Textlinguistic and Biblical Hebrew Dawson PDF

JOURNAL FOR THE STUDY OF THE OLD TESTAMENT
SUPPLEMENT SERIES
177
Editors
David J.A. Clines
Philip R. Davies
Executive Editor
John Jarick
Editorial Board
Richard J. Coggins, Alan Cooper, Tamara C. Eskenazi,
J. Cheryl Exum, John Goldingay, Robert P. Gordon,
Norman K. Gottwald, Andrew D.H. Mayes, Carol Meyers,
Patrick D. Miller
Sheffield Academic Press
This page intentionally left blank
Text-Linguistics
and Biblical
Hebrew
David Allan Dawson
Journal for the Study of the Old Testament

Supplement Series 177
Copyright 1994 Sheffield Academic Press

Published by
Sheffield Academic Press Ltd
Mansion House
19KingfieldRoad
Sheffield SI 19AS
England
Typeset by Sheffield Academic Press

and
Printed on acid-free paper in Great Britain
by Bookcraft
Midsomer Norton, Somerset
British Library Cataloguing in Publication Data

A catalogue record for this book is available
from the British Library
ISBN 1-85075-490-X
CONTENTS
Preface
Abbreviations
7
9
Chapter 1
INTRODUCTION, AND EXAMINATION OF SEVERAL

RECENT PUBLICATIONS
11
Chapter 2
EXAMINATION OF KHAN's 'STUDIES' AND LONGACRE's 'JOSEPH'
52
Chapter 3
AN INTRODUCTION TO ONE MODEL OF TEXT ANALYSIS,

AND A METHODOLOGY DERIVED FROM rr
70
Chapter 4
TEXT-LINGUISTIC OBSERVATIONS ON NARRATIVE

AND NON-NARRATIVE TEXT-TYPES
123
Chapter 5
JEPHTHAH AND RUTHREFINING AND TESTING

THE HYPOTHESES
154
Chapter 6
SUMMARY, CONCLUSIONS AND IMPLICATIONS
209
Appendix 1
THE TEXT OF JUDGES 2 (BHS) IN COLUMNAR FORMAT
220
Appendix 2
THE TEXTHE OF RUTH (BHS) IN COLUMN
223
Bibliography
Index of Authors
237
242
PREFACE
The bulk of this work originated in my doctoral dissertation (under
Professor John Gibson at Edinburgh University, Scotland, 1993),
having developed out of a marriage between my delight in textlinguistics and my love of Classical Hebrew. It quickly became
apparent that if others had discovered the same loves, few of them had
offered anything in print, fewer still had written for Hebrew scholars,
and an exceptionally limited number had written books which could
be employed by the hebraist without substantial training in modern
linguistics.
My original goal was to produce a text-linguistic commentary on a
particular text. Ruth, Jonah and the Joseph stories all came to mind,
yet no sooner had I started on this project, than Dr Robert Longacre
published his Joseph: A Story of Divine Providence: A TextTheoretical and Text-Linguistic analysis of Genesis 37 and 39-48
(Eisenbrauns, 1989). I had worked under Dr Longacre during
linguistic training for work with Wycliffe Bible Translators several
years before, and his work had always been the root of my own. I was
immensely pleased to see this material come to publication. Dr
Longacre's method is difficult to grasp, however, and difficult to put
to practical use, for its starting point is well beyond the average
hebraist's linguistic facility. As I worked with this book, and several
others, I decided that I would try to put together a book which would
form a bridge between the ratified works of the ultra-trained linguists
and the minimally trained (in linguistics, that is) Hebrew scholars.
Thus, this book is intended for those with an interest, but little
background, in modern text-linguistics. This is not to say that I have
avoided issues which generally require more finely tuned training, but
rather that I have sought, by use of explanations and examples, to
grant to the average scholar access to all the topics covered in this
volume.
I am aware that this book cannot hope to solve confusions of
Text-Linguistics and Biblical Hebrew
terminology and theory. But its principal focus is on methodology,

and issues of terminology and theory are part of methodology. I hope
to contribute to clearer thinking about these things, as I have a strong
interest in clarity and 'elegance' (a term which will recur with some
regularity in this volume, and which will, in due course, be defined).
Much has been left out; but this is intended as a starting point. The
starting point seems to have been omitted from this discipline, and
there is little communication between those whose training in linguistics
exceeds the norm and those whose knowledge of the Hebrew Bible is
beyond compare. I have sought merely to establish a little more
solidity in the link between them.
As is always the case in such works, many deserve recognition for
their role in the process of getting it 'from there to here'. A great deal
of help on this work has come from Professor John C.L. Gibson,
whose scholarly integrity and the simple human concern which
characterizes his daily life are standards which I hope to achieve in
my own life. And no amount of words could express my gratitude to
my parents and best friends, Frank and Ruth Dawson, for their
encouragement, love and support through the years.
ABBREVIATIONS
AB
AnBib
BibSac
BKAT
FOTL
JQR
JQRMS
JSOT
JSOTSup
Or
OTL
SANT
TynBul
VT
WBC
ZAH
ZA W
Anchor Bible
Analecta biblica
Bibliotheca Sacra
Biblischer Kommentar: Altes Testament
The Forms of the Old Testament Literature
Jewish Quarterly Review
Jewish Quarterly Review Monograph Series
Journal for the Study of the Old Testament
Journal for the Study of the Old Testament Supplement Series
Orientalia
Old Testament Library
Studien zum Alten und Neuen Testament
Tyndale Bulletin
Vetus Testamentun
Word Biblical Commentary
ZeitschriftfurAlthebraistik
Zeitschrift fur die alttestamentliche Wissenschaft
Chapter 1
INTRODUCTION AND EXAMINATION
OF SEVERAL RECENT PUBLICATIONS
1.1 Introduction to the Content and Goals of this Volume

In this work, I will examine several currently available works on
Classical Hebrew,1 explain a particular theory of linguistics and its
derived approach to the data, examine a series of texts, and draw
conclusions from all the above. The lynch-pin holding all this together
is the question of the role(s) of different types of clauses in different
types of texts.2
My goal is to measure some of the current work on this question
against two basic criteria: (1) integrity of description, and
(2) effectiveness of communication. In a very real sense, the author of
a study offers him- or herself (or rather, the work) as a bridge
between the topic studied and the reader; if the connection between the
author/work and either the topic or the reader is inadequate in some
way, then the author's purpose will not be accomplished in a
satisfactory way. So, if a researcher approaches data with a faulty
theoretical base, or faulty methodological procedures, then that end of
1. I use this term to refer to Hebrew of the biblical period (extending to roughly
200 BCE); it is more or less synonymous with the term Biblical Hebrew, but allows
for the inclusion of non-biblical materials in the corpus; although these are not
relevant for the present study, I adopt the term on principle: we are dealing with a
language per se, not simply a set of features of some strictly limited corpus of data.
2. Jones, in 'A Synopsis of Tagmemics', commenting on Longacre's
contribution of this type of enquiry, has written, 'Longacre has studied the effect of
discourse types on units both large and small in texts. "In effect once a discourse
type is chosen, many decisions as to structure of very small parts of it are already
made [1972, p. 133]"' [p. 94]; this, in brief, is the justification for my interest in this
approach: I seek to describe the motivating factors at the macro-syntactic level for the
employment of micro-syntactic features.
12
the bridge is insecure; likewise, if one's communication of truly valid

insights about the data is imprecise, or impenetrable (perhaps owing
to use of unfamiliar and poorly defined vocabulary), then the other
end of the bridge is insecure. Both must be solid if the work is to
advance the study of the language. With these two concerns in mind, I
will survey five contemporary works, and will then turn my attention
to samples from the data.
Before much can be said about the data, however, my own approach
to ittheory, methodology and working hypotheseswill require
elucidation. In Chapter 3, therefore, I will present definitions, relate
concepts and justify procedures, which will come into play in my
examination of the data. This will be followed by examination of six
extended sections of text from the Hebrew Bible,3 culminating in an
integrated reading of the book of Ruth, where I will examine the text
in light of hypotheses developed out of the previous examinations of
data. The final chapter will summarize the material contained in the
previous five, and will look forward toward further application of this
kind of research. The layout is as follows:
Chapter 1 Introduction and objectives, and examination of the first three
works (Niccacci, Syntax, Eskhult, Studies, and Andersen,
Sentence)
Chapter 2 Examination of Khan, Studies, and Longacre Joseph
Chapter 3 Presentation of theoretical and methodological approach to the data
Chapter 4 Presentation of worked material from Judges 2, Leviticus 14, and
6-7, and the 'Tabernacle texts' (sections from Exodus 25-40)
Chapters Presentation of Judg. 10.6-12.7 (with particular interest in
Reported Speech sections), and the whole text of Ruth, integrating
the hypotheses examined
Chapter 6 Summary, practical applications, and review of objectives
Chapter 1 introduces the present research project, and will provide a

summary of some of the major recent contributions in the field of
Hebrew text-linguistics; a brief introduction to the principal theoretical questions involved in this study will be followed by an in-depth
3. Judg. 2; Lev. 14.1-32; Lev. 6.1 [Heb.]-7.38; sections of Exod. 25-40 (these
four data-samples are dealt with in Chapter 4]; Judg. 10.6-12.7; and the whole of the
book of Ruth (these are found in Chapter 5).
1. Introduction
13
analysis of one of the currently most influential works in this field,4

and of two other significant works.5
The principal theoretical elements that I will explain in this chapter
are these:
the necessity of theoretical and methodological integrity and
clarity;
the primacy of the data;
language universals, and their significance for Hebrew studies;
text-level structures as describable features of language;
bag-of-tricks languages, and, with such languages, the necessity
of studying all possible features that perform the same
function, in order to describe accurately the function of
any one of these features;
'text', providing a basic definition for this term as I use it, and
'discourse', and why I avoid using it;
text-type, and its influence on the distribution of clause-types;
and, finally the interrelationship between micro-syntactic and
macro-syntactic levels of language.
This theoretical section is followed by reviews of works by Niccacci,
Eskhult and Andersen.
The following quotations from David Crystal's Dictionary of
Linguistics and Phonetics will illustrate some of the difficulties we
will face in Hebrew text-linguistics:
One sign of immaturity [in a science] is the endless flow of terminology.
The critical reader begins to wonder if some strange naming taboo attaches
itself to the terms that a linguist uses, whereby, when he dies they must be
buried with him.6
DISCOURSE A term used in linguistics to refer to a continuous stretch of
(especially spoken) language larger than a sentencebut, within this
4. A. Niccacci, Syntax of the Verb in Biblical Hebrew Prose (trans.
W.G.E. Watson; JSOTSup, 86; Sheffield: JSOT Press, 1990).
5. M. Eskhult, Studies in Verbal Aspect and Narrative Technique in Biblical
Hebrew Prose (Acta Universitatis Upsaliensis: Studia Semitica Upsaliensia, 12;
Uppsala: Uppsala University, 1990), and F.I. Andersen, The Sentence in Biblical
Hebrew (The Hague: Mouton, 1974).
6. D. Bolinger, Aspects of Language (with D.A. Sears; Harcourt Brace
Jovanovich, 3rd edn, 1981), p. 554; quoted by D. Crystal, Dictionary (The
Language Library; Oxford: Basil Blackwell, 1985), p.v.
14

broad notion, several different applications may be found. At its most
general, a discourse is a behavioural unit which has a pre-theoretical status
in linguistics: it is a set of utterances which constitute any recognizable
speech event (no reference being made to its linguistic structuring, if any),
e.g. a conversation, a joke, a sermon, an interview. A classification of
discourse functions, with particular reference to type of subject-matter, the
situation, and the behaviour of the speaker, is often carried out in sociolinguistic studies (of primitive societies, in particular), e.g. distinguishing
dialogues v. monologues, or (more specifically) oratory, ritual, insults,
narrative, and so on. In recent years, several linguists have attempted to
discover linguistic regularities in discourses (discourse analysis or DA),
using grammatical, phonological and semantic criteria (e.g. cohesion,
anaphora, inter-sentence connectivity). It is now plain that there exist
important linguistic dependencies between sentences, but it is less clear
how far these dependencies are sufficiently systematic to enable linguistic
units higher than the sentence to be established.7 The methodology and
theoretical orientation of discourse analysis (with its emphasis on wellformedness and rules governing the sequence of permissible units, in both
spoken and written texts) are often contrasted with those of conversation
analysis. Some linguists adopt a broader, psycholinguistic perspective in
studying discourse, which they view as a dynamic process of expression
and conmprehension governing the performance of people during linguistic interaction. Some adopt a sociolinguistic perspective, in which the
purpose or function of the discourse is emphasised. These emphases
distance the subject from 'text linguistics', when this is seen as the formal
account of the linguistic principles governing the structure of texts. But
there is considerable overlap between the domains of discourse analysis
and text linguistics (for example, the notion of cohesion is prominent in
both), and any attempt at a principled distinction would be premature.8
We may read these quotations with some amusement, for the first
draws into focus a problem that affects the term 'discourse' in
particular, with acerbic clarity, as can be seen by the definitionfs] that
Crystal offers. When we get into the realm of linguisticsand in
particular, text-linguisticsthe humour of these statements begins to
7. These 'linguistic dependencies' are the focus of 'text-linguistic' analyses; that

these are 'sufficiently systematic to enable linguistic units higher than the sentence to
be established' is indeed a matter for some debate, yet is also one of the
presuppositions that undergirds the text-linguistic undertaking as a wholeand it is
one that is being amply substantiated by data from the widest possible array of the
world's languages, both living and dead.
8. Crystal, Dictionary, p. 96.
1. Introduction
15
darken, for linguistics generates new terminology at a mind-boggling

rate.
There are at least three things to which a reader should find
accessin addition to the results of the work undertakenin a
research publication: the author's presuppositions, theoretical perspective and methodology. This is because the only means a reader has
of assessing the trustworthiness and applicability of a researcher's
conclusions is by evaluating them through that researcher's approach
to the data. The problem, however, is that many researchers either
have not learned the technique of self-analysis with regard to these
concerns, or consider them of too little importance to be included in
published results.
Presuppositions must be uncovered, examined, and put to the most
difficult challenges throughout a researcher's work with the data, for
these, if left unchallenged, can govern one's work to the point that it
produces nothing of value. They must either (1) be recognized as
false, and rejected, (2) be acknowledged as unproven, hypothetical,
and therefore remaining suspect, or (3) be proved to be in concordance with the data, at which point they cease to be presuppositions,
and become part of the researcher's theoretical base.
That theoretical base (which is, in effect, a summary of, and a set of
generalizations from, data already processed) in turn gives the
researcher meaningful ways of interacting with the data; that is to say,
the theoretical base suggests a methodology. This methodologyor
the procedures one will implement in analysing the datawill in its
turn assist in organizing the data in a way that leads to solid scientific
conclusions (even where these conclusions be that the processing of
data according to one's methodology has led to no solid conclusions).
These conclusions are then compared with the theory's projections.
Where the two are in accord, the theory is strengthened; where they
are not, both must be re-examined in order that fault be found in
either the methodology or the theoretical base. It must then be
determined whether the inadequate approach should be altered to
make room for the new results, or whether it should be jettisoned
altogether in favour of another, more workable one. Any less
rigorous approach to these concerns is the less objective and the less
trustworthy.
16
In the long run, all but the data must be considered suspect.9 This is
true particularly when dealing with a restricted corpus of language
material, since there are insufficient data to substantiate hypotheses
fully. Only where the data appear to be in contradiction with all other
possible explanations can it be permitted that the data be called into
question. It is rare, however, that scholars go to these lengths, and
many settle for 'improving' the text as a path of lesser resistance.
With regard to this tendency, Niccacci writes,
It is, in any case, a duty to presume that even the various kinds of
'glosses' or inserts also follow the rules of grammar and syntax. I think it
injudicious to adopt the principle which unfortunately so many scholars
follow that so-called 'difficulties' or 'mistakes' of grammar and syntax are
indications of later reworking. In effect this would mean that the writers
of such glosses either did not know the language or at the least were inept.
I wish to reiterate here a caution against the danger of making syntax as
arbitrary as literary criticism... I prefer to follow this method closely
rather than 'correct' the texts using 'rules' even if difficult cases remain
which require further study... !
I present here, in this chapter and in the following, several relatively
recent works, which have to do with the assessment of syntactic
features in prose. I will sketch my own theoretical and methodological
bases more fully in Chapter 3; for the moment, however, the following
will serve to introduce these features sufficiently to provide the reader
with a starting point for understanding my assessment of these works.
9. I acknowledge that our corpus may contain errors of transmission, and that
this places the analyst in a different situation from one working in a living language,
where the data can be reconfirmed when doubt exists about their accuracy. This
possibility is not open to those working on Classical Hebrew, for example. However, cautious textual criticism and emendation is not what is at stake in the above
statementrather, it is that too many people engaged in analysis of this language
come to it with inflexible theories and/or ideologies, which they are unwilling to reexamine in light of the data. Radical restructuring of the text is, for some, only a
starting point in their protection of theory or ideology: difficulties in the text lead to
rewriting the text. I will overstate this point at several stages in this volume. I
recognize that the current trend in biblical scholarship is away from radical rewriting
of the text, and in the direction of the position I hold. Yet much of what is available
to the researcher and student, originating as it does from a different perspective,
affirms this more invasive approach to the data. I emphasize the issue in this work to
provoke the reader to conscious examination of his or her own approach to the data.
10. Niccacci, Syntax, p. 13.
1. Intro
17
It will be clear from the comments above, and to follow, that I

consider it of prime importance to import as few changes into the data
as possible, and to regard the data as the anchoring point of the
research process. Nonetheless, there is a rather small amount of data
available to us in this language (by contemporary linguistic standards),
and we are somewhat at a disadvantage. Modern descriptive
linguistics,11 however, has discovered that, despite the great number
of different languages in the world (over 5000 at latest count, not
including languages of the antiquities), and despite the enormous
diversity exhibited among these languages, there is a remarkable
degree of consistency in language features; that is to say, the world's
languages demonstrate a limited number of possible variants.
To take a rather mundane example, one might expect, given the
three main constituents12 of a transitive clause (Subject, Object, and
Verb), that we should be able to find the following combinations:
S-V-O, S-O-V, V-S-O, V-O-S, O-S-V and O-V-S. In fact, only the
first three combinations occur with any frequency as the standard
word order for transitive clauses. The latter three are extremely
uncommon, the last being virtually unheard of.13 Thus when we look
at a new language, we can start from a certain confidence that O-S-V
is not a strong option for the normative word order of a transitive
clause, and that any occurrences of the O-S-V clause-type should be
examined closely, since they are highly unusual. This particular
example is not of a great significance for the present topic, but a more
relevant one is at hand.
One of the 'discoveries'rather I should say 'new emphases'of
contemporary linguistics has been the analysis of large structures in
texts. Thus, just as one can describe the constituent structure of a
transitive clause as S-V-O (where each constituent in this formula
represents a set of possible 'fillers' in the formula, for example, the
11. Please note that I make no attempt at precision in the use of the term
'descriptive'; it has been used in the past as a label for certain brands of 'structural'
linguistic theories, which is not my intent here. Here, I merely mean linguistics
which has as its principal concern the empirical analysis and description of the
features of the data.
12. 'CONSTITUENT (CONSTITUENCY) A basic term in grammatical analysis
for a linguistic unit which is a component of a larger construction' (Crystal,
Dictionary, p. 68); I will return to this topic in Chapter 3.
13. On this topic, see, for example, Khan, Studies, p. 225, and his notes.
18
' V represents a collection of verbsbut not just any verb: these must
be transitive verbs, and so on), so one can also describe a story as a
sequence of constituent units, each constituent having its proper place,
and each section filled with its proper sort of 'filler'. This has allowed
scholars to observe that certain features occur regularly in the world's
languages at these 'larger' levels; for example, it is common, if not
universal, for languages to mark the most significant event or events
in a story, so that this material stands out from the rest of the story.
Likewise, distinctions are normally made between background information and foreground information (regardless of the basic text-type
in question), and so on.
'Language universals', as these general tendencies of human
language have been dubbed,14 give language workers the 'basic
starting point' kind of information that a traveller would hope to find
in a good guide-booknot the specificities like 'If you have time to
kill, and wish to see a humorous spectacle, watch the changing of the
guard in front of the Parliament building on Syntagma Sq. Every
hour on the hour two sets of enormously tall evzones (guards) slowly
wind up like toy soldiers, kick their heels about, and fall backwards
into symmetrical little guard houses on either side of the Tomb of the
Unknown Warrior...',15 but rather general facts like 'Pharmacies:
Indicated by a Byzantine-style red cross. Open Sat.-Sun. on a rotating
14. 'In their broadest sense, then, LANGUAGE UNIVERSALS are equivalent to
the general design features of human language... FORMAL UNIVERSALS are the
necessary conditions which have to be imposed on the construction of grammars in
order for them to be able to operate. They include such notions as the number of
components, types of rules, ordering conventions (e.g. cycles), types of transformations, and so on. SUBSTANTIVE UNIVERSALS, on the other hand, are the
primitive elements in a grammar, required for the analysis of linguistic data, e.g. NP,
VP, [+ grave], [+ abstract]... Some of these categories may actually be found in
every language, but it is not crucial to the notion of substantive universal that they
should be. All that is required is that they be constructs which need to be denned by
linguistic theory to enable cross-language generalisations to be made, i.e. they are not
terms established for the analysis of just one language, but are capable of general
application... ABSOLUTE UNIVERSALS are properties which all languages share;
there are no exceptions. RELATIVE UNIVERSALS are general tendencies in
language; there may be principled exceptions' (Crystal, Dictionary, p. 321,
emphasis added).
15. L. Shang-Huei Chao (ed.), Let's Go: The Budget Guide to Greece & Turkey
1992 (London: Pan Books 1992), p. 87.
1. Introduction
19
basis; signs are posted in the windows of each to indicate the nearest
open pharmacy. You can also find out by phoning 166.'16 Now, the
material we are dealing with, of course, does not usually generate the
kind of urgency exhibited in my analogy, but the researcher will find
that a list of 'most likely language features' derived from language
studies around the world will be of some considerable help in trying
to answer questions about Classical Hebrew. Longacre writes,
The linguistic specialist who takes out even a minimum of his time to read
the writings of his colleagues who work in other fields of specialization,
is often pleasantly surprised to find that the exercise is relevant and
helpful. For one thing, languages around the world involve not only
particulars but universals. The universals are partially masked by the
particulars. Consequently, one often finds in comparing two languages
(typically of two different language areas), a feature which is somewhat
latent and covert in language B is marked and overt in language A. The
student of language A, having learned of the presence of this feature in
language B, returns to his own speciality with his attention now directed
to a feature which had not formerly received sufficient attention from
him.17
Another language universal is that a language will employ one of two

