not delegate) to the narrator his power of focalizing or not focalizing (Genette 1988: 73).
Focalization, Subject and the Act of Shaping Perspective
185
The attempt to reconcile these opposing views of focalization leads to paradoxical claims, also exemplified in Abbotts Cambridge Introduction to
Narrative (2002). Here, focalization is a strategy at a higher level than either narrator or characters, but it is important to keep in mind that focalizing is not necessarily achieved through a single consistent narrative consciousness. Focalization can change, sometimes frequently, during the
course of narrative, and sometimes from sentence to sentence (Abbott
2002: 190); Abbott consequently associates focalization with the narrative
voice (In this study I present focalization and voice as companion
concepts, Abbott 2002: 190). While voice, for Genette, is the consequence of the strategy of focalization, for Abbott voice is focalization itself.
I would like here to adhere to the definition of Genette, for whom focalization is closely connected with the overall semantic construction of
the narrative. It constitutes a sign of a strategy which distributes, and,
from our point of view, generates, meaning in a literary text. I am interested in the moment at which, says Genette, only narration itself can be
focalized, and the possibility of focalization is open only to that, who focalize the narration (or not focalize): this is the narrator, orignoring the
convention that evaluates the fictionthe author himself, who delegates
(or does not delegate) his task of focalization to the narrator (cf. Genette
1988: 73). To recapitulate: for Genette, focalization is connected with the
level at which narration itself is focalized.
Before turning to my main topic, I would like to draw attention to another productive area which opens up when the concept of focalization is
so defined, and which connects Genettes theories with those of the Prague structuralists, and of Jan Mukaovsk. For Mukaovsk, too, it was
the question of intention that constituted the central question in connection with the generating of meaning: the text is the vehicle of the intention, and the intention at the same time refers to the situations of author,
text and reader. For this reason, he introduced the notion of the subject in
this context: it is an abstract subject, contained in the structure of the
work itself, which is merely a point from which the whole structure can
be comprehended (Mukaovsk [1937] 2000: 258). Mukaovsk then
focused his attention on understand-ing the production, and the textual
location, of the subject, at a time (the first half of the 1940s) when he was
at the peak of his powers. And the essential content of the notions of
focalization in Genettes writings, and subject in Mukaovsks, must
186
Tom Kubek
prompt the question why the former did not need to use the term implied
author, and why the latter came so close to this concept.
In his Poetics of Composition, Boris Uspenskij distinguished a number
of structural levels on which a narrative point of view is formed as a
function, that he delimitated at a pragmatic level of the act of narration.
The function of this point of view is to lead the narration to fulfil a certain
purpose, and to understand this purpose it is necessary, according to
Uspenskij, to investigate the principles of its constitution. Uspenskij convincingly demonstrated the way in which the processes through which the
narrative point of view is manifested are closely connected with the formation of values in the fictional world. And his concept is in its own way
confirmed by Schmids model of the construction of perspective, which
amplifies Uspenskijs four structural levels (ideological, phraseological,
temporal-spatial and psychological) with a perceptual level that is hierarchically superior to them.
So if we perceive this point of view as a basic means of constructing
value in the fictional world, then it is necessary to reformulate the question of the character and intention of a textual realization that produces
the situation of focalization, or, more simplyof the character and intention of a textual instance of focalization. Therefore we are interested not
only in the range and depth, but also in the quality of a focalization
of the fictional world. This quality cannot be connected merely with the
density of the information that we acquire concerning the fictitious
world, but also with its value. However, in the present context our questions concerning the principles through which meaning is generated in a
text, and the principles through which the fictional world is constructed,
come, by virtue of the strategy of focalization, very close to the questions
of the intention of the narrative act, of the reading public, of the context,
and of the principles through which the reader re-produces meaning in the
form of a unique sense. Our point of departure is the area of semantics,
and we set out for the area of the pragmatics of a narrative act. Therefore
the procedure is fully in the tradition of the Prague school, for whom the
semantic gesturethe overall construction of meaning in a workis a
question of pragmatics.
1 The Subject According to the Prague School

The Prague structuralists connected the question of the subject closely
with the concept of the semantic gesture, as the point of maximum mean-
187
ing in the work where, according to Mukaovsk, both author and recipient participate. The subject will then be the
point from which the works structure can be perceived in all its complexity and in its
unity. It is therefore a bridge between poet and reader, who can project his own ich
into the subject and thus identify his own situation in relation to the work with that of
the poet. The subject may remain hidden in (but in no way absent from) the work, as
for example in the objective epic, or, on the contrary, be realized more or less
strongly (through first-person narration, the emotional cast of the work, the identification of the poet with one of the characters within the work, and so on). Therefore
the subject cannot be identified with the poet a priori, even when the work seems to
express the poets feelings, his relation to the world and to reality in a direct way.
(Mukaovsk [1941] 2000: 264)
And intentionality requires a subject, from which it proceeds and which is its source;
thus it presupposes a human being. The subject is in no way located outside the work
of art, but within it. It is a part of it. [...] The person who has worked out the words
and their import is the subject; the person who is addressed by them is also the subject. And these are not in essence two subjects, but one. (Mukaovsk [1944] 2000:
286)
The subject is something other than a concrete individual [...]. As long as we remain
within a work, the subject is a mere epistemological will-o-the-wisp, an imaginary
point. When it is made concrete, this point can be occupied by any individual at all, no
matter whether this is the originator or the recipient. In any event, the individual is
something that remains outside the scope of the work. (Mukaovsk [1946] 2000:
30708)
The extracts above are from various studies in which Mukaovsk deals
with the subject, cited here to provide a more focused idea of the form in
which the subject is perceived by him, and the areas of discussion with
which it is connected. But the basic features of the concept of the subject
do not change: it is an entity (or point) realized by the work. At the same
time, the work represents a boundary dividing the subject from its specific
product or producer, author or readeror rather a meeting-point between
the intentions of reader and text. According to Mukaovsk, the subject is
a mental construct uncovering the intention of the construction of
meaning and the unification of all its component parts. This point is fully
realized within the work, but its recognition (the fulfilment or creation of
meaning) depends on the activity of concretization, and therefore on the
activity of the recipient. Its result is then the subject, which is a product
of the intention embedded in the work. Mukaovsk speaks of a point,
which is the same term that he uses also in defining the semantic gesture.
Therefore, but not only for this reason, his definition of subject and
188
Tom Kubek
semantic gesture can be apprehended simultaneously. That is confirmed

also by the similarity of their formulations and the identity of the terms
used for their formation.
So if Mukaovsk perceives the subject as a point, it is possible to understand it as a kind of understanding, closed in terms of process, or a
subjectivization. On the other hand, however, Mukaovsk regards this
point as unattainable through any concretization whatsoever. So it is necessary to consider the nature of a semantic process where the task of delineating meaning lies constantly ahead of us. Here the subject is at one
and the same time a unifying principle underlying the semantic structure
(intention) as well as its realization in the form of a unique meaning, although it does not intersect with any potential unique sense. The temptation to perceive the subject univocally is caused by its delimitation as a
point, yielding a unique formation of meaning; a point from which it is
possible to have a single birds-eye view of the structure of the work. It
is a point at which the whole artistic structure of the work converges, and
in relation to which it is assembled, but into which any personality may
be projected, whether the perceiver or the author (Mukaovsk [1940/41]
2000: 15).
From the outset, Mukaovsks subject overlapped in part with the
narrator of a literary work (as emerges from his opening quotation), but
later its definition moved fully into the area in which the strategy of the
overall construction of meaning is locatedinto an area which is used also by the narrator as an instrument of his intention. We are then able to
perceive the subject in an analogous area where focalization (Genette)
or an implicit, implied, abstract author is defined. The semantic gesture
and the subject are then quantities of a pragmatic instantiation (situation),
together constituted by all the components of the work, which it unifies,
and which are related to it as their source.
The problem of the subject becomes central for Mukaovsk in his
later study, Zmrnost a nezmrnost v umn (Intentionality and lack of
intentionality in art [1943]). In it he reopens the problem of the identity
of the literary work, which now becomes for him the problematic unity of
sign and object, where the sign tends towards unity in meaning, and
therefore towards concretization, and the object constantly resists concretization because, as a thing, it belongs to the world of natural facts
which we cannot determine. In this productive tension Mukaovsk is trying to redefine the identity of the work, and to establish the limits of the
189
role and identity of the subject within this never-ending debate. He concludes accordingly that
it is not only the poet and the structure he imposes on the work that are responsible for
the semantic gesture that the recipient perceives in the work: a significant part is played also by the perceiver, and [] the perceiver often substantially modifies the semantic gesture, in contradiction to the poets original intention. (Mukaovsk [1943]
2000: 373)
In this, the semantic gesture, as a principle of semantic unity, is recognized as both intentional and unintentional.
The concept of the subject as a construct dependent on the intention
underlying the work and at the same time on a unique concretization
(with which it, however, does not quite overlap) on the part of the individual recipient, is indicated by the mechanism of this production, which
takes place in the area of intersubjectivity that we have recognized. Thus
Mukaovsks concept of the subject has shifted from its original delimitation as the intention generating the work (the authorial intention) in favor of the intention that is fully realized in the work, and an increased focus on the act of concretization that recognizes and generates the subject.
A similar shift also affects the semantic gesture, originally conceived by
Mukaovsk (in relation with the activity of the author) as a significant
process through which the work originates, and which is re-established in
the reader by reading (Mukaovsk [1933] 2001: 451). The reader therefore becomes primarily a passive solver of puzzlesthe addressee of a
code. But the semantic gesture later shifts entirely into the framework of
the work, defined in a broad sense, in favor of its own intentionthe intention which is generated in the conflict between sign and object, and in
which an essential role is now played also by the recipient.
Mukaovsks followers adopted this concept and virtually settled it in
the form in which it was suggested in the essay Intentionality and lack of
intentionality in art. So to prevent the work dissolving in the multiplicity
of its concretizations, ervenka (1992) sets up the authority of the work
as their original stimulus. He then defines the work as an organization of
linguistic signs, and at the same time as a structure of stimuli for further
linguistic and extra-linguistic activities on the part of the perceiving
subject (within the field of concretization). ervenka realizes that in
the process of concretization the work enters a broad context that has an
essential influence on the form of its unique concretization, and for this
reason he also considers sociological problems, which are reflected in his
concept of norms. (These norms help us understand the work, and are
190
Tom Kubek
used by the work uses in order that it may be understood; they are a product of the work and themselves produce it.) For him, the norms are beyond mans reach but reflect human interests and values. So these norms
are produced by history and develop in time. ervenka here develops
Vodikas ideas (see his Vznamov vstavba literrnho dla [The
Meaning Structure of a Literary Work (1992)]), and his conception of
the significance of context for the character of the work and for its
semantic development.
Similarly, Milan Jankovi considers the relationship between a unique
concretization and the intention of the work, and concludes that semantic
motion in a work is not given and does not achieve closure. For this reason, the signified can never be definitively established in a work. This
non-closure and non-givenness mean that the work constantly changes its
meaning while still maintaining its identitybecause possible meanings
at the same time intersect in it. So for Jankovi the work is situated at the
focal point of its interpretations (concretizations), behind which we identify its source, although this cannot be unambiguously designated, and it
therefore becomes abstract, a mere procedural motion. Like Barthes, Jankovi in consequence inclines to dismissing a unique interpretation (concretization) as unimportanthe recognizes a specific message as irrelevant, even if it is the only possible one (see his Dlo jako dn smyslu
[The work as a semantic process]).
From the above there emerge two important general questions: the social grounding of the work, and the connection between its semantic
process and time, including its attachment to time.
2 How Narrative Models the Perspective

Our central question is the manner in which the perspective with which
we attribute meanings to the work is modelled. If we define this perspective as an intersubjective space where the intention of the work meets a
unique concretization, it will be necessary to observe two phenomena in
succession: the activity of the text, and our reaction to this activity in this
context 3 . Cognitive semantics can be invoked here.
3
This unifying perspective, which we connect with the meaning of the notions of subject (as was specified by the Prague structuralists) or focalization (as understood by
Genette), is the result of a pragmatic situation that is constructed by the text, but the
reader creatively participates in it. However, it is necessary to distinguish this unifying
191
Let us now briefly refer to one specific type of narration from the
Czech literature and let us examine the text written by Josef apek Stn
kapradiny (Shadow of a fern) (specifically the beginning of the novel)
to discover the manner in which the perspective of reception is modelled
through entering a fictional world.
Rudolf Aksamit and Vclav Kala, comrades through thick and thin, bent over the
prey. You blue, black and green forest; you, forest brown and misty! Wild joy runs
through their poachers nerves; under their fingertips they had the carcase of an
animal, that beautiful carcase of a roebuck. He was theirs.
Vaek Vaek! hissed Aksamit. Rudy, oh Rudy! breathed Vclav Kala. They
were trembling, spellbound, an ecstatic passion seething within them, a drunken giddiness coursing through their veins. Oh, my goodness, what luck we had today! There
are no words to describe it.
Vaek and Rudy were bending over the roebuck, under their fingers there was the
carcase of the animal, yielding, still warm, still marvellously tense; and then a gamekeeper burst in from the thicket and roared: Dont move! Those were old unsettled
accounts, the gamekeepers voice was choking with fury. You generous, wild forest!
That roebuck carcase, still warm and tense. The joy of the poachers was cut short in
an instant, and in a sudden eruption it boiled over in the red lava of anger. Rudy
crouched behind the roebuck, an enraged beast raising its hackles within him; Vaek
found himself being flung at the gamekeepers throat. And now the fire of revenge has
blazed up: a gun has gone off, and that is Rudy shooting the gamekeeper. Bastards!
screams the enemy, and topples into the grass, head on one side. []
You gave me onethe body gasps, but there is no stopping the boiling lava, it
blazes volcanically and runs everywherebeneath the fingernails, up to the hot earlobes, full to the height of the eyes.
Hes had enough, wail Rudolf Aksamit and Vclav Kala, its had enough, that
corpse, still warm, still tense, that yielding corpse that will never be a gamekeeper
again. He wont take away that roebuck from us again, hell never strut about the
woods again, hell never go out to get his tobacco! (apek 1930: 5) 4
The first view is presented from the perspective of an outside observer

who at the same time confirms his knowledge of a wider context (comrades through thick and thin). Accordingly, we adopt an external perspective, at the center of which there appear two characters (identified by
rigid designators). But we are quickly invited to amplify this view: the
immediate surroundings of the scene are introduced, in the form of an expressive invocation (you blue, black and green forest). In the next sec-
perspective from those perspectives that mediate the story to us. Therefore, it is again
the perspectivization of the narrative space (as its value anchoring) in relation to the
narration (rcit) on the one hand, and the story (histoire) on the other.
Tr. Tom Kubek.
192
Tom Kubek
tion, we are encouraged by the text to alter our perspective or to unify it

with the perspective of the poachers (see the evaluative term beautiful,
which is subsequently confirmed also by the tactile implications of
warm). And in the section immediately following, the perspective
moves through the space that is described (and represents it at the same
time); it is amplified by the perspective of the gamekeeper, from which
we perceive the scene once more (with identical motivesroebuck,
warm, tense carcase), and as the internal perspective changes, its value
criteria also change.
Clearly, thus, the space is modelled before us through a shifting perspective, which allows access to a great deal of information about the fictional world. Within an operation of consolidation, we can then create the
complex perspective of the narrative scene. Conversely, we can then
check the relevance of the expressive statement addressing the forest and
judge it from the standpoint of any of the possible internal perspectives.
Then we note that in this sentence, which represents a transition between
the two internal perspectives, observed reality is strongly subjectivized,
and we are invited to perceive it through the eyes of another, without
being able definitively to refer the statement either to the poachers perspective, or to an overall narrative perspective. The characterization of the
two main characters (rigid designators), by means of the relationship
between part and whole in the form of the expression poachers nerves,
together with the unstable perspective, then suggests the form of the cognitive processes that we are to use for our operation of understanding. In
this way, we are returned to the area of our own experience, with a similar
type of literature and narrative style.
The unstable perspective that we encounter here guarantees that the
mediation will involve a large number of elements, from which the fictional world is constructed, and at the same time the proximity of the
point of view that we adopt in relation to this world (according to the demands of the text), giving the fictional world a strong granularity. The
perspective of the two poachers becomes the central axis of the narrative,
and determines the selection of the elements of the action and the manner
in which they are ordered. At the same time, however, it is continuously
controlled by the authority of the superior narrator, one which therefore
generates a perspective. The presence of this authority also confirms its
interpretative activity in relation to the object of its observation, and
therefore in relation to the perspective of the two poachers. It then uses
linguistic means to establish the receptive perspective of the addressee.
193
The narrative world thus has located within itself a dialogue between
both perspectives. It does not encompass merely a reciprocal confrontation of these two frameworks of knowledge and experience, even though,
from the standpoint of the modelling of the readers perspective as a cognitive action, it is precisely this confrontation that is its most important
property. The framework that we identify as belonging to the two poachers conversely reveals the framework of the superior narrative authority as
insufficient or incomplete, or breaches it (for example, it reveals the
reluctance of the superior narrator to provide some information, or even
reveals gaps in this superior framework). The reader then brings this dialogue up to date in the co-ordinates of the dialogue between the textual
situation and his own, during this cognitive unifying operation he activates his experience with similar frameworks, and on the basis of them
he creates the specific characteristics of the updated frameworks.
As for the modelling of the readers perspective, the entry into the narrative space mentioned above also uncovers three further processes: (1) it
shows that a high capacity to combine narrative elements will be necessary for the cognitive processes controlling the understanding and construction of the fictional world, in which (2) it will be necessary to refer to
our cultural encyclopedia (to interpret the notions of poacher, gamekeeper, forest, roebuck and prey, as well as their mutual combinations),
which also contains knowledge about social roles and possible relationships between individual elements within the narrative (for example,
poacher and gamekeeper) and which shows us that the lexis that we use
emerges from an environment of social interaction, in which is reflected
not only the capacity to use it, but also more generally (3) its capacity for
cognitive evaluation of our knowledge of the real world and our experience of it, as well as of the problem of the context to which the narrative
refers. These three processes then control and determine the meaning
which we assign to the narrative as its possible framework, and under the
influence of which, during the course of the narrative, we decode both the
partial and the more complex messages.
The perspective that is modelled in the narrative is totally dependent
on the grammatical resources of language. With their aid, it determines its
(and our) location in the narrative space, its distance from the object
depicted, the manner in which it is represented (whether the perspective is
stable or unstable) and at the same time, the logic of this representation
(the subsequent move to a close perspective and a local space of ob-
194
Tom Kubek
servation 5 ) and its direction in time (its concentration on the present temporal moment 6 ). The intensity of the perspective is also established (not
only through the expressive diction, but also in a number of the observed
details) and the levels of the perspective are established hierarchically,
together with the areas in which the fictional world is mapped (and their
density) from the point of view of narrative strategy, and the manner in
which this mapping is accomplished (the character of the information).
Within the framework of a cognitive operation of understanding, the
conceptual connections are then made, which is a process in which a variety of otherwise disconnected conceptual material is brought together.
This process draws on two basic overall forms of realization: connecting
above the scene and connecting in time. In it, we determine which elements specify the structure of cognitive representation evoked by the given narrative. Leonard Talmy speaks in this case of a scaffolding or an
axis around which linguistic material can be distributed or folded (cf.
Talmy 2000: passim). But as it is a proposal (although we have established the manner in which the perspective is modelled by the narrative
text), it is a subjective act, in which there occurs a preference for possible
frameworks and in consequence a preference for the possible elements
producing this meaning. The individual elements are then judged from the
point of view of their capacity to be inserted in some meaningful way
into this framework as a unifying complex. Without this operation, which
is a parallel structuring of the fictional world, we would, in the case of the
narrative, be dealing merely with an assemblage of individual juxtaposed
elements and not with a universe that is being united as a meaningful
complex of ideas.
To achieve this complex, it is necessary to supplement (concretize) it
with certain actions or conceptual networks at the same time. The narrative challenges us to adopt this behaviour, whose consequence is an individual realization of the supplementation which is the basis of our interpretative activity. Within the framework of this operation of supplementation, we can distinguish between the elements (relationships and phenomena) that are obligatory, which must be supplemented, those which
are optional, which it is possible to supplement, and those that are redundant. In the brief extract here quoted, trespassing can be seen as an obligatory element, the situating of the scene in the morning, for instance, can
5
6
The opposite would be a summary or synoptic perspective.

Alternatives would be retrospection and anticipation.
195
be seen as optional, and the question of animal rights is redundant.

Whereas the former two operations of supplementation aid an understanding of the narrative or the scene of the narrative, and its possible
meanings, the third leads us astray into misinterpretation. But supplementation is an operation that is evaluated during the course of time, and
for this reason elements that have been considered obligatory can become
less important during a subsequent reading, or vice versaoptional elements can become obligatory. It is also one of the ways through which
unreliability is constructed. In such a case, the text challenges the reader
to undertake a certain operation of supplementation in the framework of
the obligatory area which will be recognized later as redundant, or vice
versa. To make the unreliability a recognizable textual strategy and dominant within the semantic construction, it is of course necessary that this
operation, intended by the text, be carried out.
But the reader does not undertake only this operation of supplementation, broadly conceived, when challenged to do so by such textual
signals, but also a number of other operations, such as comparison (see
above, a mutual comparison of perspectives formed in the text of narrators or reflectors), categorization (again determined in terms of the relationship between one perspective and another), abstraction or schematization, summarization, and so forthin other words, operations that bear
on the basic conditions of reading depend on the individual capacity of
the recipient to carry out these operations, and on his widely based experience. Every lexical unit contains or evokes a series of cognitive domains
(foundations of its meaning)as an invitation to produce certain
conceptualizations, which should lead to a certain level of understanding,
and therefore to a certain form of the cognitive complex (but not, of
course, its totality). In this process, the flexibility of lexical units, in the
sense of their potential for incorporation in various complexes, is large,
and the literary narrative of this capacity of theirs is often used also in dependence on the level of their literary or experimental qualities.
Interpretation is the basis of this operation, as is stressed for instance
by Ronald W. Langacker, and is central to both semantic and grammatical
structures 7 . Langacker further states that linguistic significance resides
not only in the content of a lexical unit, but is a multidimensional
phenomenon, whose individual aspects reflect some basic cognitive capa7
Although long overlooked in traditional semantics, it is crucial to interpret (construct)

for both semantic and grammatical structure. (Langacker 1999: 5)
196
Tom Kubek
cities, that can be summarized under five general headings: specificity,

background, perspective, scope and prominence. In narrative, specificity
refers to our capacity to denominate the narrative entity which evokes
meaning on the basis of the information in the text, as well as our capacity
to distinguish variant meanings of this entity. The background can be seen
as the wide context (as well as the intertext) to which the given expression
refers. We have already discussed the meaning and form of perspective,
that bears closely on the value arrangement of the fictional world
discussed above. The scope has a similar character, as well as the location
of a narrative element within the scope or the distribution of the narrative
scope, which then significantly models the hierarchy of values of the unit
of meaning. So this might be constituted by a repetition of a certain
expression at the beginning of the article, or its positioning in some key
location in the text. For example, in Jan eps story Do msta (Journey
to Town) we encounter three colours (gold, blue and red) that are
individually varied in the text (for example, as corn, sky and poppies), but
always in the same order and therefore constituting a hierarchy, so that
these can be combined in the course of the text to construct an interpretation parallel to the three main characters of the story (father, mother
and son), and also a parallel to a more complex cultural interpretation in
which the colours are combined with the hierarchy of the family (God,
Mary and Christ). The beginning and the end of the narrative are framed
by the combination of father, mother and son (maintaining this order) and
the colours (which acquire the character of symbols in relation to the
above expressions) then appear in the same order in the center of the text.
In this manner, the given terms emphasize their key semantic position.
Prominence is a question of the denomination of the given element, for
example on the basis of social experience or class structure (e.g. king,
father, man, lad, human being). It is clear that the above categories refer
immediately to the formation of value within the cognitive space of the
statement (see below).Something substantial can now be said. The
operations that we have mentioned enable us to understand narrative,
indeed to read in the first place. But we should constantly be conscious of
the general nature and indeed the production of these operationsonly
thus can we guarantee that during our reception of the narrative we will
not be tempted by mimesis to replace the complex narrative structure with
a unique ideological interpretation of our own. As we shall assert further
below, but as we have already done in part in the introduction to this
chapter, our interpretation must also be perceived as part of a productive
197
dialogue between the potential capacities of the meaning and its realization in practice. This dialogue should lead to the communicative
situation which the narrative makes possible, initiates and controls at the
same time. Within the scope of the narrative, the causal (and temporally
determined) perspective in this operation of connection is recognized and
realized; it is not only questions of purpose but also questions of value
that come into play here in relation to the meaning. Value, and evaluation,
are relevant not only to the result of a narrativized process (poacher
preygamekeepercarcase/corpse), but also to the perspectivization of
this space, to each of its individual partsand therefore also to the determination of the hierarchy of values through the allocation of perspective
to it. In the opening quoted above, three perspectives are encountered
(those of the poachers, of the gamekeeper and of the overall narrator),
which impose a dialogue on this space and impart to it a three-fold set of
values that the reader must unify. Boris Uspenskij earlier noted that value
(the question of value as a structural and structuring element) is one of the
basic properties or qualities of perspective (cf. Uspenskij 1975).
Although, as we have established, the narrative text plays a considerable role in achieving unification by issuing a challenge to undertake this
cognitive operation, it is the recipient (the addressee) of the narrative text,
who realizes it definitively, by selecting a specific framework
scaffolding or axis, and the individual operation he carries out includes processes of combination and selection that happen in time and individually vary and combine the general scopes or frameworks. Here we
are already in the area of the unique semiotic process, and the text holds
controlling authority. This reciprocal activity points to the relevance to
the process of intersubjectivity, and thus to an intersubjective construction
of the fictional world. Hilary Putnam writes: The elements of what we
call language or mind penetrate so deeply into what we call reality,
that the very project of presentation of ourselves mapping something
independent on the language is fatally half-hearted (Putnam 1990: 57).
This observation, together with what has been said above, can be regarded
as defining our position as the subject of reception.
References
Abbott, H. Porter (2002). The Cambridge Introduction to Narrative. Cambridge:
Cambridge UP.
Abel, Gnter (1999). Sprache, Zeichen, Interpretation. Frankfurt a. M.: Suhrkamp.
198
Tom Kubek

Toronto P.
(1991). On Story-Telling: Essays in Narratology. Sonoma, CA: Polebridge Press.
(1988). Framing the Sign: Criticism and its Institutions. Oxford: Blackwell.
apek, Josef (1930). Stn kapradiny. Prague: Aventinum.
ervenka, Miroslav (1968). Vznamov vstavba literrnho dla. Prague: Karolinum,
1992.
Chamberlain, Daniel Frank (1990). Narrative Perspective in Fiction: A Phenomenological Mediation of Reader, Text, and World. Toronto: U of Toronto P.
Chatman, Seymour (ed) (1971). Literary Style. A Symposium. London: Oxford UP.
Cornell UP.
Christensen, Carleton B. (1991). Language and Intentionality. Wrzburg: Knigshausen & Neumann.
Culler, Jonathan (1981). The Pursuit of Signs: Semiotics, Literature, Deconstruction.
Ithaca: Cornell UP.
Doleel, Lubomr (1960). O stylu modern esk przy. Prague: eskoslovensk akademie
vd.
(1973). Narativn zpsoby v esk literatue. Praha: eskoslovensk spisovatel.
(2003). Heterocosmica. Prague: Karolinum.
Genette, Grard (1983). Nouveau discours du rcit. Paris: Seuil.
(1988). Narrative Discourse Revisited. Tr. J. E. Lewin. Ithaca: Cornell UP, 1998.
Herman, David (ed) (1999). Narratologies: New Perspectives on Narrative Analysis.
Columbus: Ohio State UP.
(1995). Universal Grammar and Narrative Form. Durham, NC: Duke UP.
Jakobson, Roman (1995). Dva aspekty jazyka a dva typy afatickch poruch.
M. ervenka (ed). Poetick funkce. Prague: H&H.
Jankovi, Milan (1992). Dlo jako dn smyslu. Prague: Prask imaginace.
Lamarque, Peter (1996). Fictional Points of View. Ithaca: Cornell UP.
Langacker, Ronald W. (1999). Grammar and Conceptualization. Berlin: de Gruyter.
Lanser, Susan S. (1981). The Narrative Act: Point of View in Prose Fiction. Princeton:
Princeton UP.
Meggle, Georg (ed) (1979). Handlung, Kommunikation, Bedeutung. Frankfurt a. M.:
Suhrkamp.
Mhl, Martin (1997). Die Handlungsrelativitt der Sinne. Bodenheim: Philo.
Mukaovsk, Jan (1933). Vznamov vstavba a kompozin osnova epiky Karla
apka. Studie II. Brno: Host, 2001, 45180.
(1937). Individuum v umn. Studie I. Brno: Host, 2000, 25558.
(1940/41). Strukturalismus v estetice a ve vd o literatue. Studie I. Brno: Host,
2000, 925.
(1941). Bsnk. Studie I. Brno: Host, 2000, 25974.
(1943). Zmrnost a nezmrnost v umn. Studie I. Brno: Host, 2000, 35388.
(1944). Osobnost v umn. Studie I. Brno: Host, 2000, 27590.
(1946). Problm individua v umn..Studie I. Brno: Host, 2000, 30335.
(1966). Studie z estetiky. Prague: Odeon.
(1971). Cestami poetiky a estetiky. Prague: eskoslovensk spisovatel.
199
(1982). Studie z poetiky. Prague: Odeon.