basic techniques for marking peak situations, and other such 'textlevel' features. Some languages will have oneand only onemarker
for a specific function. Others chose from a variety of possibilities
(Longacre calls languages of this latter type 'bag-of-tricks' languages,
for obvious reasons), and the choice of which item is very often a
'stylistic'18 one.19
16. Chao, Greece, p. 76.
17. Longacre, Theory and Application in Processing Texts in non-Indoeuropean
Languages (Papiere zur Textlinguistik, Bd. 42; Hamburg: Helmut Buske Verlag,
1984), p.v.
18. I won't attempt to define this term closelyI tend to use it to describe the
situation where a language user is faced with more than one option to perform a
function, but where the differences between the options are not great; in such a case,
the language user makes his choice on the basis of any number of factors, many of
which are not solely related to semantics, etc. (such as energy level of the speaker),
which are often too subtle to be understood well by the non-native user, for example.
19. In English, for example, we can connect two paragraphs by the words 'thus',
'therefore', 'and so', etc. without any great change in meaning or macro-syntactic
significancethough the choice may still be rather rigidly controlled by other factors,
such as the socio-linguistic setting.
20
A language with only one way to mark a peak event, (to cite an
arbitrary example) in no way exhibits an impaired stylistic realm, it
must be said; those nuances which indicate the 'style' of a text are
simply apportioned to other facets of the language.
The role of language universals in the present study is all the
greater, in that the limited scope of our data is rendered, by language
universal sign-posts, not as significant a hindrance as we might
otherwise have been forced to conclude; for, just as comparison of
Hebrew with Ugaritic and Arabic can help us elucidate lexical and
other difficulties, these language universals can indicate to us certain
features that may occur in our texts, or they may indicate to us
possible solutions for difficult syntactic problems. One may approach
the language with a sense that one does not need to carry out an
analysis with no clues whatsoever, as if performing an autopsy on an
extra-terrestrial. Rather, the language lies before us (to mix
metaphors) as a countryside already vaguely familiar. In the same way
that we would suspect, based on our knowledge of general geography,
that abundant vegetation in the cleft between two hills may indicate a
water-source, we can construct initial hypotheses about our data,
based on language universals, where otherwise we might not notice
enough of the signs in our language to make any such observations.
Languages all over the worldand from every age from which we
have language datahave, to a great extent, had to perform the same
kinds of tasks (this is confirmed by the very existence of language
universals as well as by the evidence they present to us). This is to be
expected: all humans have the same size neurological language centres,
and social interchange tends to demand certain general things20 of the
speaker, which individual languages must accomodate. Psycho- and
neuro-linguists will affirm a certain consistency of psychological and
neurological language patterning, which does not vary greatly from
culture to culture, and so on.21
At the structural level, then, all language systems tend to have a
20. Can the reader imagine a language with no facility at all for discussing 'time',
'movement', 'food' or 'relatives'?
21. The question of whether we can apply such up-to-date research results to
languages from millennia past is one that we cannot ever fully answer; studies in
early language data (as early as we have it, that is) confirm, however, that 'language'
as a human tool has not altered so significantly in the time elapsed that today's
knowledge could not elucidate yesterday's data.
1. Introduction
21
concept of syllable structure, of intonation patterns, of hypothesisconclusion sentence structure, and so on; this should not evoke a
sceptical response from the reader. It may, however, be less evident
that languages have a strong tendency to structure larger units as
wellthat is, systems exist within all languages so far studied, which
serve to indicate to a reader or hearer22 that a speech has ended, or
that a new scene in the story has begun, or that the key point in an
exhortation has been made. These structures are sometimes subtle,23
and will often occur in overlapping fashion to give a cumulative
effect, but they exist nonetheless, and can be catalogued and codified
in the same way that 'hypothetical sentence' structures (English 'if X,
then Y') can be catalogued and codified.
'Universal syntactic structures', that is, those found in the vast
majority of languages usually include, from smallest building block to
largest unit: MORPHEME-WORD-PHRASE-CLAUSE-SENTENCEPARAGRAPH-TEXT.24 We are accustomed to analysing phrase-,
clause-, and to some extent sentence-structure, but paragraph- and
text-structure are relatively new to us.
I have avoided using the term 'discourse' here. My term 'text' is
very nearly equivalent to the former, as it is used in Longacre's brand
of 'Tagmemic' linguistics,25 that is, a text (or 'discourse') is a unit of
speech, whose constituents are paragraphs, and other, shorter, units;
texts exhibit consistent tendencies in internal development, which
features can be described linguistically. This definition, though inadequate, will suffice for the time being. It can be seen from the dictionary entry quoted at the beginning of this chapter that the term
'discourse' has a variety of uses; in the following material, I will
identify each author's own use of the term, and will be specific with
regard to the meaning I intend, if my own use of the term might lead
to confusion. In the long run, however, I try to stay away from
'discourse' as a technical termtoo many people have used it in too
22. In this work, reference, by one of these two terms, to the transmission of
language material, will not exclude the otherfor my purposes, the same language
processes are going on whether the communication is oral or written.
23. For example, falling intonation may indicate the end of a speech.
24. The definition of those terms requiring it will be given, either where first
encountered, or in Chapter 3.
25. A working introduction to this theory will be presented in Chapter 3.
22
many applications for it to be free enough of unwanted connotations

to serve my purposes.26
In addition to the universal tendencies mentioned above, all
languages need to be able to relate sequences of historical events;
likewise, all need to be able to give lists of instructions; all need to be
able to predict, exhort, describeand all languages need to be able to
distinguish these text-types one from another. The theoretical system
presented in the following chapters is a most helpful tool for looking
at this facet of language; it provides a framework for analysing and
codifying the features seen at the 'text' level of language.
Words are usually the building blocks of phrases, phrases the
building blocks of clauses, and so on. Texts rely on the content of
their constituents in order to convey their own content; it is to be
expected, then, that text-type distinctions depend heavily on the type
of information encoded in clauses, and so forth, by verbal aspects,
time-frames and the like, to expound the text-type being worked with.
Micro-syntax and macro-syntax work hand in hand: proposing that a
certain feature has macro-syntactic significanceas I will not
infrequently do in the following pagesin no way deprives it of its
micro-syntactic value and significance.
In addition, there are a limited number of grammatical and
syntactic options in the verb system, and, therefore, these options will
sometimes be called upon to perform more than one function (e.g. a
clause-type that serves as the foreground form27 in one text-type may
be the furthest from the foreground in a different text-type).
Nevertheless, it is importantif not essentialthat the researcher
acknowledge these text-type distinctions and their related verb usages
in order to make sense of the data; without such distinctions, no
patterns will surface to aid the researcher to understand the features
being examined.
26. This, too, is a fairly standard language trait: development of new applications
for a term is often followed by abandonment of it in other circlescf. the same
process with the word 'gay', whose more general definition as 'light-hearted' has
been more or less completely abandoned by much of the English-speaking populace,
owing to its more recently developed application as a term describing a sexual
orientation.
27. 'Foreground' and 'background' will be defined in Chapter 3.
1. Introduction
23
The essential theoretical elements that inform my methodology are:

1.
2.
3.
that identifiable structures exist at the paragraph- and textlevels, and that these can be described;
that 'text-type' is one of the strongest motivating factors at
macro-syntactic levels in the deployment of micro-syntactic
constructions;
that the positing of macro-syntactic roles for certain
constructions in no way lessens their micro-syntactic
identities, but that both layers work hand in hand to convey a
wide variety of necessary information.
These basic elementslike the dry bones in the valleywill be knit

together with sinews and flesh in Chapter 3. The basic features I have
described above will be sufficient for the reader to make sense of the
comments that follow, on the five works to be examined. They will
serve to create a reasonable context within which we can operate until
such a time as a further definition of theory and methodology becomes
more appropriate.
The works I will be examining in detail in this chapter and the next
are:
Andersen, F.I., The Sentence in Biblical Hebrew (The Hague: Mouton,
1974)
Eskhult, M., Studies in Verbal Aspect and Narrative Technique in Biblical
Hebrew Prose (Uppsala: Uppsala University Press, 1990)
Longacre, R.E., Joseph: A Story of Divine Providence: A Text-Theoretical
and Text-Linguistic Analysis of Genesis 37 and 39-48 (Winona Lake,
IN: Eisenbrauns, 1989)
Khan, G., Studies in Semitic Syntax (Oxford: Oxford University Press,
1988)
Niccacci, A., The Syntax of the Verb in Classical Hebrew Prose (trans.
W.G.E. Watson; JSOTSup 86; Sheffield: JSOT Press, 1990)
Each of these books contributes to a greater understanding of the textlevel features of Classical Hebrew, and I will be evaluating them on
the basis of the significance of that contribution (in particular, with
reference to the specific concerns of this study).28 It is only fair to
acknowledge that these books may not have intended to contribute to
28. Which, as I mentioned above, are (1) the interrelationship between clausetypes and text-types in Classical Hebrew, and (2) the question of integrity of theory
and methodology in enquiries of this sort.
24
an enquiry of exactly this sort, and I acknowledge at the outset that, in

some ways, I will not be 'giving them a fair reading'. Yet, these
provide an excellent working collection for assessing the kinds of
presuppositions, theoretical bases, and the like, which one finds in
contemporary literature that has a bearing on the description of the
text-level features of this language.
I will examine these books in two stages: the first group includes
Niccacci, Eskhult and Andersen, all of whom approach their topic
with macro-syntactic considerations in mind; the first two (Niccacci
and Eskhult) are concerned broadly with description of the verb in
Hebrew; Andersen, on the other hand, focuses on particular syntactic
features and presents a fuller description of their variants and
functions.
The second stage will consist of an examination Khan's Studies, and
of Longacre's Joseph.29 Khan and Longacre have more in common
than their particular theoretical background (which is shared by
Andersen), in that they each offer something unique in their work. In
Khan's case it is clarity of presentation, matched by solidity of
theoretical grounding; in Longacre's, it is his astute insight into the
macro-system of language as a whole, and of Classical Hebrew in
particular. As we shall see, these works also have their weak points,
which are likewise instructive.
My purpose in looking at these books will be to sample the variety
of theories and methodologies current in the academic world of the
hebraist, with regard to analyzing larger bodies of Hebrew text.
Excursus on B.K. Waltke and M. O'Connor, An Introduction
to Biblical Hebrew Syntax (Winona Lake, IN: Eisenbrauns, 1990)
Most works produced at present tend to steer rather wide of the kind
of work I am engaged in in this study; those which do address the
issue often do so with a wariness of the intellectual 'investment and
returns' involved. Those which do attempt this work are often marred
by an inadequate grounding in empirical theory, or never break out of
the world of 'linguists communicating with linguists'.30
29. Longacre's approach to language has informed my own to a great degree, and
his principles, for the most part, are my own. This work is dealt with in Chapter 2.
30. As I will reiterate below, in my examination of the various works in this and
the next chapters.
1. Introduction
25
One book that adopts the first approach is Waltke and O'Connor's
Introduction. I have a mixed response to this impressive 'syntax'. On
the one hand, it is a remarkable work, and welcome; yet it is also
seriously disappointing on several levels: in the first place, the authors
spend a vast amount of time on semantic evaluation of forms (which is
grammar, not syntax), and, in all honesty, very little time on syntax
itself. Although they do occasionally discuss word order and the like,
in clauses containing various forms of the verb, they spend far more
time on discussing the meaning rather than the function of those
forms.31 I have yet to find, for example, a discussion of what elements
may be found, and in what order and with what significance, in a
participial, or an infinitive, phrase. All this underlines a certain
confusion about what constitutes grammar, and what constitutes
syntaxby far, most of what purports to be syntactic description is
little more than grammar where the nuance of the form in question is
fine-tuned by a look at the context. This is not to say that I denigrate
Waltke and O'Connor's volumeit is a treasury of information; my
complaint on this score is solely that they don't end up doing as much
syntax as one is led to expect.
In the second place, Waltke and O'Connor blow hot and cold on
'discourse analysis or text linguistics',32 opting for 'the more
traditional path' of old-style phrase and clause analysis (and which
they do not do very systematically at the syntactic level). They write,
We have resisted the strong claims of discourse grammarians in part for
the theoretical and practical reasons mentioned earlier: most syntax can be
and has been described on the basis of the phrase, clause, and sentence.^
Further, it is evident that the grammatical analysis of Hebrew discourse is
in its infancy. As an infant, it offers little help for the many problems of
grammar which have not been well understood. Most translators, we
think it fair to say, fly by the seat of their pants in interpreting the Hebrew
conjugations. Hebrew grammarians have only recently come to appreciate
morphemes as diverse as the 'object marker' 't and the enclitic mem. No
modern grammar, further, has begun to gather together the wealth of
individual studies that have been carried out in a more traditional framework; thus it is not surprising that some students know little about the
31. E.g. 'A non-perfective of instruction expresses the speaker's will in a context
of legislation or teaching', Waltke and O'Connor, Introduction, p. 510.
32. Waltke and O'Connor, Introduction, pp. 53-54the entirety of the section.
33. Emphasis added.
26

case functions and some commentators make egregious errors in their
interpretations of prepositions. For our purposes, therefore, we are
content to stay with more traditional bases than those of discourse
grammar.34
And yet they also write,

If we seek to systematize our understanding of textual organization, we
need to introduce the notion of different levels and types of organization.
Not every verse, for example, works in the same way in itself and in relation to the verses around it. We may recognize a class of major textual
markers or macrosyntactic signs, by which we mean conjunctions and
other expressions that bind together the sentences constituting a larger
span of text.
Here follows a quote from W. Schneider, Grammatik des biblischen
Hebraisch,35 pointing out that certain 'signs' have macrosyntactic
significance. Waltke and O'Connor follow this with a very lucid
summary, then a disclaimer:
Like the Masoretic accent system, this method of analyzing textual
organization requires independent study (cf. 3.3.4).
A simpler approach may be offered... 36
In fact, their 'simpler approach' is not simpler, but muddier, and the
simpler approach is the one they have forsaken.37
They detail in their book what has been written about Hebrew; at
times this appears to exclude the idea of actually describing the
language itself. For example, it is nearly universally accepted, nowadays, that 'narrative' is a category with particular features in Biblical
Hebrew; Waltke and O'Connor accept it as such, without ever
examining what that acceptance implies for language in general, nor
for their description of Biblical Hebrew, as the subject of their
work.38 They have taken it on board because it is a matter of general
34. Waltke and O'Connor, Introduction, p. 55.
35. (Munich: Claudius, 1974), p. 261.
36. Waltke and O'Connor, Introduction, p. 634.
37. Waltke and O'Connor, Introduction, pp. 634ff. I do not intend this as
negatively as it might sound; the fact remains that text-linguistics has already made
significant contributions to the organization and simplification of Hebrew grammar
and syntax. These Waltke and O'Connor eschew in favor of a method that focuses
on particulars rather than patterns, to the complication and detriment of their language
description.
38. See, for example, Waltke and O'Connor, Introduction, p. 647.
1. Introduction
27
acceptance. And yet, although there is likewise a consensus of opinion

that acknowledges TP to be an intriguing and significant feature of the
language, the index directs the reader to only three locationsof
these, two are tangential mentions in the context of a brief reference
to someone else's work, and the third is to tell the reader that one
should translate the clause nrn Tm into English as 'One day'; it is
accompanied by one example. This glossing over of such a fascinating,
and textually significant, feature of the language is not helpful; they
choose to do so, I believe (given their own statements in their
introductory sections), because if they were to engage in a fuller
treatment of the issues involved, it would take them out of their
realm, and into areas they have chosen, for personal, rather than
academic, reasons, to avoid.
It would be irresponsible of me to give a purely negative review of
this book;39 but to be honest there is something in the underlying
principle of the work to which I take serious exception: this is a textus
receptus approach to describing the language; it has chosen to include
only what has been accepted previously, and to break no new
groundnor even to summarize current developments in a way that
encourages the reader to look forward to progress and development.
It is not likely that another work of a similar nature and scope will
receive, for a long time to come, the support and prominence that this
has received; therefore, this book will be one of the standards for this
generation, and possibly more, of Hebrew students;40 it has had an
opportunity to bring the fascinating array of advances in description
of the language to the broadest possible readership, in a way that
could have moved the study of this language forward exponentially,
yet it has chosen a safer, simpler (for the authors, but not for the
student) approach. In my opinion, had the authors taken the task more
'holistically',41 and had done the four, or seven, or ten, more years'
39. The introductory overviews are exceptional, for example, despite the restrictions of space which must curtail most of these discussions; and, for the most part, it
is not what they cover that engenders disagreement, but, rather, what they have
omitted.
40. Re-editing of some of the older standard grammars and syntaxes, with a view
to bringing in modern linguistic insights, is presently underway, and promises to
combine the best of both worldshow much more, then, should we be able to find
this in a work like Waltke and O'Connor!
41. I recognize that this excursus may sound rather shrill, yet I believe these
28
work, which would have enabled them really to write a 'syntax' based
on the work available, the book would have been shorternot because
less would be said, but rather because more would have been said,
with greater clarity, elegance and simplicity.
2. Niccacci, The Syntax of the Verb in Classical Hebrew Prose
Niccacci's Syntax is an attempt to describe the variety of uses of the
Hebrew verb from a text-linguistic perspective. In several ways it is a
strong beginning, but also exhibits several flaws.
The greatest strength of this book is that it eschews older styles of
Hebrew grammatical/syntactic description,42 in favour of an approach
drawn from modern linguistics. Niccacci points out the weak points of
these more 'traditional' grammars on occasion, in hopes of converting
the reader to this new 'text linguistic' approach to language analysis,
and in justification of his own approach:
While it is true that Hebrew had only a limited number of verb forms at its
disposal, it still seemed odd that, for example, WAYYIQTOL could be
translated by virtually all the finite tenses of modern languages, as would
appear from classical grammars. Nor is it easy to accept the view that
QATAL, which was supposed to be the form for beginning narrative in
Hebrew, could have been replaced so often in that position by the
WAYYIQTOL by customary misuse. It was obvious to me that the
lengthy catalogues of special cases and exceptional uses listed in the
grammars only show how difficult the problem is.
In turn, translators select the equivalent tenses of modern languages
somewhat at random, applying their own interpretation and sensitivity.43
and,
It is clear, then, that text linguistic analysis enables us to formulate a set of
rules concerning the use of WAYYIQTOL and so considerably lessen the
frustrating impression gained from leafing through traditional grammars:
that almost any tense of modern languages can be used to translate it.44
complaints are inherent weaknesses in the work, and yet are often overlooked owing
to the otherwise justifiably impressive nature of the volume.
42. Derived originally from grammars of Latin, hence references to 'genitive
constructions', etc., which are out of place for a language with no case system.
1. Introduction
29
His general statements with regard to language principles are often

excellent. Consider the following:
From the aspect of text linguistics every construction which breaks the
narrative chain belongs to the background.45
As in pure narrative the chain of WAYYIQTOLs is not interrupted without
a reason.46
It is clear, then, that the tense of an individual form and construction and
therefore the most suitable tense for translation into modern languages is
not tied to that actual form or construction but to its function in the text.47
In addition, this book contains some remarkably succinct statements

about the syntax or functions of particular features. For example,
In respect of linguistic attitude, WAYYIQTOL is the tense for narrative
(s. 81); in respect of emphasis it denotes the foreground (s. 86); in
respect of linguistic aspect it denotes degree zero (s. 88).48
Essentially, Niccacci's starting point, and conclusion as well, is that

there are two basic features which, when examined together, provide
the key to understanding the syntax of the Hebrew verb (in prose, as
his title indicates). The first consists of a three-fold set of mutually
exclusive categories:
Linguistic attitude: Narrative/Commentary (or Discourse)
Emphasis/Highlighting: Foreground/Background
Linguistic Perspective: Retrieved Information/Degree Zero/Anticipated
Information
The second feature is word order within the clause (he will usually
refer to this as 'position in the sentence', but for the most part, he is
dealing with clause-level syntax rather than sentence-level syntaxor
higher).49 These two sets of criteria are his analytical parameters.
After laying out his basic theoretical ideas in chs. 1, 2 and 3,
Niccacci looks at the two 'tenses' WAYYIQTOL and QATAL in
ch. 4.50 Here, Niccacci looks at the distribution of these two
47. Niccacci, Syntax, pp. 164-65.
49. See Niccacci, Syntax, pp. 20ff., 23, 173, et passim.
50. I take exception to the use in this book of the term 'tense'though I
acknowledge that some of the problem may have been engendered by the difficulties
30
conjugations, in narrative, and in 'discourse'.51