Palmer, Alan (2003). The Mind Beyond the Skin. D. Herman (ed). Narrative Theory
and the Cognitive Sciences. Stanford: CSLI, 32248.
Prince, Gerald (1987). A Dictionary of Narratology. Lincoln: U of Nebraska P.
Phelan, James (1981). Worlds from Worlds: A Theory of Language in Fiction. Chicago:
U of Chicago P.
(1989). Reading People, Reading Plots: Character, Progression, and the Interpretation of Narrative. Chicago: U of Chicago P.
(1996). Narrative as Rhetoric: Technique, Audience, Ethics, Ideology. Columbus:
Ohio State UP.
Phelan, James & Rabinowitz, Peter J. (eds) (1994). Understanding Narrative. Columbus:
Ohio State UP.
Putnam, Hilary (1990). Realism with a Human Face. Cambridge, MA: Harvard UP.
Rabinowitz, Peter J. (1987). Before Reading: Narrative Conventions and the Politics of
Interpretation. Ithaca: Cornell UP.
Rescher, Nicholas (1973). The Ontology of the Possible. M. K. Munitz (ed). Logic and
Ontology. New York: New York UP.
Richardson, Brian (ed) (2002). Narrative Dynamics. Essays on Time, Plot, Closure and
Frames. Columbus: Ohio State University.
Ronen, Ruth (2002): Representing the Real. Amsterdam: Rodopi.
Ryle, Gilbert (1949). The Concept of Mind. London: Hutchinsons University Library.
Shibatani, Masayoshi & Thompson, Sandra A. (eds) (1996). Grammatical Constructions:
Their Form and Meaning. New York: Oxford UP.
Singer, Alan (1993). The Subject as Action: Transformation and Totality in Narrative
Aesthetics. Ann Arbor: U of Michigan P.
Sternberg, Meir (1978). Expositional Modes and Temporal Ordering in Fiction. Baltimore: Johns Hopkins UP.
Talmy, Leonard (2000). Toward a Cognitive Semantics, Vol. 1. Cambridge, MA: MIT
Press.
Uspenskij, Boris A. (1975). Poetik der Komposition: Struktur des knstlerischen Textes
und Typologie der Kompositionsform. Tr. G. Mayer. Frankfurt a. M.: Suhrkamp.
Verhagen, Arie (2005). Constructions of Intersubjectivity. Oxford: Oxford UP.
Vodika, Felix (1948). Potky krsn przy novoesk. Prague: Melantrich.
CHRISTIAN HUCK
(London)
Coming to Our Senses: Narratology and the Visual

1 Introduction
Marco Polo is believed to have traveled about 14.000 miles during his
lifetime; Ibn Battuta, the great Arab explorer of the middle ages, managed
about 75.000. But both were dwarfed by the Englishman James Holman, a
retired naval officer, who traveled roughly 250.000 miles in the first half
of the 19th centurybefore the arrival of trains, steam boats and planes.
He trekked deep into Siberia, sailed to Brazil, rode through southern
Africa, explored unmapped parts of Australia and survived the banditinfested Balkans. However, the most remarkable thing about all this is
that Holman had been blind since the age of twenty-fourhe made all his
travels without seeing where he was going: he heard, smelled and felt his
way cautiously through the world. While vision gulps, tactility sips, his
biographer notes, an object yields up its qualities not all at once, at the
speed of light, but successively over time, and in sequence of necessity.
(Roberts 2006: 69) However, despite his obvious achievements, Holman
was never taken seriously by his contemporaries, and was soon forgotten.
His experiences were deemed invalid for the simple reason that he could
not use his visual sense: His sightlessness makes genuine insight impossible (Roberts 2006: xii). The Enlightenments epistemological paradigm
of the eyewitness did not allow for other sense data to become the basis
for new knowledge.
In this article, I want to compare two travelogues that mark the submission of the travel report to the paradigm of the eyewitness. The two
texts in question are Daniel Defoes Tour through the Whole Island of
Great Britain (172426) and Edward Wards account of his ramblings
through London in The London Spy, originally published as a periodical
between 1698 and 1699. While the two texts deal with roughly the same
subject matter, London around the year 1700, they present two very different accounts of it. In line with the centurys empiricist imperative to
202
Christian Huck
observe, both emphasize that they will only report those things they have
personally witnessed. However, the resulting reports could not be more
unlike. Defoes calm, plain, and objective description of the streets and
buildings of the city is contrasted by Wards rushed, exuberant and excited account of its inhabitants. How can the two descriptions be so different, when the perceived object is basically the same?
A literary historian might credit this difference in description to different political aims: the Whig Defoe is trying to present an economically
progressive Britain, while the Tory satirist Ward attempts to ridicule the
human follies of his fellow citizens suffering the consequences of (early)
modernity. Narratologically speaking, they consequently show very different points of view, they reveal a markedly different perspective on
things, they focalize different aspects of the city. However, instead of
explaining the differing accounts with reference to the ideological backgrounds of the authors and thus making only metaphorical use of the
terminology, I want to analyze a difference manifested in the creation of
two specific narrator-figures, the employment of their senses, and the relation between perception and reporting which these narrators reveal.
It becomes obvious, when analyzing the two texts more closely, that
while perception in Defoes text is restricted to the visual, the narrator in
Wards text employs all kinds of sensory perceptions. The attempt to describe and theorize the different narrators, then, leads to the question,
whether there is an aural, olfactory or even a haptic equivalent to a point
of view: a point of smell, maybe, or a point of taste? What would be the
difference between these? And could a specific mode of perceiving (a
story) influence the mode of reporting (in discourse)? As there are few
predecessors which to build on, and as studies of the impact of perceptual
regimes on modes of writing are still rare, all I will be able to offer here is
a tentative investigation of what is at stake in the relation between
perspective and the senses, and a few suggestions concerning how and
why this relation could and should be further explored.
2 The Rise of the Visual

However persistent and/or ambivalent the classical Greek privileging of
vision (Jay 1993: 33) and however ocularphobic (36) the Middle Ages
might have been, the ocularcentrism of post-Renaissance culture would
be difficult to deny: vision, aided by new technologies, became the dom-
Narratology and the Visual
203
inant sense in the modern world (45). The importance of the visual soon
became pervasive:
From the curious, observant scientist to the exhibitionist, self-displaying courtier,
from the private reader of printed books to the painter of perspectival landscapes,
from the map-making colonizer of foreign lands to the quantifying businessman
guided by instrumental rationality, modern men and women opened their eyes and
beheld a world unveiled to their eager gazes. (69)
Although by no means a homogeneous field, the visual sense came to be

dominated by the particularly influential scopic regime of linear perspective, embodied by the technical device of the camera obscura (cf. Crary
1990: 2729). Ldemann outlines how this scopic regime establishes a
specific observer position: It gives
the observer the illusion he could see without being involved, that he could see, without being seen, without changing the observed through observing and without himself
being changed by the act of observing: The subject that sees by means of linear perspective installs itself behind the window of the peep show [] in the position of a
secret, for himself and others invisible voyeur. Consequently, he is an empirical subject only in a very limited sense. While he is in the world in the emphatic sense that
the things of the world organize themselves according to his perspective [], he is at
the same time distanced from the world by this very act. Like the Cartesian cogito the
observer is bereft of his body. (Ldemann 1999: 66) 1
As I want to argue in the following, it is such an observer position that a

text like Defoes ascribes to its narrator, a narrator curiously situated at
the same time in and out of the world he describes. But it is also the observer position that forms the basis for the concept of the perceptive/reflective figure in (classical) narratology 2 .
Throughout the nineteenth century, a new mode of observing evolved:
the mirror was replaced by the lamp as the paradigm for (artistic)
vision (cf. Abrams 1953). This new scopic regime was one of subjective
vision, a vision that had been taken out of the incorporal relations of the
camera obscura and relocated in the human body (Crary 1990: 16). Two
aspects of this new development appear crucial. On the one hand, the replacement of the mirror by the lamp, or of the camera obscura by the la1
2
Here, as in the following, English translations of German texts are mine.

On perspective see also Jay (1993: 5155). Crary emphasizes that linear perspective
does not necessarily lead to the observer position embodied in the camera obscura, and
that the two are similar but not identical (cf. Crary 1990: 34). The observer position
described above is the effect of a scopic regime influenced by linear perspective and
the camera obscura.
204
Christian Huck
terna magica, is a replacement of one scopic regime for another; the

dominance of the visual remains unaffected. On the other hand, the new
scopic regime has to be interpreted in a specific way so that it can be integrated into the narratological framework. As Klepper (2004) has recently
argued, the central shift from the old to the new scopic regime is based on
the deconstruction of transparency: while older texts assumed the possibility of an impartial observer, later ones reveal the partiality of every
(subjective) observation. It seems to me that narratological theory takes
this later, adaptive, Jamesian stance as its starting point and reinterprets
earlier narratives accordingly, i. e. that they, also, were biased. However,
this theory inherits or adopts both the visual bias and the epistemological
model of the older scopic regime, because it interprets a constructive
mode of observation within the wider framework of perspectivism, of
which the (Cartesian) linear perspective is understood to be only one particular instance. In the framework of narratology, the observer is, as I will
argue, still watching from inside a camera obscura, albeit one which has a
distorting prism in its hole.
3 Ut pictura poesis: Narratology and the Visual

Classical narratological theory, from Henry James to Franz Stanzel and
Grard Genette, was developed in response to the novel of the 18th and
19th centuries. Given that these centuries mark the heyday of the primacy
of visual observation, it comes as no surprise that the classical texts of this
era and subsequently the theories concerned with these should also show
a strong visual bias (cf. Klepper 2004). The narratoror character whose
perceptions the narrator reportsis generally conceived as a subject that
perceives its (fictional) world almost exclusively visually.
The question whether such visual bias poses a problem for narratology
did not seem important to most theoreticians, who touch on it only slightlyif at all. Bals definition of focalization, for example, could not be
more visual: Whenever events are presented, they are always presented
from within a certain vision. A point of view is chosen, a certain way of
seeing things, a certain angle. Focalization is, she continues, the relation between the vision and that which is seen, perceived (Bal 1985:
100). Without further ado, she makes seeing stand in for all forms of
perception. Bal seems to follow Genette, who thinks it enough to take up
[] the slightly more abstract term focalization, to avoid the too specifically visual connotations of the terms vision, field, and point of view
205
(Genette 1980: 189). However, when revisiting his theory, Genette claims
that his only regret is that [he] used a purely visual, and hence overly
narrow, formulation. Consequently, he wants to replace who sees? with
the broader question of who perceives? (Genette 1988: 64). Similarly, in
their chapter on Focalization Martinez and Scheffel appear to realize
the reductive pairing of who sees and who speaks, but think it enough
to add in brackets: (seeing should be understood here in the more
general sense of perceiving) (Martinez & Scheffel 1999: 64). Finally,
Rimmon-Kenan also hopes with Genette and Bal that the more abstract
term of focalization can avoid the specifically visual connotations of
point of view, but admits that even this new terminology is not free of
optical-photographic connotations and proclaims that its purely visual
sense has to be broadened to include cognitive, emotive and ideological
orientation (Rimmon-Kenan 1983: 71). But although she declares her
intention to transgress the limits of the purely visual sense of
focalization and acknowledges that perception also includes hearing,
smell, etc., all her examples remain within the realm of the visual (77).
Quite obviously, this visual bias of narratological terminology and the
failure to amend it have not gone unnoticed. In what might be called postclassical narratology, I found at least two possibilities to interpret these
findings. The first follows the line set out already by Rimmon-Kenan and
claims, in the words of Niederhoff, that the metaphorical character of a
scientific term does not diminish its suitability (Niederhoff 2001: 45).
The conceptual model, this suggests, remains unhampered by the terminology. Chatman, for example, claims: Genette has always seemed to
mean more by focalization than the mere power of sight. He obviously refers to the whole spectrum of perception: hearing, tasting, smelling, and
so on (Chatman 1986: 192). Prince takes the substitution of seeing for
perceiving even further:
Note [] that the verb perceive is to be taken in a broad rather than narrow acceptation: to apprehend with the senses (to see, hear, touch, etc.) or with the mind, or with
something like their equivalent. In other words, what is perceived may be abstract or
concrete, tangible or intangiblesights, sounds, smells, or thoughts, feelings, dreams,
and so on. (Prince 2001: 44)
According to this line of thinking, one can amend the terminology and
leave the underlying model untouched. Consequently, Nelles, following
Jost, distinguishes between ocularization, the visual element of focalization, auricularization, the aural point of view (cf. Jost 1983), gustativization, olfactivization, and tactivilization (cf. Nelles 1997: 9596).
206
Christian Huck
The problem I have with such supplementation lies in the subordination

of different senses under a model that was quite obviously developed with
the visual in mind. When Prince defines point of view as yielding that
which might be perceived from a certain perspective (Prince 2005: 442),
he is simply substituting the wider term perceiving for the old seeing,
but consequently must suggest that we smell or taste from a certain
perspectivewhich, I think, already stretches the metaphor, and the
model, a bit too far, as does the idea of an aural point of view. I will
return to the problem of perspective in regard to other senses later.
A second line of response to the visual bias of the terminology seems
to accept that the terminology is not just arbitrary, but a metaphor we live
by, not a surface problem, but one that conceptually frames our thinking.
Consequently, Lanser affirms the visuality of the concept of point of
view by conceding, with John Berger and others, the primacy of the
visual over all other senses: perception is always structured upon a relationship of perceiver and perceivedupon a point of view (Lanser 1981:
4). In a similar way, Nnning and Nnning affirm the visual bias of their
term perspective: The traditional correlation of visual-optic and cognitive aspects, which is already conditioned etymologically, is as much a
constant of the term perspective as is the close relation to epistemological
dualism. (Nnning & Nnning 2001: 8) Quite obviously, this is a perfect
tool for analyzing works created within the 18th and 19th century framework of representational realism. For Nnning and Nnning, however,
perspective is, systematically, the prism through which all environmenttal stimuli are refracted (12)thereby turning a historically and culturally situated philosophical framework into a given premise for narratological reasoning. Neither can this model incorporate a sense like tactility, which defies a neat compartmentalization of object, idea and
subject-observer, nor is it suitable for radical forms of subjective perception, where the creative act goes beyond the refracting of given
stimuli. As I understand it, Nnning and Nnnings metaphor of the
prism allows them to include a subjective/constructivist perspective
into an otherwise Cartesian epistemologyby putting a prism into the
hole of the camera obscura, and leaving it otherwise intact. As a consequence, the narratologically conceived observer sees the (fictional) world
through a prism even if he is smelling or hearing.
Is point of view, as much as focalization and perspective, then, just
another example of the primacy of the visual in our culture and the hegemony of the scopic regime of perspective within this culture? Is nar-
207
ratology simply mirroring what a plethora of recent studies have identified as the dominance of the visual in modern culture? Is narratology,
then, just another instance of what McLuhan understands as a central
consequence of the rise of the Gutenberg Galaxy, that is, the reduction of
experience to a single sense, the visual, as a result of typography
(McLuhan 1962: 125)? Is there, as Uspenskij claimed in his article, a
Structural Isomorphism of Verbal and Visual Art (1972)?
To a certain degree I would answer these questions positively. Consequently, a use of terms such as point of view or perspective, which
affirms its visual bias and consequently limits itself to analyses of visual
perception within a certain cultural framework, is surely appropriate; also,
such analyses should do justice to the bulk of mainstream 18th and 19th
century novels. When it comes to dealing with other than visual sense
perceptions, though, I would disagree with Nelles that we can successfully examine these within the given framework. In the following, I will
attempt to exemplify the limits of the visual narratological terminology
(and framework) in a comparative examination of the above mentioned
texts by Defoe and Wardand their differing perceptual and narrative
modes. Here, McLuhans claim of the relation between seeing and printing will also have to be re-examined.
4 A Terminological Re-Approximation
The fact that I am dealing with two factual texts seems to by-pass large
parts of what is normally discussed under the terms perspective, focalization, or point-of-view, and what the title of this book reveals as the
central function of these terms: mediation. As Nelles defines it: Focalisation is a relation between the narrators report and the characters
thoughts (Nelles 1997: 79). Or, as Jahn elaborates in more detail:
Focalization denotes the perspectival restriction and orientation of narrative information relative to somebodys (usually a characters) perception, imagination, knowledge, or point-of-view. Hence, focalization theory covers the various means of regulating, selecting, and channeling narrative information, particularly of seeing events
from somebodys point of view []. (Jahn 2005: 173)
Stanzels Typenkreis, Genettes tripartition and Bals refinement as

much as Nnnings Perspektivenstruktur, all deal, essentially, with the
informational relation between a character and a narrator. So what if
208
Christian Huck
there is no character-perspective from which to distinguish a narrator-perspective, and consequently no mediation between the two?
At first glance, there seems to be no immediate distinction between a
subject of perception and a subject of narration in factual texts. Who perceives? Daniel Defoe. Who narrates? Daniel Defoe. Genettes famous incentive for dealing with focalization in the first place, the intention to distinguish between who perceives and who speaks, seems to become
rather irrelevant. But is that true? It is quite obvious that what Daniel
Defoe perceives is not the same as what he narrates: the diegetic world,
although not fictional, is still a version of the real world. It is as
unlikely that Defoe never smelled anything in the whole of Britain 3 as it
is that he never interacted with anyone on his travelsand of neither of
which does he tell us. But that does not mean he is lying, he does not
necessarily hold back information.
It is my conviction that both Defoe and Ward create a specific
narrator-figure whose conception is responsible for the selection of perceptions. The relation that I want to focus on, then, is the relation between
the bias of perception and the bias of narrating. As mentioned above,
I am going to concentrate on sensory differences of perception, leaving
ideological questions aside 4 . Also, I will leave aside the question what
degree of reality the perceptual position of the narrator actually had for
Defoe: was he so convinced of this perceptual position that he actually
masked any smell, sound, etc., so that his conscious perception actually
became purely visual? Or is it just a conceptual constriction of which he
was well aware? Was his perception determined by the discursive cultural
framework or did he simply write what he thought was expected from
him?
Putting aside questions like these and despite the dangers of adding
even more narratological terms to an already well stacked pile, I want to
distinguish the two separate acts involved here as slanted perception on
the one hand and narrative focalization on the other 5 . I think it important to uphold a distinction, terminologically and conceptually, between
the act of perception and the act of reporting (cf. Schmid 2008: 129
3
See further Cockaynes (2007) timely reminder of the sensual assaults the eighteenth
century provided.
However, as might be deduced from the following, certain ideological positions seem
to go hand in hand with certain perceptual positions.
Etymologically, the optical connotation of focus supplanted the older sense of
hearth. For me, then, focalization means concentrating on the heated center.
209
30)even when dealing with factual accounts. My proposed distinction

seems close to Jesch and Steins contribution in this collection, although
mine is not restricted to fictional texts, and neither does it deal with the
perceptual act of characters as such: The first element is the perception
of the world invented by the author through narrators and other agents
also invented by the author; the second element is the regulation of
narrative information within the communication between author and
reader (59).
However, against Jesch and Stein I would argue that both entities are
always present: never is there a perception without slant, and never is
there a narration without focalization. While this, I assume, should be
undisputed, as Schmid has successfully argued (cf. Schmid 2008: 120
21), the much more interesting question for me is whether there is a
connection between the way perception is slanted and the way reporting is
focalized. And more specifically: how does the way perception is
conceptualized influence what one has to say about the world one travels?
In my analysis of Defoe and Ward I will argue that there are culturally
and historically specific models that suggest specific relations between
the two acts, limiting the systematically available possibilities.
However, while the proposed terminology might be better adapted to
diferent sense perceptions, it still remains within the realms of epistemological dualism. As long as the conceptual framework of diegetic
world-making, of the distinction between narrator and story-world, forms
the foundation of narratological theory (and it might turn out to be indispensable), point of view and related concepts remain central: The
novel and other narrative genres cannot escape the question at stake because they necessarily model a world and afford a specific viewpoint on
this world (Klepper 2004: 460). As will become clear, texts like Wards
reveal the limits of such frameworks; and all I can offer here is a pointer
towards these limits from within this framework.
5 The Travelers Senses

Daniel Defoes A Tour through the Whole Island of Great Britain formed
part of a new vogue in travel writing, dealing with Britain instead of faraway and exotic places on the one hand, and discarding scholastic accounts on the other. A predecessor to Defoe admits in his preface that
voluminous Treatises of this Nature already seem to exist. But: what
so eminently distinguishes our Ingenious Author from most, if not all, is
210
Christian Huck
that he presents you here with nothing but his own Ocular Observations.
Older authors, confining themselves to their Studies, can only report
what they have taken upon the bare Credit of those, who were, perhaps,
more slothful than themselves (anon. 1694: n. pag.).
Defoe follows this new tradition. He also promises to report nothing
but what he has been an Eye-witness of himself (Defoe 172426: I, 48),
and he, too, praises his own work for not being raisd upon the burrowd
lights of other Observers (48). When he relates a long Fabulous Story
that some Historians (108) tell, he discards the fable with the following
words: I satisfy myself with transcribing the Matter of Fact, and then
leave it as I find it (108). However, this commitment is at the same time
the source of a central problem in Defoes book. Defoes use of letters,
which are supposed to be reports of several separate circuits, is to ensure
his status as an eyewitness. In order to prove that his report is accurate,
Defoe creates an easily discernible narrator figure who gives detailed descriptions of the traveled topography. In keeping with the empiricist doctrine of the age, the subjective point of view is to guarantee an objective
account 6 . However, as we know today, the (empirical) author Daniel
Defoe not only collected information on diverse travels that failed to
match the reported circuits, he also used several secondary sources, and
only much later brought the collected information into a coherent form.
The narrative account, it appears, was created at another place, and another time, than the diverse perceptions.
What I want to argue now is that the temporal and spatial detachment
of the act of perception and the act of reporting is mirrored in the perceptual position Defoe ascribes to his narrator. In whatever way the real
authors perception was slanted, the narrator in the text has a peculiar and
easily discernible slant of perception. In alliance with the 18th centurys
predominant concept of visual perception, Defoe seems to be traveling
within a transportable walk-in camera obscura; he poses as a distanced
observer to whom the world presents itself as if through an incorruptible
machine. The following depiction of a camera obscura represents this
conception perfectly.
See Jay (1993: 64): Intersubjective visual witnessing was a fundamental source of
legitimation for scientist like Robert Boyle. See further Crary (1990: 41).
211
Figure 1: Athanasii Kircheri, Ars Magna Lucis et Umbrae, 1671: 709 (extract).
The distance between Defoes narrator, who poses as an experiencing

figure, and the described objects correlates with the effects of the process
of writing. The distance which the writer, sitting in his study, experiences
in relation to the described objects finds its perfect embodiment in the
idea of a transportable walk-in camera obscura. The camera obscura, as
Crary has analyzed, performs an operation of individuation; that is, it
necessarily defines an observer as isolated, enclosed, and autonomous
within its dark confines (Crary 1990: 3839). The observer is cast as a
free sovereign individual and a privatized subject confined in a quasi-domestic space, cut off from a public exterior world (39). Incidentally, this
is also a perfect description of the situation of writers and readers emerging in the eighteenth-century 7 , sitting alone in their respective domestic,
private closets (cf. Heyl 2004: 50626): The reader can open the door of
a novel, enter, and quietly shut the door behind him (Zimbardo 1978: 8).
He or she is alone and not alone at the same time: A thousand readers
indeed stare, from their closets, into a single mirror of print, and each of
them does it alone (Hunter 1984: 285). The author, similarly writing on
his or her own, by means of this device, can, as Fielding has it, hold the
7
The changes in question here are to be seen in relation to the older paradigm of the
scribecollectively writing in a monasteryand the paradigm of the audienceexperiencing collectively in the theatre.
212
Christian Huck
Glass to thousands in their Closets (Fielding 1742: 6). Here, scientificphilosophical empiricism, the technique of the camera obscura and the
new situation of the reader seem to converge: The camera, or room, [or
book; C. H.] is the site within which an orderly projection of the world, of
extended substance, is made available for inspection by the mind (Crary
1990: 46).
However, the concept of the camera obscura has consequences not
only for the perceiving subject, but for the perceived object, too: a multifaceted thing is turned into a purely visual semiotic signsound, smell,
taste, touch; nothing of this can be reproduced within the box. Finally,
objects can be looked at without having the chance of looking back; perception is bereft of any reciprocity. Correspondingly, in the act of reading, the object of observation is present only as mediated and physically
absent: one can observe the object, without having to experience it in its
full presence and without having to fear that it might stare back. (And if it
does, as in some printed pictures, this feels uncanny.) Therefore, the sort
of actual and symbolic distance involved when perceiving an object visually through the camera obscura makes possible the uninvolved stance
of Defoes account. Narratologically, this position is embodied nowhere
better than in the heterodiegetic narrator of the classical realist novel, and
although Defoes narrator is strictly speaking homodiegetic, i. e. a part of
the story of Britain, he nonetheless appears to remain external to this
world. From this position, Defoe develops his calm and objective mode of
writing, his now legendary concise, clear prose, his plain, easy,
straightforward style (Backscheider 1986: 46, 53). However, such narrative mode would seem quite at odds with an observer who claims to be
in the thick of it, interacting and turn-taking. Rather, this mode of narrating is only credible in relation to the peculiar narrator-observer position
developed by Defoe, being there but not there at the same timelike the
camera (and the audience) in a classical Hollywood production, protected
by the fourth wall.
The only time Defoe gets carried away is when describing the society
at Tunbridge-Wells, a place full of Fops, Fools, Beaus, and the like
(Defoe 172426: I, 165), where you are surprizd to see the Walks
covered with ladies compleatly dressd and gay to profusion; where rich
Cloths, Jewels, and Beauty [] dazzles the Eyes (164). Bedazzled by
such spectacle Defoe rants about the dangers at such places, and the slander that increases such dangers, and finally has to cut himself short before
becoming too agitated: But this is a digression (166). Apart from this
213
perplexing encounter, Defoe appears to have avoided every contact with

living human beings while on his travels. Although the subtitle of his
book promises an account of the Customs, Manners, Speech, as also the
Exercises, Diversions, and Employment of the People, the inhabitants of
Britain receive little attention beyond the remark that the population is
increasing (Feldmann 1997: 37). When dealing with the capital, Defoe
admits that by London [], I mean, all the Buildings, Places, Hamlets,
and Villages containd in the line of Circumvallation (Defoe 172426:
II, 74)an imaginary line Defoe has drawn in order to measure the city.
The creation of this line seems to guarantee a vantage point, a perspective
from which the monstrous City (74) that London is for Defoe can be
tamed, that is, ordered, chartered. For him, London is nothing more than a
great Mass of Buildings (74), and consequently he describes the appearance and function of every important building, market etc.but never does he stoop to describe anything that cannot be contained within his
line of circumvallation and would suggest a reciprocal, interactional
approach: human beings, for example.
Ned Wards London Spy develops a markedly different narrator-figure.
In stark contrast to Defoes distanced view of Londons buildings, Ward
constantly reports on not only seeing other people, but also hearing,
smelling, and touching them. His slant of perception is noticeably different from Defoes vision. Instead of looking down from aboveDefoe
variously describes ascents to specific vantage points in order to have a
better (over-)view, the Spy is at eye-level. Instead of being distanced,
he is close. As a consequence, the Spy looks and is being looked at, he
hears other people and is heard, he touches them and is touched. Unlike
the linear perspective/camera obscura visual observer, Wards narrator is
all too aware of his own physical presence, aware of the effects his presence has on the observed objects and aware of the consequence such observations have on him. Finally, he interacts with peopleand his report
is full of people.
However, this experience of the social does not necessarily depend on
which senses are used 8 , but rather on how they are employed. The mode
of observation, whether technically or discursively formed, determines
whether co-presence, reciprocity and interaction, key elements of the social, are allowed for or not (cf. Bohn 2000). The perceptual stance of the
8
Kant thought the ear to be the privileged sense when it comes to the social, whereas
Simmel opted for the eye; see Bohn (2000: 32122).
214
Christian Huck
narrator-observer predetermines, at least to a certain degree, the selection

of as well as the relation to the objects described. And sometimes these
objects even cease to be mere objects.
But Ward is no complete exemption from the 18th centurys craze for
all things visual. Rather, the Spy employs his visual sense without completely subjecting his environment to the demands of unilateral perspectival spectating. Because the Spy includes other than visual sensations, it
seems his mode of visual observation is also different from Defoeshe
sees differently, because his relation to the objects of perception is formed
by several senses. Most importantly, the Spy remains receptive on all
channels. This, however, bears certain dangers. Again and again he seems
overwhelmed by his sensory experiences: nothing I could see but light,
and nothing hear but noise (Ward 1709: 29). He experiences the city
with a very acute, and sometimes over-powering awareness of [] sensory experiences (Hyland 1993: xv). His ears hear the sundry passingbells, the rattling of coaches, and the melancholy ditties of Hot Baked
Wardens and Pippins! (Ward 1709: 29); he sees the dazzling lights
whose bright reflections so glittered in my eyes (29); his nose smells the
narrow lane, as dark as a burying-vault, which stunk of stale sprats, piss
and sir-reverence (39).
In the realm of seeing there seems to be a strong preference for the observed object over the observer, especially under the scopic regime described in connection with Defoe. As described above, in the process of
observation, the observed is turned into an independently existing object,
while the observer is fashioned as a separately existing subject, uninvolved in the creation of the perceived object. On the other hand, most
of the non-visual senses require a closer relation to the object of perception; hearing and tasting, for example, are often conceptualized as taking
in the perceived object. The perceived object, finally, takes up such
presence that a specific perspective, defining the individuality of the subject, is not necessarily easily to be made outthe distinction between observer and observed, subject and object threatens to collapse.
The completely opposite mode of writingexcited, emotional, exuberant, with which Wards perceptions are reported, seems to result
from this closeness to the perceived objects and the way this closeness
affects him. Instead of traveling in a confined, distancing and sensually
diminishing camera obscura, Wards experiencing figure walks among
his fellow citizens. Being so close, he cannot help but experience other
sensory experiences than purely visual ones. The writing style that results
215
from this might be best described as linguistically overstuffed (Wall