Perhaps Niccacci's strongest moment is in ch. 5, where he looks at
Hebrew Narrative, and where he truly begins to engage in textlinguistic analysis, yet this is also one of his more dangerous points as
well, for he is not entirely thorough, and draws conclusions that do
not accurately describe the data.
He attempts a similar 'text-type-distribution' analysis in ch. 6, with
reference to 'discourse', but I find his conclusions unsatisfactory,
owing to his definition of this category.
Chapter 7 deals with Tense Shift', in which he looks at the motivation for moving away from the standard tense for the text-type being
examined; this again is a strong point, in particular where he refers to
narrative material; however, I take exception to some of his
findingsonce again with reference to 'discourse', but here with
reference to other conclusions as well.
Chapter 8 laysout the details of what he terms the 'two element
syntactic construction'. In my opinion, however, this whole concept is
based on some of his weakest prior conclusions, and therefore does
not exhibit a great deal of internal strength and consistency.
Chapter 9 is Niccacci's summary of findings. He is definitely at his
best when summarizing, and, where his findings are solid, his summary statements are brilliant. Where he is on less solid ground it is
often difficult to make sense of his statements, particularly in
comparison to alternative explanations of the data.
Niccacci's most striking contribution to the study of Hebrew syntax
of translation. Tense', according to Crystal, is 'a category used in the grammatica
description of verbs (along with aspect and mood), referring primarily to the way the
grammar marks the time at which the action denoted by the verb took place'
(Dictionary, p. 306, a fairly standard contemporary understanding of the term). It is
more or less universally agreed that Classical Hebrew does not have 'tense'-based
conjugations. Niccacci himself points out on several occasions that Classical Hebrew
does not exhibit tense-like features (see, for example, the above-quoted material from
pp. 9 and 177), and yet he chooses this term to describe the Hebrew 'conjugations'
which term seems to me better suited to his needs; failing that, he could always
choose the option of such phrases like 'QATAL-forms'. This inconsistency in
adherence to the principles of modern linguistic description (which prefers labels that
describe the form, rather than the function, when the form is the distinguishing
feature) is only one of several minor difficulties that can be found in this work.
51. I will reserve my comments on Niccacci's category of 'discourse' until
further on.
1. Introduction
31
is, in my opinion, his assessment of the foreground/background

opposition, as being indicated by shifts from one verb-form or
syntactic construction to another. And he is very much on the right
track with his distinction between 'narrative' and 'discourse', but his
handling of 'discourse' is considerably muddied by his definition of
this category. And again, his handling of the three-way distinction
between 'recovered information', 'degree zero' and 'anticipated
information' is going in the right direction, but is weakened by certain
inadequacies related to other linguistic questions.
The major 'flaws' that I find in Niccacci's work are as follows:
1.
2.
3.
He has lumped all conversational material (reported speech)

into one category called 'discourse'regardless of differences of text-typeand expects this category to show internal
consistency and predictability, in spite of the fact that he does
not provide parameters for distinguishing, say, between
exhortations and reported history.
Although he opts for a text-linguistic analysis of the verb
system of Classical Hebrew, he almost never gets beyond the
clause level.
He is not rigorously thorough in his application of linguistic
principles, and permits himself both short-cuts and
inconsistencies.
The first 'flaw' is at the theoretical level; the second two are
methodological.
With reference to my first objection: we require that our language
give us many clues as to the type of material we are hearing or
reading, and these clues must be more or less instantly recognizable.
So, for example, we need to be able to differentiate the following
phrases, in order to know what sort of social interaction is involved
and expected:
He went to the store, and he bought bread.
Let him go to the store, and let him buy bread.
The clause types used in the two sentences immediately indicate to us

the text-type we are dealing with in each case (the first is historical
narration, the second is exhortation). However, if both the
above sentences occurred in Reported Speech, nothing in Niccacci's
32
description would distinguish the first from the second.52 This is

inadequate.
Niccacci recognizes the principle that each type of speech
(narrative, exhortation, exposition/explanation, etc.) must use a
different verb-form or construction as its main-line form. One can
hardly expect the main form for story-telling to be an imperative;
and, in the same way, exhortation does not rely on historical past as its
main verb form; yet each occurs with regularity in 'conversational'
material/reported speechNiccacci's 'discourse' category is incapable
of accounting for this.53
Were he to have described these syntactic structures in terms of
their text-type, his analysis of 'discourse' would have greater simplicity and clarity. As it is, his analysis is over-complicated and sometimes
more than a bit confusing. He concedes, however, that this category
requires further work:
The forms of discourse [conversation] still require study, though, in
particular the indirect volitional forms.54
With reference to the second so-called 'flaw': Niccacci only once

presents a longish section of Hebrew text for analysis, and this is with
reference to narrative material embedded within 'discourse';55 his
analysis of this section of text is intended to discover whether or not
embedded narrative (his 'discourse narrative') follows the same
pattern as narrative which is not related in Reported Speech. And yet,
although he insists that units exist beyond the sentence level, and hints
that these units also have identifiable structures,56 he never looks at a
complete text in an attempt to isolate such structures before trying to
identify similarity or dissimilarity to a normative pattern. This
prevents him from seeing that a variety of features may, for example,
serve to introduce a narrative unit, which discovery would in turn
inform and streamline some of his other language descriptions.
The third difficulty with Niccacci is a certain tendency toward
52. See his Table 2, Niccacci, Syntax, pp. 168-69.
53. No fewer than fourteen possible syntactic constructionsmore than half
those he listsare given as possible foreground 'discourse' clause-types, with no
real indication of under what conditions they perform this function.
55. Judg. 11.12-28, Niccacci, Syntax, pp. 102ff.
56. See Niccacci, Syntax, ch. 5.
1. Introduction
33
overstatement of 'rules' of syntax, some of which are occasioned by

the theoretical and procedural difficulties mentioned aboveand, as a
result, he occasionally contradicts himself. For example, Niccacci says
of the 'report QATAL' that it 'never heads a sentence' as well as that
it 'is a form with first position in the clause'.57 These two statements
cannot be reconciled with each other, and do not describe the data.
And again, he writes, 'a very important fact concerning the use of
QATAL in discourse is that it always comes first in the sentence; this
never occurs in narrative'.58 This statement is in fact wrong on both
counts: clear exceptions to both facets of this rule occur in Ruth, for
example.59
I do not often find myself in agreement with Niccacci's rules,
particularly with regard to distribution of specific syntactic constructions. Non-linguistic readers of Niccacci's work may be tempted to
think that the difficulties in understanding this work stem from their
own lack of sophistication in this discipline; Niccacci, however, must
share the responsibility for the difficulties encountered, since his work
exhibits a certain internal inconsistency.
I hasten to add, however, that Niccacci is not so much 'wrong' in
his conclusions, as he is 'only partly right'theoretical factors that he
has overlooked provide contextual clues that clarify the distribution of
these syntactic elements, and permit a more restrained and 'more
elegant' formulation of 'rules.' Niccacci is working toward the
description of the system that governs the deployment the verb in
different clause-types in Classical Hebrew; and he is correct, both in
supposing that it is there to be discovered, and in seeking many of the
clues to it at the macro-syntactic levels. The difficulties encountered
59. Cf. Ruth 3.17.2 [the three indices in this form of reference identify
Chapter. Verse.Clause; this last refers to the clause within specific versein this
case, the second clause of Ruth 3.17] "D ^ p rfron cnwomra..., contrary to his
first assertion, and Ruth 4.7.2-3 inm1? ]rm // fpun BTK ^...(1 will use this
convention [//], from time to time, where it is useful to see the text in continuous
sequence, but where it is nonetheless advantageous to bear in mind the clause
divisions of the text), contrary to his second assertionthis latter I would not call
"Narrative' and therefore the second assertion might be valid according to my criteria,
but it is not true according to Niccacci's, for he would not distinguish this material as
being of a separate text-type (see chapters 3, 4 and 5 in this volume, for the theory
and examples relating to this issue).
34
here arise out of a partially insufficient theoretical base and a few

procedural short-cuts, with the result that the conclusions he comes to
are less elegant than they could be, and do not fully match the data.
There are several other comments I wish to make with regard to
Niccacci's analysis. First, he recognizes and focuses on macrosyntactic markers (for example, run and nn,u); this is a strong move
forwardhere again, however, he will benefit from isolating
distribution of these elements according to text-type.
Second, he examines forms of rrn as a possible macro-syntactic
markers, and makes some cogent observations. This is a significant
improvement over Waltke and O'Connor (and others), and the
traditional grammarians, who limit themselves to only the briefest
allusions to this feature's (and others') macro-syntactic significance.
On the other hand, while he acknowledges that semantics must play
a role even in macro-syntactic description,60 he restricts his semantic
analysis to grammatical forms. He overlooks the fact that, for
example, rrn never represents a 'full' event, by virtue of its inherent,
'stative', meaning, and thus can never be a foregrounded narrative
verb. His explanation of the distribution of this verb is fairly
convoluted, and could have been simplified with recognition of the
stative nature of rrn.61
One of the difficulties Niccacci runs into here is due to his apparently not having examined larger units for structural similarities.
This sort of investigation would have shown that 'interruptions' to the
60. 'We have already stated that semantics is of importance, even if only
secondary, in determining the function of a verb form or grammatical construction'
(Niccacci, Syntax, p. 165).
61. I am aware that a case can be made for a dual semantic nature of rrn, with the
'be' and the 'become' meanings treated independentlythis may have significance
for the proposed macro-syntactic functions of the word, with the 'be' occurrences
having a stronger macro-syntactic effect in narrative, for example, than the 'become'
occurrences. I do not see the need for this, at this point, as neither of them appear
ever to be main-line verbs in the Narrative History data I have examined (definitions,
and further discussion of the issue, will follow). In addition, I am sceptical of this
distinction at another level: is this not simply a matter of difficulties in translation?
What evidence do we have that this was a significant distinction for native users of
the language? Until I have seen empirical evidence that the 'be' sense and the
'become' sense of this verb have distinctive distributions in the text itself, I see no
need to treat them separately. 'Do not explain with two things what can be explained
with one', to paraphrase the 14th c. epistemology of William of Ockham.
1. Introduction
35
main narrative are marked as more or less significant on the basis of

(1) how far removed the interruption is from the normal semantic,
temporal, and aspectual (etc.) qualities of the main-line of the text,62
and (2) how many non-main-line clause-types occur in tandem. This
second factor is significant for evaluation of rrn formsif Niccacci
were to have worked from the assumption that Classical Hebrew is a
'bag-of-tricks' language, the idea that 7 m clauses are among the
possible options in the bag of tricks available for a specific task, and
he would not have to claim a distinction between 'm as a 'full form of
the verb' and 'm as a macro-syntactic marker, since 'm as a 'full
form' could still function as a macro-syntactic marker. Comparison of
several large units of text would lead as well to the realization that
narrative units can be broken down into smaller units without
endangering the integrity of the whole, and he would not need to insist
that -m is always a marker of continuity. (In my reading of the data,
TTI seems almost without exception to function as a paragraph-break
marker, and as such is often a marker of discontinuity.)63
Third, had Niccacci analysed the structure of larger narrative units
more carefully, and had he acknowledged the integrity of other texttypes, he would not need to posit a difference between narrative
embedded in discourse and 'narrative proper'.64
When analyzing the Judges 11 text as an example of narrative
within 'discourse', Niccacci makes several surprising statements:
I use the term 'narrative discourse' for this type of narrative in which the
events are not reported in a detached way, as in a historian's account, but
from the speaker's point of view. Naturally, verbal forms in the first and
second person predominate.65
The claim that 'verbal forms in the first and second person
predominate' is odd, particularly in light of the fact that, in the
62. We will return to rework this concept more deeply in Chapter Three.
63. Perhaps some of the difficulty here is related to the term 'continuity' and its
cognates; I perceive an element of continuity in the function of vn, due to its we +
Prefix formit is certainly more indicative of 'continuity' than the same clause
would be if it contained a Suffix form of rvnyet I also want to highlight that 'm
appears to have a consistent role of the 'hinge' between paragraphs, etc. In this
sense, which I consider dominant, it is an indicator of 'discontinuity'.
36
example Niccacci himself gives, in the narrative in question,66 all

narrative clauses are in third person; two cohortative clauses do
occur, which record the messages sent to the two kingsthese are not
narrative, but exhortation: 'Let us pass through your land' (in vv. 17
and 19).
In addition to this, Niccacci claims that 'narrative discourse' and
'discourse proper' obey different rules, and that
The text of Judges 11 just examined shows that QATAL is the verb form
for beginning 'narrative discourse'... Now here a basic fact must be
emphasized. No 'narrative discourse' begins with a WAYYIQTOL; the
WAYYIQTOL is always the continuation form of an initial construction
typical of discourse (cf. s. 24)... This fact reveals the fundamental
difference between the WAYYIQTOL of narrativewhich is either initial
or the continuation of another initial WAYYIQTOLand the WAYYIQTOL
of discourse which is never initial but always the continuation of a
non-narrative initial construction, different, that is, from WAYYIQTOL.67
Niccacci is at ease with the idea that a narrative unit can (or perhaps,
'must') be preceded by an 'antecedent construction' sectionthat is, a
section that provides background information such as location, time
reference, and so on.68 Why is it, then, that he discounts this explanation of the opening temporal clause of the Judges narrative on the
grounds that it occurs in recorded speech? This explanation actually
fits the data, and eliminates the need for multiplex descriptions of this
verb-form and text-type.69 In my opinion, Niccacci considerably
overworks the 'linguistic perspective (retrieved information/degree
zero/anticipated information)' factor. This is in compensation for the
too-broadly defined 'discourse' category; for, where he has failed to
distinguish the integrity of individual text-types and the consistency
with which clause types function within them, he has been forced to
attempt another explanation for the bewildering remainder of material
left unexplained by his other two categories (linguistic attitude:
narrative/discourse; and prominence: foreground/background)both
of which categories are quite solid, apart from the ineffectively broad
'discourse' category.
66. Judg. 11.16-22, 27 clauses in all, by my count.
68. Cf. Niccacci, Syntax, pp. 36-37 and p. 48.
69. I will address this problem again in Chapter 5, after we have examined some
data, for I will propose another explanation for this phenomenon.
1. Introduction
37
This 'linguistic perspective' category likewise has promise, but does

not rise to the occasion, since it does not explain all the occurrences it
purports to explain; this factor does not appear to play a major role in
verb and clause usage in Classical Hebrew, and Niccacci does not seem
to be on the right track in assuming it as a motivating factor for a
wide array of features; other motivating factors can be demonstrated
to have a far greater influence over the choice of clause type, resulting
in a much clearer presentation of the micro- and macro-syntax of this
language.
My last objection to Niccacci's description is again one questioning
the degree of influence of a certain factor, in this case word order,
within the clause. Here, I am largely in favour of Niccacci's
conclusion: that the emphasis of the clause is determined by what has
first position in it (excluding conjunctions such as -i and o); however,
I feel he takes this too far. Permit me to quote at length:
At this point we can state our opinion concerning the traditional definition
which runs as follows: a clause is verbal when the predicate is a finite
verb and nominal when the predicate is a noun. For this definition to be
valid it should also be specified that in Hebrew a finite verbal form is
predicate when it comes first in the clause. When, instead, it is preceded
by an element of any kind (other than WAW) the verbal form is not the
predicate and therefore the clause is nominal (CNC)...
By definition, the 'subject' is the topic spoken about (usually a person or
animate being) and the 'predicate' is what is said about the subject...
Now the subject is a noun or noun equivalent ('noun-phrase') while the
predicate is a verb ('verb-phrase'). According to modern linguists the
'noun-phrase' has first position in the sentence. It should be noted,
however, that this statement does not suit Hebrew, for two reasons. First,
in Hebrew the first position in the sentence is filled by the predicate, not
by the subject. Second, when a Hebrew sentence begins with a noun or
an adverb the predicate is not identical with the verb but in actual fact with
that noun or adverb. Accordingly, what is normally the 'subject' becomes
the 'predicate' and vice versa.
This transformation is not exclusive to Hebrew as it occurs in other
languages, ancient and modern. It is effected by nominalisation of the
verbal form. To this type belong the first two constructions in Hebrew,
listed above:
1)
2)
x-participle with article

x-'asher + QATAL
38

Typically, in Hebrew there is a way to effect the change from 'subject' to
'predicate' without nominalisation of the verb. Instead, the noun is simply
placed at the head of the sentence. To this type belong the other two
constructions discussed above:
3)
4)
x-QATAL
x-YIQTOL70
This opposition of subject and predicate seems overstated. In truth, all

the information in a 'sentence' (by which he here means 'clause')
and not just the predicateis vital to the meaning of the sentence, and
though one element may have greater importance, this is only on a
sliding scale; to label one element 'subject' and another 'predicate' in
this fashion is to put into black and white that which should be
described in terms of greys. Rather, is this not a case, not of the transformation of subject to predicate and vice versa, but more simply of
the de-emphasizing of the predicate and the emphasizing of the
subject, without the categories themselves being reversed? Niccacci's
approach oversimpifies the phenomenon, while creating a new range
of concepts and definitions for terms which served adequately in their
old guise. I feel that nothing is gained in calling a noun which
precedes the verb 'the predicate', rather than 'sentence-' or 'clauseinitial'. Such a redefinition is extraneousone accomplishes the same
thing by giving the label 'subject-initial' clause, or some other similar
(and more empirically descriptive) label; this 'transformation' of
subject and predicate is not necessary, and complicates description of
the language.
He is also not accurate in his assessment of 'modern linguists'; it is
incorrect to say that they believe the noun phrase to be 'first in the
sentence'. Perhaps this is true of those who do not engage in
linguistics as a description of data (though I cannot imagine what
other sort of linguistics would exhibit any integrity); it is certainly not
true of linguists whose goal is to describe 'what is there'. The syntax
of Classical Hebrew must be described in such a way that all syntactic
options are accounted for; this is easiest if one specifies as accurately
as possible what sorts of clauses (for example) have what sort of
constructionfor Classical Hebrew, for example, one might say that
main-line narrative transitive clauses are normally V-S-O, while
1. Introduction
39
background narrative transitive clauses are S-V-O and, rarely,

O-S-V.71
Niccacci suggests that this transformation from subject to predicate
and vice versa is found in modern languages, and cites the French
'c'est moi qui...' as an example.72 This in fact is simply a stative
identificational clause (identical to the likes of 'c'est un homme', 'il est
professeur', 'he's my friend', and so on) which follows its standard
rule for stative identificational clauses in French:
S + stative-Verb [= etre] + Noun Phr. or equiv.
In this case the predicated Noun Phrase is a personal pronoun, which

is modified by a relative clause (which is left incomplete in our
example, but which will follow its own rules of constituent structure).
There is no need to propose a shift of predicate to subject and vice
versa: properly written rules forestall the need to redefine basic
terms.
I do not share Niccacci's confidence in the finality of his results, as
evidenced by his statement, 'scrutiny of a wider selection of texts
might contribute further refinements but I do not envisage major
modifications'. 73 I believe he is well on the right track, yet I do
envisage modifications, some of which will indeed be major.74
71. This is, of course, a vastly oversimplified 'rule', taking no account of

sentences with conjunctions, temporal phrases, adverbs, etc., each of which can be
described accurately, without requiring a redefinition of terms, and without becoming
too cumbersomebut this illustrates well enough the benefits of describing a
feature by its own features: 'S-V-O' is a much more 'accurate' description than
'nominal clause' as opposed to 'verbal clause'. These other labels may have some
value in other facets of language description, but at the level at which Niccacci,
Eskhult, myself, and others, are working, it is far clearer to employ the former
option.
72. Niccacci, Syntax, p. 29 n. 18.
74. Lest this sound condescending, I must acknowledge my debt to Niccacci for
several insights and characteristics, not the least of which, in addition to several
observations on clause function in Classical Hebrew, is his example and encouragement in challenging outmoded ways of thinking.
40