1998: 137)drawing attention to the utterance itself rather than the utterer. The attempt to render the vast amount of multi-sensorially experienced
details into language leads to an exuberant style that constantly escapes
into similes and analogies when a complete rendering of the multisensorial experience becomes impossible. As a consequence, the reader
has great difficulty locating a similarly well-defined point of view to the
one we find in Defoe. Only rarely can the reader follow where the narrator is, and whom he is speaking to. Often, the reader learns more about
how the narrator is affected by his experiences than about the object that
(apparently) emanates the stimuli. And only rarely is the localization of
the experiencing figure possible: from where is an event heard, or smelled
in the dark? A well-defined, easily locatable perceptual stance appears to
be the privilege of the visual. Impressions, otherwise, do not seem to add
up to a well-defined diegetic world. And although I think there is no
doubt that Ward is narrating, it is not clear whether he is actually involved
in world-making.
6 Consequences
The slant of perception of every experiencing figure is heavily influenced
by the inclusion or exclusion of specific senses, by their emphasis or suppression. The perceptual position that results from such a slant, in turn,
influences what a report can include, and how the report is fashioned;
this, then, is what I termed narrative focalization. No report, obviously,
can render all sensory experiences. Therefore, every narration needs a
specifically equipped and positioned experiencing figure, which filters
what can be experienced and consequently determines, at least to a certain
degree, what can be reported. In turn, every form of report needs an
accompanying slant of perception. And while there is no strictly causal
relation between a certain slant of perception and narrative focalization,
there appear to be some culturally suggested default cases at least. Of
course, there is no inherent superiority among different possible positions;
Defoe is clearly able to see something that Ward can not, whereas Ward
can render experiences that Defoe remains blind to. However, the
specificity of these positions warrants close observation.
In the case of Ward and Defoe, attention to their perceptual stance
helps understand their peculiar positions. Both pose as participant-observers, roaming the world they observe and subsequently describe.
216
Christian Huck
Defoes perceptual stance, however, reveals him as an observer rather

than a participant, while Ward, on the other hand, is marked as a participant rather than a mere observer. The visual sense, and especially the
scopic regime of the camera obscura, appears to appeal to those who try
to be observers first, whereas a multi-sensorial approach seems to suggest
participation. Closer attention to the perceptual situation and its technical
and discursive determinations, then, might be able to distinguish historically and culturally specific embodiments of different narrative positions.
Defoes heterodiegetic narrator, who is at the same time close enough to
see everything but distanced enough not to be seen, as well as Wards
specific homodiegetic narrator, can be more closely analyzed with regard
to the use of their senses. And finally, looking at the cultural models or
regimes of perception and reporting in factual accounts might also reveal
what limitations and possibilities fictional narratives encounter.
References
Abrams, Meyer Howard (1953). The Mirror and the Lamp: Romantic Theory and the
Critical Tradition. New York: Oxford UP.
Anon. (1694). An Historical Account of Mr Rogers Three Year Travels over England and
Wales. Giving a True and Exact Description of all the Chiefest Cities, Towns and
Corporations. London.
Backscheider, Paula R. (1986). Daniel Defoe: Ambition and Innovation. Lexington: UP of
Kentucky.
Toronto P.
Bohn, Cornelia (2000). Sprache, Schrift, Bild. B. Heintz & J. Huber (eds). Mit dem
Auge denken: Strategien der Sichtbarmachung in wissenschaftlichen und virtuellen
Welten. Vienna: Springer, 32145.
Chatman, Seymour (1986). Characters and Narrators: Filter, Center, Slant and InterestFocus. Poetics Today 7:2, 189204.
Cockayne, Emily (2007). Hubbub: Filth, Noise and Stench in England, 16001770. Yale:
Yale UP.
Crary, Jonathan (1990). Techniques of the Observer. On Vision and Modernity in the
Nineteenth Century. Cambridge, MA: MIT Press.
Defoe, Daniel (172426). A Tour through the Whole Island of Great Britain (17241726).
3 vols. Ed. J. McVeagh. London: Pickering & Chatto, 2001.
Feldmann, Doris (1997). Economic and / as Aesthetic Constructions of Britishness in
Eighteenth-Century Domestic Travel Writing. Journal for the Study of British
Cultures 4, 3145.
Fielding, Henry (1742). The History of the Adventures of Joseph Andrews. London.
217

Heyl, Christoph (2004). A Passion for Privacy: Untersuchungen zur Genese der brgerlichen Privatsphre in London, 16601800. Munich: Oldenbourg.
Hunter, J. Paul (1984). The World as Stage and Closet. S. S. Kenny (ed). British
Theatre and the Other Arts, 16601800. Washington: Folger, 27187.
Hyland, Paul (1993). Introduction. E. Ward. The London Spy. East Lansing:
Colleague P, xi xxiii.
Jahn, Manfred (2005). Focalization. D. Herman et al. (eds). The Routledge Encyclopedia of Narrative Theory. London: Routledge, 17377.
Jay, Martin (1993). Downcast Eyes: The Denigration of Vision in Twentieth-Century
French Thought. Berkeley: U of California P.
Jesch, Tatjana & Stein, Malte. Perspectivization and Focalization: Two ConceptsOne
Meaning? An Attempt at a Conceptual Differentiation. In this collection, 5977.
Jost, Franois (1983). Narration(s): en dec et au-del. Communications 38, 192212.
Klepper, Martin (2004). The Discovery of Point of View: Observation and Narration in
the American Novel 1790-1910. Unpublished Habilitationsschrift, University of
Hamburg.
Lanser, Susan S. (1981). The Narrative Act: Point of View in Prose Fiction. Princeton:
Princeton UP.
Ldemann, Susanne (1999). Beobachtungsverhltnisse. Zur (Kunst-)Geschichte der Beobachtung zweiter Ordnung. A. Koschorke & C. Vismann (eds). Widerstnde der
Systemtheorie: Kulturtheoretische Analysen zum Werk von Niklas Luhmann. Berlin:
Akademie Verlag, 6375.
Martnez, Matias & Scheffel, Michael (1999). Einfhrung in die Erzhltheorie. Munich:
Beck.
McLuhan, Marshall (1962). The Gutenberg Galaxy: The Making of Typographic Man.
London: Routledge & Kegan Paul.
Nelles, William (1997). Frameworks: Narrative Levels and Embedded Narrative. Frankfurt a. M.: Peter Lang.
Niederhoff, Burkhard (2001). Fokalisation und Perspektive. Ein Pldoyer fr eine
friedliche Koexistenz. Poetica 33:12, 121.
Nnning, Ansgar & Vera Nnning (2001). Von der Erzhlperspektive zur Perspektivenstruktur narrativer Texte: berlegungen zu Definition, Konzeptualisierung und
Untersuchbarkeit von Multiperspektivitt. A. Nnning & V. Nnning (eds). Multiperspektivisches Erzhlen: Zur Theorie und Geschichte der Perspektivenstruktur im
Englischen Roman des 18. bis 20. Jahrhunderts. Trier: WVT, 338.
Prince, Gerald (2001). A Point of View on Point of View or Refocusing Focalization.
(2005). Point of View (Literary). D. Herman et al. (eds). The Routledge
Encyclopedia of Narrative Theory. London: Routledge, 44243.
Methuen.
Roberts, Jason (2006). A Sense of the World. How a Blind Man Became Historys
Greatest Traveller. London: Simon & Schuster.
218
Christian Huck

Uspenskij, Boris (1972). Structural Isomorphism of Verbal and Visual Art. Poetics 5,
539.
Wall, Cynthia (1998). The Literary and Cultural Spaces of Restoration London.
Cambridge: Cambridge UP.
Ward, Edward [Ned] (41709). The London Spy. Ed. P. Hyland. East Lansing: Colleague
Press, 1993.
Zimbardo, Rose A. (1978). Imitation to Emulation: Imitation of Nature from the Restoration to the Eighteenth Century. Restoration: Studies in English Literary Culture,
16601700 2:2, 29.
Part III: Transliterary Aspects of Mediation
ROLAND WEIDLE
(Hamburg)
Organizing the Perspectives: Focalization and the

Superordinate Narrative System in Drama and Theater
1 Introduction
Epic elements in drama and theater have been the subject of wide interest
in the last decades and especially in the last few years. Particular attention
has been paid to so-called narrator figures who present the audience with
embedded narratives. These teller-figures have been compared to intra-,
hypo-, meta-, homo-, hetero- or even privileged intradiegetic (Korthals
2003: 310) narrators in narrative fiction. Approaches like these have
opened up a new branch of academic interest toward a narratology of
drama (Sommer 2005: 123). Research has led to valuable insights into
the various dramatic functions and effects of such embedded narrators and
narratives (such as addressing the audience, distancing, alienation, irony),
the underlying assumption behind these approaches being that narrative
storytelling is not only to be found in narrative literature but in all literary
genres, and thus also in poetry and drama (Nnning & Sommer 2002:
108)1. This statement requires no further explanation and is amply
illustrated by the publication of whole monographs and anthologies
devoted to the analysis of embedded narratives with a transgeneric and
even transmedial perspective2. In the following, however, I would like to
pursue a somewhat different line of inquiry which so far has only
attracted the interest of a limited number of critics3. These scholars promote the investigation of extradiegetic narration of drama (as opposed
to the focus on intradiegetic narration in drama). Given the wide interest
that has been bestowed upon extradiegetic narration in other genres and
1
2
3
Here, as in the following, English translations of German texts are mine.

See Nnning & Nnning (2002); Herman (1999; 2003; 2004); Ryan (2004).
Such as Brian Richardson, Manfred Jahn, Richard Aczel and Holger Korthals, and, to
some extent, Rolf Fieguth, Patricia Suchy and Horst Spittler.
222
Roland Weidle
media, such as film and poetry, narratologys neglect of the narrating

voice in performance is actually quite surprising (Richardson 2001:
682). It is my contention that narratological concepts such as extradiegetic
narration and focalization can be fruitfully applied to, or at least tested in,
the analysis of drama in order to show that drama, like narrative literature
and poetry, is also narrated, if only in a broader meaning of the term.
However, I do not wish to focus on locating the extradiegetic narrator in
the communicative system of drama and clarifying his ontological status.
It is far more rewarding to concentrate on the question how and to what
effects drama is narrated than to concern oneself with the question as to
who narrates, that is, whether the extradiegetic narrator in drama is
synonymous with the author, the implied author or yet another agency.
When applying transgeneric narratological terms to drama one has to
be aware of several methodological pitfalls. Before setting out on the
course just laid out, it is therefore necessary to tackle briefly some of the
essential problems associated with such an approach.
2 Story, Narration, and Narrating

At the end of their helpful explication of the various layers of fictional
narration, Martinez and Scheffel come up with a table of six main planes:
event (Ereignis [Motiv]), story (Geschehen), plot (Geschichte), plotscheme (Handlungsschema), narration (Erzhlung) and narrating (Erzhlen)4. Whereas the first four levels belong to the realm of Handlung
(action, understood as the sum of all the elements that constitute what is
being told), the latter two belong to the category Darstellung (presentation), meaning: they deal with the how of telling5. For the purpose of
this analysis I would like to argue that Geschehen (story), as understood
by Martinez and Scheffel, constitutes the most relevant layer in the realm
4
Martinez and Scheffel understand Erzhlung as the narrated events in the order of their
representation in the text and Erzhlen as the actual presentation of the plot in a
manner particular to a specific language, code or media (cf. Martinez & Scheffel 1999:
25). For my choice of using the English narrating for the German Erzhlen (instead
of discourse, as one might also suggest) and its significance for my intermedial
approach see below.
See Martinez & Scheffel (1999: 25). In the subsequent chapters the authors refer to the
aspect of representation (Darstellung) as the How (Wie) and to the aspect of the represented (Handlung) as the What (Was).
Organizing the Perspectives
223
of the represented, whereas on the side of presentation the two categories

narration and narrating are of equal importance for our perspective.
Martinez and Scheffel understand story in a similar vein as Edward Morgan Forster, as a chain of chronologically ordered events without causal
relationships. Such a broad conception of story may not be necessary
when discussing plays of Shakespeare, Dryden, Lillo or Shaw. Here the
objects of presentation are never interconnected without causality. But
what about the drama of Beckett, Pinter, and Stoppard? Is causality a
useful concept when trying to analyze the actions of Stanley, Goldberg,
and McCann in The Birthday Party? Where is the causality between
Vladimir and Estragons final decision to leave and their subsequent not
moving in Waiting for Godot?
VLADIMIR: Well? Shall we go?
ESTRAGON: Yes, lets go.
[They do not move.] (Beckett 1990: 88)
With regard to the narrative contents of the plays of Beckett and Stoppard
it becomes evident that Genettes relations of linking, opposition, repetition (Genette 1983: 25) are just as valid in constituting stories as is causality for pre-modernist drama6.
I am aware that a strict separation of narrative content and narrative
presentation is problematic. After all, even our minimal definition of story
as events linked by repetition, chronology, similarity or any other relationship presupposes an agent that does the linking, a consciousness that
generates and/or identifies these links7. As mentioned earlier, Martinez
and Scheffel define two aspects of presentation in their theoretical overview: narration and narrating. The former is more or less synonymous
with Genettes rcit and Rimmon-Kenans text, designating the narrated events in the order of their representation in the text (Martinez &
Scheffel 1999: 25), differing from the story above all in the ways the
6
Whereas the contrast between Vladimir and Estragons words and their subsequent action, or rather: non-action, at the end of Waiting for Godot constitutes an oppositional
relation, the beginning of Stoppards Rosencrantz and Guildenstern are Dead may
serve as an example of a repetitional relationship between events. The first scene of the
play shows the two protagonists flipping coins with Rosencrantz always winning (cf.
Stoppard 1966: 1112). The coin-flipping-business, which continues well into the first
act, is of course only one of the many instances that contribute to the repetitional character of the play (cf. also the verbal and gestural repetitions in the play).
Jonathan Hart argues with reference to Ross Chambers and Seymour Chatman that it
is difficult to separate stories from their telling (Hart 1991: 140).
224
Roland Weidle
events are temporally restructured and regrouped8. The latter category

(narrating), refers to the actual process and form of narrating, including
the different possible forms of presentation as regards language, code or
media (cf. Martinez & Scheffel 1999: 25). Martinez and Scheffels inclusion of the aspect of narrating seems to be a very useful approach in applying narratological concepts to the analysis of drama. It draws attention
to the fact that narrations are always linked to actual narrating processes
that can occur in different semiotic systems, codes, languages, genres, and
media, and not only in narrative literature. Bearing in mind the dramatic
focus of this article, this inevitably leads us to reconsider our notions of
the text containing the narrated events, and even more so to reconsider
our ideas of the conditions that are constitutive of the narrating process. I
will first deal with the conditions defining the dramatic narrating situation
and then comment on the textual character of dramatic narration.
In the Routledge Encyclopedia of Narrative Theory Marie-Laure Ryan
posits six dilemmas in trying to define narrative. Apart from discussing
the problematic notion of narrative universals and other issues, she also
mentions the aspect of mediation:
Does narrative presuppose a verbal act of narration by an anthropomorphic creature
called a narrator, or can a story be told without the mediation of a narratorial consciousness? [] Some scholars have attempted to reconcile the narrator-based definition with the possibility of non-verbal narration by analysing drama and movie as
presupposing the utterance of a narratorial figure, even when the film or the play does
not make use of voice-over narration []. (Ryan 2005: 346)
In her summary of approachesnarratives are either verbally mediated,

told without mediation, or non-verbally mediatedonly the last definition
seems to offer a viable approach when applied to drama and its enactment
on stage. The scripts for Becketts mimes Act Without Words I and Act
Without Words II, which consist solely of secondary text, are neither
verbally mediated on stage, nor is dramatic representation in general
devoid of mediation9. Both as playtext and as performed drama the story
always depends upon mediation, either via the arrangement of written
8
For Genette's definition of rcit see Genette (1983: 27 and passim). For RimmonKenans definition of text see Rimmon-Kenan (1999: 3 and passim).
Stefan Schenk-Haupt drew my attention to the fact that mimes do not consist of secondary text, but contain solely directive information as to how to play the mime.
Yet, as is the case with secondary text proper, the directive text is not verbally
mediated on stage but itself mediatesvia other semiotic channelsthe story of the
mime.
225
words on the page (primary and secondary text in the script), or through
various linguistic and non-linguistic sign-systems on stage. Thus dramatic
stories are selected, arranged, mediated, and, in a wider meaning of the
term, narrated to us. Ryans third option, which allows for an utterance
of a narratorial figure behind stories, points in a more helpful direction,
althoughas so often in recent discussionswith an undue focus on the
narratorial agent. I would therefore like to postulatein slight
modification of Jahns concept of a superordinate narrative agent (Jahn
2001: 672)the existence and working of a superordinate narrative
system in drama with an anonymous and impersonal narrative function
controlling the selection, arrangement, and focalization (674) of the
story-data. Before illustrating some of the levels on which this superordinate narrative system operates, I would like to discuss the textual
character of dramatic narration and the problems associated with an unduly narrow focus on the dramatic playtext.
Playscript Mode
Holger Korthals German monograph Zwischen Drama und Erzhlung:
Ein Beitrag zur Theorie geschehensdarstellender Literatur (Drama and
Narration: A Contribution to the Theory of Story-representative Literature), published in 2003, is, to date the only monograph on the subject of
superordinate narrative structures in drama, or rather poetic drama, because Korthals makes it very clear that he is solely interested in the dramatic playtext. By displacing the literary text of drama from the theatrical performance we arrive not only at a seeming but at a genuine comparability to other story-representative genres such as the novel or novella (Korthals 2003: 60). According to Korthals, such a concentration on
the written text endows the reader of drama with privileges formerly ascribed only to readers of narrative literature, such as the suspension of the
irreversible linearity (Pfister 1977: 63) of dramatic performance. As
practical as such a comparison of dramatic and narrative texts may be, it
neglects the essential and defining feature of drama: namely, that it is
written to be performed. Although Korthals later concedes that there is a
link between the literary dramatic text and its performance (Korthals
2003: 74) and that the play-text always has to be seen with a view to its
performance, he ascribes a greater presentational force to the written sec-
226
Roland Weidle
ondary text than to its enactment on stage10 and concludes that the presented story can in principle be deduced from reading the dramatic playtext without having any notion about the norm of theater (cf. 5859).
The question as to whether drama should be approached mainly as a
dramatic text or a theatrical performance has been widely (and controversially) discussed. For the present purpose I will refer to Manfred Jahns
short and useful summary of the main views on this subject. He differentiates between three interpretive approaches. Whereas the school of
Poetic Drama roundly prioritizes the dramatic text, whose main interpretive strategy is a close reading which aims at bringing out the dramatic
works full aesthetic quality and richness, the school of Theater Studies,
by contrast, privileges the performance over the text (Jahn 2001: 661).
The main strategies of this approach include, according to Jahn,
considering a performance as the product of historical and cultural theatrical conditions, describing the sociology of drama, analyzing stage codes and semiotics, stage
histories, and the dynamics of collaborative authorship. (661)
This approach attacks the school of Poetic Drama for its academic isolation, and it considers the performed play as really the only relevant and
worthwhile form of the genre (661). The third school of Reading Drama, however, is an approach that Jahn favors and that I also would like
to take as a basis for my understanding of drama. It combines the approaches of Poetic Drama and Theater Studies:
Reading Drama is a school that envisages an ideal recipient who is both a reader and
theatergoera reader who appreciates the text with a view to possible or actual performance, and theatergoer who (re)appreciates a performance through his or her
knowledge (and rereading) of the text. Its interpretive strategies include performanceoriented textual analysis, paying particular attention to the secondary text of the
stage directions, and comparing the reading of plays to the reading of novels. (662)
Such an approach requires careful reading and analysis of the play text,
but it also challenges us in our imaginative efforts to visualize what we
read: if we are to make sense of the play, we must read with especially
active visual imagination (Campbell 1978: 187). Not only do we have to
pay particular attention to the secondary text of the stage directions, we
10
See for example his disputable comment on the final lines of Waiting for Godot, already referred to above: Thus the note in the secondary text makes the contradiction
between talking and acting far more visible for the reader than for the spectator who
can only be made aware of the importance of this contradictory behaviour through the
players demonstrative and conspicuous performance (Korthals 2003: 67).
227
also have to pay particular attention to the performative messages encoded in the primary text (the implied stage directions) that tell us something about how the story is to be enacted on stage. It is, however, important to draw a line here between drama and performance analysis. I am
not suggesting the inclusion of every possible way of visualizing and enacting the play-text on stage, nor do I propose to engage in an analysis of
individual performances by production collective[s] (Pfister 1977: 11)
with a specific aesthetic and/or ideological motivation. Instead, the playtext has to be understood as the playwrights instruction of how to present
or envision things on the stage. The written text, be it secondary or
primary text, does not only narrate (as Holger Korthals argues) but it also
has a clear referential and sometimes even imperative function in providing the recipient of the dramatic text (reader or production collective)
with the minimum information necessary to visualize or to enact the textual data11. The imperative function can be aptly illustrated with the opening passage from Becketts Krapps Last Tape:
A late evening in the future.
KRAPPS den.
Front centre a small table, the two drawers of which open towards the audience.
Sitting at the table, facing front, i.e. across from the drawers, a wearish old man:
KRAPP.
Rusty black narrow trousers too short for him. Rusty black sleeveless waistcoat, four
capacious pockets. Heavy silver watch and chain. Grimy white shirt open at neck, no
collar. Surprising pair of dirty white boots, size ten at least, very narrow and pointed.
[]
KRAPP remains a moment motionless, heaves a great sigh, looks a this watch, fumbles
in his pockets, takes out an envelope, puts it back, fumbles, takes out a small bunch of
keys, raises it to his eyes, chooses a key, gets up and moves to front of table. He
stoops, unlocks first drawer, peers into it, feels about inside it, takes out a reel of tape,
peers at it, puts it back, locks drawer []. (Beckett 1990: 215)
Whereas the first line A late evening in the future does not provide
concrete information as to its enactment on stage, the description of the
room, of Krapps appearance, and his movements are very specific and
tell us exactly how Beckett wanted the plays beginning to be staged12. It
is of course a far more difficult matter to talk about an authors intention
ifas for example in Shakespeares casewe do not have an original
11
12
For a discussion of the illocutionary force of stage directions see Jahn (2001: 66369).
The importance of the detailed stage-directions becomes evident in the course of the
play, as the desk, its position on the stage, and the drawers play an integral role in the
plays story.
228
Roland Weidle
text, licensed by the author. As this is not the moment to engage in a postmodern debate about the valid or invalid concept of the author, let it suffice to say that even dramatic texts, where it is difficult to arrive at the
authors intentions, provide the reader with signals as to their scenic enactment. Thus, reading and analyzing plays in this playscript mode
(Jahn 2001: 673) entails performance-oriented textual analysis (662). It
avoids treating the play-text as a purely textual phenomenon without
taking into account its theatrical orientation.
3 Superordinate Narrative System

Repudiating the idea of narratorless narratives, which holds that certain
sentences of fiction do not occur in the spoken language and cannot be
said to be enunciated by a narrator (Banfield 2005: 396), Richard Aczel
prefers to see
the narrator as an umbrella term for a cluster of possible functions, of which some
are necessary (the selection, organization, and presentation of narrative elements) and
others optional (such as self-personification as teller, comment, and direct reader/narratee address). (Aczel 1998: 492)
Such a view goes back to Chatmans differentiation between diegetic and

mimetic narration13, which in turn derives from a broader interpretation of
Aristotles concept of mimesis not only as the imitation of words but also
as the production of larger structuresin particular, structures of plot
(Chatman 1990: 110). According to Aristotle (Poetics 9.1451b) the poet
is a maker of plots [] and his mimesis is of actions (Aristotle 1995:
61; emphasis added), which holds true for both the epic genre and
tragedy. Aczels differentiation between necessary and optional functions
of the narrator is interesting and to some, like Genette, inacceptable because it renders actual telling as a possible but not necessary requirement
for the act of narration. Shakespeares Pericles presents an interesting
case in point for both the necessary and optional functions of a narrator.
Before the play proper begins, an actor impersonating John Gower, the
14th-century poet and author of the Confessio Amantis, one of the sources
of Shakespeares play, appears and speaks the following lines:
13
See Chatman (1990: 10923). To show a narrative, I maintain, no less than to tell
it, is to present it narratively or to narrate it (113).
229
To sing a song that old was sung