3. Eskhult: Studies in Verbal Aspect and Narrative
Technique in Biblical Hebrew
Eskhult's book has a narrow focushe looks principally at a single

clause-typebut he attempts a great deal within these narrow
confines. He writes,
The aim of this study is to shed light on the system of aspectual contrasts
in Biblical Hebrew, and more particularly how aspectual contrasts are
used as a device in narrative technique.75
His basic approach is solid: he looks to the world of theoretical

linguistics for direction in resolving the opacities of Hebrew syntax.
Although his is essentially a micro-syntactic enquiry, with a shy
courtship of macro-syntax (he does not really step outside of microsyntax even to examine macro-syntactic function), he asserts the easily
overlooked fact that 'there are factors on the clause, sentence and
episode level that bear on the verbal aspect'76 and, further, that
it is decisive for the interpretation of classical Hebrew prose, that one is
clear about which clauses carry on a narrative, and which depict a background state of affairs. In narrative discourse the skeleton is made up of
sequential wayyiqtol-clauses, while the subj-qtl clause almost exclusively
furnishes some background information.77
These are insightful observations, solidly rooted in good theory, and

they lead him to posit the (we)subj-qtl clause as a feature that
separates sections of main-line narrative clauses, providing a structuring mechanism for narrative texts.78
Eskhult is hoping to contribute to solutions in the realm of Hebrew
macro-syntax; his approach is to perform a 'cross-over' attempt,
75. Eskhult, Studies, p. 9.
78. He suggests that this feature is particularly suited to a largely illiterate society,
with its emphasis on the oral, rather than the written, use of narrative (Eskhult,
Studies, p. 43); but this is somewhat irrelevant, since literate societies, with their
written narrative, have the same tendency to use clause-types to endcode episode
margins, etc., as do illiterate ones (see the example from Mary Shelley's
Frankenstein, below). This foreshadows a theoretical weakness which will be
betrayed more clearly by other difficulties; his theoretical base is too far removed
from 'real language'.
1. Introduction
41
whereby he looks to micro-syntax for help in determining the function and distribution of macro-syntactic features. He forcuses strongly
on the aspectual (and thus micro-syntactic and semantic) nature of the
verbal clauses under consideration.
The first of three major difficulties that I perceive in this work,
however, is that Eskhult attempts much in too little space, and leaves
the sceptical reader largely unconvinced.79 There is no doubt that
micro-syntax and macro-syntax interpenetrate, and he is right in
assuming that the aspectual nature of these clauses is the means by
which they perform their macro-syntactic functions, but the
connections that Eskhult proposes are not so apparent that his
substantiation of these claims is sufficient. Eskhult has left it unclear as
to whether he sees himself doing clause-level analysis, or ' discourse'level analysis;80 and though this type of study is welcome, it does not
really belong in either category, and his proposal requires a more
lucidly presented, and much more thoroughly substantiated, work than
this if it is to be accepted by his readers.
In addition, the extensive micro-syntactic description of the
aspectual features of a few clause-types (the scope of this work
extends to cover only (we)subj-qatal clauses, with only the very
briefest examination of other clause-types), with occasional reference
to possible functions at higher levels, does not really constitute a
macro-syntactic analysis, norowing to lack of thoroughly reasoned
explanations for connecting micro-syntax so closely to macrosyntaxdoes it convince the reader to do the work more thoroughly
him- or herself.
This emphasis on micro-syntax is not prerequisite to a description
of macro-syntactic functions, for once one accepts that the difference
in meaning conveyed by the various clause-types enables the hearer/
reader to register its macro-syntactic signals, the aspectual nature of
specific clause-types ceases to be central to macro-syntactic analysis.
That is to say, a syntactic study needs to identify and describe the
structures and functions of elements at whatever syntactic level one is
working at; it does not require us to do all other levels of analysis at
the same time.
79. This need not be the case; a more in-depth examination of such a broad topic
has been successfully undertaken: compare the thoroughness of M. O'Connor's
Hebrew Verse Structure (Winona Lake, IN: Eisenbrauns, 1980).
80. Eskhult uses the term 'discourse' to refer to large units of text.
42
This may seem at first blush either obvious or irrelevantor

simply wrong; yet macro-syntactic descriptions will benefit from the
clarity gained by omitting secondarily related material. The macrosyntactic function of a particular clause-type may, for example, be
more simply apprehended if we stay within the syntactic level where
the feature is to be found, as opposed to stopping frequently to assess
its aspectual connotations. We expect a book on Hebrew syntax to
forego in-depth analysis of grammatical constructions, except where
they substantiate a point under examination; likewise we would find a
lexicon 'obsessive' which makes constant reference, on even the most
minor points, to the phonological system of the language. In each of
these cases, we accustomed to the fact that features that can be
analyzed in their own right can become (more or less) unquestioned
building blocks at levels further up the hierarchy; if we don't take at
least some things as tentative starting points, we will never get off the
ground.
The (we)subj-qatal clause does warrant approach from both the
macro-syntactic and the micro-syntactic directions; however, Eskhult
attempts to cover too much ground in too few pages (objection no. 1),
and he (like many others) is working from a theoretical base that does
not direct him to focus on text-type as a conditioning factor in
deployment of clause-types (objection no. 2); in addition, he does not
attempt to examine this clause-type as one of several possibilities for
this particular macro-syntactic function (objection no. 3)for these
reasons he describes successfully neither the aspectual relationships,
nor the macro-syntactic distribution and functions, of this clause-type.
And yet, it must be said that, like Niccacci, he makes some very
astute and helpful observations. Some of his summary comments are
direct and incisivethe model of elegancefor example,
Generally speaking, a clause expressing a completed action with reference
to the present is static, since it refers some state to a preceding situation.
Therefore, it seems natural that such a clause is construed as a descriptive,
not a narrative-sequential, clause.81
81. Eskhult, Studies, p. 39; it should be noted, however, that such nuances of
time reference are decidedly difficult to pin down. This is a constant thorn in the flesh
of this study: starting from the point of aspectual contrasts (which tend to be too
evasive of secure identification and codification to serve well in this role), enforces
on this study a constant weakness.
1. Introduction
43
With regard to my second objection, Eskhult, like Niccacci, fails to

take into account that different text-types will use verbs and clauses
differently; their attempts to generalize on the distribution of clausetypes in Reported Speech material, in particular, are confused and
unsuccessfullacking entirely the elegance visible in their more
solidly grounded observations. Weeding out mutually exclusive
occurrences, predictable on the basis of differing text-types, would
enable each of them to describe much more clearly the harmony and
symmetry of the Hebrew verb system.
Eskhult writes,
from a discourse perspective the constrast between 'state' and 'motion' is
also a contrast between 'background' and 'foreground', which means that
aspectual contrasts are liable to be used by a narrator in order to facilitate
the apprehension of the structure of a story on the part of his audience.
This is, however, fully pertinent to narrative portions only; in dialogue
this contrast is less apparent.82
82. Eskhult, Studies, p. 121; Eskhult's summary of the role of aspectual contrasts
is very much on the mark, yet as I hope to illustrate beyond doubt below, factoring in
the role of text-type in the deployment of clause-types will improve the description
even further. It is a feature of the general state of affairs in Classical Hebrew textlinguistics that Reported Speech is evaluated as a single, unified text-typeit is rare
that hebraists interested in text-linguistics look at Narrative texts in Reported Speech,
and compare them with Predictive, with Exhortation, with Instructions, and with
Expository texts in Reported Speech, to identify features peculiar to each, and traits
common to all. Another recent contribution to the discussion falls afoul of the same
insufficiency: A. den Exter Blokland's 'Clause-Analysis in Biblical Hebrew
NarrativeAn Explanation and a Manual for Compilation' (Trinity Journal 11
[1990], pp. 73-102) is a bit of a rarity, in that its goal is to present a methodology to
the reader, so that one can retreat to the study and 'do it' after him (this methodology
is computer-based, however, and requires materials prepared by J.H. Sailhammer of
Trinity Evangelical Divinity School, Deerfield, Illinois, which restricts its
practicability). This study makes the same mistake of lumping all 'discourse' material
into a single category; in the context of pointing the reader in the direction of
language universals, den Exter Blokland writes, 'Linguistic reseach of the past thirty
odd years has shown that texts themselves show patterns in their composition that
arise from the nature of the medium language [I suspect his point would have been
more clearly made had he enclosed in quotation marks the word 'language'] rather
than from text-genre combined with the author's literary freedom alone. A case in
point is the fact that in many languages the distribution of tenses differs according to
whether the text-type is discourse or narrative. There are rules of language that
determine in what tense a sentence will be cast, depending on whether the sentence
forms part of a discourse or a narrative text-segment, whether it is background or
44
The difficulties alluded to in this reference to 'dialogue' settle calmly

into place with a proper appreciation of the influence text-types exert
on linguistic material. This subject will occupy our attention at
considerable length in Chapter 3.
Thirdly, Eskhult no more than cursorily evaluates the range of
possible features that could replace his chosen clause-type as an
'Episode Marginal Circumstantial'.83 This is inadequate. Saturation
with a range of features is the proven tendency for marking the
macro-structure of a vast number of the world's languages (in
particular, those that fit into the bag-of-tricks category).84 Therefore,
any macro-syntactic markers in Classical Hebrew (since it is a bag-oftricks language) need to be examined in company with other options
that can perform the same function. Eskhult attempts to discern the
macro-syntactic contribution of the (we)subj-qatal clause on the basis
of its aspectual identity, without doing a thorough survey of its
distribution as one of a collection of forms that can perform the same
function, and therefore he presents a rather limited perspective on the
matter.
In Mary Shelley's Frankenstein, to take a random example, we find
the following clause distribution:
foreground, whether the statement is general or specific' (p. 73). Here again, the
only substantial weakness in this statement is the definition of 'discourse' as a texttype parallel to 'narrative'. The dismantling of this presupposition is one of the main
goals of the present volume.
83. Eskhult, Studies, p. 9, etfreq.
84. There is no compendium of 'text-level features of the world's languages',
and therefore these statements which refer to 'a vast number of the world's
languages' may seem to beg the question. It would be irresponsible of me, however,
to leave these claims totally unsupported, though I can direct the reader to no more
than a token substantiation. In Mandarin (Chinese), for example, peak events are
marked by a single particlethis is not a bag-of-tricks language; in colloquial
American English (I have noticed this feature in Britain as well, though it seems less
socially acceptable there), a story-teller may shift from simple past to (historical)
present tense, to highlight the peak episode of a story; verb saturation, staccato clause
structure, and other such features may also mark the peak episode, but none of these
is required as is the particle in Mandarin. These languages in which one particle has
one function, which none other can replace, are documented from around the world,
in linguistic journals and monographs. Beyond this very short summary, I must refer
the interested reader first to Longacre's Grammar Discovery Procedures and
Grammar of Discourse, then to the language-specific literature listed in his
bibliography, and those of others.
1. Introduction
45
a long string of 'simple past' clauses;

a string of 11 clauses built on the verb 'to be';
a long series of clauses with 'past perfect' forms, and the occasional
'modal clause' ('that I should...') and 'stative clause';
then, another long string of 'simple past' clauses.
This is the distribution of clauses found at the end of Chapter 2 and

the beginning of Chapter 3.85
The readers know that they have come to a significant break in the
action of the story; they do not tend to register consciously the fact
that these 'to be' clauses provide stative description, or that the pastperfect clauses are requiring them to examine prior events. It is
possible (if not probable) that the same unconscious relationship to the
syntax (macro-syntax) of this material is true of the author as well.
An analysis of the aspectual values of these clauses is illuminating, but
it is not required in order to understand the macro-syntactic
significance of these features. It is clear, in addition, that a study of
only the stative clauses, in this example, will not lead us to a very
well-rounded understanding of the macro-syntax of this textnor, for
that matter, will it tell us all that much about the stative clauses, since
we do not gain an understanding of how they work with other clausetypes to perform their own macro-syntactic functions.
I hold to a basic assumption that informs and controls my own
approach: when assessing the use of a specific linguistic feature, it is
important to seek first its strongest motivating or conditioning factor;
this will provide the observer with the clearest and most uncomplicated data from which to describe the particular feature. It is possible
that Eskhult and Niccacci would affirm this as well, but their
theoretical starting point directs neither of them to all the possible
motivating factors.
A goal of any linguistic description is 'elegance'that is, a linguist
seeks to describe data accurately and fully, yet as economically as
possible. An accurate and readable analysis of three pages in length is
considered more elegant than an equally accurate, and equally readable,
analysis of thirty pages in length. This goal of an elegant description
reflects an underlying presupposition: that the human brain requires
organization in order to encode thoughts into languagewe are not
capable of synthesizing vast quantities of only marginally related data
85. The chapter break occurs after the 11 'to be' clauses; cf. pp. 41-42 in the
Oxford University Press's 'World Classics' edition (1980).
46
without some sort of organization being applied to itand, therefore,

that each language can be reduced to its principles of organizing raw
data. Both Niccacci and Eskhult are shy of the mark with regard to
this elegance, in that they have overlooked factors that would have
streamlined their results.
Eskhult's work does contribute to our knowledge of the Hebrew
verb; it does not, however, greatly advance our understanding of
Hebrew macro-syntax. Analyses such as this, which focus on the
inherent meaning of a form,86 are essential for an understanding of
the Hebrew verb. However, the inherent meaning of a form should in
theory be described only in conjunction with the contextual meaning
and function. The inherent sense of the verbless clause, for example,
is exactly that which makes it the most appropriate form to use for
Expository texts,87 and the inherent sense of wayyiqtol forms make
them the most appropriate for narrative material.88 When a verbless
clause is used in an expository text, it should raise no eyebrows; it
imports little or no added nuance to the text. However, if a verbless
clause occurs in a narrative text, this should incite comment, since it is
very different from the standard form, and thus is used to contribute
something to the text, by virtue of its being 'unexpected'.
When a writer seeks to explain certain syntactic features, but does
so only from the perspective of the inherent meaning of that feature
or form, regardless of its function in the text, we do not get a true
picture of things; it is untrue that a syntactic feature always imposes
its inherent meaning upon a text (for the reasons stated immediately
above), and therefore an analysis from such a perspective is
misleading in that we are told that these nuances are present in all
occurrences of the feature or form being examined.
4. Andersen: The Sentence in Biblical Hebrew
What has been said in the introduction to my analysis of Khan
regarding the narrowness of his topic, and the thoroughness with
which he describes itcan also be said of Andersen, although I must
open my remarks on this landmark in Biblical Hebrew studies with the
86. As opposed to its function in its context.
87. See Chapters 3, 4 and 5, for a fuller discussion of this text-type.
88. Or it might be the evolved function which has given rise to the significance of
a particular form, though this is less likely... yet another 'chicken-or-egg' situation.
1. Introduction
47
comment that the average hebraist will find this one of the most
unreadable books available to the Hebrew scholar.89 For example,
where the author begins to explain 'deep' grammar versus 'surface'
grammar, the text reads as follows:
Consider the phrases in English:
a man,
whose name is/was Job
the king, whose name is David
Esau,
whose name is Edom
you,
whose name is Yahweh
These all have the structure:90

Nph
(antecedent) Head: NEq<A>Modifier:RelClJ
This is a 'Tagmemic' encoding of a description like the following:

'this type of Noun Phrase is composed of an antecedent (that is, a
definite or an indefinite article, or the like), which is optional,
followed by the 'Head' of the phrase (that is, the core unit), which can
89. One of the benefits of the 'Tagmemic' school of linguisticswhich has given
us both this book, and Khan's, and Longacre's various works, among othersis
that it is empirically rooted. It has provided a systematic theory which has opened the
door to a great deal of understanding of language; its theoretical underpinnings are
sufficiently economical, yet all-encompassing, to permit all facets of a language to be
assessed in relation to all other facets of that language (from extra-linguistic signals,
such as gestures, to the structures of the largest bodies of text, and again to the most
minute details of pronunciation). The drawback is that such a wealth of interrelating
information requires to be put into print, and is often summarized in very dense
jargon; linguists familiar with this theoretical system tend to try to include as much of
this related material as possible in any given analysis, and (always trying to produce
the most 'elegant' solution possiblewhere 'elegance' sometimes seems to refer to
least number of pages rather than merely clarity and economy of logic) they tend to
try to present it in a minimum of space. This leads to a vast amount of material being
presented in a short-hand form resembling mathematical calculusit is not
coincidence that Andersen's and Longacre's works are alike in this.
90. Andersen, Sentence, p. 29.
48
be composed of a Noun or its equivalent (a pronoun, personal name,

etc.); this is followed by a modifier, which is composed in this
instance of a Relative Clause, and which is tied together with the Head
section by grammatical agreement'.91 While this approach provides
the benefit of an enormous amount of linguistic analysis in a nutshell,
it is all but impenetrable to one who has not mastered the theory with
its idiosyncratic short-hand.
This is a significant drawback, yet Andersen's book has nontheless
had widespread effect on Hebrew studies; it is cited in nearly every
text dealing with the syntax of Biblical Hebrew since its publication. It
is an example of the highest calibre of scholarship and of scholastic
integrity, informed by language universals, and thus enabled to
describe in penetrating detail a particular level of the Hebrew
language. However, its transmission to the average interested hebraist
has not, altogether, been a success.
Andersen engages in both 'from the bottom up' description and
'from the top down' descriptionthat is to say, he describes the
Biblical Hebrew sentence (1) in terms of what units make it up, (2) in
terms of what other options might replace it in the various slots which
it can fill. In addition, he describes it in terms of what effect it has on
its wider context or contexts.
One of the great strengths of this book is that it does recognize the
strong influence of text-type on other levels of the language, for
example the author frequently deals with 'precative discourse'92 and
'predictive discourse' in separate sections, since to mix them together
would not allow certain patterns of usage to appear. In addition, the
reader will find practical justification for this kind of research in
statements like the following:
The connections secured by the hierarchical grammar of Hebrew
discourse are shown on p. 138. The example is instructive, because the
results are different in translation when these structural signals to the
higher levels of discourse structure are not heeded. RSV, to look at only
one attempt, takes considerable liberties with the conjunctions, thus
throwing the paragraphs into quite different arrangements, breaking some
91. Chapter 3 will deal at length with the principles and particulars of Tagmemic
linguistic theory, but this volume will not require the reader to 'learn' any of its
idiosyncratic calculus.
92. Andersen uses the term 'discourse' to refer to a longer, structured, section of
text.
1. Introduction
49
close linkages, and creating others where there should be a break. It omits
the conjunction at the beginning of 8.18, making a break where there is
sequence. It misses completely the trio of clauses governed by pen, which
as we have seen, are unified by chiasmus. Instead, it adds a gratuitous
beware lest to verse 17, severing its sequential connection with verse
14b.93
Andersen writes,
Without explicit and methodologically rigorous definitions of basic units
and relationships the classification of a linguistic datum remains whimsical,
and the same clause will often be described differently by different
writers, with no discussion of the reason for doing so.94
In his opening pages he provides a detailed introduction to his topic