From ashes ancient Gower is come,
Assuming mans infirmities
To glad your ear and please your eyes.
It hath been sung at festivals,
On ember-eves and holy-ales,
And lords and ladies in their lives
Have read it for restoratives.
[]
This Antioch, then; Antiochus the Great
Built up this city for his chiefest seat,
The fairest in all Syria.
I tell you what mine authors say.
This king unto him took a fere
Who died, and left a female heir [] (I.18, 1722)14
The prologue introduces himself as the resurrected poet (Assuming

mans infirmities) and author of the story of Pericles that the audience is
about to witness. Does this fulfil Aczels optional criterion of the selfpersonification of the teller? Do we really have with Gower the superordinate narrative agent of Shakespeares play Pericles personified on
stage? Does the fact that Gower reappears throughout the play as a perceptive moderator who introduces each of the remaining acts make him,
as Jahn states, the behind-the-scene show-er agency in control of selection, arrangement, and presentation (Jahn 2001: 671)15? I agree with
Jahn when he concludes (although with caution) that Gowers discourse
acquires the status of an inset (672), that is, in narratological terms,
Gower has to be understood as an intradiegetic narrator, who is part of the
diegetic level16. The prologue appears at regular intervals in the play, announcing, summarizing, and commenting on the hypodiegetic story enacted by the actors who play the roles of Pericles, Marina, and the rest. The
interplay between the different diegetic levels of the prologue on the one
side and the playworld of the other figures on the other side is an integral
part of the plays composition and bespeaks a behind-the-scene show-er
14
15
16
Quotations from Shakespeares plays are from the Norton-Edition (1997), unless otherwise indicated.
Jahn also uses the terms narrating instance (660), dramatic narrator (669), superordinate narrative agent (672), and impersonal narrative function (674).
I disagree, however, with Jahns statement that Gowers diegesis is a first-degree narrative (672) which is shadowed by the superordinate narrative agents first-degree
narrative. Gowers narrative is secondary to the extradiegetic narrative of the superordinate system.
230
Roland Weidle
agency that stands also behind or rather above17 Gower. Gowers referring to his own sources in the quoted passage (I tell you what mine
authors say) aptly illustrates his embedded position and dependence on
other authorities.
3.1 Overt vs. Covert
Can one therefore argue that the narrative superordinate agent in drama,
unlike its extradiegetic counterpart in narrative literature, is always covert
and never visible in the play-text or the enacted play? The production
Isabellas Room (2004) of the Belgian theater group Needcompany provides an interesting case in this respect. The protagonist Isabella, aged 94,
sits in her room in Paris, which is filled with archaeological objects.
These objects assist her in reflecting on her past life. At the beginning of
every performance the author of the play, Jan Lauwers, who is also the
founder and artistic director of Needcompany, appears on stage and makes
an announcement:
Good evening ladies and gentlemen,
Welcome,
We will perform for you tonight Isabellas room. But before we start with the performance, I would like to tell you a bit more about the text I have written. You see,
my father died a few years ago. He left me a huge collection of more than 4000
archaeological and ethnographical objects. Some of them are presented here on stage.
[]
And so, I wrote the story of Isabella Morandi, performed by Viviane De Muynck.
[]
Next to Isabella, on my left, we have her dream: the desert prince, [] played by
Julien Faure and born out of a lie. [] The music is written by Hans Petter and
Maarten Seghers. Maarten also plays the grandson of Isabella, Franky. Can you still
follow?
[]
And in the corner, Misha Downey, the narrator, whoand this is unique in the history
of theatrewill play Isabellas erogenous zone for you.
[]
Sound: Dr Schneider, Light: Jeroen Wuyts.
Oh, I will play the man in the white suit. Now we can start. Misha, its all yours.
(Lauwers 2006b)
17
Or below in Genettes taxonomy.
231
Lauwers then remains on stage for the whole performance, sitting at the
upper side stage, dressed in white, observing his play, and from time to
time handing props to the actors18. At the end he takes part in the performance of a song. The matter becomes even more complicated as we
also have a narrator on stage, Misha, who fulfils the traditional functions of epic commentary, interaction with the audience, and standing in
for other characters. One has seen similar metalepses where a narratorfigure interacts with the audience and the other actors. The stage manager
in Thornton Wilders Our Town or Scullery in Jim Cartwrights Road are
a case in point. However, the situation here is different. Lauwers is the
author and superordinate narrative agent of Isabellas Room and he is
present on stage in every performance of the play. But he is also, as prologue and observer, part of the presented story-world and visible throughout the performance, although his function is restricted primarily to behind-the-scene tasks as handing props. In the prologue, Lauwers conflates the levels of author and character. Whereas the introduction of the
actors and the production ensemble and the teasing of the audience (Can
you still follow?) fulfil more or less conventional functions of the prologue, Lauwers references to his biographical background (his fathers
death, the collection of objects) and the comments on his motivation to
write the play bring together the worlds of the extra-textual and the diegetic. In other words: the superordinate narrative agent assumes an overt
presence in the play, or at least almost, because in the end Lauwers prologue is part of the written play-text, performed in the same waywith
minor alterationsin every performance, and thus acquires the status of
an inset. His interfering with other characters of the play-world, his visible presence, and the ensuing tension between the different narrative levels are an integral part of the play and contribute to its intricate design in
which times and places dissolve into another (Lvesque 2005). Yet
again, there is one aspect which supports the view that it is Lauwers as
superordinate narrative agent, and not as intradiegetic narrator figure, who
is overtly present in the performed play. In the published version of the
18
For a detailed account of the performance I am very grateful to Felix Sprang, an ardent
supporter and follower of Needcompany and their productions. This paper benefited
greatly from his knowledge and viewing experience. Felix Sprang also was so kind as
to provide me with the manuscript of his article on Turns on the Narrative Turn.
Showing and Telling in Needcompanys Early Shakespeare Productions and Isabellas
Room (cf. Sprang 2007).
232
Roland Weidle
play19 Lauwers does not appear; except for a short introductory note he
remains absent from the play. In the play-text the author and behind-thescene show-er remains hidden. Only through Lauwers self-personification on stage are we made aware of the actual presence of the superordinate narrative agent. That it is in fact in this function that Lauwers
appears on stage, and not as author, is confirmed by Lauwers himself:
It may seem paradoxical, but it is precisely for this reason that I take part in the performance myself this time. [] You might say that the simple fact that I am there onstage without taking part in the action makes sure that it is no longer about me.20
Taking Isabellas Room then as an exception to the rule that the superordinate narrative system is generally covert, the question remains, how
does this system manifest itself, or rather, how does it operate and how
can it be described?
3.2 Analepsis, Prolepsis, and Syllepsis
Korthals sees three possible planes on which dramatic texts reveal their
narrativity: (1) figural speechwhat I prefer to call intradiegetic narration, which so far has been the major focus of narratological analysis;
(2) secondary text, which explains the temporal and causal relationships
between individual scenes; (3) narrative-analogous structures constituted by the interplay between figural and authorial speech, primary and
secondary text (Korthals 2003: 186). In the following, I would like to focus on the latter, narrative-analogous structures, and more specifically
on dramatic examples of narrative order and narrative mood.
Most of the discussions of temporal relations between story and narration tend to focus on intradiegetic manifestations of analepsis and prolepsis, such as messengers reports, flashbacks of epic narrators or prophecies. To name only a few: Salieris retrospective showing, telling, and reenactment of his life in Mozarts Vienna in Shaffers Amadeus, Enobarbus account of Cleopatras arrival at Tarsus in Antony and Cleopatra
(II.ii,196 and passim), and the witches prophecies in Macbeth. Drawing
19
20
A French edition of Isabellas Room was published along with Lauwers The Lobster
Shop in May 2006 (cf. Lauwers 2006a).
Lauwers in an interview with Pieter TJonk in the De Tijd (Because Women are Tremendously Important [Omdat vrouwen ontzettend belangrijk zijn], 21.9.2004),
quoted in the English translation from Needcompanys website <www.needcompany.org>.
233
on Bernhardt Asmuths classification of four possible types of dramatic

prolepsis (authorial, plot-engendered, mantic, and presentational)21, Korthals privileges authorial prolepsis as constitutive of extradiegetic narration and sees its main realization in the secondary text and prologues (cf.
Korthals 2003: 215)22. Because of my doubts mentioned above regarding
the extradiegetic qualities of written secondary text and prologues I would
like to focus on Asmuths fourth type of prolepsis, neglected by Korthals:
the presentational prolepsis. Presentational prolepsis functions on two
levels, plot-externally and plot-internally, by drawing our attention to particular props, characters, words or moments, and thus endowing them
with specific (proleptic) meaning. On a level external to the plot this can
be done by virtue of specific lighting techniques, sound, music, and other
theatrical means. As this type of external prolepsis is most often employed by the production collective and not the author, it is of minor relevance for this analysis. The plot-internal prolepsis, however, is part of the
plot and depends to a large extent on the conventional expectations of
the audience concerning the genre, story, motives, props, characters and
manners of speech (Asmuth 1997: 121). Thus, for example, the repeated
appearance, hiding and finding of parts of clothing in Ortons farce What
the Butler Saw anticipate the great finale of un- and cross-dressing in the
second act23. Or to take an example on a more microscopic level: Faustus
last hour in Marlowes play, in which he awaits the arrival of the devils.
The striking of the clock at eleven, eleven-thirty, and midnight at shortening intervals mirrors (and heightens) Faustus and our own anticipation of
his death. Another manifestation of the temporal relationship between story and narrative is defined by what Genette calls syllepsis, the ana21
22
23
Asmuth (1997: 11422) differentiates between authorial (auktoriale Vorausdeutung)

and plot-internal prolepsis (handlungsinterne Vorausdeutung) subsuming under the
latter the certain (zukunftsgewisse) and the uncertain (zukunftsungewisse) variants of
plot-engendered (handlungslogische Vorausdeutung) and mantic prolepsis (mantische
Vorausdeutung). As a last type he defines presentational prolepsis (darstellende Vorausdeutung).
Korthals implies that authorial prolepsis and analepsis can also be achieved by means
of an interplay of secondary text signals, story-constitutive speech and narrative figural mediating speech (218). In his subsequent discussion of an obscure play, however, he falls short of illustrating this point convincingly (cf. 21820).
See also the ominous small cardboard box (act I; Orton 1976: 7) carried by
Geraldine at the beginning of the play. It is repeatedly referred to (and handled), and
the disclosure of its contents ends the play.
234
Roland Weidle
chronic groupings governed by one or another kinship (spatial, temporal,

or other) (Genette 1983: 85).
Shakespeares A Midsummer Nights Dream (ed. 1993) illustrates
variants both of spatial and thematic syllepses. Variants, becauseapart
from a minor anachronistic slip at the beginning of the play24the play
adheres to a strictly chronological sequence. As Pfister has shown, the
four plot lines of the Theseus-Hippolyta wedding, the four lovers, the
fairy world, and the mechanicals are grouped and thus narrated along
the spatial axis court-city-wood (cf. Pfister 1977: 259). However, these
plot lines are also arranged around a central motif, the moon. Not only is
the moon referred to in almost every scene, it also plays a prominent part
as an appointed date for the royal wedding (I.i,13), as the appointment
time for Lysander and Hermia (I.i,20913), as a source providing the necessary moonlight (cf. I.ii,83) for the mechanicals rehearsal of Pyramus
and Thisbe, and in setting the atmosphere in Oberon and Titanias first
meeting (cf. II.i,60)25. Finally, the moon even appears on stage, albeit not
personified very successfully by Starveling. The moon suffuses the plot
lines of the play with its polyvalent, ambiguous, and often contradictory
meanings, consequently establishing a sylleptic, superordinate narrative
web26.
A similar narrative and structuring function is fulfilled by the tape in
Becketts Krapps Last Tape. The three time-levels of Krapp at the age of
69, 39 and 29 are all arranged thematically around the actual process of
listening to and narrating on tape27. The important narrative role of the
tape is also supported by the prominent position of the tape recorder on
the table situated front centre28 and by the title of the play. Especially
with this type of thematic syllepsis, as Korthals poignantly states, it is
not the memories of a participating character but the arrangement of the
24
25
26
27
28
Theseus announces at the beginning of the play that his wedding will take place in four
days, at full moon (cf. I.i,13). The subsequent events until the wedding, however, take
up only three days.
According to Harold F. Brooks, the moon in her many aspects is regent of the Dream
(Shakespeare 1993: cxxviii).
In his Dictionary of Symbols and Imagery Ad De Vries lists a variety of different
meanings with which the moon in western culture has been associated. One of the
striking features of the meaningsbesides their richnessis their often contradictory
character. Thus, the moon stands for opposing values: female and male, fluid and volatile, constancy and inconstancy, etc. (De Vries 1976: 326).
For a helpful visualization of the time structure of the play see Jahn (2003: D 7.3).
Stage direction at the beginning of the play (Beckett 1990: 215).
235
narrative discourse (Korthals 2003: 193) that performs the functions of

the narrative agency.
4 Focalization
Genette separates the notions of voice and mood and argues that we have
to differentiate between the agent who narrates and the agent who perceives29, or, as he reformulates in his Narrative Discourse Revisited, between the question who narrates and the question where is the focus of
perception? (Genette 1988: 64) I will briefly repeat the three well-known
types of focalization (1): zero focalization: the narrator knows and says
more than any of the characters knows; vision from behind; (2) internal
focalization: the narrator says only what a given character knows; vision
from within; (3) external focalization: the narrator says less than the
character knows; vision from without.
For Korthals external focalization constitutes the default mode of drama. Characters, actions, and events are shown to us on stage and we perceive them from the outside without a narrator giving us additional information (cf. Korthals 2003: 27374). Zero focalization on the other
hand occurs, according to Korthals, through explanatory secondary
textswhat Issacharoff calls autonomous stage directions (Issacharoff
1989: 20), providing the reader with additional information that helps
him or her to transcend the perceptive horizon of the characters (and
eventually of the audience sitting in the theater). The following stage direction from Stoppards Travesties (act I) may serve as an example:
A note on the above: the scene (and most of the play) is under the erratic control of
Old Carrs memory, which is not notably reliable, and also of his various prejudices
and delusions. One result is that the story (like a toy train perhaps) occasionally
jumps the rails and has to be restarted at the point where it goes wild. (Stoppard
1974: 11)
29
It is interesting to note that there is still no general agreement on the difference between the narratological notions of focalization, point of view, and perspectivesee
Prince (1988: 31, 73); Niederhoff (2001); Surkamp (2005: 424); Prince (2005: 442
43). Genette explains the introduction of the term focalization on two grounds:
(1) to avoid the too specifically visual connotations of the terms vision, field, and
point of view (Genette 1983: 189), and (2) to distance himself from earlier confusions [] between mood and voice (Genette 1988: 64) connected to the term point
of view.
236
Roland Weidle
The problem with autonomous stage directions of this kind is that they do
not reach the audience, only the reader of the play-text. Thus they cannot
contribute to the narrative set-up of reading drama as sketched out
above. Korthals second manifestation of zero focalization, prologues,
and epilogues, similarly falls short of proving the existence of zero focalization in drama, because they are not, contrary to Korthals argument, external to the presented story and they do not deliver quasi secondary text
made audible on stage (Korthals 2003: 278). Does this inevitably lead to
the conclusion that zero focalization is not possible in drama? The following scene from Patrick Marbers Closer (1997) proves otherwise. The
play is about four characterstwo men, Larry and Dan, two women,
Anna and Alicewho attempt to find intimacy, but after various attempts
and partner changes in a pass-the-lover-fashion fail to get any closer
to each other. The following passage is from a scene (scene 8) which begins at a restaurant in the evening. Anna tells Dan about the encounter she
has had with her freshly divorced husband Larry, which had taken place
only hours before at the same restaurant:
DAN So has he signed?
ANNA Yes
DAN Congratulations. You are now a divorcedouble divorce. Sorry. (Dan takes
her hand.) How do you feel?
ANNA Tired. (Dan kisses her hand, Anna kisses his.)
DAN I love you. And I need a piss. (Dan exits. Anna reaches into her bag and pulls
out the divorce papers. Larry enters.)
LARRY (Sitting.) Afternoon.
ANNA Hi. (Larry looks around.)
LARRY I hate this place.
ANNA At least its central. [] (Marber 1999: 57).
With Dan leaving the table, Anna pulling out the divorce papers, and
Larry joining Anna at her table Marber goes back in time half a day to
Annas meeting with her soon-to-be ex-husband (Afternoon). In the
following dialogue Larry offers to sign the divorce papers under the condition that Anna sleeps with him for one last time: Be my whore and in
return I will pay you with your liberty (58). Before Anna can respond,
Larry goes to the bar (he exits) and Dan appears again and resumes the
previous dialogue he is going to have with Anna that evening. She tells
him that she in fact did sleep with Larry, a piece of information which
upsets Dan and occupies most of the ensuing dialogue. The scene ends
with a fusion of both time levels and all three characters present on stage:
237
DAN [] I think you enjoyed it; he wheedles you into bed, the old jokes, the strange
familiarity, I think you had a whale of a time and the truth is, Ill never know unless
I ask him.
ANNA Well why dont you? (Larry returns to the table with two drinks. Vodka tonic
for Anna, Scotch and dry for himself.)
LARRY Vodka tonic for the lady.
ANNA (To Larry) Drink your drink and then well go. (Larry looks at her. To Larry.)
Im doing this because I feel guilty and because I pity you. You know that, dont you?
LARRY Yes.
ANNA (To Larry.) Feel good about yourself?
LARRY No. (Larry drinks.)
DAN (To Anna.) Im sorry
ANNA (To Dan.) I didnt do it to hurt you. Its not all about you.
DAN (To Anna.) I know.
Lets go home (Dan and Anna kiss.) [] (61)
By leaving Dan on stage during the analepsis the audience is able to assume a perspective superior to the characters Larry and (especially) Dan,
because weunlike Dando not have to rely on Annas account and her
internal focalization of her meeting with Larry. We can witness the encounter at the same time from a superior perspective. Or, in other words:
when viewing this scene the audience is able to share in the superior
knowledge that results from being able to compare simultaneously Annas
version with reality, something that cannot be done in narrative fiction.
Another example of zero focalization occurs in the third scene of the same
play, in which Dan and Larry chat on the internet. It is a split-scene
showing Dan sitting at his computer in his flat and at the same time Larry
at his computer at work. They type the words as they speak and their
dialogue appears on a large screen simultaneous to their typing it (22).
The audience obtains a birds-eye view sharing the superordinate
narrative instances vision but also each of the characters internal focalization of the screen with the words appearing as they are typed (and spoken). The contrast between the limited perspectives of the figures on the
one hand and the superior perspective is even heightened by the fact that
Dan pretends to be Anna, which of course Larry cannot see. These are
extreme examples of zero focalization in drama and one could in fact argue that the physical nature of the performance and the fact that weunlike readers of narrative fictionare always able to see and compare per-
238
Roland Weidle
spectives for ourselves, move drama in general closer to zero than to external focalization30.
This still leaves Genettes third type unaccounted for, internal focalization, of which Korthals says that it can appear either on the plane of the
figures speech such as in soliloquies, asides, and in narrative mediations
like reports and teichoscopies, or in the form of theatrical stagings of
mental events (Korthals 2003: 282)31, such as apparitions, dreams, memories, and the like (cf. 28295). It is almost self-evident that, for example,
Katherines dream in Shakespeares Henry VIII (cf. IV.ii,8084), which is
being performed while she is shown sleeping on stage, presents an instance of internal focalization. However, one may ask whether the fact
that we can see the focalizing agent on stage, the sleeping Katherine,
would not also allow for zero and external focalization. In this respect, it
is worth noting that the simultaneous presence of focalizing subject and
focalized object on stage seems to be the standard case in drama. Due to
the corporeal nature of scenic presentation and the immediacy of visual
perception the focalizing subjects on stage are always at the same time focalized objects32.
Hamlets vision of his fathers ghost in the closet-scene is another example, although more complex because of the fact that the ghost was in
fact seen by Horatio, Barnardo, and Marcellus in the first act, but is not
visible to Gertrude in this scene. If Hamlet were not the main character of
the play, one could actually view this scene as an illustration of Gertrudes, and not of Hamlets internal focalization (cf. Korthals 2003: 283).
The matter becomes even more complicated once we turn to modern and
postmodern drama. In Stoppards Travesties the old Carr remembers, or
thinks he remembers, the events that occurred in the years 1917/18 in
Zurich when he was in his twenties, employed at the consulate and when
he met James Joyce during his involvement in an amateur production of
Oscar Wildes The Importance of Being Earnest. His faulty memory
twists the facts and conjures up fictional encounters not only with Joyce,
but also with the Dadaist Tzara and Lenin. In a stage direction (act I)
Stoppard says that during the first encounter between Joyce, Tzara, and
30
31
32
For a brief discussion of the differences between Pfisters concept of perspective structure, Spittlers presentational perspectives, and Genettes types of focalizations see
Korthals (2003: 27476).
Korthals borrows this phrase from Richardson (1988: 204).
For a differentiation between focalizing subject and focalized object, and between
focalization from within and without see Rimmon-Kenan (1999: 7576).
239
Lenin it is possible that CARR has been immobile on stage from the
beginning, an old man remembering (Stoppard 1974: 5). If that
suggestion is taken up by the director and old Carr is on stage while we
see his faulty recollections, we have a similar case to Katherine in Henry
VIII, who is presented sleeping on stage while the audience sees her
dream enacted, namely: a case of impure internal focalization. Both
focalizing agent and the focalized are visible on stage. If, however, Carr is
absent from the stage or is part of the recollected memories as young
Carrand this is the case several times in the play, the presented memories clearly seem to be a case of internal focalization. But that leaves us
with another problem and lays open the unreliability of Carrs memories.
One may ask how Carr can focalize anything which he was not a part
of33? Another scene (act II) from the same play shows young Carr denying having fantasies about the librarian Cecily, followed by an enactment
of exactly these fantasies from Carrs-mind view:
[CECILY: ] dont talk to me about superior morality, you patronizing Kant-struck
prig, all the time youre talking about the classes youre trying to imagine how Id
look stripped off to my knickers
CARR: Thats a lie!
(But apparently it isnt. As CECILY continues to speak we get a partial Carrs-mind
view of her. Coloured lights begin to play over her body, and most of the other light
goes except for a bright spot on Carr.) (Faintly from 1974, comes the sound of a big
band playing The Stripper. CARR is in a trance. The music builds. CECILY might
perhaps climb on to her desk. The desk may have cabaret lights built into it for use at
this point.) [] (Stoppard 1974: 52)
Not only do we have here Young Carrs impure internal focalization

(both the focalizer and his fantasies are focalized), but this scene also
represents Old Carrsprobably faultyinternal focalization of Young
Carrs internal focalization of his fantasies about Cecily.
5 Conclusion
The narratological concept of focalization as a filter through which the act
of narrating takes place is problematic when applied to the analysis of
drama. Because of dramas physical and visual nature and the material
presence of the actors, focalization in drama, or to be more precise: in
33
Young Carr leaves the stage for the first time well into the first act on p. 34 and returns
as Old Carr on p. 42, shortly before the end of the first act. Thus, the intermediate
scenes are focalized by someone who actually was not in the position to focalize them.
240
Roland Weidle
reading drama, appears to be less dependent on the mediating process.

Of course, even in drama, focalization does not take place without narration (by means of a superordinate narrative system), but the relation between narration and focalizing seems to be less prescriptive and more
flexible than in narrative fiction. Mieke Bals formula for fiction x relates that y sees what z does (Bal 1985: 45) turns into x relates while y
sees what z does, and here the y refers both to characters on stage and
to the audience. As for the extradiegetic narrative voice of drama, the superordinate narrative system, I hope to have shown that a transgeneric and
transmedial application of a broader conception of narration lays open the
narrative planes and channels of drama and directs more attention to the
how of dramatic mediation than to the question, who narrates drama.
References
Aczel, Richard (1998). Hearing Voices in Narrative Texts. New Literary History 29:3,
467500.
Aristotle (1995). Poetics. Ed. & tr. S. Halliwell. / Longinus. On the Sublime. Tr. W. H.
Fyfe. / Demetrius. On Style. Ed. & tr. D. C. Innes. Cambridge, MA: Harvard UP.
Asmuth, Bernhardt (1997). Einfhrung in die Dramenanalyse. Stuttgart: Metzler.
Toronto P, 1997.
Banfield, Ann (2005). No-Narrator Theory. D. Herman et al. (eds). Routledge Encyclopedia of Narrative Theory. London: Routledge, 39697.
Beckett, Samuel (1990). The Complete Dramatic Works. London: Faber and Faber.
Campbell, Sue Ellen (1978). Krapps Last Tape and Critical Theory. Comparative
Drama 12, 18799.
Chatman, Seymour (1990). Coming To Terms: The Rhetoric of Narrative Fiction and
Film. Ithaca: Cornell UP.
De Vries, Ad (21976). Dictionary of Symbols and Imagery. Amsterdam: North-Holland
Publishing Company.
Fieguth, Rolf (1973). Zur Rezeptionslenkung bei narrativen und dramatischen Werken.
Sprache im technischen Zeitalter 47, 186201.
Hart, Jonathan (1991). Narrative, Narrative Theory, Drama: The Renaissance. Canadian
Review of Comparative Literature 18:23, 11765.
Herman, David (ed) (2003). Narrative Theory and the Cognitive Sciences. Stanford: Center for the Study of Language and Information.
(ed) (1999). Narratologies: New Perspectives on Narrative Analysis. Columbus,
Ohio: Ohio State UP.
241
(2004). Toward a Transmedial Narratology. M.-L. Ryan (ed). Narrative Across

Media: The Languages of Storytelling. Lincoln, NE: U of Nebraska P, 4775.
Herman, David et al. (eds) (2005). Routledge Encyclopedia of Narrative Theory. London:
Routledge.
Issacharoff, Michael (1989). Discourse as Performance. Stanford: Stanford UP.
Jahn, Manfred (2001). Narrative Voice and Agency in Drama. New Literary History
32:3, 65980.
(2003). Poems, Plays, and Prose: A Guide to the Theory of Literary Genres.
<http://www.uni-koeln.de/~ame02/ppp.htm> (6.10.2006).
Korthals, Holger (2003). Zwischen Drama und Erzhlung: Ein Beitrag zur Theorie geschehensdarstellender Literatur. Berlin: Schmidt.
Lauwers, Jan (2006a). La Chambre dIsabella / Le Bazar du Homard. Arles: Actes Sud.
(2006b). Isabellas Room. Script (Work Version), 33 pp. Supplied by Needcompany
on 21.9.2006.
Lvesque, Solange (2005). An Ode to Life, Disguised as Kaddish. Le Devoir, Montreal
(3.6.2005), quoted on Needcompany.org (s. below).
Marber, Patrick (1999). Closer. New York: Dramatists Play Service, 2000.
Martinez, Matas & Michael Scheffel (1999). Einfhrung in die Erzhltheorie. Munich:
Beck.
Needcompany.org. <www.needcompany.org> (9.10.2006).
Niederhoff, Burkhard (2001). Fokalisation und Perspektive. Ein Pldoyer fr eine friedliche Koexistenz. Poetica 33:1, 121.
Nnning, Ansgar & Sommer, Roy (2002). Drama und Narratologie: Die Entwicklung
erzhltheoretischer Modelle und Kategorien fr die Dramenanalyse. V. Nnning &
A. Nnning (eds). Erzhltheorie transgenerisch, intermedial, interdisziplinr. Trier:
Wissenschaftlicher Verlag Trier, 10528.
Nnning, Vera & Nnning, Ansgar (eds) (2002). Erzhltheorie transgenerisch, intermedial, interdisziplinr. Trier: Wissenschaftlicher Verlag Trier.
Orton, Joe (1976). The Complete Plays. New York: Grove.
Pfister, Manfred (1977). Das Drama: Theorie und Analyse. Munich: Fink, 1997.
(1993) The Theory and Analysis of Drama. Tr. J. Halliday. Cambridge: Cambridge
UP.
Prince, Gerald (1988). A Dictionary of Narratology. Aldershot: Scolar Pr.
(2005). Point of View (Literary). D. Herman et al. (eds). Routledge Encyclopedia
of Narrative Theory. London: Routledge, 44243.
Richardson, Brian (1988). Point of View in Drama: Diegetic Monologue, Unreliable
Narrators, and the Authors Voice on Stage. Comparative Drama 22:3, 193214.
(2001). Voice and Narration in Postmodern Drama. New Literary History 32:3,
68194.
Routledge.
Ryan, Marie-Laure (2005). Narrative. D. Herman et al. (eds). Routledge Encyclopedia
Shakespeare, William (ed. 1993). A Midsummer Nights Dream. Ed. H. F. Brooks.
London: Routledge.
The Norton Shakespeare (1997). Eds. S. J. Greenblatt et al. New York: Norton.
242
Roland Weidle
Sommer, Roy (2005): Drama and Narrative. D. Herman et al. (eds). Routledge Encyclopedia of Narrative Theory. London: Routledge, 11924.
Spittler, Horst (1979). Darstellungsperspektiven im Drama. Frankfurt a. M.: Peter Lang.
Sprang, Felix (2007). Turns on the Narrative Turn: Showing and Telling in Needcompanys Early Shakespeare Productions and Isabellas Room. C. Stalpaert et al.
(eds). No Beauty For Me There Where Human Life Is Rare. On Jan Lauwers
Theatre Work With Needcompany. Gent: Academia Press and International
Theatre & Film Books, 13248.
Stoppard, Tom (1966). Rosencrantz and Guildenstern Are Dead. New York: Grove, 1967.
(1974). Travesties. New York: Grove, 1975.
Suchy, Patricia A. (1991). When Words Collide: The Stage Direction As Utterance.
Journal of Dramatic Theory and Criticism 6:1, 6982.
Surkamp, Carola (2005): Perspective. D. Herman et al. (eds). Routledge Encyclopedia
SABINE SCHLICKERS
(Bremen)
Focalization, Ocularization and Auricularization

in Film and Literature
In the following, I will limit my discussion of focalization to fictional
films although it cannot really be limited to these only. But fictional films
are more complex than factual ones or documentaries, because, like literary narrative texts, they are based on a double communication situation:
the implied director or author is not identical with the narrative entity of
the film, which I term camera1. Therefore, we have to differentiate between the implied director and viewer on the one hand and the camera
as intermediator of visual and acoustic information as well as the narratees of this information on the other.
In her study of point of view in three Spanish films adapted from literary texts, Susan Rubio Gribble (cf. 1992: 2526) argues that the camera
cannot be equated with the narrator of a literary text. What is at stake,
however, is not what we call this narrating agent in the film, but the fact
that, in film too, the act of narrative mediation is located on the
extradiegetic level. This is why we have to speak of the filmic narrator instead of the author as does Rubio Gribble, who is conscious of introducing a polyvalent and therefore problematic term: author means (a)
the actual author or director, (b) the implied author (whom she fails to
mention), (c) that agent which Michel Foucault terms author and which,
when we inspect it more closely, turns out to be a mixture of implied
author2 and narrator3 and, finally, the auteur of New Wave Cinema. In
1
In this article, I will use quotation marks (camera) in order to differentiate this narrative agent from the technical instrument. For more details on the following modeling
see my book Verfilmtes Erzhlen (Schlickers 1997).
Everyone knows that, in a novel narrated in the first person, neither the first-person
pronoun nor the present indicative refers exactly either to the writer or to the moment
in which he writes, but rather to an alter ego whose distance from the author varies,
often changing in the course of the work. It would be just as wrong to equate the
244
Sabine Schlickers
my argumentation, however, I will not focus on the entity of the author,

but rather on perspectivation. In film, as I will show, perspectivation is
mediated in the form of focalizationby the help of the camera as a
filmic narratorand its interplay with seeing (ocularization) and
hearing (auricularization) as well as through editing and montage. On
the level of the intended meaning, perspectivation has to be connected
with the agent of the implied author or the implied director.
As in literary narrative texts, the camera as narrator can be located
on the extradiegetic, intradiegetic, hypodiegetic, etc. levels. Nevertheless,
in the vast majority of cases, the camera is located on the extradiegetic
level, as illustrated in the figure below. We can locate it on the lower levels only when the camera is visible as a camera on these lower levels of
the shown world serving a diegetic function, for example, when a film is
being shot within the film.
(1) extratextual level: actual director and team
(2) intratextual level: implied director
(3) extradiegetic level: heterodiegetic camera + subjective camera" + voice-over
(4) intradiegetic level: shown world, characters
(5) hypodiegetic level: visualized narrations of the characters
(6) hypohypodiegetic level
(7)
(4) intradiegetic narratee

(3) extradiegetic narratee
Fehler!
(2) intratextual level: implied spectator
(1) extratextual level: actual spectator
Table 1: Model of Narrative Levels in Fiction Film
author with the real writer as to equate him with the fictitious speaker; the author
function is carried out and operates in the scission itself, in this division and this
distance. (Foucault 1984: 112)
The text always contains a certain number of signs referring to the author. These signs
[...] are personal pronouns, adverbs of time and place, and verb conjugation. (Foucault
1984: 112)
Focalization, Ocularization and Auricularization in Film and Literature
245
What is more complex is the case of the so-called subjective camera

or point-of-view shot. Here, the takes are recorded (virtually) from the
point of view of one characterthe subjective camera, as an agent of
enunciation, is located on the extradiegetic level. As a consequence, what
it is showing, i.e. the enounced or rather the view of the character, must
be located on the intradiegetic level4. This interplay of enunciation and
enounced can be compared to the way speech is rendered in narrative
literary texts: what the characters say is located on the intradiegetic level,
yet it is the extradiegetic narrator who quotes, selects, and condenses
these items of speech.
As a narrative agent, the camera can track sound as well as images,
which means it is not limited to showing only. What is more, film as a
plurimedial semiotic system is capable of simultaneously transporting
visual and acoustic information about the fictional world. Decoding this
fundamental double perspectivation is not an easy task, particularly because the perspective and the flow of information keep changing incessantly and hardly ever remain constant over a longer period of time.
Moreover, pieces of visual and acoustic information can be concordant or
discordant and likewise, discrepancies regarding information can be
created among the various characters. And there is yet another factor:
only the selection and composition of the individual shots and takes or
rather the mise en scne create the meaning of the filmic narration5.
Before I go on to illustrate these possibilities with examples of films, I
would like to present a typology of filmic perspectivation which is based
mainly on Franois Josts critique of Grard Genettes typology devel4
At this point, I wish to correct the thesis I put forward in 1997 in Verfilmtes Erzhlen
(Filmed narration) according to which the subjective camera is homo- and intradiegetic, because this would mean that there are two cameras in every fictional film,
an extra- and an intradiegetic one, which keep taking turns.
In earlier publications (cf. Schlickers 1997), I located the artistic-compositional organizing principle of the montage on the level of the implied director, who, analogous
to the implied author, governs the entire dimension of image and sound but possesses
no semiotic signs of his or her own to articulate him- or herself and therefore has to
utilize the camera as a narrative agent. Contrary to this earlier description, I would
like to agree with Markus Kuhns conception (see his contribution in the current volume) and locate the montage also on the extradiegetic level. After all, filmic narration
only comes into existence through the combination of the different takes and shots, and
filmic perspectivation is the result of this composition, in which the sequence of the
individual takes determines the type of focalization.
246
Sabine Schlickers
oped for literary perspectivations. I developed this typology of literary

and filmic perspectivation in Verfilmtes Erzhlen (Filmed narration,
1997). In this contribution, I will simplify and abridge it as well as, regarding certain points (see above), correct and supplement it.
Genettes distinction of focalization and voice was intended to contribute to the analysis of the fundamental option of double perspectivation
in terms of knowing versus showing/telling which we always have in
narrative literary texts too.
However, to my mind most of the theoretical works on this subject [perspective] (which
are mainly classifications) suffer from a regrettable confusion between what I call here
mood and voice, a confusion between the question who is the character whose point of
view orients the narrative perspective? and the very different question who is the
narrator?or, more simply, the question who sees? and the question who speaks?
(Genette 1980: 186)
Reducing the question who is the character whose point of view orients
the narrative perspective? to the formulation of who sees? limits
focalization to the visual aspect, which is what Genette aims to avoid. Reducing focalization to seeing precludes analyzing those passages in
which the narrator is not viewing through the eyes of a character but nevertheless is able to penetrate into his or her interior, for instance: As she
was crossing the bridge, she was thinking of her ailing mother. At a later
point in his works, Genette expanded the category of focalization in order
to include the knowledge of the narratorand thus merged what he had
originally set out to distinguish the regrettable confusion between []
mood and voice (Genette 1988: 64). I therefore aim to distinguish once
more between seeing and knowledge. After all, especially when we
consider film, it becomes clear that this is a medium which employs
double perspectivations even more so than does literaturewhich, moreover, are usually being mediated simultaneously.
Franois Jost defines focalization as the knowledge of the narrator in
relation to the characters (cf. Jost 1989: 71). Unlike Genettes characterization of focalization, in his definition, seeing and hearing are replaced by the categories of ocularization and auricularization respecttively. Admittedly, these are awkward categories, but after years of consideration, I have not been able to come up with a better idea, so that I
247
will carry on using these categories6. Jost adopts the forms of focalization
which Genette introduced drawing on Jean Pouillon and Tzvetan Todorov.
In spite of the fact that the term zero focalization has been rightfully
criticized in narratological discussion for implying that there is an absence of
perspectivation, I would like to continue employing this term, because it is
well-established internationally. Therefore I will continue to distinguish:
zero focalization: narrator > character (n > c)
internal focalization: narrator = character (n = c)
external focalization: narrator < character (n < c)
Table 2
Even without a narrator, various forms of focalization can be used. This is

why it seems necessary to expand this well-known typology in order to include focalization on the level of the characters (i.e. the intradiegetic and
lower levels). Thus we render more dynamic the interplay of knowledge,
lack of knowledge, and the conscious withholding of information of the
various characters from each other as well as in their relationship to the extradiegetic narratee and the implied reader/viewer. On the intradiegetic level,
the narrator should be replaced by cn for characternarrator, the former
character becomes character2 (c2).The formulas are thus:
cn < c2
cn = c2
cn > c2
Table 3
The second expansion is implied in table 3 because it refers to an entity

which Jost and other narratologists and scholars of film leave out: the extradiegetic narratee (nee, table 4) or the intradiegetic narratee (cee, table 5), who
also plays a significant role in the reconstruction of focalization:
6
Lintvelt (1981) coined the term plan psychique, which could be equated with the
term of focalization employed here, his plan perceptif to ocularization and auricularization.
248
Sabine Schlickers
n + nee > c
n + nee = c
n + nee < c
Table 4: Focalization of Narrator and Narratee on the Extradiegetic Level
cn + cee > c2
cn + cee = c2
cn + cee< c2
Table 5: Focalization of characternarrator and intradiegetic narratee
Alfred Hitchcocks Notorious (1946) serves as a particularly intricate example: character c1, played by Ingrid Bergman, is a secret agent. Her mother
in law (c2) and her husband (c3) have found out about this. C2 is planning to
kill slowly her sons wife and therefore puts every day some poison in the
young womans coffee. The extradiegetic narratee (nee) is aware of the fact
that c2 wants to see to that problem, yet knows nothing about her concrete
plans. The extradiegetic narratee then experiences how c1 is getting a little
sicker every day (c1 + nee < c2 + c3: external focalization). The very moment,
however, that c1 realizes that she is being poisoned, her cup of coffee is
frozen in an extreme close-up that lasts several seconds, to the effect that, at
this short and slowed down moment, the narratee goes through the same
terror that c1 is experiencing. We can thus detect, from the side of the
camera and its extradiegetic narratee, internal focalization regarding c1 (n +
nee = c1). At the same time, however, we can detect zero focalization
regarding the other characters of the world shown (c1 + n + nee > c2 + c3),
because those that have been administering the poison have no idea that they
have been found out.
Jost postulates three forms of ocularization, which, for practical reasons, I
have chosen to reduce to two:
zero ocularization = nobodys shots, hetero- and extradiegetic camera
internal ocularization = subjective camera and mindscreen
Table 6
249
The most common case is that of zero ocularization, or nobodys shots,

as Nick Browne (1976: 5) terms them: the camera is showing the character from the outside and does not count as a diegetic agent within the
fictional world. It is thus marked as hetero- and extradiegetic and does not
usually mark its enunciation. Unlike Jost (1989) I also count the marking
of enunciation by the help of peculiar camera movements or changes in
color as zero ocularization if this makes the camera self-referential or
rather foregrounds its hic et nunc. In this case, the narrative mode of the
camera can be regarded as authorial. The use of an unsteady hand-held
camera may be involuntary as a result of the lack of better equipment, for
instance in the production of cheap documentaries. It may also be a result
of aesthetic and intentional principles, as is the case in the Dogma
Films, whose directors have subscribed to a vow of chastity7.
Unlike zero ocularization, internal ocularization refers to subjective images which can be ascribed to a diegetic agent. Here we can often find the
use of the subjective camera mentioned above, in which recording takes
place almost completely from the point of view of one character. However,
subjective images can also be created by the use of montage, lap dissolves,
fade-over, color fade, slow-motion, distortion of sound, music and other
devices. Although ocularization can always be defined in purely technical
terms, when looking at the specific semiotic-narrative context we can only
classify it in its interplay with focalization. Thus a segment of what a character remembers is not tied to internal ocularization. Instead, conventional
flashback-structure first shows the character in zero ocularization and only
then slides into this characters visualized flashback which itself is mediated
in zero ocularization
This conventional structure is dissolved in Francesco Rosis adaptation of
Gabriel Garca Marquezs short novel Chronicle of a Death Foretold when
at first somebodys dream is shown and we see how the dreamer wakes up
only after that. In film, dreams and memories are thus visualized from the
characters interior perspective while the characters see themselves as acting
characters from the outside. We cannot describe ocularization as internal in
this context (since then the character of the dream would not be visible).
Nevertheless, this flashback-structure has become conventionalized as a
subjective form of remembering or dreaming and must therefore be
described as internal focalization.
7
See < http://www.martweiss.com/film/dogma95-thevow.shtml>.
250
Sabine Schlickers
Analogous to ocularization, Jost distinguished three forms of auditory

point of view, that is, auricularization, which I have also chosen to reduce to two:
zero auricularization: the sound cannot be ascribed to a shown agent
internal auricularization: auditory subjectivity
Table 7
Zero auricularization refers to those cases in which the sound cannot be

located and is indeterminate and extradiegetic. This is quite obvious when
we have off-screen sound and voices which are mediated by the help of the
camera, but in most cases zero auricularization is a sound whose source is
on-screen. Conversely, internal auricularization constitutes auditory subjectivitythe sound track mediates what the character is hearing in the
shown world. It must therefore be located on the intradiegetic levels or those
below, just like the enounced of the subjective camera. An unambiguous
ascription can be undertaken when a character is eavesdropping at a door: in
this case, the sound can be ascribed to this characters auditory perception.
Alejandro Amenbars film Tesis (1996) manages to mediate internal
auricularization in a more unique way. Both protagonists are sitting in the
university cafeteria, a few meters from each other, and are listening to
their walkmans. She is listening to classical music, he is listening to loud
rock. Whenever she is looking at him, we can hear classical music
(internal ocularization + auricularization), conversely, we can hear
strident rock when he is looking at her (internal ocularization + auricularization). At one point, the camera shows him jumping up angrily
and running over to her (internal ocularization zero ocularization; his
internal auricularization). Only as he comes to a stop at her table, does he
take out his headphones, she takes out hers simultaneouslyand the
viewer sighs with relief because the noise has finally stopped.
Cases of change of both forms of auricularization are very rare. In
El Sur, a fictional film of Spanish filmmaker Vctor rice, which I will
discuss in more depth below, we can nevertheless find an example:
Estrella, a young girl, is hiding under the bed and hears intimidating, repeated pounding from the room above (internal auricularization). The next
shot shows where the sound originates from (zero auricularization): her
father is sitting in an armchair in his chamber in the attic, gazes into
space, and, with a stick, beats the wooden floor at regular intervals. The
251
noise grows louder as it is carried on and resonates through the whole

house. The next cut leads us back to Estrella, who is lying under the bed
crying. Her mother enters the room, stops, and hears the intimidating
pounding (internal auricularization).
Establishing those sounds of which we do not know for sure to which
part of the shown world they belong is more difficult. In this case, their
point of origin can be either mental (as in an acoustic hallucination) or be
located outside the diegesis. Thus the visual sphere gains significance in
determining auricularization, particularly because our sense of hearing is
less evolved than our eye sight. On the other hand, the sound quality of a
human voice is very distinctive and unambiguous, and can therefore be
more easily ascribed to a character in a film than a particular way of looking could be.
Making use of a voice coming from off-screen is a popular device of
auricularization. Often, a homo- or autodiegetic narrator-character is introduced at the beginning of the frame story. This narrator then leads the
viewer through the film, just like in Francesco Rosis Chronicle of a
Death Foretold. Sometimes it only returns towards the end of the film,
just like the (fictitious) writer Marguerite Duras in the filmic adaptation of
her novel LAmant (which has the same title as the novel). I would now
like to put forward the thesis that it is irrelevant for determining auricularization whether the extradiegetic voice from off-screen is part of the
shown world8: although speakers of the voice from off-screen are often
homo- or autodiegetic, i.e. they are usually characters of the world shown,
they might just as well be heterodiegetic. In this case, the speaker of the
off-screen voice takes on the function of the classical heterodiegetic
literary narratorfor instance in Y tu mam tambin, where the
elaborated speech of the off-screen narrator forms a sharp contrast with
the colloquial Mexican Spanish of the on-screen characters.
Luchino Viscontis Ltranger (1967), based on Albert Camuss 1942
novel of the same name, demonstrates the option of combining zero auricularization with zero ocularization, which has nevertheless to be described as internal focalization: after Meursault killed an Arab and was
arrested, he is now sitting in his prison cell brooding. While the camera
closes up on Meursaults silent face (zero ocularization), Meursaults
autodiegetic voice is rendered in zero auricularization. In the original
8
This thesis is contrary to my earlier descriptions in Verfilmtes Erzhlen (Filmed Narration, 1997).
252
Sabine Schlickers
Italian version, it is actually actor Marcello Mastroiannis voice coming

from off-screen. Although the acoustic and visual perspectivation chosen
is not internal, this interior monologue of Meursault must be established
as internal focalization. The main difference from the autodiegetic narrative mode of the novel consists in the use of such objective forms of
auricularization and ocularization which paradoxically effect subjective
perspectivation. Viscontis film does not only use auricularization and
ocularization in interior monologues in order to effect subjectivization. In
fact this strange way of rendering the outside world through the protagonists internal perspective thus objectivized is characteristic of the
entire filmwhich, therefore, adapts the similarly strange focalization of
the novel adequately9.
In the context of this article, I cannot elaborate further on the characteristics of the three types of focalization and their interplay with the two
types of ocularization and auricularization which we find in filmic and
literary texts (cf. Schlickers 1997: 15367). Instead, I would like to examine more closely another film adaptation in order to exemplify further
options of perspectivation in film and to investigate more profoundly the
above mentioned determination of focalization in terms of the extra-, intra-,
and hypodiegetic narratees and the implied viewer.
Victor Erices film El Sur (1983) is based on Adelaida Garca Morales
short novel of the same name (written in 1981, published in 1985). Fifteen-year-old Estrella is the autodiegetic narrator-character of the film.
Following the opening scene, the film is presented as an extended flashback which sets in as Estrella is eight years old and which continues till
the fictional present time of the narrative situation, thus covering seven
years of narrated time. In the novel, the protagonistwhose name there is
Adrianaaddresses her second person narrative to her father, trying to
shed light on their (platonic) incestuous relationship10. Nonetheless, what
9
10
Genette reconstructs the perspectivation in Camuss novel at first as internal focalization with an almost total paralipsis of thoughts (Genette 1988: 124). Yet then he moves
away from this formula because it implies that Meursault was actually thinking
something we cannot see in the text. Genette thus comes to the conclusion that in
LEtranger we have a homodiegetic narrating that is neutral, or in external focalization
(124; cf. Schlickers 1997: 16566).
See Rubio Gribble (1992) and Nimmo (1995) as well as the scene of initiation with the
pendulum in the novel which reads like a sexual act: When I had it in my hands, [...]
its stillness discouraged me. I was scared that it would never move in my hands. Now
you were telling me in a whisper [...]. When you turned off the lights, without ceasing
the soft murmur which was more and more occupying my mind, I felt my heart beat
253
is covered up by the constant presence of the intradiegetic narratee is the

factual absence of the father: we learn soon that he committed suicide and
that Adriana is directing herself at a dead person. Neither of the media
tells us how much time has passed since the fathers suicide and the
protagonists return to the house of her father.
In the film, however, the role of the narratee Estrella addresses as narrator remains unspecific. While working through her memories, Estrella/Adriana is continuously growing aware of the fact that her father
(Rafael in the novel, Agustn in the film) was not only the magician who
used to swing his pendulum, who was from the other Spain, that of the
south, and who withdrew to the rough north because of his longing for
solitude and for political reasons. Rather, he was a lonesome person, a
coward, who betrayed the love of his life for reasons of bourgeois conventions, and slowly perished because of it. After reading this gloomy
epistolary narrative for the first time, one is led to think that the challenge
of a film adaptation would mainly be the lack of dialog and the brevity of
the text. In fact, the last 12 pages of this 47 page long novel, that is one
quarter of the text, which cover Adrianas journey to the south, constantly
evoked in its mythical quality, do not even form part of the filmthey
were left out in order to keep costs low. The film was very successful at
the 1983 Cannes Film Festival, although the originally planned sequence
was never made and the film thus remains uncompleted.
The perspective in the novel is limited to the perception and knowledge of auto(intra)diegetic narrator Adriana. In the film, too, the protagonist is the character that is directing the narratees attention: as the one
who is remembering (in the extradiegetic voice-over) and the one who is
being remembered and who is acting in the world shown (intradiegetic).
In terms of its aesthetics, the film follows a structural principle which is
not very common today: each intradiegetic memory sequence is framed
by a black screen: the world shown is being opened by the help of a slow
fade-in to one of the images that constitute it; the sequence ends with a
slow fade-out to black-screen. We must, however, be wary of the world
shown. After all, some of the images that we see derive from Estrellas
imagination and bear no relation to the fictional reality. For instance, right
at the beginning of the film, the camera shows 15-year-old Estrella,
who is holding her dead fathers pendulum. She closes her eyes, a tear
with violence, my breath going faster and that I was beginning to tremble [...]. (Garcia
Morales 1985: 1112; tr. S. Sch.)
254
Sabine Schlickers
runs slowly down her face (zero ocularization). This is followed by black
screen, a slow fade-in shows the father, sitting beside his pregnant wife
and using his pendulum in order to determine the babys sex. On-screen
he tenderly speaks to his wife: Her name will be Estrella.11 A slow
fade-out leads back to a black screen that finishes this memory sequence.
Meanwhile, the narratee can hear Estrellas voice from off-screen (zero
auricularization): They told me that my father had predicted that I would
be a girl. Thats the first memory that comes to my minda very intensive image which, in fact, I invented. The fictitious character of this pendulum-scene is thus established by the autodiegetic narrators voice-over
onlya scene that works beautifully in order to illustrate the power of the
word or to oppose the theory of the primacy of the image. At the same
time, this scene demonstrates that Estrella as narrator is capable to control
the images of the camera and, therefore, that she is the one carrying out
the central narrative function in this film.
Another scene, however, illustrates the narrative power of images and
contains a congenial time lapse: the narratee sees eight year-old Estrella
as she leaves her parents house by bicycle and rides down the long avenue, her little dog, almost a puppy, trotting behind her. The camera remains in the same position and shows the empty avenue. An unnoticeable
lap dissolve takes us to the next scene: the camera, still in the same
position, shows Estrella as she is returning on her bicycle. She is still accompanied by her dog, now fully grown, and has herself become a fifteen
year-old teenage girl.
In a key scene of the film, eight year-old Estrella discovers that the
mysterious Irene Ros, her fathers former lover, does in fact exist. The
narratee shares the experience of this process of recognition purely
through the images shown: one late afternoon, Estrella sees her fathers
moped parked in front of the Arcadia cinema. She looks at the poster of
the movie which they are showing (zero ocularization). Over the following two minutes, the camera keeps changing, through shot-reverseshots, between Estrellas looking at the poster (internal ocularization) and
backward-moves to show her face (zero ocularization), continuously enlarging the position of the internal ocularization. At first, we can see the
movie poster in its full size, as though looking at it with Estrellas eyes.
The movie is titled Flor en la Sombra (Flower in the Shade), and the
poster shows the drawing of a woman who, from a very short distance, is
11
Here, as in the following, English translations of Spanish texts are mine.
255
looking at a man yearningly. Another woman can be made out in the

background; she is essentially overseeing the entire scene. In the right
hand lower corner, the names of the main actors are shown, the first one
on the list being Irene Ros. The camera rolls back to Estrellas incredulous face; closes up more on the letters of the actresss name, and
then rolls back again to Estrella. She enters the lobby of the movie theater
where she can see more movie posters. She eyes these black-and-white
publicity stills critically, the camera pans to follow her gaze from the
right to the left till it gets stuck on a picture of Irene Ros. In two reverseshots, the camera zooms in till, in an extreme close-up, it fills the entire
screen. Estrella walks over to the box office and asks the cashier there
whether the blond actress is Irene Ros but the woman cannot help her.
Estrella leaves the movie theater and observes its entrance from the other
side of the street. Another internal ocularization transports her gaze over
to the movie theater; a travel shot slowly glides up the faade and comes
to a rest on a brightly illuminated round glass arch. A humming noise can
be heard from off-screen, without any mediation, a cut leads to a scene of
the hypodiegetic film within the film: Irene Ros is humming a song and
looks straight into the camera while she is brushing her hair. A cut shows
Estrellas father from his side in a close-up, as his gaze is glued to the
screenhe is thus the intradiegetic narratee or the viewer of the film
within the film. Another cut jumps back to Irene Ros, whose direct look
into the camera seems to be addressed directly to Estrellas father.
Since Estrella is waiting outside while this happens, these and the following passages, which uncover the film within the film, cannot be ascribed to her as a narrator. If at all, then they must be ascribed to her imagination as Martn-Mrquez suggests (cf. 1994: 134)an explanation
that is not very convincing when we consider her young age. Rubio
Gribble (1992: 185), on the other hand, recognizes that the narration has
shifted to another narrative level (another focalization). Instead of reconstructing, thus, the depicted scene as Estrellas internal focalization,
we have to attribute the narrative function in this sequence to the extradiegetic camera, which consequently leads us to replace the internal focalization hitherto dominant by zero focalization (n +nee > c).
When we refer to the implied viewer, we can attest internal focalization also in this scene, because he or she decodes the character of miseen-abyme of this melodramatic sequence of a film within a film with
which Estrellas father identifies: when, in the melodramatic film noir,
Irene Ros is shot dead by her jealous lover, the camera pans to show
256
Sabine Schlickers
Agustns very sad face from the front. The extradiegetic narratees
knowledge regarding Estrella, however, corresponds to zero focalization,
because, unlike her (whose focalization with respect to her father is external), the extradiegetic narratee sees the passages of the film within the
film and, just a moment later, hears her fathers voice coming from offscreen, as he sits in a caf, writing a love letter to the actress after he has
left the theater. Agustns voice-over can be heard by the extradiegetic
narratee only, yet it cant be perceived by intradiegetic Estrella, who is
curiously peeking at the writing paper through the window as her father is
coming out of the caf towards her. Significantly, the extradiegetic narratees knowledge is more substantial even than that of Estrellas extradiegetic voice when she formulates ideas about the content of what seems
to be a letter.
In the scene following the next, Agustn sits in another caf and reads
the letter which his former lover sent to him as a reply. This time, the extradiegetic narratee hears the actresss voice from off-screen. Agustn is
shown as he is reading the letter (zero ocularization) and is imagining the
voice of the woman who wrote it. In spite of this internal process, the
auricularization in this scene, analogous to the example given above, that
is Viscontis Ltranger, must be described as zero. With regard to
Agustn, the narratees and implied viewers focalization is therefore internal; with regard to Estrella, however, focalization must be considered
external also in this scene, because Estrella does not know the letter of the
former lover. It is therefore the camera that takes on the narrators
function in this sequence and eliminates Estrella as a narrator: the narratee
learns at the same time as Agustn why Estrella never heard of Irene Ros
againshe gave up acting when Agustn left her. Now she cant
understand why Agustn is contacting her again, after so many years, she
wants to let bygones be bygones and does not want to hear from him
again. The scene foreshadows Agustns own end as in reaction to the
actresss refusal to take up again the love story of the past, he will commit
suicide.
In the novel, the second person narrative, directed at an extradiegetic
narratee, which Adriana uses to address the intradiegetic You of her late
father, works to exclude him or rather to push him into the role of a voyeur (cf. Rubio Gribble 1992: 17475). In the film, however, these newly
added scenes do not only bestow on him a certain presence but also give
him superior knowledge. At the same time, these scenes illustrate the perfect interplay of image (camera/ocularization) and sound (Estrellas
257
voice-over/auricularization), yet also the competition between the two

modes of expression. In each sequence, we can detect anew the dynamic
and often changing focalization in terms of the extra- and intradiegetic
agents. Nevertheless, this does not only hold true for El Sur but for every
intellectually and artistically demanding fictional film and literary text.
Translated from German by Katharina Kracht.
Films cited
Amenbar, Alejandro (1996). Tesis. Video.
Annaud, Jean-Jacques (1992). L'Amant. Video.
Cuarn, Alfonso (2002). Y tu mam tambin. Video
rice, Victor (1983). El Sur. DVD.
Hitchcock, Alfred (1946). Notorious. Video.
Rosi, Francesco (1985). Chronicle of a Death Foretold. Video.
Visconti, Luchino (1967). Ltranger. Video.
References
Browne, Nick (1976). The Rhetoric of Filmic Narration. Ann Arbor: UMI Research Press.
Camus, Albert (1942). Ltranger. Paris: Gallimard, 1957.
Foucault, Michel (1984). What Is an Author? P. Rabinow (ed). The Foucault Reader.
New York: Pantheon Books, 10120.
Garca Morales, Adelaida (1985). El Sur. Barcelona: Anagrama.
Jost, Franois (1983). Narration(s): en de et au-del. Communications 38, 192212.
(1984). Le regard romanesque. Ocularisation et focalisation. Hors Cadre 2,
6784.
(1987). L'il - Camra. Entre film et roman. Lyon: Presses Universitaires, 1989.
Lintvelt, Jaap (1981). Essai de typologie narrative. Le point de vue. Paris: Corti.
Martn-Mrquez, Susan L. (1994). Desire and Narrative Agency in El Sur. G. CabelloCastellet et al. (eds). Cine-Lit II. Essays on hispanic film and fiction. Oregon:
Oregon State U, 13036.
Nimmo, Clare (1995). Garca Moraless and Erices El Sur: Viewpoint and Closure.
Romance studies 26/26, 4149.
Rubio Gribble, Susana (1992). Del texto literario al texto flmico: Representacin del
punto de vista en tres adaptaciones del cine espaol de los ochenta. Dissertation,
New York State University.
258
Sabine Schlickers
Schlickers, Sabine (1997). Verfilmtes Erzhlen: Narratologisch-komparative Untersuchung

zu El beso de la mujer araa (Manuel Puig/Hctor Babenco) und Crnica de una
muerte anunciada (Gabriel Garca Mrquez/Francesco Rosi). Frankfurt a. M.:
Vervuert.
MARKUS KUHN
(Hamburg)