(the sentence), giving definitions arising out of concerns raised by
modern linguistics. So, rather than saying that a sentence is 'a
complete thought expressed in words' (which definition is found
wanting on several counts by modern standards), he sets out what
exactly can and cannot be termed a sentence for purposes of linguistic
assessment.95 This section is not vastly easier to understand than the
remainder of the book, but amply repays careful study. Working
through this material gives the reader great insight into the kind of
analysis of data, and the kind of language description, that one must
cultivate in order to do justice to the task we undertake.
The main impediment in this book is that it is exceptionally difficult
to read. One is constantly faced with the difficult task of trying to
decipher tagmemic calculus; and there are minor weaknesses where a
particular remark is not clarified for the uninitiated, for example,
'such a participial phrase (incomplete clause) generally functions as a
NOUN EQUIVALENT (NEq); the participle is verbal within the
phrase, nominal outside it'.% The reader is left to divine the sense of
this summary, and while the meaning is not impenetrable, it is
difficult to determine that one's interpretation is exactly what the
author had intended.
This book does not require further attention in this particular work,
as the shortcomings of the work are immediately apparent, and the
95. Waltke and O'Connor, Introduction, pp. 77ff., borrow heavily from
Andersen in their discussion of the same topic.
50
assets of the work have long been recognized. In addition, the principal
relevance for the present study is its incorporation of the 'higher
levels' of the linguistic hierarchy, and its theoretical foundations, and
to these latter we will return for a more penetrating discussion in
Chapter 3.
5. Conclusions
In this chapter I have introduced the general plan of this book: its
concern that contemporary works on Hebrew text-linguistic studies
fail to complete the task of communicating solid results in a form that
is accessible to the readership, and the intention to provide an
evaluation of some of those works, an introduction to text-linguistic
theory from the 'Tagmemic' perspective, and a presentation of some
worked examples as an illustration of a methodology derived from
that theoretical base.
In addition, I have presented an analysis of several recent Hebrew
language studies. Earlier in this chapter I used the metaphor of a
bridgethe two sides of a chasm representing (a) the material to be
conveyed, and (b) the readership, the bridge itself representing the
author and the workto illustrate the two elements of the
communication process I would be evaluating in each of our authors. I
have found each of them wanting to some degree with regard to one
or the other of these two elements.
This, I want to emphasize here, does not mean I reject these works,
nor that I do not appreciate their worth; I have found each of them
helpful, and many others as well, whose weaknesses and failings have
been far more significant. The choice of these authors' works was not
because they are particularly glaring examples of these 'weaknesses',
but rather because they are so close to a proper balance. Permit me an
illustration from a different realm altogethera sporting competition,
one in which the performance is measured aesthetically: In a 10-metre
diving competition, for example, two divers might do an identical
dive with exactly the same execution, both making the same error
(say, a large splash instead of a small one, on entering the water); the
only difference between them is that one of the divers is more
graceful in the air than is the other. In theory, the more graceful dive
should receive the higher award. In fact, this is often not the case, and
the less graceful one often receives the higher award. The reason for
1. Introduction
51
this is that the errorthe large splashis the more obvious in the
dive which is the closer to 'perfection'. That is, the more perfect the
'ensemble', the more striking are the exceptions to that perfection.
And that is the case here: the books examined in this chapter and the
next are among the best text-linguistic studies available to the hebraist.
Their inadequacies are the more obvious because they are seen against
the background of their significant contributions to Hebrew studies.
Niccacci's book has been welcomed as an excellent study on Hebrew
macro-syntax. Its major failing is its difficulty in handling 'discourse'
material; the category is too broadly denned, the patterns within it do
not emerge, and false conclusions are drawn. Eskhult shares the same
weakness, but the final result is slightly different. Where Niccacci
concludes that 'discourse' is a category with rules different from those
of narrative, and much more nebulous ones, Eskhult concludes that it
is the inherent 'aspect' of a form, rather than a shift to 'discourse'
text-type, which governs the contribution a particular form makes to
the 'whole' in which it is found. Both authors miss out principal
conditioning factors, and focus on secondary ones.
Andersen, on the other hand, does not exhibit theoretical weaknesses, but rather 'fails' to communicate his results clearly. His
impressive work is all too readily shelved without thorough study
because it demands so much of the reader. Linguists with a rich
background in theory often forget that the average reader will not
understand their jargon, and will therefore not benefit from the
exceptional terseness that is frequently the goal of linguists. A treatise
on putting on a pair of socks, or unstopping a blocked drain, could be
written in such a scientific way as to be utterly incomprehensible to
the untrained readerand, at least in the case of the blocked drain,
the untrained person would be the worse off. If research is intended to
make a contribution to the non-specialist's world, it must be
communicated in a way that makes it accessible. Andersen certainly
cannot be singled out as the worst offender in this, but it is a mark
against this book that it does not communicate readily to those who
would benefit most from its insights.
In the following chapter I will present analyses of two more works,
one of which is a model of good scholarship and good communication,
but which only lightly touches on our topic, and the other, which is
also a model of good scholarship, but which is less than completely
successful in communicating to Hebrew scholars.
Chapter 2
EXAMINATION OF KHAN'S 'STUDIES' ANDLONGACRE'S 'JOSEPH'

1. Introduction
I move on now to two other recent works: Khan's and Longacre's; I
have mentioned above my reasons for examining them together.1
Khan's work will not occupy our attention for any great length, since
his topic does not feature highly in the data that I myself examine;
Longacre's will require the bulk of this chapter.
2. Khan: Studies in Semitic Syntax
This study adopts a 'comparative Semitic' approach, examining data
from Arabic, Biblical Hebrew, Biblical Aramaic, Syriac, Akkadian
and Amharic. The breadth of data consulted is compensated for by the
narrowness of topic: Khan treats only the subjects of 'extrapositional
sentences (Ex)' and 'pronominal agreement constructions with resumptive pronouns (PAR)', which he defines as follows:
By the term 'extraposition' I understand the syntactic construction in
which a noun or nominal phrase stands isolated at the front of a clause
without any formal connection to the predication. The initial 'extraposed'
nominal is not adjoined to any relational particle such as a preposition or
an object marker and in those languages which have a case inflection it is
generally in the nominative. The grammatical relation of nominal in the
predication is usually indicated vicariously by means of a co-referential
resumptive pronoun...
1. In brief, I consider these two be the best examples available to the hebraist of
works with a positive view of text-linguistic researchKhan, for his clarity of communication, and integrity of methodology, and Longacre, for the rich potential of his
theory and methodology, and the astute insights he offers into the text-level structures of Classical Hebrew.
2. Examination of 'Studies' and 'Joseph'
53
'Pronominal agreement' is a construction where a noun or nominal

phrase whose grammatical relation is indicated by its case inflection or by
an adjoining relational particle is accompanied in the same clause by a coreferential pronoun agreeing with it in number, gender, person, and grammatical relation.
Unlike extraposed nominals, nominals which are accompanied by such
'agreement pronouns' are not restricted to initial position but may occur
anywhere in the clausethe front, the interior, or the end...
Pronominal agreement constructions in which the 'agreed with'
nominal occurs at the front of the clause are closely allied to
extraposition.2
This book is well-organized, and likewise very well-written. The

introductory chapter opens with some general statements on language
functions, then moves on to definitions of terms, choice of languages
(and sources from which data were taken), a brief yet thorough examination of previous literature on these two constructions, and a section
on contemporary linguistic theory. Khan then outlines his own presuppositions and methodology,3 with definitions given for more specific technical terms. The term 'discourse' is once again in prominence, and in this case it is defined obliquely as a 'certain stretch of
text'; 4 and although Khan does not clarify this further, it is to be
understood as a rather vague category of a large body of texthis
term 'span of discourse' seems more akin to Longacre's 'discourse'.5
One of the points made in Khan's section on methodology is that,
with each separate language, he divides his treatment of these constructions into two sections: the first deals with the internal structure
of the constructions, and the second deals with their functions within
their wider context. Each of the languages mentioned above is
examined in this manner, in chs. 1 through 5. This is one of the
greatest strengths of this book: Khan separates the 'observable' from
the 'interpreted'.
The book concludes with a chapter on Comparative Semitic Syntax,
which examines the principles and merits of this discipline. Khan
writes,
2. Khan, Studies, pp. xxvi-xxvii; Khan's n. 2 adds here, 'Indeed in some cases
the two constructions are indistinguishable, cf. 108, 160'.
3. Khan, Studies, pp. xxxiii ff.
4. Khan, Studies, p. xxxiii.
5. Or my 'text'; see below, this chapter, and the following.
54

Comparative Semitic syntax is not only of interest to the general language
typologist but is also of direct relevance for the Semitist in so far as it
helps to clarify the understanding of features in the individual Semitic languages.6
Khan's work is restrained, and, although he does not claim to be

describing the macro-syntax of any of the languages he works on, he
describes very well the text-level functions of these features. Had he
chosen to work 'from the top down',7 he would have had to consider
many more syntactic features than he has done, describing each as an
option available for use at the higher levels; however, working 'from
the bottom up' 8 allows him to describe only these particular features;
he need only describe the function of his chosen constructions, noting
their effect onand their function inhigher-level units. One might
say that Niccacci, for example, is engaged in this same 'from the
bottom up' sort of work; he, however, has set a much broader scope,
attempting to describe the macro-syntax of the language by describing
its micro-syntax, while Khan, on the other hand, describes a specific,
limited, and closely related pair of features in terms of both its structure (micro-syntactic components) and its function (which includes its
macro-syntactic significance).
Khan moves very well from micro-syntax to macro-syntax. The
organization and lucid style of this work contribute a sense of 'flow'
from the level of internal structure to that of contributions to macrostructure. Khan looks first at the bits and pieces making up his two
contructions, and then looks at how these two constructions become
'bits and pieces' which help to make up other, larger, constructions.9
One is instructed first in how a certain drawer is made, and then how
that drawer forms an integral part of the desk.
6. Khan, Studies, p. 233; I believe that the same is true of what I have been
calling 'language universals', for the more we understand the working of the human
mind with regard to its use of language, the better we will understand the vagaries of
Classical Hebrew.
7. That is, were he to start with a text or a text-type (both of which I am doing),
and describe the variety of macrosyntactic features found therein, then moving from
this level to the level below it (to paragraph), and to the one below that (to sentence),
and so on.
8. Starting, in this case, at the clause-level, and moving from there to sentence,
and paragraph, and beyond.
9. This is one of the basic concepts of 'Tagmemic' linguistic theory; it will be
unpacked in Chapter 3.
55
This is a model presentation of research; it demonstrates clearly the

elegance of description that is the ideal of a descriptive linguist. The
clarity of this book, and the thoroughness of its treatment of this
topic, particularly with the comparative methodology, are its principal
strengths. If one must cite a draw-back, it would be that the topic is
narrow, and provides little general insight into the system as a whole,
in any of the languages examined; however, by contrast with Niccacci,
for example, Khan does not attempt this goal. Would that this were
but a chapter in a much larger work detailing a wider range of syntactic and macro-syntactic features.
Ex and PAR are constructions that can be considered 'exceptional'
clause-types in the same sense that a Hebrew noun phrase that includes
an attributive adjective or participle can be considered exceptional;
both are developments away from the simplest and most common
constructions of their type. Since these constructions occur but rarely
in the data that I examine, there is little overlap between Khan's work
and my own.10
Khan notes that Ex and PAR 'perform the same functions and are in
most cases interchangeable', and that 'one of the most widespread
functions of Ex and PAR clauses is to mark the boundaries of spans of
discourse. They coincide with either the onset or the closure of a
span.'11 My own findings agree with these conclusions: exceptional
clauses (of which Ex and PAR clauses are examples) always mark in
some way an interruption of the continuity of the material being presented. The effect that these particular constructions produce does not
appear to be linked to the text-type in which they are used; this may
be traceable to the fact that they are exceptions more to microsyntactic than to macro-syntactic (and, therefore, text-type-based)
standards.
Khan's Table of Contents is perhaps the best summary of his
approach to this material;12 his excellent organization allows the
reader to see at a glance the way he works through the functions of
these two constructions in Biblical Hebrew, and his awareness of other
macro-syntactic structures:13
10. Therefore my principal interest in this book, here, is with reference to its
methodology and its presentation.
12. In particular, Khan, Studies, pp. xv-xvi.
13. Cf. pp. 86ff. (sections II. 1.3 under 'Biblical Hebrew'); I find his interest in
56
Khan adjoins an appendix dealing with the special case of the use of
these constructions in 'legal precepts', and writes,
Extraposition occurs particularly frequently in the structure of legal formulae in the Old Testament. This is also the case with regard to post-Biblical
law corpora which were composed in Hebrew, e.g. the Qumran text serek
hayyahad (The Rule of the Community) and the halakic works of the
Tannaim. It is convenient, therefore, to devote a separate section to extrapositional structures which are characteristic of this genre of text, bringing
together for the sake of completeness both Biblical and post-Biblical law
formulae.14
Khan has done well to separate out this particular context for special
attention owing to the frequency of occurrence in these texts. However, I am not convinced that this actually forms an alien usage of this
construction; rather it seems to me simply another context in which
the construction performs the function of 'change of topic'a
function he has noted and described on pp. 79ff. Owing to the relatively 'topic-intensive' nature of legal material,15 the topics of legal
regulation change from one to the next fairly rapidly; that Ex and
PARpossible topic-changing mechanismsare found in greater
frequency in legal material than in other text-types is not enormously
surprising.
Khan's work is a model of good scholarship and excellent presentation. I can find no flaws in it that impede its usefulness. We can hope
that such successful research and communication becomes more and
more common in our literature.
3. Longacre: Joseph: A Story of Divine Providence: A TextTheoretical and Text-Linguistic Analysis of Genesis 37 and 39-48
This book represents the most significant advancement in Hebrew
textlinguistics seen to date; it contains much of near-revolutionary
value to the student of Classical Hebrew syntax. Several of Longacre's
comparing the features on which he chooses to focus, with those which can perform
the same function, to be a great strength of his approach. We will look at this
methodological principle in greater detail in the rest of this volume, in particular in
Chapter 3.
14. Khan, Studies, p. 98.
15. In the sense that it tends to incorporate a lot of information into as little space
as possible.
2. Examination of'Studies'and'Joseph'
57
contributions are of the sort that are immediately accessible to the

readership, and which prepare the way for a much clearer (more
'elegant') description of the language. Others are more esoteric, but
equally important.
On the other hand, this volume suffers from some of the same difficulties as does Andersen's Sentence, in that it contains so much information that it becomes difficult to sort through. Where Andersen's
difficulty, however, was that of relying too heavily (for the average
reader) on linguistic jargon, Longacre's is more a matter of leapfrogging over too many steps in the theoretical logic, thus leaving
behind all but the most astute reader, and fellow linguists (although
even fellow linguists may not follow, if they are not familiar with the
tagmemic approach to language).16
3.1 General Background to the Book
At the outset of this examination of Joseph, it will not be out of place
to mention one of the reasons why this work is particularly valuable
a reason that is not related to the study of Hebrew in particular. Much
of the development of the tagmemic school of linguistics has taken
place 'in the field'. Literally. The three or four primary movers in
this theoretical development have spent, between them, well over a
century (a most conservative estimate) working among indigenous
populations, and with previously unanalysed language materialand
following this, perhaps the double of that in processing data from as
wide a range of the world's languages as possible, for theoretical purposes. Tagmemics, as an approach to the data, developedfrom its
'birth' in the fifties17out of dissatisfaction with the fact that previously accepted theoretical models of language analysis did not deal
well enough with living language data. A new model, which had an
increased capacity to 'learn from' discoveries in real language data,
was developed, which worked from simple basics, but which allowed
for the description of finer details of language use. This capacity to
16. Longacre does include a 3-page appendix on tagmemic theory (pp. 311-13),
from which I will quote at length in the next chapter, but this is inadequate to the task
of 'educating' the readership up to a standard where the book becomes easier to
follow.
17. K.L. Pike's Language in Relation to a Unified Theory of the Structure of
Human Behavior (first published in 3 vols., 1954-1960, Summer Institute of
Linguistics) is the seminal work of this approach.
58
'learn from' real language data has produced a theoretical model that
finds the basic patterns, and the permutations thereof, within any
language studied; and the results of these analyses are particularly
well-suited for comparative linguistic research. The end result is that
'language universals' are constantly compared with data from specific
languages in a way that advances the study of both. Longacre has long
been at the forefront of this development, and is without doubt the
text-linguist with the greatest exposure to the breadth of the world's
language data, of anyone in print today.18
A sample of Longacre's collation of language data resulting in an
appreciation of language universals is the following:
The successive events of a narrative paragraph may be given as a series of
what I call build-ups, each in a separate sentence, or the whole narrative
action may be expressed in one long run-on sentence with but one final
verb at its end, i.e. in a one-sentence paragraph. What is the rationale of
this choice? Is this pure caprice?
At this point the study of whole discourses is helpful. In fact, a perusal
of the Wojokeso corpus of text material (Longacre 1972a, Text)19 suggests a resolution of the problem. There is a narrative discourse which has
narrative paragraphs composed of fair-to-middling-length sentences until
one reaches what is really the denouement of the whole story. At this
point, we find a long run-on structure in which all the events are lumped
together in one paragraph-length sentence. Similarly, we have a procedural discourse on housebuilding, in which likewise we find sentences of
fair-to-middling length until we reach the target procedure where the
house is finished and the couple move in to spend their first night in it.
Here again we find a long run-on one-sentence paragraph in which all the
steps of the paragraph are in one page-length sentence. Sentence is here
being used in both narrative and procedural discourse to mark the peak of
the discourse in the surface structure, which corresponds to either
denouement in narrative discourse, or to target procedure in procedural
discourse.
18. This is not an idle claim: Longacre, after working for many years in the field
on the languages of indigenous people groups of Mexico, has gone on to work with
Pacific, North American and African languages (in each case, he has worked with
not one, but scores, of languages from these regions), in addition to the biblical languages; the bibliography in Joseph lists 20 articles and booksonly a fraction of his
publications.
19. I.e. Longacre's Hierarchy and Universality of Discourse Constituents in New
Guinea Languages. I. Discussion. II. Texts (Washington: Georgetown University
Press, 1972).
59
Parallels are not lacking elsewhere. Thus, Charles Green has pointed
out to me that a not dissimilar phenomenon is found in Hemingway's
story, 'The short happy life of Francis Macomber'. Here, at the climax of
the story where the main character of the story is shot in the back of the
head (accidentally?) by his wife, we find a long run-on, rollicking
sentence not unlike in kind from what we have mentioned in the
Wojokeso discourses. Something similar is found in the text of the Greek
New Testament. We find in the account of the feeding of the five
thousand (Matthew 14:13-21) an absolutely unparalleled string of
participles in sentence-initial position precisely at the denouement of the
account (Matthew 14:19), where Jesus takes the loaves, multiplies them,
and feeds the people.20
My praise of Longacre's depth of experience and knowledge may

seem out of place; his credentials, however, are not likely to have
preceded him very far into this discipline (as he writes primarily for
linguistics publications)though not for want of influenceKhan, for
example, writes,
My methodology has been inspired mainly by the work of Joseph
Grimes, Robert Longacre, and Teun van Dijk. The value of their approach
to the analysis of discourse is that it is largely empirical. This applies
especially to the work of Grimes and Longacre, who have both carried out
extensive field work in a large number of language communities.21
Most of Longacre's writing has been for linguistsin particular, field

linguiststhus it is not surprising that few hebraists are aware of his
work; thus, the idea that this book needs to be given considerable
attention will not be immediately evident to the average biblical
scholar.22
20. Longacre, Discourse, pp. 30-31.
21. Khan, Studies, p. xxxiv; In a note on this section, Khan writes, 'Andersen's
study of the sentence in Biblical Hebrew was in fact based mainly on the methodology which was developed by Longacre'; he states later, in the text, 'I have, however,
rejected the more abstract methods of discourse analysis, many of which are still
embroiled in theoretical preliminaries and show little concern for the direct analysis of
texts' (Studies, p. xxxiv, reflecting in his own approach the priorities of the
Tagmemic school of linguistics).
22. In fact, only one of the entries under his name in the bibliography of Joseph
was published in the literature hebraists are most likely to consulthence my
'touting' of his background and achievements: although his is not a name on the lips
of most biblical scholars (they will not have seen much of his published material), he
is one who has earned the right to speak with authority on the subject at hand.
60
3.2 Contents ofLongacre 's Joseph

Joseph is organized into four parts: Part One introduces the text-unit
(Gen. 37 and 39-48) and gives the broader interpretative and textlinguistic features of the story; Part Two looks at the unit in terms of
textlinguistic observations (observations more specific to the language
features, such as 'off-the-line' material, TH and its text-level functions,
etc., than his comments in Part One) and sociolinguistic observations
(relating primarily to questions of mitigation and deference in
Hortatory portions of Reported Speech); Part Three focuses on
Participant Reference, and Dialogue; and Part Four is a presentation
of the entire text examined in this study, presented in a typographical
arrangement intended to highlight the text-linguistic features noted in
the preceding analysis (particularly those commented on in Part Two).
This is followed by a short appendix on tagmemic linguistic theory.
3.2.1 'Pan One'

The broader movement of this book is discernible from the Table of
Contents, once the intricacies of the individual chapters are understood. Longacre starts by setting the broader contextthe whole of
the historical material, from Genesis 1 through to the end of
2 Kingsand presenting part of his methodology (Introduction), then
he moves to the narrower context of the nnbn of Jacob.23 Here he
looks at the articulation of episodes in the Joseph story, examining
devices that signal the boundaries of the major sections (ch. 1). In
ch. 2 he looks at 'macro-structures'by which he means 'an overall
meaning and plan'.24 In ch. 1, then, he presents the surface structure
organization of the story, and in ch. 2 he looks at the underlying flow
of issues, meaning and plan.
23. His understanding of the way the Joseph story fits into its nearer context
leads him to leap over the Judah and Tamar section (Gen. 38)an approach that will
not appeal to all; in addition the explanatory material concerning his decision is pared
to a minimum, where it would have contributed to the reader's understanding of the
material that follows to know how Longacre has come to the conclusions he does.
24. Longacre, Joseph, p. 43; I use this term to refer to the constituent structure of
the largest units of language, referring to 'episodes' of a story, etc.but in the sense
of the surface structure itself (paragraphs, etc.), rather than the 'notional' or 'deep'
(cognitive) level. To refer to what Longacre terms 'macrostructure', I would probably borrow his alternative expression: 'overall meaning and plan'. His term is
derived from van Dijk's work (see his 'References' section, p. 316).
61
3.2.2 'Part Two'