Narrative Mediation in Self-Reflexive Fiction Films
1 Introduction: Narrative and Narrative Mediation
With the exception of focalization, there is hardly a field in the area of
film narratology that has been discussed as controversially as the question
of narrative mediation. Does employing the construct of a film narrator
make sense? David Bordwell and his followers reject a narrator and
emphasize the part of reception 1 . Other scholars defend the concept of a
narrative agent and locate it within a multi-layered communication
model. Thus terms such as grand imagier (cf. Metz 1971), imagemaker (cf. Kozloff 1988), fundamental narrator (cf. Gaudreault 1988)
or cinematic narrator (cf. Chatman 1990) were coined, without, however, gaining general acceptance 2 .
Naming and locating a presumed narrative agent in the film communication model is one of the areas which should be investigated further.
Another area is that which we encounter in the context of an intermedial
narratology extending beyond narrative literature and film to include further visual, auditory, and audiovisual mediasuch as drama, cartoons,
video games, comic strips, animations, radio plays, paintings, music etc.
How can we define narrativity in such a way that the definition does not
only apply to verbal narratives? Narrow definitions of narrativity, like
those in classical German narrative theory, postulate the existence of an
instance of narrative mediation; broad definitions refer to the level of
the represented and define narrativity in terms of events or changes of
state 3 .
1
2
See Bordwell (1985: 62); Branigan (1992: 10810); Fleishman (1992: 13), et al.
Griem & Voigts-Virchow (2002: 16163) list different positions, although with too
great an emphasis on the aspect of anthropomorphization of the film narrator.
See Chatman (1990); Jahn (1995); Schmid (2005), et al. The discussion revolving
around a universal term of narrativity is extended by combinatory, functional, and
260
Markus Kuhn
It is quite obvious that a broad definition serves as a better foundation

for the application of narratology to all kinds of media. However, we have
to ask: how do such a definition of narrativity and the models derived
from it benefit the narrative theory of the media in question and facilitate
the analysis of works of art in such media? Doesnt too broad a definition
run the risk of overlooking the basic narrative patterns of the medium in
question? From the point of view of film study, for example, it is
surprising that broad intermedial narratological approaches often ignore
the film apparatus that mediates the events, thus regarding film as a
mimetic narrative medium or as a medium without a mediating agent 4 .
If we decide to apply narratological concepts to film analysis, we need
to develop an awareness of the strategies of narrative mediation in film to
explain the basic structures of a certain work, genre or epoch. Narrative
mediation, if not bound to the concept of an anthropomorphic narrator, is
a tertium comparationis between narrative fiction and fiction film. Comparing the devices of narrative mediation in literature and film can offer
new impulses for the study of literary screen adaptations and prevent that
studies of crossmedial influenceswhich are encompassed in such terms
as filmische Schreibweise (filmic style of writing)remain stuck in vague
aesthetic categories 5 .
gradual definitions (cf. Fludernik [1996]; Wolf [2002]; Herman [2002]; Ryan [2005],
et al.). See Prince (2003: 12): As we know, nothing like a consensus has been reached on that subject. Some theorists and researchers believe that everything is a narrative;
others maintain that everything can be; and still others contend that, in a sense, nothing
is (because narrativity is culture-dependent and context-bound). Some define narrative
as a verbal recounting of one or more events and others as any kind of event representation (including non-verbal ones) [].
In his otherwise very convincing essay on intermedial theory of narration, Werner
Wolf (2002) omits the aspect of narrative mediation and credits drama with a higher
narrative potential than film. Although Seymour Chatman subsumes film under the
mimetic narratives (Chatman 1990: 115), he discusses the problem of narrative
mediation in film very thoroughly when he describes his concept of the cinematic
narrator (Chatman 1990: 124-38)unlike many scholars who quote the schemes he
has developed (Jahn [1995]; Bach [1999], et al.).
This does not give preferentiality to a narrow definition of narrativity within the framework of intermedial narratology. However, in the area of film narratology, which is not
as sophisticated as language-based narratologies, a systematic description of the various forms of narrative mediation has yet to be developed.
261
2 The Model of Narrative Levels in Fiction Film
Figure 1: Model of Narrative Levels in Fiction Film
The model I propose is an adaptation of the communication model of narrative theory in literary studies, such as can be found in Fieguth (cf. 1973:
186) or Schmid (cf. 2005: 4748), to the film medium. It is largely analogous to that proposed by Sabine Schlickers and in many aspects resembles Seymour Chatmans notion of levels 6 , but differs from them insofar
as I divide the category of the filmic narrative agent (i.e. Schlickerss
heterodiegetic camera or Chatmans cinematic narrator) into an audiovisual narrative instance, which I term visual narrative instance/
visual NI, and one or more facultative verbal narrative instance(s)/
6
See Chatman (1990); Schlickers (1997) and Schlickers article in this publication,
pages 24358.
262
Markus Kuhn
verbal NI(s) 7 . Thus the implied director, who does not have any
semiotic sign systems at his disposal, employs, on the extradiegetic level, the visual narrative instance as well as one or more verbal narrative
instance(s) (or none), in order to achieve filmic narration. Highly complex cinematographic narrative situations can be created through the interplay between the visual narrative instance and the facultative verbal
narrative instance(s) (voice-overs, inserted texts or intertitles), i.e. between showing and telling.
Not only the moving picture within one shot 8 (i.e. the process of selection, perspective, and accentuation by the camera, or cinematography),
but also the combination of shots into sequences (i.e. the process of editing, or montage in terms of classical film theory) should be attributed to
the visual narrative instance. That which is generally known as filmic or
cinematographic narration comes into existence through editing. Focalization can only be determined through the interplay of the edited shots.
When cinematic narration is realized through showing, there is no categorical separation between what the camera shows within a shot, and
what the editing reveals through the combination of various shots. Often
the difference from one shot to another is the only indication of a change
of state, a necessary condition of narrativity. However, we must also take
into account aspects of the mise en scne as part of the visual narrative
instance. After all, shot composition, lighting, and set design can contribute significantly to visual narration.
The implied director can be found on this level, on which the aspects
of all narrative instances in film come together. The implied director
serves as an explanation for the complex interplay of visual and verbal
narrative instances and for the analysis of certain forms of unreliability 9 .
7
The point of view shot does not represent a transition to the intradiegetic level because the camera does not become an element of the diegetic world. In the instance of
the point of view shot the extradiegetic visual narrative instance approximately shows
what a character is seeing (internal ocularization).
A shot can be defined as the time in which the camera runs without interruption or as
a continuous strip of motion picture film.
As is the case with all instances that can be derived from the structure of the work in
question, the visual narrative instance, the verbal narrative instances, and the implied author/director are instances assumed theoretically and not existing entities (and
much less anthropomorphous figures). If we use analysis only in order to prove these
assumed instances or if we abuse these instances as advocates of a certain interpretation, we run the risk of ending up in a tautological short-circuit. However, particularly in the area of film analysis, narratological categories can ward off the temptation
263
3 Focalization
A basic idea of my proposed concept of focalization is understanding it,
as Sabine Schlickers and Franois Jost do, in terms of knowledge, i.e. the
relation of knowledge between the narrative instance and the character,
and separating it from questions regarding perception in the narrow sense.
In the context of the visual aspects of perception (seeing) I will use the
term ocularization, and the term auricularization for the auditory
aspects (hearing) 10 .
This classification of focalization categorizes the relation of knowledge between the narrative instance and the character(s) into (a) zero,
(b) internal, and (c) external focalization, that is when the visual/verbal
narrative instance shows/tells the narratee (a) more than, (b) as much as,
and (c) less than the character(s) know. It has proven to be a valuable tool
in the comparative analysis of literature and film. I will thus maintain the
term zero focalization, despite its sometimes inopportune implications,
and use it only to refer to the relation of knowledge between narrative instance and character and not to the limited or unlimited knowledge of the
narrative instance per se, which would require a complex model of perspective. Focalization thus defined can usually only be classified in the
succession of shots. However, it is impossible to identify distinctly each
sequence in its focalization. It is therefore necessary to highlight clearly
ambivalences and uncertainties in focalization.
4 The Visual Narrative Instance in its Interplay with Facultative

Verbal Narrative Instances
The proposed method of distinguishing between the narrative instances
shows its value in the analysis of films such as R. W. Fassbinders
epilogue of Berlin Alexanderplatz (West Germany 1980). Next to the
10
of an approach too grounded in the aesthetics of the process. Far too often, we encounter film analyses which focus heavily on the production. The question whether we
can omit the history of the term implied author/director and its contended theoretical
implications arises quite naturally (cf. Booth 1961; Chatman 1978 and 1990; Nnning
1993; Kindt & Mller 1999 and 2006; Schmid 2005). In the context of this article, I
will use this term only referring to provable intratextual aspects.
See Jost (1987); Schlickers (1997) and Schlickerss contribution in this collection in
which she discusses the relationship of focalization, ocularization, and auricularization
in detail.
264
Markus Kuhn
visual narrative instance, various verbal narrative instances are employed

on the extradiegetic level (in the form of various voice-overs, intertitles,
and text captions). Every extradiegetic verbal narrative instance can be
both heterodiegetic or homodiegetic in its relation to the diegetic world.
They can each focalize differently and be in opposition to the visual
narrative instance, which in turn can also focalize independently 11 .
Whether a visual narrative instance can actually also be homodiegetic (i.e.
whether some form of first person narrative situation [sensu Stanzel]
can be accomplished on the visual level) must be discussed with the help
of borderline cases (part 7).
Contrary to what has been widely declared, there is no primary relationship of dominance between visual and verbal narrative instances in
film, in effect there is no primacy of the image (i.e. the narrative instances
cannot be located on two levels that stand in a hierarchic relationship).
The verbal narrative instance is not automatically located in a position
above the visual narrative instance or vice versa. The reliable extradiegetic visual narrative instance can uncover the unreliable extradiegetic verbal narrative instances (All about Eve, Joseph Mankiewicz,
USA 1950). However, the visual narrative instance can also be unreliable
(Stage Fright, Alfred Hitchcock, USA 1950; Fight Club, David Fincher,
USA 1999), or its reliability can be called into question with the help of
verbal narrative instances (Rashmon, Akira Kurosawa, Japan 1950). An
extradiegetic verbal narrative instance can dominate the visual narrative
instance and reduce it to an illustrating function (opening of Magnolia,
Paul T. Anderson, USA 1999). However, it can also serve to structure that
which the visual narrative instance shows, order it in time and space or
summarize past history (expository voice-overs, intertitles indicating the
actions setting in silent movies). This relation can be as alternating and
ironical as in Franois Truffauts Jules et Jim (France 1962) or as
ambivalent as in Alain Resnaiss Lanne dernire Marienbad
(France/Italy 1961). In silent movies this interplay is also encapsulated in
a complex way because of different methods of speech representation,
such as reports by a narrator or quoted direct speech in intertitles.
11
Just as Jost (1987) and others proposed, I assume that narrative instances are able to
focalize and therefore do without the instance of a focalizer.
265
5 The Relationship of Narrative Instances

In order to illustrate the interplay of verbal narration and visual images in
film, Sarah Kozloff (1988: 103) suggests a continuous graph comprising three areas: disparate, complementary, and overlapping.
Kozloff does not attempt to introduce either binary or clearly delimited
categories. Instead, she speaks of the degree of correspondence between
narration and imageswhich is sensible because no clearly defined dividing lines can be drawn. To use Kozloffs scheme as a framework to
describe the dynamic relationship between visual narrative instance and
verbal narrative instance, we have to complement her three categories. If
both instances are in a disparate relationship, we have to detect whether
both instances stand in evident contradiction to each other, or just tell disparately about different facts. In a complementary relationship, we are
confronted with the question as to whether both instances mesh to tell the
main story or whether they illuminate different facts or storylines which
each complement the other. In an overlapping relationship, each instance
can equally paraphrase the other, or the visual narrative instance can
merely illustrate that which the dominating verbal narrative instance reports and vice versa. Between the overlapping and the complementary relationships, we find a polarizing relationship which we refer to when one
narrative instance embeds the other, each resolving the others ambivalences or uncertainties. The resulting categories are: contradictory, disparately, complementary, meshing, polarizing, illustrating, paraphrasing.
In practice, of course, these overlapping categories have no clear limits. There could be a relation of dominance for each of these relationships, but it is impossible to point out general guidelines for the specification of these relationships and the resulting relations of dominance
which might arise. There are films in which this relationship can be detected quite obviously; others, however, show no clearly distinguishable
relations. The relationship between visual and verbal narrative instance
within one film is hardly ever static. In more complex cases (for instance,
when questions of unreliable narration arise or in films with various verbal narrative instances) we have to analyze how the interplay of the instances throughout the film has been organized by the implied director.
Even when we analyze conventional forms of transition between different diegetic (or narrative) levels, the heuristic value of the analytical
distinction between visual and verbal narrative instances becomes obvious. This applies all the more when analyzing complex forms of nar-
266
Markus Kuhn
rative mediation in films which use voice-over and have a multi-layered

structure (part 6), or the special cases of film within the film. We will
stress this highly self-reflexive phenomenon in two cases: the shooting of
a film shown in a film (part 6) and film-making as a central theme having
influence on both form and plot (part 7).
6 Narrative Mediation in Self-Reflexive Multilayered Film:

Pedro Almodvars La mala educacin
The opening film of the 2004 Cannes Festival, Pedro Almodvars La
mala educacin (Bad Education), narrates the story on three encapsulated
diegetic levels, which the following short summary will highlight. After
the credits are shown, the diegesis is located with a caption: Madrid
1980. Successful director Enrique Goded (Fele Martnez) is looking for
inspiration for his next movie. Ignacio, a childhood friend of Enrique who
is now an actor under the stage name ngel (Gael Garca Bernal),
coincidentally shows up with a short story he has written. Entitled La
visita (The Visit), the story is based on their childhood. Parts of this
story are shown as a metadiegesis: Ignacio, as a transvestite, returns to the
town where he went to a Catholic elementary school. He intends to
blackmail his former teacher, Father Manolo, with a story about his childhood. On the level of metametadiegesis, this fictitious childhood is also
shown: Enrique and Ignacio discover the first blossoming of their tender
homosexual love. Their friendship is nipped in the bud by Father Manolo,
who is in love with Ignacio and uses his position of power as school principal to abuse Ignacio. Whether, on the level of the metadiegesis, Ignacio
is successful in his attempt at blackmail, is not shown in the film. However, later on the happy ending of the short story is mentioned.
On the level of diegesis, director Enrique decides to make La visita
into a movie. His researches lead him to the house of Ignacios parents,
where he discovers that Ignacio died three years ago. Ignacios younger
brother Juan, taking the name ngel, falsely presented himself as Enriques friend Ignacio. Ignacio however, is the real author of La visita.
Ignacios mother shows Enrique a letter he never received. In this letter
Ignacio explains his attempt at blackmailing Father Manolo.
For the screen adaptation of La visita, Enrique adds a tragic ending
to the story: Father Manolo kills the transvestite Ignacio, because he is
scared he will continue to blackmail him indefinitely. Towards the end of
the film shoot, which is shown in part (the second metadiegesis), a stran-
267
ger appears: Seor Berenguer. He later unveils himself to Enrique as

Father Manolo and tells him how Ignacio really died. This is shown in a
third metadiegesis: Ignacio, as a drug-addicted transvestite, tries to blackmail Seor Berenguer/Manolo. Berenguer makes the acquaintance of
Ignacios younger brother Juan/ngel as he tries to delay Ignacios blackmail attempts and becomes romantically involved with him. Together
they kill Ignacio with an overdose of narcotics.
Firstly we can observe that however complex the structure of levels
might be, all levels are indicated unequivocally. They can be located
within the hierarchy, and all meta- and metametadiegeses are embedded
in a narrative situation, on the next highest diegetic level 12 . Telling and
adapting stories is a central topic of La mala educacin. We see Enrique
as a director collecting clippings of bizarre stories and discussing whether
they have the necessary qualities to be turned into a film. The discours
presents us with various embedded literary stories, a number of letters,
and a film within a film. Functionally the different embedded fictitious
narratives have consequences for the plot on the diegetic levels above
them. The topic of storytelling and narration is complemented by the
films reflexive way of dealing with the relationship between factual and
fictional narration. Ignacio and Enriques childhood and the decisive act
of abuse committed by Father Manolo are, for instance, only presented
through narrated narrated narration. Nevertheless, this fictitious narrative of a childhood experience is considered to be truthful enough to
serve as the basis for blackmail.
The crucial question is how the narratives within the film are being
mediated. The diegesis, i.e. the plot level that revolves around director
Enrique Goded, is mostly shown through an extradiegetic visual narrative
instance (visual NI). In the second third of the film, an extradiegetic,
homodiegetic verbal narrative instance (realized through a voice-over of
Enrique) interferes for a short while. This verbal NI meshes with the visual NI. The verbal NI summarizes the impressions during the film shoot
(internal focalization on the Experiencing-I of Enrique). This internal
focalization corresponds to the focalization of the visual NI on Enrique
(which, however, is often interrupted by zero focalization).
12
This is in contrast, for instance, to some of David Lynchs films, in which we cannot
reconstruct the level structure unequivocally, see Lost Highway (USA 1997), Mulholland Drive (USA/France 2001), Inland Empire (USA/Poland/France 2006).
268
Markus Kuhn
The most conventional change of levels in the film is the last one:
Seor Berenguer is sitting in Enriques office as he prepares to tell Enrique how Ignacio died. In this scene, Berenguer starts telling as intradiegetic, homodiegetic verbal NI: About three years ago, someone put a
copy of The Visit on my desk []. While the last words of his reply
(Ignacio Rodrguez) are spoken, the visual NI changes to the metadiegesis: Berenguer as an editor in a publishing company receives a
phone call by Ignacio Rodrguez, threatening to blackmail him. The following scenes alternate the diegetic and metadiegetic levels. Because the
voice-over of Berenguer reappears several times during this metadiegesis
and the situation of conversation on the diegetic level is frequently intercut with the metadiegesis (altogether five times), this prolonged visual
metadiegesis is linked unequivocally to the situation of conversation. In
some transitions, Berenguers voice is used in a classical overlap: it starts
in the scene on the diegetic level, is continued as voice-over (while the
visual NI changes to the metadiegetic level) and stops when the metadiegesis is (re)established visually.
The first change of levels in the film is equally conventional. The extradiegetic visual NI shows Enrique reading the title The Visit out loud.
His lip movements are synchronized with his voice. His reading of the
story, however, is then rendered through a voice-over. His lips have
stopped movingthe classical form of the filmic interior monologue
used to represent thoughts and inner voices 13 . The visual NI slowly fades
over to the first shot of the metadiegesis and the voice-over soon ends.
The metadiegesis is unequivocally linked to Enrique who is reading,
which becomes even clearer when, throughout this sequence, the visual
NI jumps back to this reading situation several times. We should take note
that it is the voice of Enrique as reader that we hear in this voice-over and
not that of the author of the text. This constellation is reversed in another
part of the film, when Enrique is reading Ignacios letter and we hear
Ignacio in the voice-over (although, on the level of plot, he is already
dead). When analyzing transitions to the level below, it is generally
advisable to observe in which situation of conversation or narration the
13
Strictly spoken, his voice is not an element of the diegetic world. Filmic interior
monologues can usually be described as extradiegetic, homodiegetic verbal NIs with
internal focalization on the Experiencing-I.
269
lower level is embedded, part of whose narrative the lower level forms,
and whose voice is heard in the voice-over 14 .
The first transition from the metadiegesis to the metametadiegesis is
somewhat more conspicuous. Within the metadiegesis shown, Ignacio is
visiting Father Manolo with his childhood story threatening to blackmail
the priest. The visual NI shows how Ignacio indicates a text passage to
Father Manolo. A childs voice (as homodiegetic voice-over) begins
reading the text (the exact part of the text which can be seen in frame).
There is a cut, and the visual NI shows the face of Manolo reading (still in
the metadiegesis) while the childs voice continues reading aloud and
without interruption. Another cut takes us to the level of the metametadiegesis and we see children playing. This meshing of verbal and visual
NI represents a rather conventional transition of levels. However, the
voice we are listening to now is not the inner voice of Manolo reading,
but the conspicuously high-pitched childs voice of the homodiegetic
author of the textafter all, the text which the metadiegetic Ignacio
threatens to use for blackmail was already written by Ignacio as a young
boy. That means that the Narrating-I of the metametadiegesis is the
voice of young Ignacio who wrote it down after the experience (indicated
by the use of past tense). The Experiencing-I of the metametadiegesis is
the young Ignacio who is shown in the scene. The visual metametadiegesis is thus linked to the narrating childs voice of Ignacio, which
14
When embedding and attributing visual metadiegeses, we can also find forms that are
far more complex and ambiguous. The mere opposition of contradictory attributions
through visual and verbal markings, missing markings, narrational forms of attribution
and contradictory denouements of metadiegeses when returning to the diegesis, result
in manifold forms. Kozloff (cf. 1988: 4953) elaborates on the various possibilities of
anchoring voice-over narration in the diegesis, however, she does not take into
consideration visual and complex forms of attribution. The first metadiegesis in
La mala educacin is additionally marked by a frame on the left and right of the film
picture. A black bar moves from the left and right into the picture when changing into
the metadiegesis, and back out of the picture when changing back to the diegesis. With
the second and third metadiegesis of the film there is no such framing (also the
transition from the metadiegesis to the metametadiegesis is not marked by additional
framing). The ambiguous framing of the first metadiegesis can be interpreted, in
respect to the relationship of factual and fictional narration, as an indicator that the
visual representation of the novel is fictional.
270
Markus Kuhn
keeps reappearing as voice-over, as well as to Manolo reading, to whom

the visual NI changes repeatedly 15 .
I have thus given an idea of how the transitions between the levels are
constructed in the film and how the respective meta- and metametadiegeses can be attributed to a reading, writing, or narrating intra- or
metadiegetic character. The question to be answered is who actually narrates the meta- and metametadiegeses. Although the meta- and metametadiegeses can be attributed to a certain telling or reading situation, and
they are usually introduced by a verbal NI, the larger part is shown by a
visual NI. Consequently, the narration relies predominantly on visual
showing (and the showing of dialogues) rather than on verbal telling
(through the short voice-overs). Logically following the proposed construction of levels, the visual NI should be classified as (a) intradiegetic or
(b) metadiegetic if it shows (a) the metadiegesis or (b) the metametadiegesis. Firstly, however, the visual NI that shows the metadiegesis is not
an element of the diegetic world that frames it, i.e. it is not a shown
showing instance (comparable to what we know in literature as narrated
narrating instances). And secondly, neither in the way and style of
showing, nor in focalization and perspectivation, nor in the pace of
editing and mise en scne, does the intradiegetic visual NI differ from the
extradiegetic visual NI (the same holds true for the level below).
This is where we can detect an inherent contradiction which can be
found not only in this film. On one hand, levels can be unequivocally
identified in terms of their diegetic rank. On the other, they are shown
through a visual NI which does not seem to be linked to one particular inner rank of levels. This is related to the fact that a film narrative embedded within a film is usually supposed to simulate a narrative in another medium. The filmic metadiegesis within the filmic diegesis is supposed to represent a written text, a novel, a letter read aloud, a stream of
consciousness, an oral account, a recorded tape, etc. It is only very rarely
used as what it really is, as a subordinate film sequence within a film. The
rarer case of the unequivocal visual change of levels, i.e. actually shown
visually-showing narration, can be found in two forms in La mala
educacin: firstly, the shooting of the film within the film and secondly,
protagonists watching a film. The second form can be found more
15
Another form of transition between levels in La mala educacin works entirely without
voice-over: at first, the visual NI shows a page of text being read and then fades over to
the next-lowest diegetic level.
271
frequently in fictional film: the extradiegetic visual NI shows the intradiegetic narratee who is watching a film, which in turn is being shown
by an intradiegetic visual NI. Film production within a film, however, is
an exception.
In the case of an ascription of a filmic metadiegesis to a verbal narrative situation of the diegesis, there is usually no unequivocal visual
change of levels. The visual NI of the diegesis cannot be distinguished
systematically from the visual NI of the meta- or metametadiegesis. While
I can only treat it as an example in this context, I will term this
phenomenon a visual short-circuit of levels. This short-circuit arises
from the fact that a visual NIwhich is not an element of the diegetic
world and must therefore be located on the extradiegetic levelnarrates
stories through showing which, with the help of specific markings, are
attributed to intra- and metadiegetic characters (although they are not told
by them, or, if so, only in part) 16 .
This visual short-circuit of levels can be irrelevant in the context of
certain films. In La mala educacin, however, it leads to a clear dominance of the visual. Most of the episodes told on the various diegetic
levels are being shown by an extradiegetic visual NI in the same characteristic style. They are thus focalized and assessed. This is how the
metametadiegesis of Ignacios and Enriques childhood is endowed with a
greater sense of reality than it would have in terms of the logic of levels
which in turn is significant if we want to assess the relationship of
fictional and factual narration.
One of several prominent sequences demonstrating the potency of the
visual NI, situated beyond the diegetic levels, is a sequence in which a
16
This phenomenon, which I subsume as a visual short-circuit of levels, has not been
extensively discussed yet. It is evident in many fictional films with multiple levels or
with embedded voice-over narration. Things are somewhat different when the whole
metadiegesis is marked visually (for example throught the use of black and white
footage within a color film, special lenses and filters, etc.). When applying the model
of diegetic levels in literature to the film medium, special mention has to be made of
the problem of the visual short-circuit of levels. How this applies specifically to film
depends on our understanding of diegetic levels. There are substantial reasons to maintain the term of levels (alternatively one could speak of pseudometadiegesis, etc.)
and to investigate which function or effect a visual short-circuit of levels can serve in
a film. Kozloff (cf. 1988: 4349) circumvents this question when she refers to the
viewer (whom, however, she does not define as a category) who attributes everything
that is shown to a character, even though she maintains that a more powerful narrating
agent, the image maker (49) is positioned above any homodiegetic voice-over.
272
Markus Kuhn
transition from the metametadiegesis to the diegesis occurs, skipping the

metadiegesis. In a shot-reverse-shot beginning on the level of metametadiegesis, showing Ignacio and Enrique as young boys looking at each
other, the visual NI shows a time-lapse of young Ignacios face maturing
to his presumed age on the level of diegesis (yet not showing an adult
Ignacio, but ngels face). Subsequently, the visual NI shows young
Enriques face changing into the face of an adult Enrique, on the level of
the diegesis. The film continues with the transformed characters on the
diegesis. This transition from the metametadiegesis to the diegesis,
created by a technical device, is not irrelevant for the assessment of the
figuration. After all, the visual NI indicates that ngel is Ignaciowhich
proves wrong in the course of the plot. The narratee of the extradiegetic
visual NI is being deceived, just like Enrique by ngel.
The dominance of the extradiegetic visual NI becomes even more
apparent in the only sequence in which a real intradiegetic visual NI is
present in form of the camera shooting the film within the film. The sequence begins with the take being shot by this intradiegetic camera:
Ignacio threatening to blackmail Father Manolo. The clapperboard onscreen and several off-screen voices (quiet, roll camera) are clear
markings of the film shoot. The sequence then cuts away to a shot of the
film team and camera on set. After two alternations between the filmed
situation (metadiegesis) and the filming situation (diegesis), a longer episode of the metadiegesis follows. In order to be consistent, this metadiegetic sequence should be shown as a single shot from a single camera
position. However, we are confronted with evident use of editing: Ignacio, Manolo and his associate are shown in a shot-reverse-shot, which
would require three camera positions in the diegesis. Following the return
to the diegesis, we are left in no doubt that there is only a single camera
on set. Thus we cannot attribute the metadiegetic sequence to this one
intradiegetic camera. What we are confronted with at this point is a latent
visual short-circuit of levels: the extradiegetic visual NI has taken over
the film within the film without breaking the illusion of the metadiegesis,
or showing the intradiegetic camera again. The extradiegetic visual NI
shows the metadiegesis of the film Enrique is directing in the same
stylistic manner as the first visual metadiegesis, namely Enrique reading
the short story. We could interpret the film adaptation as a continuation of
the short story. Alternatively Enriques reading of the story could be
interpreted as his inner film, the directors vision of the completed film.
Further indicators support both these interpretations, making La mala
273
educacin an increasingly reflexive film about visual storytelling. This

ambivalence can only be resolved through interpretation, and not by
systematic analysis. If, in this context, we decided to cross the boundaries
of the work itself, we could even go a step further and suppose an autobiographic connection between director Almodvar and his fictitious
character, director Enrique Goded. At the very least La mala educacin is
equally a result and an account of all its reflections about the issues of
adapting literary narratives for the screen.
The layering of diegetic levels in La mala educacin functions as an
important device in the creation of suspense. Playing the possibilities of
its various subplots, the film mainly builds suspense on the level of
discourse. La mala educacin can thus be viewed as part of a tendency
in contemporary cinema: films that self-reflexively put the possibilities of
cinematic narrative to use in order to build immensely dense movies and
to create suspense through their narrative realization.
7 Film Production within the Film and the Question of FirstPerson Film: Lars Kraumes Keine Lieder ber Liebe
Lars Kraumes film Keine Lieder ber Liebe (Germany 2005), literally
No Songs Of Love, presents us with an unusual example of self-reflexivity without transition between diegetic levels. The film blurs the borders
between documentary and fiction film in several points. Its a fictional
documentary about the Hansen Band, a band of real actors and musicians
cast for the film, touring Germany. This real tour serves as the set-up
for a fictitious mnage trois revolving around the films three main
characters: Tobias, his brother Markus, and Tobiass girlfriend Ellen.
Aspiring filmmaker Tobias Hansen (Florian Lukas) is making a documentary about his brother Markus (Jrgen Vogel), who is the lead singer
of the Hansen Band. Tobias takes his girlfriend Ellen (Heike Makatsch)
along on the tour, and soon discovers she cheated on him with Markus a
year earlier. The film Keine Lieder ber die Liebe purports to be Tobiass
documentary film. The development of the mnage trois increasingly
pushes Tobias into the focus of his film, turning the project into a filmic
self-portrait.
On the extradiegetic level we have a visual NI which is flexible in
terms of focalization, ocularization, and auricularization and which makes
use of various visual stylistic elements in order to suggest either a high
degree of immediacy or its presence within the diegesis. Moreover,
274
Markus Kuhn
Tobias Hansenss voice-over must be considered a homodiegetic verbal

NI on the extradiegetic level. Depending on the point of the story that is
being told, it fluctuates between internal focalization on the ExperiencingI and zero focalization when the Narrating-I tells more than the Experiencing-I knows. In many of the voice-over replies, these two Is can
easily be distinguished from one another (e.g.: I swear that at the time I
had no idea this would become a film about the three of us 17 ).
When the visual NI is showing Tobias, the shown character corresponds to the Experiencing-I of the verbal narrative situation. This
Shown-I of the diegesis is aware of the fact that it wants to turn all the
material it is collecting into a film. This is why it sometimes announces
specific information about time and setting in front of the camera or once
whispers an account which is supposed to be seen and heard only by the
camera (and not by his girlfriend who is in the adjacent room). The scenic
Shown-I that talks to the camera upfront wants to make a documentary
which isand this is where we find the first (self-)reflexive feedback
already completed at the moment it is being shown by the visual NI and
being seen by the extradiegetic narratee. The second feedback is created
when Tobias Hansen, as the fictitious filmmaker who edits all the material
afterwards, ostensibly takes on the position of the implied director who
organizes the interplay of verbal and visual NI.
This constellation brings the act of narration to attention in a selfreflexive way, and is comparable to constellations in literature where the
narrative takes the form of a text written by an extradiegetic homodiegetic narrator. This type of literary narrative situation is usually adapted to the film medium with the help of voice-overs and/or framing narratives. The film is then meant to be a novel written by a (intra)diegetic
character 18 . However, the constellation found in Keine Lieder ber Liebe
is much rarer: the film is meant to be a documentary film made by one of
the films characters.
Do we then have a homodiegetic cinematic narrator in the narrow
sense, i.e. a human character who narrates in a cinematic format? In sections of the film that use voice-over, we can detect a homodiegetic verbal
NI (and thus a first-person narrative situation in the narrow sense). This is
17
18
Here, as in the following, English translations are mine.

This attribution in a different medium is in part comparable to the attribution of a
filmic metadiegesis to a verbal narrative situation in the diegesis found in La mala
educacin.
275
particularly true for the exposition in which the homodiegetic verbal NI

dominates the visual NI. On the visual level, there are only rare instances
of a homodiegetic NI. However, the structural position of the implied director is in part filled by Tobias, a diegetic character. But there are two
things to take into account: firstly, the intratextual implied director must
be distinguished from the extratextual actual director of the film, Lars
Kraume (there is no autofictitious short-circuit). Secondly, there is a difference between the character designated as implied director and the instance of the implied director that actually structures the film itself. The
difference is hardly noticeable at times, at others it can be clearly detected, particularly if we look at the structures that establish Tobias as director within the film. The character of Tobias is not the implied director
(which is why we should rather speak of a fictitious director). Nevertheless, the processes that structure the film, which are usually attributed
to the implied director, can be attributed to Tobias, or he can be seen as
the image of the director conveyed by the film. However, the question of
whether Tobias has carried out certain filmic structures consciously or
unconsciously can only be answered through interpretation. Just as in
every self-portrait, we cannot draw clear lines between conscious, unconscious and coincidental. Additionally Tobias is just a character created by
a film purporting to be his workKeine Lieder ber Liebe is thus in a
way a cinematic counterpart of M.C. Eschers Drawing Hands.
The film alternates various narrative constellations. There are montage
sequences to the bands songs in which the heterodiegetic visual and the
homodiegetic verbal NI mesh to summarize longer developments in the
relationships (summary; both instances: zero focalization). The absence of
a verbal NI and the visual NIs stepping back through the use of inconspicuous cuts, with close shots and close-ups bringing the characters
emotions into focus, create a strong sense of immediacy during conflict
scenes. In other scenes the presence of the camera is mentioned by the
characters without a visual cue (Shall we take the camera with us to
bed?). They interact consciously with the camera, making signs to it,
speaking to it directly, or they block the cameras view, not wanting to be
filmed any longer. In some of these sequences we can detect a rare form
of the homodiegetic visual NI without use of voice-over, a sort of purely
visual first-person narrative situation 19 . In these scenes the visual NI can19
For the question of first-person film see Kawin (1978); Hurst (1996); Brinckmann
(1988); Bach (1999), et al.
276
Markus Kuhn
not be attributed to Tobias, but rather to an anonymous cameraman, or the

camera itself. The camera, which is not seen on-screen, is only part of the
scene because the characters interact with it. When the film shows what
the camera is recording at the current point in time, the camera is a part of
the diegetic world (it is therefore homodiegetic) but it also creates the
diegetic world (and is therefore extradiegetic) 20 . These few sequences in
which we can speak of a homodiegetic visual NI are unedited. The end of
these scenes is a natural result of the action, for example when Tobias
puts down the camera he has been speaking to when his girlfriend enters
the room. We can partly detect an anthropomorphization of the homodiegetic camerafor example when it is being carried around like a
video camera (idea of someone who is live on scene). In other scenes the
camera is used as a purely technical instrument which, for instance, is
being put on a tripod.
These scenes in which a camera is used as a homodiegetic visual NI
are short and in most cases, editing indicates an interference with the immediacy of the scene. Thus a conspicuous temporal difference is created:
the homodiegetic camera is located on the temporal level of the diegesis, editing on the temporal level of the narration afterwards. This time
gap corresponds to the temporal difference between the Narrating-I and
the Experiencing-I: the simultaneousness of the cameras recording corresponds to the simultaneousness of the Experiencing-I; the temporal
delay of the editing corresponds to the temporal delay of the Narrating-I.
However, this special case of the homodiegetic camera can only be
attributed to Tobias Hansens Experiencing-I in short passages of the exposition. In all other cases, it remains a homodiegetic technical instrument
or, in the case of anthropomorphization, it is attributed to an anonymous
character: the camera as an anonymous homodiegetic Observer-I or as a
first-person observer.
A scene in which Markus and Ellen flirt in Tobiass absence illustrates
the interplay of instances. Both are aware they are being filmed, the
camera as Observer-I is present, and Markus takes off his coat to give it to
Ellenjust like in one of those old movieshe looks directly into the
camera and says: Let him see this. If he doesnt like it because he is
20
The visual NI/camera is not intradiegetic because it is never seen on screen, even if the
characters interact with it (not a case of a shown showing camera). Instead, it shows
what the camera is seeing at the same moment (internal ocularization)that what
the camera is seeing in these sequences is simultaneously the diegetic world which
only exists because it is shown.
277
jealous he can cut it out. He is Tobias, who will edit the video material
afterwards. The fact that Tobias, the editor, leaves this passage in the film,
in spite of or even because of the risky flirt going on, draws our attention
to the fact that he is the organizing presence after the shooting. He saw
the scene and decided not to cut it out. The conscious nature of this
decision is emphasized by a noticeable cut immediately after Markuss
reply. Tobiass I, the Editing-I, is present, without the presence of the
scenic Shown-I of Tobias. The Observer-I, the anonymous cameraman, is
not Tobias, but he is not independent of Tobias because of Tobiass
influence as the filmmaker either.
This constellation, characterized by two exceptions, the character as a
fictional director and the camera as a homodiegetic observing instance, is
extremely rare in film. Keine Lieder ber die Liebe complements this with
highly specific narrative situations. Such constellations featuring a
diegetic character filling a position endowed with multiple functions,
metaleptic short-circuits, and a complex interplay of instances in unusual
narrative forms are also known in classical narratology. Because of their
similarities and differences to comparable literary constellations, some of
the narrative situations which we have analyzed in both films might hopefully inspire the imagination of transmedial narratology.
References
Bach, Michaela (1999). Dead MenDead Narrators: berlegungen zu Erzhlern und
Subjektivitt im Film. W. Grnzweig et al. (eds). Grenzberschreitungen. Narratologie im Kontext. Tbingen: Narr, 23146.
Booth, Wayne C. (1961). The Rhetoric of Fiction. Chicago: U of Chicago P.
Bordwell, David (1985). Narration in the Fiction Film. Madison: U of Wisconsin P/Methuen.
Branigan, Edward (1992). Narrative Comprehension and Film. London: Routledge.
Brinckmann, Christine N. (1988). Ichfilm und Ichroman. A. Weber & B. Friedl (eds).
Film und Literatur in Amerika. Darmstadt: Wissenschaftliche Buchgesellschaft, 65
96.
Chatman, Seymour (1978). Story and Discourse. Narrative Structure in Fiction and Film.
Ithaca: Cornell UP.
Cornell UP.
Fieguth, Rolf (1973). Zur Rezeptionslenkung bei narrativen und dramatischen Werken.
Sprache im Technischen Zeitalter 43, 186201.
Fleishman, Avrom (1992). Narrated Films. Storytelling Situations in Cinema History.
Baltimore: Johns Hopkins UP.
Fludernik, Monika (1996). Towards a Natural Narratology. London: Routledge.
278
Markus Kuhn
Gaudreault, Andr (1988). Du littraire au filmique. Systme du rcit. Paris: Klincksieck.

Griem, Julika & Voigts-Virchow, Eckart (2002). Filmnarratologie: Grundlagen, Tendenzen und Beispielanalysen. A. Nnning & V. Nnning (eds). Erzhltheorie transgenerisch, intermedial, interdisziplinr. Trier: WVT, 15583.
Herman, David (2002). Story Logic. Problems and Possibilities of Narrative. Lincoln: U
of Nebraska P.
Hurst, Matthias (1996). Erzhlsituationen in Literatur und Film. Ein Modell zur vergleichenden Analyse von literarischen Texten und filmischen Adaptionen. Tbingen:
Niemeyer.
Jahn, Manfred (1995). Narratologie: Methoden und Modelle der Erzhltheorie.
A. Nnning (ed). Literaturwissenschaftliche Theorien, Modelle und Methoden. Eine
Einfhrung. Trier: WVT, 2950.
Jost, Franois (1987). LOeilCamra. Entre film et roman. Lyon: PU de Lyon.
Kawin, Bruce F. (1978). Mindscreen. Bergman, Godard, and First-Person Film.
Princeton: Princeton UP.
Kindt, Tom & Mller, Hans-Harald (1999). Der implizite Autor: Zur Explikation und
Verwendung eines umstrittenen Begriffs. F. Jannidis et al. (eds). Rckkehr des
Autors. Zur Erneuerung eines umstrittenen Begriffs. Tbingen: Niemeyer, 27387.
Kindt, Tom & Hans-Harald Mller (2006). The Implied Author. Concept and Controversy.
Berlin: de Gruyter.
Kozloff, Sarah (1988). Invisible Storytellers. Voice-over Narration in American Fiction
Film. Berkeley: U of California P.
Metz, Christian (1971). Langage et cinema. Paris: Larousse.
Nnning, Ansgar (1993). Renaissance eines anthropomorphisierten Passepartouts oder
Nachruf auf ein literaturkritisches Phantom? berlegungen und Alternativen zum
Konzept des implied author. Deutsche Vierteljahresschrift fr Literaturwissenschaft und Geistesgeschichte 67, 125.
Prince, Gerald (2003). Surveying Narratology. T. Kindt & H.-H. Mller (eds). What is
Narratology? Questions and Answers Regarding the Status of a Theory. Berlin:
de Gruyter, 1-16.
Ryan, Marie Laure (2005). On the Theoretical Foundations of Transmedial Narratology.
J. C. Meister (ed). Narratology beyond Literary Criticism. Mediality, Disciplinarity.
Berlin: de Gruyter, 123.
Schlickers, Sabine (1997). Verfilmtes Erzhlen. Narratologisch-komparative Untersuchung
zu El beso de la mujer araa (Manuel Puig/Hctor Babenco) und Crnica de una
muerte anunciada (Gabriel Garca Mrquez/Francesco Rosi). Frankfurt a. M.:
Vervuert.
Wolf, Werner (2002). Das Problem der Narrativitt in Literatur, bildender Kunst und
Musik: Ein Beitrag zu einer intermedialen Erzhltheorie. A. Nnning & V. Nnning (eds). Erzhltheorie transgenerisch, intermedial, interdisziplinr. Trier: WVT,
23104.
JAN-NOL THON
(Hamburg)
Perspective in Contemporary Computer Games 1

1 Toward a Model of Perspective in Contemporary
Computer Games
While the relatively new medium of the computer game has elicited an
increasing amount of academic attention from a variety of disciplines in
the last few years, research on perspective and point of view in computer
games generally focuses on questions regarding the presentation of space,
i.e. on perspective as being determined by a point of view in the purely
spatial sense 2 . Within narratology, on the other hand, it is quite common
to conceptualize point of view and perspective as multidimensional phenomena, both with regard to literary texts 3 and, albeit to a lesser extent,
narrative films 4 . It therefore seems as if our understanding of perspective
in computer games could benefit from the complex models of perspective
that narratology has developed. Computer games, however, are neither
literary narratives nor narrative films, and although the results of narratological research on perspective are doubtlessly inspiring, most of the
models developed for the description of literary texts (or narrative films,
for that matter) cannot be directly applied to computer games without
missing some of their most central characteristics. Hence, the present paper proposes a multidimensional model of perspective in computer games
that takes into account their specific medial properties.
For this purpose, we distinguish between three dimensions of perspective. The first dimension is that of spatial perspective, which is determined by the point of view, i.e. the spatial position from which the game
1
2
3
4
A longer version of this paper was published online in 2006 as Toward a Model of
Perspective in Contemporary Computer Games. <http://www.icn.uni-hamburg.de/
images/download/beitrag_thon_bfs.pdf > (15.9.2008).
See Poole (2004); Rumbke (2005); Wolf (2001).
See Chatman (1978); Schmid (2005); Uspenskij (1973).
See Branigan (1984); Mitry (1998); Smith (1995).
280
Jan-Nol Thon
space is presented audiovisually (this includes the presentation of sound

which is often presented from the same position that the game space is
presented from). Since the presentation of space in computer games is audiovisual instead of verbal and therefore closer to the movies than to literary narrative texts, we will mainly draw on film theory and works on
perspective from computer game studies, rather than try to adopt models
developed in literary narrative theory. The second dimension is that of actional perspective, which is determined by the point of action, i.e. the position from which the player can interact with the game space. Here, we
will mainly refer to Neitzels work on the point of action in computer
games (cf. Neitzel 2002). The third and most complex dimension is that
of the ideological perspective structure, which is determined by the various positions from which the events in the game are evaluated. Although
we will focus mainly on the question of how characters in computer
games evaluate events and situations, this dimension also refers to other
positions within a game, namely that of the player and the implied game
designer. With reference to the spatial perspective determined by the
point of view and the actional perspective determined by the point of action, we will here speak of an ideological perspective that is determined
by the point of evaluation.
Before we discuss these types of perspective in more detail, it has to
be stressed that the three dimensions of perspective distinguished here are
not all that could be considered. Although the spatial, actional and ideological dimensions of perspective seem to be most central, the analysis of
particular games might well make it necessary to examine dimensions of
perspective not treated in this paper 5 . Especially the analysis of the ideological perspective structure of a game may make it necessary to describe
other forms of perspective that may be used in the presentation of fictional worlds in contemporary computer games. Our main aim, however,
lies in the introduction of the idea that perspective in computer games
consists of more than just spatial perspective, and the distinction of three
dimensions of perspective seems to be enough for this purpose.
With regard to additional dimensions that could be considered in the analysis of computer games, one can examine the narratological models of perspective already mentioned. Schmid (2005), for example, distinguishes between five dimensions of perspective in literary narrative texts, namely spatial, ideological, temporal, linguistic and perceptual perspective. Both the linguistic and temporal perspective may occasionally be
worth analysing, especially with regard to the narrative elements of computer games.
Perspective in Contemporary Computer Games
281
2 Point of View and Spatial Perspective

There is a wide variety of ways in which computer games can construct
the space in which they take place, from all text-based (Wolf 2001: 53)
via various forms of two-dimensional spaces (cf. 5565) to [i]nteractive
three-dimensional environments (65). However, since many if not most
contemporary computer games present a three-dimensional space on a
two-dimensional screen, it is this form of computer game space that the
present paper is mainly interested in. Before we can examine more closely
the various forms of spatial perspective that can be found in such games,
it has to be made clear to which parts of these games we refer. Since
many computer games are set in complex fictional worlds, one has to
distinguish between the space of the fictional world as a whole and the
spaces that the player can interact with through the interface. Jesper Juul
draws a similar distinction between world space and game space (cf.
Juul 2005: 16467). Since most of the events in computer games take
place in the game space, it seems to be mainly this part of the space of the
fictional world that is of interest with regard to the question of spatial perspective in computer games.
Such game spaces often are three-dimensional environments in which
the player can more or less freely move certain objects such as his or her
avatar (i.e. representative in the game space) as well as the point from
which the space is presented and which, in games using an avatar, is often
in some way connected to the position of the latter (thereby moving
automatically when the avatar is moved). When referring to the point of
view in computer games, one of the more commonly used terms is that of
camera position (cf. Rumbke 2005: 24445). This is not too surprising
since, according to Wolf, many contemporary computer games follow, to
some degree, the precedent set by the space represented in classical
Hollywood film (Wolf 2001: 66) and accordingly the presentation of the
game space in computer games may at first glance seem similar to the
presentation of space in film. But while terminology originating from film
theory is doubtlessly useful for describing spatial perspective in audiovisual media, it has to be emphasized that all talk of a camera or a
camera position is metaphoric when referring to computer games since
game spaces are generally not created by actual film cameras. Hence, it
seems more precise to speak of a point of view as the spatial position
from which the game space is presented aurally as well as visually and
which determines the spatial perspective of a computer game.
282
Jan-Nol Thon
One of the most common distinctions between different types of spatial perspective in computer games is that of first-person perspective,
where the game space is presented from the spatial (and sometimes even
perceptual) position of the players avatar, and that of third-person perspective, where it is not. Aside from the fact that the category of thirdperson perspective is very broad (cf. Rumbke 2005: 24648), this distinction is also inappropriate in its reference to grammatical categories
that cannot be applied to audiovisual presentations of space in such a
straightforward manner. A more appropriate and differentiated categoryization of audiovisual point of view in computer games has been proposed
by Neitzel (2002). Referring to Mitrys The Aesthetics and Psychology of
the Cinema (1998), she distinguishes between subjective, semi-subjective
and objective points of view. Although this distinction is relatively broad
as well, it provides a good starting point for a description of the spatial
perspective(s) used in actual games.
3 Subjective, Semi-Subjective and Objective Points of View

Computer games using a subjective point of view have the position from
which the game space is presented coincide with the position of the players avatar. This perspective is, most prominently, used in so-called firstperson shooter games such as Doom (1993), Halo (2001), or SWAT 4
(2005). One can, in fact, observe an increasing sophistication in the way
first-person shooter games realize their respective subjective points of
view. While early games such as Doom use nothing more than a hand
holding a weapon protruding into the presented space to indicate the existence of the players avatar, more recent games such as Halo show its
avatar on various occasions. Nevertheless, the hand holding a weapon is
still seen most of the time (figure 1). There is, however, a tendency towards an implementation not only of the spatial but also the perceptual
perspective (cf. Schmid 2005: 13132) of the players avatar that has led
to games such as SWAT 4, where grenades, pepper spray and flash packs
not only affect the avatar, but also have an effect on the audiovisual presentation of the game space. Another instance of a game that simulates the
perceptual perspective of its avatar is World of Warcraft (2004), where
the avatars drunkenness affects the presentation of the game space.
283
Figure 1: Subjective point of view in Halo (2001)
According to Neitzel, one can speak of a semi-subjective point of view

when the point of view is connected to the movements of the avatar; it is
not a substitute for the viewpoint as in case of the subjective POV, but
rather a viewing-with (Neitzel 2002: n. p.) the players avatar. The camera follows the avatar at some distance, allowing for a better sense of its
precise position in the game space than is the case in games with a subjective point of view. This form of spatial perspective is typically used in
action adventures from Tomb Raider (1996) to Grand Theft Auto: San Andreas (2005) as well as in more recent role-playing games such as Fable
(2004), Jade Empire (2005) or World of Warcraft. A closer examination
of these games reveals that although the category of semi-subjective point
of view allows for some variation as to the distance between the position
of the camera and the avatar or the angle from which the avatar is shown,
many games using a semi-subjective point of view use it in quite a similar
manner. Most of the time, the camera floats slightly above and some way
behind the avatar, showing it in relation to its surroundings (figure 2).
284
Jan-Nol Thon
Although the spatial position of the avatar is not the same as that of the
camera, the cameras position is always linked to the avatar.
Figure 2: Semi-subjective point of view in World of Warcraft (2004)
When the game space is presented from a position that is not connected to
an avatar, one can speak of an objective point of view. This oldest and
most diversified (Neitzel 2002: n. p.) perspective is used in a wide variety of games, but most obviously in strategy games such as Z (1996),
Warcraft III (2002) or Warhammer 40.000: Dawn of War (2004). The
main aim of these games is to build large armies and take control of the
game space, which normally consists of a more or less extensive landscape. Hence, the objective point of view in these strategy games offers
the possibility to observe a large game space without being constrained by
the spatial perspective of an avatar or comparable entity. The objective
point of view shows a game space from a position that is not part of this
game space (as is the case with a subjective point of view) and is not
connected to an entity in the game space (as is the case with a semisubjective point of view). However, most strategy games do not show the
285
whole game space at once, but present only a small part of it at a time,
allowing the player to determine which part is shown (figure 3).
Figure 3: Objective point of view in Warhammer 40.000: Dawn of War (2004)
4 Point of View and the Player

Although one could further distinguish between various forms of objecttive point of view (especially when attempting to describe not only computer games presenting a three-dimensional space on a two-dimensional
screen but also games presenting two-dimensional game spaces), Neitzels general distinctions that can be mixed and altered in the games
(Neitzel 2002: n. p.) seem to be appropriate for a categorization of spatial
perspective in computer games. Nevertheless, it should be emphasized
that many contemporary games not only combine various forms of spatial
perspective but also allow their players to control camera movements
(which is an essential part of the gameplay in most strategy games) and
switch between different perspectives themselves. While such a playercontrolled change in perspective is naturally rare in first-person shooter
286
Jan-Nol Thon
games that derive their name from a constant use of the subjective point
of view (although Halo switches to a semi-subjective point of view when
the avatar is controlling vehicles), it has become common in games using
a semi-subjective point of view to allow the player some degree of control
over the camera position. There are even games such as World of
Warcraft that allow their players to switch from a semi-subjective to a
subjective point of view if they so desire.
In Tomb Raider, which founded the action-adventure genre, the player
cannot change the semi-subjective point of view the game uses to present
its game space. It is, however, possible to influence the position from
which the game space is presented by way of making Lara Croft, the avatar of the game, look in various directions. Without switching to a subjective point of view, the camera will then change its position, allowing
the player to see what Lara seesor would see if she was not an avatar in
a computer game but a real person capable of seeing (figure 4).
Figure 4: Lara Croft in Tomb Raider (1995), looking to her upper-left hand side
287
Obviously, the ways in which the player can influence the camera position have evolved since 1996, the year in which Tomb Raider was published. Hence, World of Warcraft allows its players not only to change the
camera position in order to look at the avatar from virtually all angles but
also to change the distance between the camera and the avatar, which can
be adjusted on a scale of 15 steps. While the largest distance allows the
player to see the most of the surroundings of his or her avatar, the
smallest distance makes the position of the camera coincide with the spatial position of the avatar, thereby allowing the player to switch from the
semi-subjective point of view (which is the standard mode of the game in
version 2.0) to a subjective point of view.
It can be concluded that many contemporary computer games allow
their players an ever greater amount of control over the spatial perspective(s) used in the presentation of the game space. While this is particularly the case with action-adventure and role-playing games, it is also true
for most other games with the previously mentioned exception of firstperson shooters. Since strategy games do not present the player with a
single avatar, the occurrence of a genuine semi-subjective or even subjective point of view seems unlikely here. Nevertheless, most of the more recent strategy games, e.g. Warcraft III and Warhammer 40.000: Dawn of
War, allow the player not only to change the part of the game space that is
presented on the screen, but also to change the camera angle from which
it is presented. Finally, it may be noted that while players generally like
the opportunity to take control of the camera, they rarely use the possibility to change the default point of view. This has to do with the fact
that the default point of view is often best suited to the interaction with
the game space required by the game. And although the appreciation of
beautyfully designed game spaces is surely a part of the pleasure in playing a computer game, the interaction with the game space will, of course,
be more important to most players than the game space itself.
5 Point of Action and Actional Perspective

Unlike the spaces that are presented in Hollywood film, computer game
spaces allow players to interact with them through the interface. The importance of this interactive nature of computer games leads us to the question of how the interaction between player and game can be described in
terms of perspective. For this purpose, we will build on Neitzels notion
of a point of action, by which she refers to the position from which ac-
288
Jan-Nol Thon
tion can be taken, and the way it will be taken in (Neitzel 2002: n. p.),
determining the actional perspective of the computer game. So what exactly is meant by actional perspective with regard to computer games?
Neitzel describes the relationship between the seeing and acting of the
computer game player as follows: The computer takes the effects of the
actions out of the spatial-material reality of the player and distributes
them in the space of the monitor. This space, including the effects of the
actions, is observed and interpreted [by the player, J.-N.T], which then influences the subsequent actions (Neitzel 2002: n. p.). It is not, however,
the case that a player can choose freely what he or she sees or does when
playing a computer game. As we have seen, computer games present their
game spaces using different points of view that result in different spatial
perspectives and thereby determine to a great extent which part of the
game space can be seen by the player and how he or she sees it.
In much the same way, computer games use different points of action
that result in different actional perspectives and thereby determine what
the player can do in the game and how he or she can do it. Neitzel argues
that the point of action in computer games can be described using three
basic distinctions. Firstly, the point of action can reside either within or
outside the diegesis, so that one can speak of an intradiegetic and an extradiegetic point of action (Neitzel 2002: n. p.). Secondly, Neitzel distinguishes between a concentric and an ex-centric and, thirdly, between a
direct and an indirect point of action. Since an intradiegetic point of action means that the actions of the player result in actions that can be ascribed to some character or object within the game world, every game that
uses an avatar automatically uses an intradiegetic point of action. An
extradiegetic point of action means that the actions of the player result in
actions that cannot be ascribed to some character or object within the
game world. This is typically the case in strategy games that do not cast
the player in the role of some ruler character, who then guides the fortunes of his subjects (Neitzel 2002: n. p.).
The distinction between intradiegetic and extradiegetic points of action is often not very clear-cut, since games such as Warcraft III or Warhammer 40.000: Dawn of War do not in any explicit way construct a ruler
character to whom the results of the player actions could be ascribed, but
still have the player-controlled troops react to the players commands
with expressions of obedience such as Yes Sir!, thereby implying that
the result of the players actions can actually be ascribed to some entity
within the game world (the same entity that is addressed as Sir in the
289
above example). Although there seem to be considerable differences between the ways in which the points of action in these strategy games and
those in games such as Grand Theft Auto: San Andreas that have the player control the same avatar through the whole game are related to the entities in the fictional worlds of these games, one would have to describe
both points of actions as intradiegetic. Hence, the usefulness of that first
distinction may be doubted.
Neitzels second distinction is much clearer. She proposes to distinguish between a concentric point of action, meaning that the players actions are executed at only one location in the game space and an excentric
point of action, meaning that the players actions can be executed at
multiple locations in the game space. Hence, games such as Grand Theft
Auto: San Andreas which have the player control a single avatar to which
the result of the players actions can be ascribed would be categorized as
using a concentric point of action while games such as Warhammer
40.000: Dawn of War, where the player uses the keyboard and mouse to
control huge armies, taking control of individual troops or buildings as he
or she pleases would be categorized as using an ex-centric point of action.
While this distinction helps to describe which objects in the game space
are controlled by the player, it does not answer the question of how they
are controlled, i.e. how the actions of the player influence objects in the
game space.
It is this question to which Neitzels third distinction refers. Many
games using an avatar allow the player to control the avatar directly. This
means that every press of a button or movement of the mouse results in an
instant action of the avatar. Among many other games, first-person
shooters generally use such a direct point of action. On the other hand,
there are many games where the relation between player actions and avatar actions is not as direct. Strategy games such as Command and Conquer, Warcraft III, or Warhammer 40.000: Dawn of War often allow the
player to take control of many different objects in the game world. In
these games, a click with the mouse is enough to make a large number of
troops move over a large distance, and another click will make them attack the enemy. It is not necessary (or even possible) for the player to
control directly every movement of his or her troops. This also means that
there is no constant association of the pressing of a certain button with a
resultant movement of the avatar. Hence, one can say that these games
use an indirect point of action. Neitzel also notes that some games combine a direct and indirect point of action. This is the case, for example, in
290
Jan-Nol Thon
World of Warcraft, where the player controls the basic movements of the
avatar directly, but also has to employ the mouse to make the avatar use
its abilities or interact with other characters by clicking on a variety of
icons or on the character he or she wants to interact with.
6 Subjective, Semi-Subjective and Objective Points of Action

Although especially the latter two of Neitzels distinctions seem quite
useful, it is questionable if a typology as complex as the one proposed by
Neitzel is necessary. Alternatively, we propose to distinguish between
only three different kinds of point of action, applying the distinction between the subjective, semi-subjective and objective point of view to the
point of action. In games that use a subjective point of action, the action
position of the player coincides with that of the players avatar. Here, the
player has direct control over the movements of his or her avatar, every
press of a button instantly results in an action (Neitzel 2002: n. p.). This
also means that the player can control his or her avatar and nothing else.
The player cannot interact directly with the game space.
In games that use the semi-subjective point of action, the interaction
with the game world is connected to an avatar, but the player also has to
interact with the game space directly. Interaction does not exclusively
happen through the avatar, as is the case in games using a subjective point
of action. In games such as World of Warcraft, the player controls the
basic movements of the avatar in the same way as in games using a
subjective point of action. He or she does, however, also have to employ
the mouse to make the avatar use some of its various abilities or interact
with other characters. In games using an objective point of action, the interaction with the game world is not connected to a single avatar. This is,
for example, the case in strategy games such as Warhammer 40.000:
Dawn of War where the player uses the keyboard and mouse to control
huge armies, taking control of troops or buildings as he or she pleases.
Although there may be a certain tendency for the three types of point
of action to converge with the respective forms of point of view, this is by
no means generally true. Tomb Raider combines a semi-subjective point
of view with a subjective point of action, Baldurs Gate (1999) combines a
semi-subjective point of view with an objective point of action and Myst
(1993) combines a subjective point of view with an objective point of
action. Furthermore, although of central importance for the gaming experience, the spatial perspective as determined by the point of view and the
291
actional perspective as determined by the point of action are not the only
ways in which the presentation of events in a computer game is perspectivated.
7 Point of Evaluation and Ideological Perspective