In Part Two, Longacre presents a more meticulous analysis of the
surface structure of the language, and here we begin to see the revolutionary ideas he has to offer. His summary of the assumptions and
conclusions of this section is unusual, as it approaches the grammar of
this language in a manner significantly different from that seen in the
majority of current Hebrew language studies:
Typically, within a grammar of a given language all the uses of each
tense/aspect or mode of a language are listed and described en bloc in the
same section of the grammar.. .Part 2 of this volume is, among other
things, a challenge to this time-honored way of describing the functions of
the verb forms of a verb system within a language. Rather, I posit here
that (a) every language has a system of discourse types (e.g., narrative,
predictive, hortatory, procedural, expository, and others); (b) each discourse type has its own characteristic constellation of verb forms that
figure in that type; (c) the uses of [a] given tense/aspect/mood form are
most surely and concretely described in relation to a given discourse type.
These assumptions inform chapters 3, 4, and 5 of this volume, where
they are illustrated in regard to narrative, predictive, expository, and
hortatory discourse.
The constellation of verb forms that figure in a given discourse type are
structured so that one or more privileged forms constitute the mainline or
backbone of each type, while other forms can be shown to encode progressive degrees of departure from the mainline. This is developed empirically in chapter 3 in regard to narrative discourse and the verb forms that
characterize it. Here the wavv-consecutive imperfect is seen to be mainline
in that it is punctiliar, and sequential in function; the perfect is found to be
(as a whole) a non-punctiliar and non-sequential kind of past tense in narrative; the imperfect and the participles are, respectively, implicitly and
explicitly durative in the framework of the story; haya clauses and
verbless clauses represent stative elements toward the bottom of the
scheme; and negated clauses rank lowest.
In regard to these various matters, Biblical Hebrew can be shown more
and more to be a rather run-of-the-mill example of a language with a
special narrative or 'consecutive' tensea statement that can be easily
documented in regard to the narrative structures of a variety of African
languages (cf. Longacre, in press). Some of the latter have consecutive
forms that necessarily depend on a special initial form that must precede
them; other such languages simply have a special consecutive form that
need not have a special initial form to initiate the chain in narrative.
Biblical Hebrew belongs to the latter type of language, but traditionally
has been described as if it belongs to the former type. Thus the legend has
62

grown up that a wow-consecutive perfect must similarly follow an
imperfect (or some other tense/aspect/mood form).25
Longacre's ch. 3 discusses the different kinds of clauses one can find
in Narrative History,26 dividing them into 'on-the-line' and 'off-theline' options. Longacre writes,
Discourse grammarians are coming to recognize more and more that in the
telling of a story, one particular tense is favored as the carrier of the backbone or storyline of the story while other tenses serve to present the background, supportive, and depictive material in the story.27
I have assumed, then, that the storyline or the backbone of a discourse in
Biblical Hebrew is conveyed by the use of clauses that begin with a wawconsecutive verbin the balance of this book simply called the
preterite.28
Included here is a reference to rrn, which is worth quoting in toto:

It is immediately necessary, however, to qualify the above hypothesis in
one important particular. The verb hay a, 'be', even in its form wayhi,
'and it happened', does not function on the storyline of a narrative. In this
respect, the behavior of Hebrew is similar to that of a great many
contemporary languages around the world. For example, English uses its
past tense to encode the storyline of a story, but the verb be (and some
other stative verbs)even when in the past tense (for example, forms
such as was, wereis typically descriptive and depictive and does not
figure on the backbone of a story. This is simply a peculiarity of the verb
be, in many languages past and present.29
Longacre adds to these insights discussions of permutations of the

mainline clause-type, then looks at off-the-line material, giving a
description of the potential offline clause-types for Narrative History.
25. Longacre, Joseph, pp. 59-60; it is strongly tempting to go onthe material

that follows this is equally significantbut I will discuss these elements in my
Chapter 3.
26. Longacre's term is simply 'narrative'I will use my own terms for texttypes throughout this examination of Joseph; they will be explained more fully in
Chapter 3.
27. Longacre, Joseph, p. 64; this comment develops from an allusion to GKC's
recognition 'that the so-called waw-consecutive is a special narrative tense' [in foe.].
28. Longacre, Joseph, p. 65; Longacre chooses the term 'preterite' in order to
disassociate this conjugation from the 'prefixal' conjugation, on the basis that the two
are apparently historically unrelated.
29. Longacre, Joseph, p. 66.
63
In sub-section 3. of ch. 3, we come upon 'a verbal rank scheme' for

Narrative History;30 this is one of the most immediately accessible
and revolutionarycontributions of the book. Longacre here
presents, on half a page, a summary of the way clause-types function
in Narrative History, in the form of a 'cline'31 (see below).
In the next two chapters, Longacre posits verb rank clines for
Predictive and Hortatory text-types, and describes verb ranking for
Expository, as well, though he 'declines' to posit a cline for the last,
citing the need for more analysis of Expository material in the
Hebrew Bible as the reason for his hesitation.32
NARRATIVE DISCOURSE VERB-RANK CLINE33
Bandl:
1. Preterite
Storyline
Band 2:
2.1 Perfect
Backgrounded Actions
2.2 Noun + Perfect (with noun in focus)
Band 3:
3.1 hinneh + participle
Backgrounded Activities
3.2 Participle
3.3 Noun + participle
Band 4:
4.1 Preterite of hdyd, 'be'
Setting
4.2 Perfect of hdyd, 'be'
4.3 Nominal clause (verbless)
4.4 Existential clause with yes
Band 5:
5. Negation of verb clause:
irrealis (any band)
'momentous negation'34 promotes
5. to 2.1/2.2.
30. Longacre, by contrast with Niccacci, writes, 'I do not find per se that the
grand dichotomy verb clause versus noun clause is useful. Rather I absorb it into a
rank scheme that can be thought of as the verbal spectrum for narrative... In this
scheme I assume a cline, a structural slope from clauses that are relatively dynamic to
clauses that are relatively static...' (Joseph, p. 81; cf. Niccacci, pp. 23ff.).
31. This he defines as 'a scheme symbolising degrees of departure from the
storyline' (Longacre, Joseph, p. 82 n. 6).
32. Longacre, Joseph, p. 107, for Predictive; p. 121, for Hortatory; and
pp. 111-12, for Expository. I will not include these charts here, as they will be
reproduced in the following chapter.
34. 'Momentous negation' describes the situation where the absence of a certain
event carries the narrative line forward (Longacre gives the example of the failure of
the dove to return to the ark in the Flood Story, Joseph, p. 85).
64
We now have the mixed pleasure of seeing this model become even
more complex, in that we can look not only at 'discourse' identities,
but also at paragraph identities. Longacre has isolated certain language
universals concerning the nature and functions of paragraphs,35 based
on their structures and functions, and he describes nine types of paragraphs (sequence, simple, reason, result, comment, amplification,
paraphrase, coordinate, and antithetical),36 each of which may be
encoded differently according to the text-type in which they occur
that is a narrative history reason paragraph will not be encoded in
exactly the same manner as an expository reason paragraph.37
Longacre's definitions of paragraph, and especially of sentence, are
distinctive enough to warrant mention at this point. Paragraphs are
identified as such by their internal structure: 'any group of sentences
that go together by virtue of cohesion and/or coherence can be shown
to have the structure of an (embedded) paragraph of a recognizable
type'.38 'Sentences' are 'the basic building blocks of the paragraph';
and 'a sentence in Hebrew is considered to be basically a unit with a
main clause (and a main verb), to which may be attached such subordinate clauses as adverbial clauses and relative clauses'.39
In the remainder of Part 2, Longacre offers examples of the paragraph types he has found in the Joseph story. This goes at a stunning
pace, and those not already familiar with Longacre's approach to
paragraph analysis and identification will be hard pressed to make
sense of it.40 In the long run, the section on paragraphs yields little
readily accessible material; and comprehension of this material is
35. Cf. his An Apparatus for the Identification of Paragraph Types (Notes on
Linguistics, 15, Dallas: Summer Institute of Linguistics, 1980).
36. See Longacre, Joseph, p. 85ff.
37. This stands to reason, for we cannot expect narrative history (which focuses
on the past) and expository (which is in essence atemporal) to rely on the same
structures.
40. This is more evidence that the book was written with a strong bias toward
communicating with linguists, rather than hebraistsanother detail that points in this
direction is that all the Hebrew is transliterated, and this is certainly not traceable to
publishing difficulties as was the case with Andersen's work (see Andersen,
Sentence, note, p. 16). But then, Longacre's goal here is to give us the benefit of his
best thinking on the subjecthe can hardly be expected to give a full introductory
course in general- and text-linguistics at the same time.
65
rendered the more difficult owing to lack of thorough explanation of

how paragraphs enter into the interworkings of clauses and texts,
which we see so succinctly and lucidly displayed in his 'clines'.41
3.2.3 Part Three
These verb-clines are a radical innovation in the description of the
Hebrew verb-system; Longacre, however, goes beyond the verbs to
other elements of Biblical Hebrew. With reference to 'discourse-level'
analysis, he describes the verb-system as one string in a 'doublehelix' 42 the other being composed of participant reference and
speech interaction.
Part One of this book dealt with macrostructures and the larger
context of the Joseph story; Part Two analysed verbs and clause-types
according to discourse and paragraph distribution and functionsthe
first strand of the double-helix. Part Three deals with the second
strand of the double-helix, and is structured as follows: ch. 6 looks at
participant reference (nouns, pronouns and verb affixes), ch. 7 at
variants in speech-introduction formulae, and ch. 8 at the role of
dialogue in the narrative. One might say that clause-types, paragraphtypes and text-types are the bones of the text, and participant
reference, speech formulae and dialogue are its soft-tissue.
The reader will do well to keep a finger at pp. 141-43 while
working through ch. 6, for, while the theoretical base of this section is
more accessible than that of the section on paragraphs, Longacre tends
to refer to it in rather abbreviated (if not jargonal) terms. In addition,
everything in this chapter refers to the 'level' of participant reference;
and yet (by oversight, apparently) these levels are never identified and
defined as suchthree lists are given; none of these refer, in their
titles, in their contents, and in the notes following, to 'levels' of any
sort; the reader eventually discerns that it is the first list that supplies
us with the appropriate key, but such opacities as these are not
required in an already difficult book.
The analysis is rather unremarkable except for one thing; the author
writes,
41. My own focus in this work is the relationship between clause-types and texttypes; I will very rarely refer to Longacre's approach to paragraph-types; though this
material is valuable, entering that domain in this volume would require several scores
of extra pages of theoretical material, and a similar amount of processed examples.
66

One cannot read Joseph43 carefully without being puzzled by the alternation between references on level 4 (pronouns) and on level 5 (verb
affixes) in relation to expressing the object (which is almost entirely a
matter of the third person object since Joseph is a narrative told in third
person). I propose here the view that resort to level 5 (the object suffix)
rather than to level 4 (sign of accusative 'et + pronoun) has to do not with
thematicity as such but rather with dominance patterns... Succinctly, level
5 is used rather than level 4 to express a dominance pattern in which the
participant(s) referred to by the object suffix is under the dominance of
someone else.44
After reaffirming his belief that this does indeed accurately reflect the
data in the Joseph story, he goes on to state,
Further research is needed to see whether this claim can be extended to
Biblical Hebrew as a whole. Whatever the outcome of this question, it
seems clear that this claim cannot be made in regard to certain non-narrative discourses, e.g., the poetry of the Psalms.45
Longacre's ch. 6 is 16 pages long; its contents are condensed from

what might have been 40 pages, were it written in 'plain English'.
This section is meant to prepare us to make use of the participant
reference interplay, along with the paragraph structures, the
'discourse-level' peaks and boundaries, and so on, of Part Four. It
seems to me an inadequate preparation for those not already schooled
in tracing participant reference and in interpreting it along with other
clues from the 'discourse-' or text-level; it is no easy task for one with
that background.46
Next, after leading the reader through some more rather 'jargonish' analysis in ch. 7 (where, once again, the crucial information about
the 'calculus' employed is difficult to locate; in this instance it is
hidden at the end of paragraph 1.2. on p. 162), the author gives a
concise summary of his findings about speech-introduction formulae,
which are derived not only from the Joseph story, but from other
material in the Hebrew Bible as well. This summary indicates the
43. He refers here to the Hebrew text of the story.
45. Longacre, Joseph, p. 157 n. 2.
46. As I have said, the author was not engaged in writing an introductory
textbook on analysis procedures, but it is nevertheless helpful when a writer conveys
sufficient introductory material to allow interested readers to follow the development
of ideas.
67
significance, for the narration, of speech formula variants.47 Once the

jargon has been gotten through, this material could easily inspire
hebraists to further examination of data.
His ch. 8 is perhaps not as revolutionary for Hebrew linguistics as is
the earlier material on verb ranks, but it is in its own way significant.
The chapter opens with this comment:
It is obvious, even on cursory examination, that much of the narrative we
have been examining is carried forward via sections of dialogue. A story
moves along not only by virtue of what people are reported to do, but also
by what they say. Saying is indeed a special kind of doing for which we
reserve the special name speech act.4*
Longacre has two foci here: (1) he examines each dialogue for its
internal nuances, and (2) he examines this material to see how it fits
into the overall structure and flow of the narrative. The significance
of this is that we come to terms with speech material as a narrative
device for the advancement of the story-line, in addition to considering it as a stylistic technique or simply as a reflection of speech patterns of daily life, which indicates conversations are narrative units in
the same way that simple narrative statements are narrative units.
The conclusion to this chapter serves, in a way, as a conclusion to
the whole book, for what follows. Part Four ('A Constituent Display
of Joseph'), and an appendix on tagmemic linguistic theory, are
somewhat loosely tied to the rest of the work. Longacre writes,
In backing away a bit from the mass of detail presented in this chapter and
in considering again the constituent structure of the story as a whole, I
note that the narrative sequence paragraph and the simpler sort of dialogue
paragraph carry in a somewhat routine way the burden of propelling the
story forward. Most other paragraph types (excluding probably the narrative amplification paragraph) have other more specialized uses. Among the
paragraphs that picture interaction patterns, the complex dialogue, the execution, and the stimulus-response paragraphs especially serve to underscore the more dramatic parts of the story. These paragraph types have
therefore been analyzed with special attention to details of their structure;
they are too important to the structure of the whole to be passed over
lightly and summarily.49
47. Longacre, Joseph, pp. 183-84.

68
3.2.4 Part Four

Part Four is Longacre's presentation of a formatted text of the Joseph
story, with its translation. If one has internalized the system and the
calculus of Longacre's approach to the textneither of which have
been presented here for the first-time readerthere is much to be
gained from studying this section. If one is not familiar with his
approach, this section is at best fascinating, and at worst, thoroughly
off-putting. In the main, Part Four is a graphic presentation of the
text according to the analyses sampled in the preceding pages; it lays
out the text in a fashion that allows the reader (if he or she has
digested the nightmarish formatting code) to see, at a glance, the text
in terms of its structure and sub-structures, as they have surfaced
through Longacre's analysis. This is a very useful tool to have at
hand, butas is an inherent danger with all reductions from verbal
explanation to graphic representationit is difficult to make use of
unless one has mastered theory in excess of what Longacre presents in
this book.
3.3 Conclusions
Much of what Longacre has presented us with is new, or, rather, has
not been seen before in publications intended for hebraists' eyes (it is
not by any means 'new' in linguistic circles). What we have seen
herethe concept of text-types, each with their own deployment of
clause-types; the necessity of describing a form in terms of its role in
specific text-types; the concept of mainline versus offline material,
with the identification of the 'preterite' as the mainline form in
narrative history, excepting such forms of rrnthese, as well as participant reference tracking, speech formula variations, hortatory variants, and other features as well, are relatively uncharted territory for
the hebraist. And, apart from a few shortcuts in explanations, and a
few oversights in formatting, the presentation is not exceptionally
difficult to get through, provided the reader takes the time to internalize the hoard of new concepts. There is an enormous amount of
material to get to grips with, however, and it is difficult (1) to make
sense of it as a whole, and (2) to know what to do with it afterward.
A part of me feels that the only response one can offer is applause:
it's been a very fine performance, and I shall remember it fondly, but
I couldn't hope to do any of it myself. And that, in fact, is one thing I
am not happy about, for text-linguistic analysis is not innately
69
difficult, and it deserves a warm recommendation to the hebraist. This

study is so far beyond the level of language work done by most
Hebrew scholars that its effect may be minimalized as a result of the
way it has been communicated. This is a book about a text, not a
manual for analysis. And while Longacre writes sufficient introductory material about his approach to the text to enable the reader to
gain at least cursory benefit from his book, it does not inspire the
reader with confidence to continue the work him- or herself.
This point is debatable, of course; yet it is disappointing that there is
nothing available to hebraists that equips them as Hebraistsrather
than forcing them out of their fieldto do the kinds of research that
they, as hebraists, are beginning to find intriguing. That is to say,
there is no doubt that linguistics is a science, and that thoroughly
trained linguists will inevitably produce more incisive works than will
the marginally trained linguistic dabbler; nonetheless, however, there
is ample room for linguistic research among people whose training is
primarily as hebraists. It is these, in fact, who often have the deeper,
intuitive, grasp of the sense of the text. Equipped with a manageable
theoretical base, they will produce solid, valuable linguistic research.
It is time for an intermediate literature, neither 'quantumlinguistics' treatises which are only understood by other 'quantum'
linguists, nor unscientific, non-linguistic treatises which serve only to
drive the study of Hebrew further away from helpful approaches to
the languagebut rather a literature intended to provide theory and
methods that do not overwhelm the hebraist, a literature that is
unafraid to proceed slowly enough that the reader can actually become
familiar with the procedures, and return with them to the data to do
his or her own testing of hypotheses.
Khan's Studies, by contrast with Longacre's Joseph, is well
explained, and does not overwhelm the reader with its contents, in
part because it does not launch the reader into a new world quite so
thoroughly as does Longacre's.
Because I find Longacre's approach to the text so valuable, and
relatively accessible, at a logical levelif perhaps not entirely well
communicated in Josephthe remaining three chapters of this volume
will present an explanation of this approach, and an application of
some of its more basic principles to Hebrew text samples.
Chapter 3
AN INTRODUCTION TO THE TAGMEMIC MODEL
AND METHODOLOGY OF TEXT ANALYSIS
1. The Purpose of this Chapter, and its Relation

to the Rest of this Volume
I have cited the topic of this study as 'the role(s) of different types of
clauses in different types of texts';1 and if the goal of the first two
chapters has been to measure the success of those who have entered
the field (examining their 'integrity of description' and the 'effectiveness of their communication'), then the goal of the two chapters that
follow this one is to outline for the reader some of the first steps in
producing an analysis that will pass both examinations. It is necessary,
in that case, to set out some definitions, to tie together some concepts,
and to detail the procedures to be used, in such a way that the reader
is not left behind in the analyses that follow.
2. Introduction to 'Tagmemic' Theory
In this section, we will look at several of the essential elements of the
Tagmemic school of linguistics. But why should we do this? What is
to be gained by focusing on this particular approach? According to
Longacre,
Broadly conceived, tagmemics is discourse about linguistic patterns. It is
not interested merely in setting up abstract strings of symbols which will
get us to terminal strings in the most economical way. To tagmemicists
such preoccupation misses the point. Rather we want to posit in each language a system of labeled patterns that will at least to some degree parallel
the system of intuitively felt patterns used by speaker and hearer. The
1.
Chapter 1, section 1.1.
3. Introduction to the Tagmemic Model
71
patterns and parts of patterns that we posit must be labeled for the simple
reason that we want to discourse about them. We want to be able to show
the system of patterns and to contrast one pattern with another.2
And Jones writes,

Linguistic tagmemics is concerned with discovering the patterns and regularities of language, and with stating these as consistently, systematically,
and elegantly as possible.3
Of particular significance [in terms of contributions of the tagmemic
approach to language analysis], I think, has been the focus by various
tagmemicists on the discourse FUNCTIONS of various units and constructions. Longacre and his colleagues have found discourse functions
for many particles and affixes that had previously been little understood or
else simply glossed as 'emphasis'.4 I (1977)5 have suggested discourse
functions for clefts, pseudo-clefts, and various other constructions in
English. Larry Jones and I (1979)6 have suggested that
tense/aspect/mood, many particles, affixes, and some special constructions function in a number of languages to mark different levels of
significant information in texts (e.g., to distinguish peak, pivotal events,
ordinary events, significant background, and ordinary background). In
discourse analysis, as well as other areas just mentioned, tagmemics
offers a significant theory of linguistics.7
Longacre's comment about 'setting up abstract strings of symbols

which will get us to terminal strings in the most economical way' is
intended to contrast tagmemics with the majority of other approaches
to language description. He substantiates this contrast with a comparison of two different descriptions of Turkish verbs; the first is from
Gleason, and Lees ('the usual Americanist way of plotting relative
2. Longacre, 'Tagmemics', Word 36.2 (1985), pp. 137-38.
3. L.K. Jones, 'A Synopsis of Tagmemics', Syntax and Semantics 13 ('Current
Approaches to Syntax') (1980), pp. 77-96 (86).
4. Here we might refer to T. Muraoka, Emphatic Words and Structures in Biblical Hebrew (Leiden: Brill, 1985), which has examined a variety of morphemes and
structures under the banner of 'emphasis', some of which certainly have macro-syntactic significance, but which receive no consideration as macro-syntactic features.
5. L.K. Jones, Theme in English Expository Discourse (Lake Bluff, IL: Jupiter
Press, 1977).
6. L.B. Jones and L.K. Jones, 'Multiple Levels of Information in Discourse',
in L.K. Jones (ed.), Discourse Studies in Mesoamerican Languages (Dallas: Summer
Institute of Linguistics and University of Texas at Arlington, 1979), pp. 3-27.
7. Jones, 'Synopsis', pp. 94ff.
72
orders of affixes within a verb structure');8 the second is his own,

tagmemically based, description. These two show the efficacy of the
tagmemic approach in describing systems, rather than just sequences.9
This focus on systems, relative 'user-friendliness', and 'intuitively
felt patterns' has encouraged the development in tagmemic circles of
new ideas about the description of text-level features, and some of
these are particularly helpful for the kind of work we are undertaking
in this study.
The second reason for turning to tagmemics at this stage is that
which is reflected in Longacre's comments about 'intuitively felt patterns'. We readily acknowledge that we tell stories mostly in past
tense, we give predictions mostly in future, and so on. Tagmemics has
provided a means of analyzing and describing that feature of
language, in a simple and straightforward manner. The fact that these
and other features which we will address in this study are readily verifiable at the 'common sense' level means that we will not need, on the
one hand, a lengthy and heavy introduction in all the details of the
theory used, in order to see the validity of the results, nor, on the
other hand, will we be encumbered with a 'write-up' filled with jargon
and convoluted explanation. Tagmemicsfor our purposesgives us
simple tools to work with, and then turns us loose to get to work.
My results will not seem to be tied exclusively to any one theory,
and that is intentional, for the things I describe in this study are readily accessible via a common-sense approach to the language, once the
analyst has been alerted to their possible presence, and thus we do not
need to bury them in theory-specific communication.10
Jones remarks on the practical applications of tagmemics:
A primary task of a linguistics theory is to account for the empirical data.
As an important corollary, the theory ought to have practical uses in
dealing with language. There have been numerous practical applications of
8. Longacre, Tagmemics', p. 138.