While Chatman does not go into too much detail in his treatment of different dimensions of point of view, he rightly emphasizes that the term
point of view can refer not only to the position from which events are
perceived (which he calls the perceptual point of view), but also to the
position, from which events are evaluated (which he calls the conceptual
point of view). The idea that a characters world view (ideology, conceptual system, Weltanschauung, etc.) (Chatman 1978: 151) should be
conceptualized as a dimension of point of view can also be found in Uspenskijs seminal work A Poetics of Composition. Uspenskij claims that
one of the most basic aspects of point of view is manifested on the level
we may designate as ideological or evaluative (understanding by evaluation a general system of viewing the world conceptually) (Uspenskij
1973: 8). While this paper cannot hope to discuss exhaustively the question of how the events and situations in a computer game are evaluated by
the avatar and the other characters in the game (or even the game as a
system of rules), these questions are nevertheless of central importance
for the analysis of perspective in computer games. In order to distinguish
these evaluative positions from the notions of point of view and point of
action already discussed, we will refer to them as points of evaluation.
However, ideological perspective as determined by a point of evaluation
is not as easily identified in the analysis of computer games as is the case
with the dimensions of perspective in computer games already discussed.
According to Ryan, the observation that events in fictional worlds are
connected to certain goals, plans and psychological motivations, which
can be ascribed to the characters populating such worlds also applies to
computer games (cf. Ryan 2001). The fact that the player can ascribe a
specific world view to the characters in a computer game does not necessarily lead to a more compelling story, but does function as a means of
orientation for the player. The different points of evaluation and ideologyical perspectives of the characters in a computer game result in a certain
system of norms and values in which the player has to position him- or
herself. Smith notes that, for an understanding of films, it is important to
consider, first, how such systems of value are constructed; secondly, the
292
Jan-Nol Thon
range of possible types of moral structure; and thirdly, the different ways
in which a narration may unfurl these moral structures over time (Smith
1995: 189). This is also true for computer games. However, due to the
limited scope of this paper and the fact that most systems of norms and
values in computer games tend to be rather simple, we will mainly discuss
the first question, which is how these systems are constructed with regard
to the points of evaluation that can be ascribed to the various characters.
Ansgar Nnning has treated the notion of perspective within the
framework of possible worlds theory, emphasizing that it is applicable
not only to the rhetorical structure of narrative transmission, but also to
the world-models of the fictional individuals that populate the represented universe projected in narrative texts (Nnning 2001: 207). Hence, we
can describe the point of evaluation of a character in a computer game as
being determined by the characters model of the fictional world. But how
can a player ascribe a certain world view to the characters in a game?
Nnning emphasizes that in narrative texts each verbal utterance and
each physical or mental act of a character provides insights into his or her
perspective (Nnning 2001: 210). Once again this is true for computer
games. A computer games fictional world and its characters are conveyed not only through the presentation of the actual game spaces (to
which the previously discussed dimensions of perspective in computer
games mainly refer), but also through a variety of narrative techniques.
While most of the information about mental acts of characters in a computer game will be conveyed through cut-scenes and other forms of narrative techniques, the main part of physical acts will be presented in the
form of ludic instead of narrative events 6 . Therefore, in order to determine the point of evaluation of a computer game character, one has to examine the narrative as well as the ludic elements of the game.
For the purpose of the present paper, however, the actual form of these
narrative elements is less important than the function that they have for
the rest of the game, i.e. the game space and the ludic events. Narrative
events in computer games not only constitute a story and contribute to the
construction of the fictional world, but they also convey information
about the ludic structure of the game. Rune Klevjer even claims that giv6
In computer games, one can distinguish between narrative events that are already
determined before the game is played and ludic events that are determined at the moment of playing. Due to spatial limitations, the present paper cannot discuss this distinction in any detail. See Thon (2006 and 2007) for a more detailed discussion of
these different kinds of events and the narrative techniques used in their presentation.
293
ing meaning and sensation to the actions when they are performed by the
computer and the player (Klevjer 2001: n. p.) is the main function of narrative elements in computer games. He distinguishes between three levels
on which this signification of ludic events takes place. Firstly, on the
most important level, narrative (as well as ludic) events introduce a certain evaluation of possible actions. In every shooter-themed game, be it
Tomb Raider or Halo, it is important for me [the player, J.-N.T] that the
objects I [the players avatar, J.-N.T] shoot are bad guys with guns
who fight back, and who can be killed (Klevjer 2001: n. p.). This is
not a question of ethics, but of effective action. The player of Halo has to
be able to distinguish between his opponents (the bad guys) and his allies. In order to be successful he should refrain from letting his or her avatar shoot the latter. Secondly, most games will use narrative techniques to
give the player some kind of motivation for performing the specific actions that the game requires (Klevjer 2001: n. p.). In Halo, the avatar is a
(super) soldier named Master Chief who, together with his human allies,
tries to save the universe from various aliens. Here, we have a more specific level of meaning than is constituted by the mere distinction between
opponents and allies. Thirdly, many games use a chronologically and
causally ordered chain of predetermined narrative events (which is, of
course, continuously interrupted by ludic events) to present a (possibly
non-linear but nevertheless consistent) story. This is, of course, relevant
with regard to Smiths question of how a narration may unfurl these
moral structures over time (Smith 1995: 189). One example of a story
that forces us to change our initial conception of the ideological perspective structure is Halo 2 (2004), where it becomes clear during the course
of the story that certain aliens are actually allies instead of opponents in
that they help the Master Chief to save the universe.
8 Ideological Perspective Structure and the Player

Unlike the point of view and the point of action, which can both generally
be determined without too much of a problem, one has to consider the
various points of evaluation of the different characters to arrive at an appropriate description of this most complex level of perspective in computer games. According to Nnning, the term perspective structure can
be defined as the general system formed by all the character-perspectives
and narrator-perspectives as well as by the patterns of relationships between them (Nnning 2001: 214). While the present paper can only
294
Jan-Nol Thon
sketchily show how a computer games ideological perspective structure

with its various points of evaluation is constructed and can be analyzed,
this structure does indeed play a central part in the presentation of ludic as
well as narrative events in most contemporary computer games. As we
have seen, the first step in the analysis of the ideological perspective
structure of a game aims to reconstruct the points of evaluation of the
characters in the fictional world of the game. Furthermore, the characters
are generally connected to each other, either in a relation of opposition or
similarity of the respective points of evaluation. But as the above quotation from Nnning suggests it is not enough to analyze the constellation
of the various characters in a computer game. Although one would have
difficulties finding a narrator perspective in most games 7 , it is nevertheless the case that an analysis of the ideological perspective structure of a
game should also consider the choices that the player is allowed to make
with regard to his or her actions and the norms and values that are implied
by the game itself.
There is obviously a certain relationship between how the avatar evaluates the various events and situations in a game and how the player evaluates them. However, this does not mean that the player uncritically assumes the avatars position towards these situations and events. Rather,
the player will use the ideological perspective structure of a game to orient him- or herself within its ludic (as well as narrative) structure. This
also explains why the player of Halo will normally act according to the
avatars point of evaluation, and not try to befriend the aliens (which is, as
was previously mentioned, different in Halo 2). The player acts according
to the avatars point of evaluation since such action is in compliance with
the aims of the game. The game itself does not allow the player to choose
his allies freely or to decide that shooting aliens is not an action to be
evaluated positively. While the player may decide not to make his or her
avatar shoot aliens, this will most likely result in the death of the said
avatar and the player losing the game. However, we have already
mentioned that events and situations in computer games are not only evaluated on the level of character. In many contemporary computer games,
one can distinguish between the points of evaluation of the various characters in the game, the point of evaluation that the game constructs for the
7
There are certain games that use character narrators for their (at least partially linguistic) narration. Here, the notion of narrators perspective may be useful. It has,
however, to be emphasized that neither the player nor the avatar are narrators.
295
player and the point of evaluation that can be inferred from the overall
design of the game.
The relevance of a characters point of evaluation for the whole game
becomes most obvious in games with a single avatar. The avatars model
of the fictional world determines to a great extent the ways in which the
player can interact with the game world. Lara Croft, the avatar in Tomb
Raider, seems to have no doubt about the appropriateness of shooting the
various animals, humans and demons that act as her opponents throughout
the game. The game would be entirely different if Lara was a female
Hamlet, considering and re-considering the commands given by the player before finally deciding to act. It is clear that the player of Tomb Raider
is not entirely free in his or her decisions. Lara cannot be made to join the
bad guys (the main bad guy being a woman in Tomb Raider) in their
attempt at world domination. Another example previously mentioned
would be the avatar in the science-fiction-themed first-person shooter
Halo, who is presented as a soldier loyal to the human army. Here, the
player is not free to choose the alien alliance as an ally. It is true for most
contemporary computer games that many of the norms and values attributable to the avatar are not decided upon by the player. Although the
player has not much choice but to follow the avatars evaluation as far as
his (inter-)actions are concerned (since these evaluations generally define
the goals of the game), this does not necessarily mean that the player is
embracing these evaluation in any other way than with regard to the ludic
structure. The fact that a player of Tomb Raider makes the avatar of the
game shoot wolves does not imply that this player generally believes
shooting wolves to be a good thing. Indeed, it does not even necessarily
imply that the player believes that the fact that Lara Croft is shooting
wolves in the fictional world of Tomb Raider is a good thing. It is simply
a part of the game rules that Lara has to shoot wolves in order to survive.
While most computer games operate with clear-cut polarities of good
and evil, this does not mean that the player never has a choice between
the two. In games such as Fable or Jade Empire, the player can choose
which course of action to evaluate as the right one. Even in these
games, the possibilities for choice are strictly limited by the program, but
the player at least partly decides on the avatars norms and values. Another example where the player can influence the avatars point of evaluation is World of Warcraft. Here, the player gets to choose whether his
avatar is a member of the Alliance or the Horde. The players choice will
strongly influence the point of evaluation of his or her avatar, since the
296
Jan-Nol Thon
two parties are constantly at war with one another. In these cases, the
point of evaluation of the player influences how the avatar evaluates the
events in the game and what course of actions it then holds to be the
right one. However, it has again to be emphasized that what we propose
to call the point of evaluation of the player does not refer to the players
model of the actual world. Instead, it refers to the players model of the
fictional game world and his or her evaluation of the events and situations
that occur in it 8 . While some games allow their players to influence the
point of evaluation of his or her avatar, one should also keep in mind that
the choices a player can make in these games are generally choices
between narrowly defined alternatives.
We have seen that the player of a game using an avatar usually assumes that avatars point of evaluation in order to orient him- or herself
within the ludic structure of the game. This process of orientation, which
is necessary to play a game successfully, is also influenced by those
norms and values that are not directly connected to characters (be it the
players avatar or other characters) but can be attributed to the game designer(s). For the purpose of this paper, it is not relevant whether the
game designers really subscribed to these norms and values or had any intention to have them ascribed to them. If, for example, no children appear
in most parts of the game world in Fable, this is a conscious design decision that was intended to prevent the players from letting their avatars
kill children without obviously restricting their possibilities for interacttion with the game world. But, whether there was a conscious design decision behind it or not, the fact that no children can be killed may be read
as part of a system of norms and values that includes the norm that it is
not acceptable to have children killed, even in the fictional world of a
computer game. Another example is that Lara Croft can carry a variety of
weapons and kill an impressive number of various beasts in Tomb Raider
without getting problems with the authorities (or animal rights organizations). The point to be made here is that a particular ideological perspective manifests itself in the overall design and presentation of a game
8
See also Smiths discussion of allegiance. Smith assumes that something like a suspension of values must occur, if we are to explain the spectator aroused by a gangster
film, against her better (i.e. everyday) judgement (Smith 1995: 189). Although such
a suspension of values in computer games will most likely focus on the necessity to act
in compliance with the ludic structure of the game, it nevertheless occurs. See also
Schirra & Carl-McGrath (2002) on how the process of identification with characters in
computer games differs from the process of identification with characters in film.
297
world as well as in the rules and goals of the game. Here, one can speak
of the point of evaluation of an implied game designer.
A reconstruction of the system of norms and values inherent in computer games might also contribute to one of the most controversial questions concerning this relatively new form of entertainment, namely how
their often violent and politically incorrect 9 content should be evaluated
from an ethical point of view. Buchanan and Ess claim that
this debate threatens to become paralyzed on the one hand by simple-minded [...]
characterizations of e-games and their impacts, and, on the other hand, by overly
simple ethical analyses that would force us to choose between Manichean polarities of
absolute evil vs. absolute good. (Buchanan & Ess 2005: 3)
Without intending to further discuss this question here, it seems likely

that an (ethical) evaluation of the events and situations in a computer
game would benefit from considering how these events are evaluated
within the game itself. Sicart claims that players act as moral beings, that
they reflect upon those values that are contained in the system of the
game, and that they evaluate them keeping in perspective the values of the
game world (Sicart 2005: 17), but before discussing these questions, one
should probably examine exactly how values [...] are contained in the
system of the game (17).
9 Conclusion
This paper has proposed a model of perspective in contemporary computer games consisting of three dimensions. It has become clear that the
presentation of the game space in computer games differs from the presentation of space in narrative films and literary narrative texts. While the
perspective of the audiovisual presentation of the game space in a computer game is generally determined by a relatively constant point of view,
most games allow the player to control the spatial perspective at least to a
certain degree. In fact, the most obvious difference between computer
games and narrative films or literary narrative texts is the possibility to
interact with the presented space, which makes it necessary to include in a
model of perspective in computer games the notion of an actional perspective as determined by the point of action in addition to the spatial
perspective as determined by the point of view.
9
See Jahn-Sudmann & Stockmann (2008).
298
Jan-Nol Thon
Although we could only sketch the last dimension of our model of perspective in computer games, it has become clear that the ideological perspective structure that is determined by various points of evaluation and
conveyed through narrative as well as ludic elements plays an important
role in the perspectivation of events and situations in contemporary computer games. There is still some conceptual and terminological work left
to do especially with regard to the ideological perspective structure. Nevertheless, we believe that the three dimensions of perspective described in
this paper allow an analysis of the most central ways in which the events
in computer games are perspectivated.
In conclusion, it can be stated that models of perspective developed for
literary texts and narrative films cannot be directly applied to computer
games. It has, however, also become clear that the concepts and terminology developed in literary and film narratology possess considerable
heuristic value for the analysis of different media, such as computer
games. When attempting to transfer theoretical concepts such as perspective to new domains, awareness of the specific characteristics of the respective medium is of central importance. Nevertheless, differences between media do not necessarily prevent such a transfer from being successful.
Games Cited
Doom. ID, 1993. (PC)
Fable. Lionhead / Microsoft, 2004. (Xbox)
Grand Theft Auto: San Andreas. Rockstar, 2005. (PC)
Halo. Bungie / Microsoft, 2001. (Xbox)
Halo 2. Bungie / Microsoft, 2004. (Xbox)
Jade Empire. Bioware / Microsoft, 2005. (Xbox)
Myst. Cyan Worlds / Brderbund, 1993. (PC)
SWAT 4. Irrational / Sierra, 2005. (PC)
Tomb Raider. Core / Eidos, 1996. (PC)
Warcraft III. Blizzard, 2002. (PC)
Warhammer 40.000: Dawn of War. Relic / THQ, 2004. (PC)
World of Warcraft. Blizzard, 2004. (PC)
Z. Bitmap Brothers / Renegade, 1996. (PC)
299
References
Branigan, Edward (1984). Point of View in the Cinema. A Theory of Narration and Subjectivity in Classical Film. Berlin: Mouton.
Buchanan, Elisabeth A. & Ess, Charles (2005). Introduction: The Ethics of E-Games.
International Review of Information Ethics 4, 27. <http://www.i-r-i-e.net/inhalt/
004/004_full.pdf> (31.3.2007).
Chatman, Seymour (1978). Story and Discourse. Narrative Structure in Fiction and Film.
Ithaca: Cornell UP.
Jahn-Sudmann, Andreas & Stockmann, Ralf (2008). Anti-PC-Games. Exploring Articulations of the Politically Incorrect in GTA San Andreas. A. Jahn-Sudmann &
R. Stockmann (eds). Computer Games as a Sociocultural Phenomenon. Games
without FrontiersWar without Tears. Houndmills: Palgrave, 15061.
Juul, Jesper (2005). Half-Real: Video Games Between Real Rules and Fictional Worlds.
Cambridge, MA: MIT Press.
Klevjer, Rune (2001). Computer Game Aesthetics and Media Studies. Paper presented
at the 15th Nordic Conference on Media and Communication Research, Reykjavik,
Aug. 2001. <http://uib.no/people/smkrk/docs/klevjerpaper_2001.htm> (31.3.2007).
Mitry, Jean (1998). The Aesthetics and Psychology of the Cinema. London: Athlone Press.
Neitzel, Britta (2002). Point of View and Point of Action. A Perspective on Perspective
in Computer Games. Paper presented at the The Challenge of Computer Games
Conference, Lodz, Oct. 2002. Manuscript kindly forwarded by the author.
Nnning, Ansgar (2001). On the Perspective Structure of Narrative Texts. Steps toward a
Constructivist Narratology. W. van Peer & S. Chatman (eds). New Perspectives on
Poole, Steven (2004). Trigger Happy. New York: Arcade Publishing.
Rumbke, Leif (2005) Pixel. Raumreprsentation im klassischen Computerspiel (Hausarbeit Kunsthochschule fr Medien, Cologne). <http://www.rumbke.de/data/text/
pixel3%20-%20leif%20rumbke%202005.pdf> (31.3.2007).
Ryan, Marie-Laure (2001). Narrative Beyond Myth and Metaphor. Game Studies 1:1,
n.p. <http://www.gamestudies.org/0101/ryan/> (31.3.2007).
Schirra, Jrg R. J. & Carl-McGrath, Stefan (2002). Identifikationsformen in Computerspiel und Spielfilm. M. Strbel (ed). Film und Krieg. Die Inszenierung von Politik
zwischen Apologetik und Apokalypse. Opladen: Leske+Budrich, 14963.
Sicart, Miguel (2005). Game, Player, Ethics: A Virtue Ethics Approach to Computer
Games. International Review of Information Ethics 4, 1318. <http://www.i-r-ie.net/inhalt/004/004_full.pdf> (31.3.2007).
Smith, Murray (1995). Engaging Characters. Oxford: Oxford UP.
Thon, Jan-Nol (2007). Schaupltze und Ereignisse. ber Erzhltechniken im Computerspiel des 21. Jahrhunderts. C. Mller & I. Scheidgen (eds). Mediale Ordnungen. Erzhlen, Archivieren, Beschreiben. Marburg: Schren, 4055.
Uspenskij, Boris (1973). A Poetics of Composition. The Structure of the Poetic Text and
Typology of a Compositional Form. Berkeley: U of California P.
Wolf, Mark J. P. (2001). Space in the Video Game. M. J. P. Wolf (ed). The Medium of
the Video Game. Austin: U of Texas P, 5176.
Authors
DAVID HERMAN, who co-founded the Project Narrative initiative at Ohio
State University (http://projectnarrative.osu.edu) and served as its inaugural director, teaches in OSUs English Department. He has authored,
edited, or co-edited eight books on aspects of narrative and narrative theory, and he also serves as editor of the Frontiers of Narrative book series
and of the new journal Storyworlds, both published by the University of
Nebraska Press. He was recently awarded a research fellowship from the
American Council of Learned Societies for his 2009 project on Storytelling and the Sciences of Mind.
CHRISTIAN HUCK is principal investigator of the research project Travelling Goods // Travelling Moods: A Transcultural Study of the Acculturation of Consumer Goods, 19181939. He took his PhD at Tbingen
University and received his Habilitation at the University of ErlangenNuremberg. The topics of his publications range from Irish poetry and
18th-century travel literature to rockumentaries and music videos. He is
currently preparing a monograph on Fashioning Society, or, The Mode of
Modernity: Observations of Clothing in Eighteenth-Century Britain.
HHN is a professor of English Literature, Hamburg University
(retired since 2005) and member of the Interdisciplinary Center for Narratology. He has published books and articles on theory of poetry and
history of British poetry, narratology, application of narratology to poetry
analysis, and detective and crime fiction. He is author of Geschichte der
englischen Lyrik (1995), co-author of Der Entwicklungsroman in Europa
und bersee (2001), Die europische Lyrik seit der Antike (2005), The
Narratological Analysis of Lyric Poetry (2005), Lyrik und Narratologie
(2007) and co-editor of the Living Handbook of Narratology (to appear in
2009).
PETER
TATJANA JESCH is working on a postdoctoral thesis about theory and empirical experience in the field of narratology and didactics at Jena Uni-
302
Authors
versity. She is the author of Das Subjekt in Mrchenraum und Mrchenzeit (1998), co-author of Texte lesen (2008), and editor of Mrchen in der
Geschichte und Gegenwart des Deutschunterrichts (2003). In addition,
she has published several theoretical and empirical articles on (psycho)narratology and on understanding and teaching literature.
TOM KUBEK, PhD, is a researcher in the Department of the History
of Literature at the Institute for Czech Literature, Czech Academy of Sciences. Until 2008 he was a researcher in the Section Narratology, which
he directed from 2002 to 2007. He also lectures on literary theory, narratology and literary history in the Department of Czech and Comparative
Literature and Literary Theory at Charles University, Prague. He has published studies on narratology, literary theory, Czech literary structuralism,
and Czech prose. He is the author of the books: Intersubjectivity in
Literary Narrative (2007); Vyprav. Kategorie narativn analzy (2007);
Vyprvt pbh. Naratologick kapitoly k romnm Milana Kundery
(2002). He was editor in chief of the book series Theoretica and is coeditor of Library of the Possible Worlds.
MARKUS KUHN, M.A., is a visiting lecturer at the Institute of Media and
Communication (IMK) at the University of Hamburg and has just finished
his PhD thesis on film-narratology. He studied German language and
literature, media and communication studies, history of arts and journalism in Gttingen and Hamburg. He works as a freelance journalist for
print and online media. His M.A. thesis on Narrative Situations in Literature and Film was awarded the Karl H. Ditze-Preis for outstanding
Masters theses.
URI MARGOLIN is a professor emeritus of Comparative Literature at the
University of Alberta, Edmonton, Canada. He has been working for many
years in the fields of narratology and general literary theory, and has published close to 70 articles in collective volumes and professional journals
in Europe and North America.
GUNTHER MARTENS is a postdoctoral fellow of the Flemish Research
Council (FWO), affiliated with the German Department at the University
of Ghent and Visiting Professor of Literary Theory at the Free University
of Brussels. Publications on literary modernism, literature and ethics, and
the relation between rhetoric and narratology. Activities: in 2005, re-
Authors
303
search stay at the Narratology Research Group Hamburg; in 2007 (jointly

with the Interdisciplinary Center for Narratology Hamburg) and 2008,
co-organizer of the International Narratology Workshops at Ghent
University; in 2008, co-organizer of the Inaugural Conference of the
European Network for Avant-Garde and Modernism Studies (EAM).
Most recent publications: E. Dhoker & G. Martens (eds): Narrative
Unreliability in the Twentieth-Century First-Person Novel (2008);
G. Martens & B. Biebuyck: On the narrative function of metonymy in
Heines Ideen. Das Buch Le Grand (Chapter XIV). Style 41:3 (2007).
JAN CHRISTOPH MEISTER has taught at the University of the Witwatersrand, Johannesburg, and at the University of Munich. He is currently
professor of German Literature, Literary Theory and Literary Computing
at the University of Hamburg. His publications include Computing Action. A Narratological Approach (2003). Among his major research topics
is the computational modelling of narrative structures and narrative competence in Story Generator Algorithms.
ALAIN RABATEL is a professor of Language Sciences at the University of
Lyon 1 (University Institute for Teacher Training). He specializes in discourse analysis, particularly in literary, media, religious and political text
and discourse. He is the author of several books including Une histoire du
point de vue (1997), La construction textuelle du point de vue (1998), and
Homo narrans (2008) as well as of some one hundred articles setting out
an enunciative-interactional narratological approach, working within a
dialogical and polyphonic framework.
BRIAN RICHARDSON is a professor in the English Department of the University of Maryland. He is the author of Unlikely Stories: Causality and
the Nature of Modern Narrative (1997) and Unnatural Voices: Extreme
Narration on Modern and Contemporary Fiction (2006), winner of the
Perkins Prize for the years best book on narrative studies. He has written
numerous articles on different aspects of narrative theory, including plot,
time, cause, closure, character, narration, reader response, reflexivity, and
the narratives of literary history. He is the editor of two books, Narrative
Dynamics: Essays on Time, Plot, Closure, and Frames (2002) and Narrative Beginnings: Theories and Practices (2008). He is currently completing a book on modernism, misreading, and the theory of the reader.
304
Authors
SABINE SCHLICKERS is a professor of Spanish and Latin American Literature at Bremen University, Germany. She is the author of Verfilmtes Erzhlen: Narratologisch-komparative Untersuchung zu El beso de la mujer
araa (Manuel Puig/Hctor Babenco) und Crnica de una muerte
anunciada (Gabriel Garca Mrquez/Francesco Rosi) (1997), El lado
oscuro de la modernizacin: Estudios sobre la novela naturalista hispanoamericana (2003) and, most recently, of Que yo tambin soy pueta. La
literatura gauchesca rioplatense y brasilea (siglos XIX-XX) (2007) as
well as of numerous articles on narratology, literature and film.
WOLF SCHMID is a professor of Slavic Literatures at the University of
Hamburg. He founded the Hamburg Narratology Research Group
(www.narrport.uni-hamburg.de) and is currently director of the Interdisciplinary Center for Narratology (www.icn.uni-hamburg.de) and executive editor of the series Narratologia. With his Hamburg colleagues,
he founded the European Narratology Network (www.narratology.net).
He has authored Elemente der Narratologie (Russian 2003, 2008; German
2005, 2008) and edited two collections on Slavic narratology.
JRG SCHNERT is a retired professor of Modern German Literature,
Hamburg University, and member of the Interdisciplinary Center for Narratology. He has published on the theory and practice of the social history
of literature (with an emphasis on structural and functional text-theoretical models), on the history of the humanities and on problems of literary
theory and methodology. Co-author of Lyrik und Narratologie (2007) and
co-editor of the Living Handbook of Narratology (to appear in 2009).
VIOLETA SOTIROVA is a lecturer in stylistics at the University of Nottingham. She has published articles on narrative point of view in the journal
of the Poetics and Linguistics Association, Language and Literature
(Connectives in free indirect style: continuity or shift?, 2004, for which
she was awarded the PALA prize for best first publication), in Style
(Repetition in free indirect style: a dialogue of minds, 2006) and in
Poetics (Reader responses to narrative point of view, 2006). Her
publications also include articles in English Studies (2006), tudes Lawrenciennes (2007; 2008) and a chapter in Contemporary Stylistics (2007).
She is currently working on a monograph on consciousness presentation
in modernist fiction.
Authors
305
MALTE STEIN, former member of the Narratology Research Group, Hamburg University, is a teacher at the Hansa-Kolleg in Hamburg, Germany.
He wrote his dissertation on family violence in the novellas of Theodor
Storm (2006), is co-author of Lyrik und Narratologie (2007) and has published several articles on the intersection of literature, narratology and
psychoanalysis.
JAN-NOL THON is a PhD student at Hamburg University, working on a
project in the field of transmedial narratology. He has authored several
conference papers, articles, and book chapters on the theory and aesthetics
of contemporary computer games, focusing mainly on space, interaction,
simulation, narration, communication and immersion.
ROLAND WEIDLE presently holds a position as substitute professor for
English Literature at the University of Hamburg. He has published on
Shakespeare, drama and theatre from the early modern age to the present,
contemporary fiction and transmedial and transgeneric narratology. Most
recently, he co-edited a volume on contemporary British literature (Cool
Britannia, 2006) and was the focus editor of the issue Transmedial and
Transgeneric Narration for the journal Anglistik. International Journal of
English Studies (2007).

Point of View, Perspective and Focalization

Uploaded by

Copyright:

Available Formats

Point of View, Perspective and Focalization

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Point of View, Perspective and Focalization

Uploaded by

Copyright:

Available Formats

Point of View, Perspective, and Focalization

Jose Angel Garca Landa, Peter Hhn, Manfred Jahn

Walter de Gruyter Berlin New York

Point of View, Perspective,

Walter de Gruyter Berlin New York

Printed on acid-free paper which falls within the guidelines

of the ANSI to ensure permanence and durability.

Library of Congress Cataloging-in-Publication Data

Part I: Re-Specifications of Perspective

TATJANA JESCH, MALTE STEIN

Part II: Some Special Aspects of Mediation

Part III: Transliterary Aspects of Mediation

character(s). The underlying premise is that the rigorous distinction

perspectivization is ultimately governed by the author, it can be used by

within a directional grid); degree of objectivity vs. subjectivity. This

reflex of the dominant tradition of privileging vision in Western culture.

focalization as they also occur in literary texts, although more restricted in

Part I: Re-Specifications of Perspective

JAN CHRISTOPH MEISTER, JRG SCHNERT

The DNS of Mediacy

Jan Christoph Meister, Jrg Schnert

differentiated analysis. The most advanced approach in this regard is that

The DNS of Mediacy

in paying due attention to the cognitive logic of narrative processing, has

Jan Christoph Meister, Jrg Schnert

The DNS of Mediacy

found in Genettes descriptive apparatus5. Some would indeed warrant at

Jan Christoph Meister, Jrg Schnert

The DNS of Mediacy

external vision (Auen[an]sicht) onto n characters

introspection of n characters (full introspection)

narratorial, partially restricted perception (co-vision /

external vision onto the p- introspection of x (partial

actorial, partially restricted

actorial, restricted perception

external vision onto the pset of characters

1.2 Before Terminology

Jan Christoph Meister, Jrg Schnert

sual input significantly overshadows the mental activities by which we

The DNS of Mediacy

2 The Model and its Terminology

Jan Christoph Meister, Jrg Schnert

3 The DNS (Dynamic Narrative System)

The DNS of Mediacy

Jan Christoph Meister, Jrg Schnert

The DNS of Mediacy

Our term mediation refers to the process dimension of narrative representation,

Jan Christoph Meister, Jrg Schnert

What these dimensional parameters are, which of them are compulsory

Figure 1: DNS (Dynamic Narrative System) Model of Representation

The DNS of Mediacy

In a systems perspective, setting or changing one particular parameter is

Jan Christoph Meister, Jrg Schnert

ontological constraints under which the representational system operates

Figure 2: Dimension of Perception

In the DNS model we define the dimension of perception as one which is

The DNS of Mediacy

Jan Christoph Meister, Jrg Schnert

Figure 3: Dimension of Reflection

What is contributed to representation in this dimension is a sort of mental