9. Longacre, Tagmemics', pp. 138ff.
10. Of course, my write-up will bear the imprint of my own theoretical base, but
this, it is hoped, will be relatively transparent. Likewise, it is obvious that in certain
applications, the tagmemic approach will result in highly tagmemic-specific analytical
write-ups (theoretical questions tend to require more jargon, and more short-hand
type graphic representations than others, for example); nevertheless, in the case of a
rather elementary description of the language features we are here examining,
uncomplicated non-technical language will suffice for the presentation of results.
73
tagmemic theory, and I believe that the great applicability of tagmemics is

a significant strength in its favor. One application has been as a heuristic
for helping students to understand the nature of language. But this is not
all. There have been numerous successful applications of the theory to
translation, language learning, literacy, and linguistic analysisparticularly in the study of unwritten and 'exotic' languages. In addition, there
have been many applications in other disciplines, including, most notably,
anthropology (since tagmemics is a theory not only of language, but also
of human behavior).11
For the analysis of unwritten languages, discovery procedures have
evolved, perhaps better called ANALYTICAL TECHNIQUES. These do
NOT, however, constitute a mechanical algorithm that automatically outputs an analysis. Practical goals of a theory hold pure theory-building in
check, since 'fruitful theory must to some extent be limited by analytical
techniques for processing or evaluating data. Tagmemics has oscillation
between theory and method rather than a one-way priority...' !2
2.1 Foundations
This enquiry has to do with texts, and text-level features; and in this
chapter I will be working through sufficient text-linguistic theory to
enable us to approach the data (in the following two chapters) in an
informed manner. Yet it must be said that many in the Tagmemic
School of linguistics have eschewed text-linguistics in the same way as
have classical hebraists: much of the material published by tagmemic
linguists stays within the traditional bounds of clause-level analysis;
this fact limits, from one direction, the amount of published material
that can be drawn into the discussion of Classical Hebrew text-linguistics from a tagmemic viewpoint. The second limiting factor is that
a great deal of tagmemic publication takes the form of language
'write-ups'reports on features of a specific language, or grammars
of specific languagesthe majority of tagmemic publications are
examples of language analysis, rather than explanations of it.13 These
11. One might add here that one of the more surprising applications of this theory
has been in the field of music, or, more specifically, ethnomusicology, and has
resulted in a variety of analytical apparatuses with a very profitable degree of flexibilityan essential feature when dealing with non-western music systems.
12. Jones here refers the reader to Pike, Language, p. 509; Jones, 'Synopsis',
pp. 86ff.
13. For example, the list of references in Longacre's 'Discourse' (in Brend and
Pike (eds.), Tagmemics, pp. 1-44) includes 65 articles or books on tagmemic
analysis; 39 of these (60%) are reports on specific languages (Western Bukidnon
74
are occasioned largely by the work of Wycliffe Bible Translators and

Summer Institute of Linguistics workers, whose Bible translation
responsibilities require the parallel production of language descriptions. Since this group of linguists comprises the greatest number of
users of this linguistic school, so this format has become the standard
one for publication of new results. Those trained in this school of
linguistics are accustomed to looking for new ideas in reports on
Bolivian ritual poetry, to help them solve difficult problems in the
Narrative History of a Philippine language, for example, and are able
to sort through such reports in a way that others do not attempt.14 It is
regrettable that there is little attempt at collation of this material for
more ready use by other language students.15
Tagmemics has nonetheless produced one of the most functional and
profitable text-linguistic theories available to those who take up an
interest in the issue.16 Here, we will look at some broader concepts
which undergird the material that follows; this will be divided into
two sections. I will first deal with several basic principles, with which
Manobo, Kaiwa Guarani, Daga, Isthmus Zapotec, and so on), and many address
only specific features of these languages (e.g. H. Popovich, 'Large Grammatical
Units and Space-Time Setting in Maxakali'); see his pp. 4Iff.
14. These factors account for the relative paucity of references in my own bibliography and notes to tagmemic works; there is little point in extending the material to
include such works as Harold Popovich's (cited in the previous note), since these do
not bear to any great degree on the things that we must discuss here. I have included
in this bibliography those works which are the most directly relevant to this work,
and while I have not avoided language-specific materials, the actual number of books
and articles on tagmemics is rather smaller than one might expect.
15. Longacre's 'Discourse' is a rare example of a tagmemic article which is a
collation and synopsis of language features rather than a description of languagespecific features.
16. For a closer look at the origins and development of this school of language
analysis, I refer the reader to V.G. Waterhouse's The History and Development of
Tagmemics (Janua Linguarum, Series Critica 16; The Hague: Mouton, 1974), to
L.K. Jones' 'Synopsis of Tagmemics', and to Longacre's 'Discourse' (in Brend and
Pike [eds.], Tagmemics), and 'Tagmemics' Word 36.2 (1985), pp. 137-77. The
former two are partial to Kenneth Pike's approach, while I follow more closely
Longacre's; those interested in comparing the two would do to start with Longacre's
'Tagmemics', followed by idem, with K. Pike and E.G. Pike, Grammatical Analysis. The main differences are in Pike's use of a 'four-cell tagmeme', whereas Longacre employs a 'two-cell tagmeme', with the remaining material from Pike's other
two cells dealt with as 'Deep-' or 'Notional Structure' (see below, this chapter).
75
I will not interact to any significant degree once they have been
explained, but they are requisite elements of the endeavour. I will
follow this with explanations of other features of the model that either
are particularly important to the theoretical base (e.g. 'Syntagmeme')
or will be referred to with some frequency in the analyses that follow
(e.g. 'Exponence').
2.1.1 Two Fundamentals: Empirical Analysis and 'Language as a Part
of Human Behaviour'
Before turning to these elements of the theory, however, I would like
to single out two concepts that are particularly foundational. The first
is 'Empirical Analysis'.
It is important to let the data define the questions asked of it, rather
than a theoretical model. It has been found that the Tagmemic model
is sufficiently 'alive' to the features of 'real language' that it does not
need to impose structure on the language being analysed; rather, the
structure that surfaces as one works with language inevitably fits
within the range of possibilities anticipated by the theory. The fact
that the theory concerns itself with 'deep' (cognitive) structure(s) as
well as with surface structures means that there is no preconceived
idea of how a deep structure 'notion' 'must' be encoded at the surface
structure level. Yet this interest in deep structure also informs the linguist of the things that will need to be expressed, as a general rule, in
any language. The wide-ranging research that has been conducted in
real language also informs the linguist about the range of options for
encoding these 'notions'. In the long run, language data is considered
the unchangeable truth; it is what exists as languageeven were it
found to be an ungrammatical sample, for example, tagmemic theory
enables the researcher to propose (if he or she so choose) explanations
as to 'why' that particular ungrammatical construction was elicited.
For the most part, tagmemicists process vast amounts of data in their
analyses, in order to 'discover' the systems inherent in the language,
and to secure their descriptions thereof. This is the only way to work
with real language data, it is presumed, because an alleged pattern
seen in four texts may no longer hold true after fifteen have been
examinedif it doesn't, some other pattern will surface with the
greater amount of data; if the alleged hypothesis does hold true, then
the data substantiate the hypothesis. The data constitute the only
unquestionable 'fact' of the language.
76
The second of these two concepts is 'Language as a part of human

Behaviour'. No one would deny that language is a feature of human
behaviour; and none would deny that language is expressed within a
context. Tagmemics, however, takes these two facts strongly into
consideration in its analysis. A political speech makes sense within its
political and social context, for example. A series of instructions is
elucidated by the details of its setting. These 'truths' are, for the most
part, self-evident. Tagmemics, however, makes more consistent reference to them as explanatory contexts than other approaches tend to do.
Often, it will be setting and context that help sift out patterns in
difficult data. This is not to say that every bit of data must be
explained exhaustively according to its context; this would be an
overstatement. Features can be examined apart from their context, of
course, but a full explanation will include reference to setting, appropriate sociolinguistic usage, and so on. The Gettysburg Address, for
example, can be examined as a unitto discover its internal structure,
its lexical cohesion, and so onand a great deal can be learnt of its
use of language; but no one would suggest that an understanding of the
context in which it was composed and first delivered brings no greater
understanding to the meaning, andyesstructure, of the text. These
'context concerns' form the centre of attention in such disciplines as
psycholinguistics and sociolinguistics, yet never should they be relegated entirely to these disciplines. Just how much of this needs to enter
the level of grammatical and syntactic description is a matter for
debate, but the principle stands: Language exists as a part of a wider
context of human behaviour, and this context bears on the use of
language features.
2.1.2 'Sub-Basic'Concepts
The concepts dealt with above are likewise 'sub-basic'17 in that they
are fundamental to everything in this model; those which follow are
no less fundamental, but they will not require further attention once
their place in the theory is seen. We cannot really 'deal with them
briefly, then forget them'but, in fact, once we have examined them,
they will not command our attention again except as undergirding for
other features, and at these points I will address them only briefly.
17. I do not recall seeing this term elsewhere; it seems an appropriate coinage for
the task it must perform, however. No significance should be deduced from it
beyond a label for an evanescent category.
77
2.1.2.1 Patterning
The concept of 'Patterning' is basic to all theories of description. That
a piece of data can be compared to other pieces of data, and that
knowledge can be gained from doing so, presupposes that a unifying
pattern can be sought and described. Longacre writes,
Central to human behavior is PATTERNING. A noted encephalographist
has written astutely about patterning. 'The first significant attribute of a
pattern is that you can remember it and compare it with another pattern.
This is what distinguishes it from random events or chaos. For the notion
of random implies that disorder is beyond comparison; you cannot
remember chaos or compare one chaos with another chaos; there is no
plural of the word. Pattern is a quality of familiar things and familiar
things are comparable. It is much nearer the truth to say that man abhors
chaos than to say that nature abhors a vacuum... Broadly speaking one
may say that the sciences derive from pattern-seeking, the arts from pattern-making, though there is a much more intimate relation between the
seeking and making of patterns than this would suggest.'18
Granted the centrality of patterning in human behavior it follows that
we should require that a linguistic theory give centrality to linguistic
patterns. In measuring the fit of a theory with the empirical facts of
individual languages we should require that a theory lead to a description
in which patterns are thrown into bold relief. Or, in terms of evaluating
two grammars of the same language, one important criterion of evaluation
is that we recognize as superior the grammar which sets forth the patterns
of a language in the more straightforward and direct manner.19
So what we are about as text-linguistic grammarians is the clear

description of the patterns that occur at the 'text' level. Some language
scholars may be sceptical about such a level; the next section will
address some of the elements that help identify its existence.
2.1.2.2 Closure and Choice
I will borrow wholesale, here, from Longacre:
Aside from this general argument in favor of linguistic patterns, we argue
that CLOSURE and CHOICE, observed in people's use of language, also
point to the reality of linguistic patterning. The speaker acts as if he were
using units which start and stop. He backtracks and corrects himself if
18. Quoted by Longacre from W. Grey Walker, The Living Brain (New York,
1953), p. 69.
19. Longacre, Discovery, pp. 13-14; this last 'criterion of evaluation' is termed
'elegance' (see below).
78

proper closure is not given. He hesitates at certain points as if he were
confronted with a choice of item or construction. After partially or wholly
articulating one item or construction he may backtrack and correct himself
by choosing another item or construction. The hearer likewise demands
closure of units and has opinions about choicesas is often evident by
his responses. Speaker and hearer alike seem to be doing something more
direct than applying a complicated series of rules to speech; rather they
seem to be referring to an inventory of patterns.20
In other words, there are such things as linguistic units, which have
beginnings and ends. Texts, for example, begin and end; this justifies
our examining them as units. If patterns surface, their existence as
functional units is confirmed. That a speaker may backtrack to correct
something suggests that it is possible to say something in more than
one way, hence the question is not only one of 'grammaticality' versus
'ungrammatically', it is also one of appropriateness.
The concept of 'choice'where the speaker, for example, searches
among alternatives for the best way of saying somethingimplies that
there are points where the language speaker may choose between permissible options, which in turn implies 'sub-units'. If one may hesitate
over the choice of whether to employ a pronominal substitute or the
full noun phrase, this indicates, on the one hand, that both options are
permissible (though contextual, or perhaps 'stylistic', factors may
exert considerable pressure on the choice to be made), and on the
other hand, that the point of hesitation marks a functional point of articulationthe material preceding it does not unerringly determine
what follows it. That written language is generally edited before
transmission to its intended receptor in no way challenges this description of language processes; the writer, just as much as the speaker,
goes through the same backtracking and choosing processes. These
processes, whether written or spoken, underline the 'articulated'
nature of language.21 I will return to this concept of 'options' shortly.
20. Longacre, Discovery, pp. 14-15; see also his Tagmemics', esp. p. 137.
21. 'Articulated', that is, as in 'segmented', or 'jointed'. It is worth noting
another point here, as well: In this age of printed text, it is often overlooked that such
conventions as spacing between printed words, punctuation to indicate phrase, clause
and sentence divisions, paragraph indentation, and such things as chapter headings,
are artificial, yet they represent linguistic realities. These distinctions exist at the
spoken level as well, yet they are often rendered into spoken language by such
'suprasegmental' features as intonation patterns, hesitations, and the like. No one
would deny the functional reality of the 'word level', but this, in some languages is
79
2.1.2.3 'Notional' or 'Deep' Structure versus 'Surface' Structure

The level of 'Patterns' is found not exclusively in the outward structure of each individual language, but rather firstly in the language
centre of the human brain. There are discoverable tendencies at this
level, which begin the process of language structuring. This level is
called 'Notional Structure' by Longacre, et al., and 'Deep Structure'
by others.22 People who speak several languages sometimes find themselves unable, in trying to express a thought, to find words and structures all of which belong to the same language; they know what they
want to say, but are unable to get it out.23 They are sure of what they
want to say (the deep structure level provides no difficulties), but the
framing of their thought in real speech doesn't come together
properly (the surface structure poses problemsin this case, the
competing surface structure rules of different languages intermingle
and become confused). Crystal gives an example that helps to clarify
this concept of 'deep structure':
This level provides information which enables us to distinguish between
the alternative interpretations of sentences which have the same surface
form (i.e. they are ambiguous), e.g. flying planes can be dangerous,
where flying planes can be related to two underlying sentences, Planes
which fly... and To fly planes... It is also a way of relating sentences
which have different surface forms but the same underlying meaning, as
in the relationship between active and passive structures, e.g. The panda
chased the man as opposed to The man was chased by the panda.24
not easily pinned down in definition (are the Hebrew elements -3, -1?, and the like, to
be considered words? and what about n, which is sometimes independent,
sometimes a bound form? And perhaps most significantly, why or why no/?)in the
same way, text, paragraph and sentence units are difficult to pin down. It is my
contention that the major reason these are viewed more skeptically by the traditional
syntactician is that there is no inherited concept that has been handed on into the
standard grammars, etc., in the same way as were the concepts of syllable, word,
phrase and clause. Yet if we were to carry out linguistic analysis of the Ionic Greek
dialect before the period of word separation, would we have the same confidence in
word separation as we now claim to have? Language scholars have long been guilty
of a certain kind of imperialism which receives warmly that which it has been taught,
but rejects as sub-standard that which others have been taught, or have discovered.
22. My tendency is to use the two terms interchangeably, with a slight preference
for the latter. No ideological or theoretical significance need be attributed to this,
other than my own preference for more self-explanatory labels.
23. This difficulty is not limited to multi-lingual people...!
24. Crystal, Dictionary, p. 85.
80
Longacre, in a recommendation of the strengths of the tagmemic

approach to language analysis, writes,
It is precisely because tagmemics, however much as it may be become
interested in deep structure, continues to be very interested in surface
structures that the previous sections of this article, although referring to
studies completed some time ago, are valid today. Devices for investigating and describing surface structures in the world's languages (as given
above) rest on a considerable backlog of experience in languages around
the world and apparently do not need much immediate modification. As
the schools of grammarnoticeably stratificational, transformational-generative, and tagmemicdraw together in a common focus on meaning, it
will still remain that, of these various schools, the school which is most
interested in surface structure is the tagmemic school. Tagmemicists
should continue to unapologetically be interested in the way in which a
language of necessity must express itself and encode the deep structures
that people use as the natural apparatus of human thought.
Tagmemics offers more than a quick-and-convenient approach to
language analysis. It opens up vistas of researchsome of which are still
relatively unexploredand a program for linguistic activity for many
years to come.25
2.1.2.4 'Particle', 'Wave' and 'Field'

Of these, Crystal writes,
PARTICLE... (2) A term used in tagmemic analysis to refer to a linguistic
unit seen as a discrete entity, definable in terms of features. It is contrasted
with 'wave' (where the unit's contextual variability is analysed) and
'field' (where its distribution is described).26
Jones describes this set of concepts in terms of 'Perspective':

25. Longacre, 'Tagmemics', pp. 174-75; the directions for new 'vistas of
research' that are indicated by the tagmemic school of linguistics are all the more
varied and the richer for hebraistson the one hand, because of the wealth of help it
offers (through its copious processing of surface structure analysis of living language
data) in suggesting solutions to difficulties of Hebrew language description, and on
the other hand, through its relatively practical nature (a combination of its commitment to 'intuitively felt patterns' ['Tagmemics', p. 138] and its ability to communicate details of form and function without heavy reliance on jargon and idiosyncratic
calculus).
26. Clearly, then, this is not the more familiar definition of the term 'particle',
which is defined in Crystal as '(1) A term used in grammatical description to refer to
an invariable item with grammatical function, especially one which does not readily
fit into a standard classification of parts of speech'; both, Crystal, Dictionary, p. 222.
81
Tagmemic theory formally recognizes a varying observer PERSPECTIVE.

At least three different, but complementary, perspectives may be used to
view the same items. In a STATIC view ['particle'], items as individual,
discrete things are in focus. A DYNAMIC view ['wave'] focuses on the
dynamics of items overlapping, blending, and merging with each other.
Finally, in a RELATIONAL perspective ['field'], focus is on the
relationships between units, noting networks, fields, or matrices. Any one
of the three perspectives may underlie a particular linguistic description. A
description from one perspective complements, and adds to, a description
from one of the other perspectives, but does not replace it. Here also
tagmemics leaves room for more than one correct description or
grammar.27
If one analyzes the clause mrr -DN, in terms of its components, we

would describe it as a verbless clause with the structure 'Pronoun as
Subject, plus a Proper Noun as Predicate Nominal'. We have just
analysed it as a 'particle'. If we look at it from a 'wave' perspective,
we would examine it in terms of all possible variants (one might
decide to look at all simple S-P Verbless clauses, with no marginal
phrases (temporal, attributive, appositional, etc.) added, or one might
look at all Verbless clauses together;28 this is an examination of variants of the clause-type. If we begin to examine how S-P Verbless
clauses, or all Verbless clauses together, function in the language (for
example, as boundary markers in Narrative History), then we begin to
look at Verbless clauses from a 'field' perspective. These terms will
rarely recur in this volume, but the concept is basic to everything
forward from this point.
2.1.3 'Basic' Concepts
The preceding material is vital; without it, what follows will not make
sense. The terminology presented above is less important, however,
for I will not belabour it in what follows. The next set of concepts, on
the other hand, will be referred to by their 'labels' with some regularity in the following material, therefore it is distinguished from that
27. Jones, 'Synopsis', p. 80.
28. As does Andersen, in The Hebrew Verbless Clause in the Pentateuch (JBL
Monograph Series, 14; Nashville: Abingdon, 1970); cf. Waterhouse, History, who
describes the concept 'wave' as the 'DYNAMIC view', in which 'structure is made
up of waves, with nuclei and margins' (p. 6)'nuclei' being the essential part of the
unit, and 'margins' being the optional extras, thus accounting for the sum total of
possible permutations of the unit's structure.
82
which I have just covered. These concepts have a more direct relevance to the analyses I will carry out below than do those introduced
above.
2.1.3.1 Hierarchical Linguistic Structure
In his Appendix on tagmemics, in Joseph, Longacre deals briefly with
the three 'crucial' concepts of tagmemics.29 One of these is
'hierarchical linguistic structure'.30 This refers to a series of levels of
language: [MORPHEMESTEM] WORDPHRASECLAUSESENTENCEPARAGRAPH'DISCOURSE'.31 These are surface
structure terms, which are roughly paralleled in deep structure.32
Apart from the last one or two, these distinctions do not usually raise
questions among hebraists, since we are still accustomed to a more
'intuitive' approach to language description, and we tend to be aware
of such things at an intuitive level. Longacre writes,
Hierarchy is the spacing of constructions on levels from morpheme (level
of zero internal grammatical construction) up to discourse (level of maximal grammatical construction). With these two levels as lower and upper
bounds of hierarchy the other levels take their place as intermediate levels
of combination: stem, word, phrase, clause, sentence and paragraph.
Stems are derivational units. Words are inflectional units. Phrases express
modification or linkage. Clauses express predications. Sentences are
propositions which may concatenate, oppose, balance, or report predications. Paragraphs are units developing a discourse. The levels are partly
defined by such internal characteristics as these, partly by their hierarchical placement on the scale from morpheme to discourse.33
2.1.3.2 'Slot/Class' and 'Filler/Set'

These are more often referred to in the literature as 'slot' and 'filler'
than as 'class' and 'set' (or 'filler' set), and this is my convention
as well. This concept is one of the most central to the Tagmemic
29. Longacre, Joseph, pp. 31 Iff.
31. His 'discourse' equals my 'text'.
32. For a complete treatment of these two 'layers'beyond what I have given
aboveI must direct the reader to Longacre's Grammar, particularly pp. 322-27 and
273-74.
33. Longacre, 'Tagmemics', p. 143; those who are looking for a fuller, more scientific, definition of these levels would do well to consult Longacre's linguistic
books (e.g. Philippine Languages, pp. vi-xxvii, 1-3, 53, et passim; Grammar,
pp. 77-80, 269-336).
83
School of linguistics, and plays a central role in the shaping of other

characteristics of the theory.
A man gets dressed to go to work in a bank in Edinburgh, Scotland.
There are certain things he must wear, of course: shoes, socks,
trousers, shirt, tie and jacket being the basics. Now, he has a choice of
which shirt to wear (it must be white, howeverthis is Edinburgh,
not Rome), and he can choose from among several dark suits, or
combinations of trousers and jackets; he has a fairly wide variety of
socks, and more than a few ties. He has a choice between three different pairs of black shoes. Yet the ensemble is more or less dictated by
custom.
There is a typical 'slot-filler' relationship involved here. In each
case, a required piece of clothing is supplied from a collection of suitable options: the man undoubtedly has several good shirts which are
not white, yet these are not appropriate under the circumstances; he
has many pairs of socks (some of which will go very well with the suit
and shoes he has chosen), but it would be socially unacceptable for
him to wear one of these in place of the required necktie. In the same
way, grammatical (and other) relations can be described as a 'slot'
into which an item is fitted; the item is chosen from among a set of
acceptable 'fillers'.
The following clauses appear very similar, yet they reflect different
patterns:
She is my aunt.
She is fun.
She is driving.
Subj. + Cop.Vb.(Vl) + Pred.Nominal34

Subj. + Cop.Vb.(Vl) + Pred.Complement
Subj. + Pred.(V2)
The first clause is a 'stative identificational' clause; the second, a

'stative descriptive' clause; the third, an intransitive clause. Although
the word 'is' occurs in all three samples, the conjugated verb is interchangeable. This can be expressed by defining two different sets of
verbs which provide the acceptable options for these two different
slotsfor example, 'VI' could be defined as any finite form of the
verbs 'to be' and 'to become'35 (this permits the construction of a
34. The '+' sign is used to indicate that a certain element is required rather than
optional (in which case it would be marked with ).
35. The acceptable way of conjugating these verbs would be dealt with elsewhere
in the grammar.
84
clause like 'she had become my aunt'36), while 'V2' might be defined
as any finite form of any intransitive verb,37 granting the possibility
of clauses like 'she sighed', and 'she will elope', or even 'she exists'.
The slot is called Verb in both cases, but in a tagmemic approach they
would be given labels that distinguish them one from another. The
'filler' set would be likewise labeled so that it matched up accurately
with its corresponding 'slot'.
2.1.3.3 'Constituent Structure' Analysis
The sort of description of clauses that I have undertaken superficially
in the above section is often called constituent structure: when one
looks at the way a unit is composed, taking it apart to identify its bits
and pieces, one is engaging in constituent structure analysis. Such an
approach allows us to detail the patterns that surface as we compare
like with like, and divide the unit into the elements which make it up.
2.1.3.4 'Tagmeme' and 'Syntagmeme'
All tagmemic textbooks I have come across tend to deal with this concept (the identifying feature of 'Tagmemic' theory) in a perfunctory
way, assuming that it is a concept easily grasped. And, when one has
seen countless tagmemic grammars, phonologies, scientific papers and
so on, the existence of these two categories as language realities is
clear beyond questionit does indeed become a concept easy to
grasp.38 Yet, the first time around is not so simple.
In short, a SYNTAGMEME is a formula, representing a language
unit; this formula is made up of sub-units, which are called
TAGMEMES. Each tagmeme refers to that set of options which can
function in the slot represented by the tagmeme; and each option will
be described by a syntagmeme (a Clause syntagmeme will be composed of tagmemes, some of which might represent Phrases, and a
Phrase syntagmeme will be composed of Word-level tagmemesand
so the arrangement progresses through the grammatical hierarchy.
Waterhouse describes these categories as follows:
36. An odd example, perhaps, but grammatically correct.

37. These 'rules' make no effort to be grammatically complete.
38. Compare the ease with which one recognizes the color 'yellow', or the sound
of a violinand the difficulties one has in trying to describe them.
85
[The tagmeme can be defined as] a functional slot in a grammatical

construction correlated with the lexical item or class of items which could
be said to fulfill that function... A TAGMEME, then, is the correlation of a
functional slot on a specific level within a hierarchy with the class of items
that fill the slot. Thus a class is said to fill a slot. A class may also be said
to manifest a tagmeme, or a specific member of a class may be said to
manifest the tagmeme. The slot is not viewed as a location in a linear
sequencealthough it may be thatbut rather as a function in a construction type. Neither the slot nor the class that fills it is considered to be the
tagmeme by itself. For this reason, the basic tagmemic notation is dual,
with labels for both slot and class, and a colon inserted between them,
without intervening space. The colon is to be read 'filled by'. Tagmemes
may be obligatory, marked +. Optional tagmemes are marked . The
occurrence of an optional tagmeme does not thereby constitute it obligatory;
hence it is still marked . The marking as either obligatory or optional is
part of the notation for the tagmemic unit itself; thus there is no
intervening space between it and the slot label. A typical notation for a
tagmeme would be + S:n ( = 'obligatory subject slot filled by noun'), or
L:loc (= 'optional location slot filled by location word').39
A SYNTAGMEME is a construction on a given hierarchical level whose
constituent parts are tagmemes.40
Thus, we could say that mrr ] is a simple realization of the Hebrew

'Naming'-type Stative Identificational Clause syntagmeme:
Verbless Cl.: { +S : PPron +P : Pnm }41
where the subject 'slot' of the clause can be filled with a selection
from the 'filler' set 'Personal Pronoun' (Ppronwhich will be defined
elsewhere); the predicate 'slot' can be filled by any member of the
'filler' set 'Personal Name (Pnmlikewise defined elsewhere),
39. Waterhouse, History, pp. lOff.
40. Waterhouse, History, p. 11.
41. Items within brackets { . . . } purport to be tagmemes. I make no claims,
however, to grammatical completeness. Waterhouse points out that Longacre's notation 'has used the slot label alone; this does not however, mean that the filler class is
overlooked. Rather, the simpler notation is merely a convention for ease of transcription' (in loc.I tend to use this simplification as well); Elson and Pickett likewise
simplify the calculus for representing language features [B. Elson and V. Pickett, An
Introduction to Morphology and Syntax (Santa Ana, CA: Summer Institute of
Linguistics, 2nd edn, 1964). It is helpful to remember that it is the commonality of
theoretical starting points which unifies the tagmemic school of language analysis and
description, rather than a specific set of notational conventions.
86
although it is clear that this must be an 'open' set, allowing for an

unlimited number of possibilities, unlike 'Ppron', which will be a
'closed' filler set.
Our Edinburgh banker will serve to illustrate these concepts
further. The syntagmeme '[Male] Banker's Uniform' might be written
as follows:
B's Uniform: {+ Suit(Sl) + Shirt(Srl) + Tie(Tl) + Socks(Sol) +
Shoes(Shl) }
where 'SI', 'Sri', "IT, 'Sol', and 'Shi' are tagmemes of the [Male]
Banker's Uniform syntagmeme. 'Suit(Sl)' can also be analysed as a
syntagmeme:
Suit(Sl): { + Dark Jacket + Dark Trousers }
'Suit(Sl)' is a tagmeme when analysed as part of the Banker's

Uniform; it is a syntagmeme when analysed in terms of what elements
make it up.
This brief discussion has presented what is known as 'two-cell tagmemics', which records the 'slot' (e.g. Subject) and the 'filler' (e.g.
Personal Pronoun); other versions of tagmemics exist, and are superior to this simplified model for the reduction of language rules to a
shorthand 'calculus'.
The standard approach adopted by a large percentage of Summer
Institute of LinguisticsWycliffe Bible Translators personnel is that
of four-cell tagmemics, where, in addition to the two cells representing
'slot' and 'filler', two more describe the 'role' (e.g. Subject as Agent)
and the 'cohesion' (this can be explained best by avoiding the calculus
and jargon of the system; it refers to such things as 'agreement in
gender, number and case', and thus defines the kind of relations that
exist between the different tagmemes of the syntagmeme).42
The reader may be permitted a sigh of relief that precise understanding of all this is not required in order to follow the remaining
material; one might be the more grateful at not being required to
42. The definitive work for learning the mysteries of four-cell tagmemics is Pike
and Pike, Grammatical Analysis; a briefer, and much more accessible, explanation of
this apparatus is offered in Jones, 'Synopsis', section 1.3, Terms and Relations',
pp. 80ff.this also contains a very lucid summary of the basic concepts of syntagmeme and tagmeme.
87
digest the 'nine-cell' tagmeme, which was the literal 'grand-daddy' of

them all.43
The interaction between 'tagmeme', 'syntagmeme' and 'structural
hierarchy' is what interests us in particular at this point. These three
elements Longacre set out as the crucial features of tagmemics for
following such a work as Joseph', they are the core of what I am
presenting as well.44
In our Intransitive clause sample, above ('she is driving'), the
'tagmeme' that fills the Predicate slot in the Intransitive Cl.
'syntagmeme' is an Intransitive Verb Phrase. The Subject slot is filled
by a Noun Phrase or equivalent (in this case, a Personal Pronoun).
The Intransitive Clause syntagmeme could be set out as below:
Intransitive Cl. = { + Subject + Predicate}
where
Subject
Predicate
{Noun Phrase}
{Pronoun}
{Proper Name}
(etc.)
{Intransitive Verb Phrase}
'Noun Phrase' and 'Intransitive Verb Phrase' are themselves units that
can be described as syntagmemes with 'slots' into which other
'syntagmemes' fit as 'tagmemes' within the phrase ('fillers'). Each of
these tagmemes ({Noun Phrase}, (Intransitive Verb Phrase}, etc.)
may perform widely differing functions in different contexts; these
functions are defined by the syntagmeme for each of those contexts.
So, we can say that the syntagmeme defines the function of certain
tagmemes in certain specified contexts, and the tagmeme describes the
constituent structure of the syntagmeme.
This is a radically simplified demonstration of how the 'syntagmemes' of one level of the hierarchy tend to look for fillers from a
43. Even the greatest of the creative masterminds behind this theory found that
one unwieldy. The desire to describe as many features of a unit in as short a space as
possible was the motivation for these creations; but they also must be usable and at
least somewhat self-explanatory, and this one 'went the way of all flesh'. The tagmemic model is constantly evolving under pressures from new language data, and
from field linguists who require manageable tools. This is one of its greatest
strengths.
88
lower level of the hierarchy. 'Syntagmeme' can be roughly defined as

a linguistic unit, while 'tagmeme' can be roughly defined as a sub-unit
within a unit.
2.1.3.5 Exponence
We can now begin to reconstitute45 Longacre's incredibly dense
summary of this material:
Tagmeme and syntagmeme are correlative concepts in that (1) the functions of the various tagmemes are expounded by sets of syntagmemes
(including those of zero internal structure), and (2) a syntagmeme is composed of tagmemes. I call the first relation exponence and the second
composition. Thus, not only can syntagmemes of zero internal grammatical structure, such as John and Bill, occur as subject and object in a transitive clause, but compound words such as (the) redcoat, (the) gentleman,
or (the) oarsman may occur there as well; phrases such as the oldest man
in the group, an imprudent young officer or a crippled sailor can occur;
and even a few clauses such as That he came early (impressed Bill).46
If 'syntagmeme' is roughly equivalent to 'unit', and 'tagmeme' is

roughly equivalent to 'sub-unit within a unit', we must also say that
when we begin to look at a sub-unit on its own (as opposed to within
its unit/syntagmeme), then we are looking at it as a syntagmeme in its
own right. We view it now as something with its own structure, rather
than viewing as something that has horizontal relations with other
tagmemes, inside another unit. In most cases this means moving down
one level in the linguistic hierarchy.47
Longacre has written that 'the functions of the various tagmemes
are expounded by sets of syntagmemes (including those of zero internal structure)'; in our earlier example of the Intransitive Clause
syntagmeme, the 'functions' are 'Subject' and 'Predicate'; these are
tagmemes, and are 'slots' into which certain appropriate 'fillers' can
be inserted. The fillers themselves, however, are also syntagmemes
45. As one would 'reconstitute', say, instant coffee.
46. Longacre, Joseph, p. 312; his explanation in 'Tagmemics' is fuller, and more
readable, but relies heavily on Turkish and Trique language data which he presents in
the text, in graphic representations which are not reproducible here; in addition to this
fact, any meaningful quotation from that source on our subject would extend to several pages of our textnevertheless, these pages are the most helpful summary I
have yet seen of the twin concepts of 'tagmeme' and 'syntagmeme' (pp. 138-43).
47. A point to which we will return shortly.
89
(e.g. 'Noun Phrase' or 'Personal Pronoun')they are only considered

'tagmemes' when examined in terms of their function within the
Intransitive Clause syntagmemeand, as syntagmemes, can be looked
at in terms of their own tagmemes. And so a syntagmeme of the Noun
Phrase might look like this:48
Noun Phrase = { Det. Modifier (1) + Substantive Modifier (2) }
where
Determiner
Modifier (1)
Substantive
Modifier (2)
{Cardinal Number}
{'the'}
{'a'}
(etc.)
{Adjective Phrase}
{Ordinal Number}
(etc.)
{Noun}
{Present Participle}
(etc.)
{Possessive Noun Phrase}
{Verb Clause}
(etc.)
With this array of syntagmemes, the following are accounted for as

possible Noun Phrases:
'six yellow dogs'
'the first five hundred winners'
'a growing distrust of their frequent prevaricating'
'the last time I was there'
and so on (and all this is in order to explain the term 'exponence').49

The section above dealing with Modifier (l)'s 'filler' options, for
example, could be articulated in lay terms as 'Modifier (1) may be
expounded by an Adjective Phrase, and Ordinal Number...' These
exponential relations can be broken down and qualified further.
48. Again, I must stress that these 'syntagmemes' make no claim to complete
accuracy; thek purpose is to demonstrate features of this modelwere they grammatically complete descriptions of the real-language clause-type, their complexity might
muddy rather than clarify the points under consideration.
49. I am presuming that the concept of 'composition' as introduced above is more
or less self-explanatory.
90
2.1.3.5.1 Primary Exponence

I have said above that a slot tends to be filled by a syntagmeme from
the next lower level of the linguistic hierarchy; Longacre terms this
'primary' exponence.50 A hypothetical sentence51 tends to be composed of a pair of clauses; the clauses will likely be composed of
phrases, and the phrases themselves, of words. This is sort of exponence is the most readily understood.
2.1.3.5.2 Recursive Exponence
Longacre writes:
Recursive exponents of tagmemes are from the same level as the tagmemes
themselves. Thus, stem can occur within stem as in ungentlemanliness
which is a noun stem52 the exponent of whose theme tagmeme is an
adjective stem53 ungentlemanly. In turn the exponent of the theme
tagmeme of the latter is another adjective stem gentlemanly. The exponent
of the theme tagmeme of the latter is a noun stem gentleman which has
two theme tagmemes whose exponents are morphemes gentle and man.
Three derivative morphemes -ness, un-, and -ly manifest the tagmemes
nominalizer, privative, and adjectivizer.54
'Recursive' exponence occurs when a unit fills a slot in a syntagmeme

of that same level, for example, 'a growing distrust of their frequent
prevaricating', where the modifier 'of their frequent prevaricating'
itself defined by the Noun Phrase syntagmemeoccurs as a tagmeme
within a Noun Phrase: a Phrase within a Phrase.
On the above two kinds of exponence, Longacre makes these
comments:
Secondary exponence (recursion) does not vitiate the witness of primary
exponence to hierarchical structure in language. Recursion is identifiable
as something apart from primary exponence. Thus, derivative affixes in
English are a recognizable category of affixes. When we find several of
these affixes occurring together in the same form,55 then we know that
we have an instance of recursive exponence on the stem level. On the
phrase-level recursion is identifiable by the occurrence of prepositions or
51. E.g. 'If he wanted to come, he should have said so'.
52. 'Stem' is the hierarchical level of the tagmeme in question.
53. I.e. the same hierarchical level; 'theme tagmeme' is the nuclear element of a
stem, and, in primary exponence, contains a morpheme.
54. Longacre, Tagmemics', p. 145.
55. That is, in the same linguistic unit.
91
noun-phrase initial items (of, the, that) in what is apparently phrasemedial; these relators or initial items signal onset of a phrase acting as
recursive exponent of a tagmeme within another phrase. On the sentence
level tell-tale distribution of such conjunctions as and, but, or, if and
unless mark recursion.
While primary exponence gives strings in n-ary relations, recursive
exponence creates nests of constructions which can never be successfully
analyzed as simple linear strings and often are binary. Thus the English
sentence quoted above [Had they taken a sword and threatened to run him
through or held a club ready to dash out his brains, he would have died
saying, 'No. Never.'] is not a simple chain of clauses: (1) Had they taken
a sword, and (2) had they threatened to run him through, or (3) had they
held a club ready to dash out his brains, (4) he would have died saying,
'No. Never.' A nest is a structure amenable only to some sort of immediate-constituent analysis. To analyze it as a linear string with order classes
is to understructure it. The lowest layer in a nest is composed, however,
of descending exponents. Ultimately, then, a nest of phrases is composed
of words and a nest of sentences is composed of clauses.56
Recursion is more frequent on the stem, phrase, sentence, paragraph and
discourse levels, and less frequent on the word and clause levels, which
tend to be linear strings. Recursive or non-recursive propensities of a
given level constitute a further characteristic of that level. We have already
illustrated recursion for stems, phrases, and sentences. In regard to
discourse it is necessary to note only that scarcely any discourse of much
length and complexity is a simple sequence of paragraphs.5^ Rather such
a discourse contains subdiscourses, subplots, and subnarratives. In brief,
it has discourse level tagmemes (e.g., episodes) whose exponents are
themselves discourses. Paragraphs can likewise contain subparagraphs.58
56. His own note on this section reads, 'Some languages contain certain syntagmemes (e.g. numeral or adjective phrases) that occur only as secondary exponents.
Thus if a numeral phrase occurs only as a modifier of a noun then it is always a
phrase-level recursive exponent, that is, while itself a phrase it occurs only as an
exponent of a phrase-level tagmeme (Longacre, 'Tagmemics', p. 175 n. 3).
57. Italics my own; this is the principal reason for the complexity of Longacre's
graphic presentation of the text of the Joseph story (Joseph, Part Four, pp. 209310), and one of the principal reasons for my presenting so much tagmemic theory at
this stageour investigations below will reveal patterns that are recognizable when
these tendencies of language structuring have been explained to some degree.
Tagmemics is not the only school of linguistics that is capable of describing these
patterns, but it is one of the best, not least because, once the principles have been
understood, jargon can be left behind, and explanations be given in lay terms.
58. Longacre, 'Tagmemics', pp. 146-47.
92
2.1.3.5.3 Back-Looping Exponence

In this case, 'a higher-level unit fills a slot in a lower level'.59 We have
seen an example of this in 'the last time I was there'; 'I was there' is a
Stative Descriptive Clause, which occurs as a 'Modifier' tagmeme in a
Noun Phrase: a Clause within a Phrase.
Likewise a sentence can occur within a Phrase: 'his "if-we'd-doneit-my-way,-it-would-have-worked" comments'a Sentence within a
Phrase.
2.1.3.5.4 Level-Skipping Exponence
This is slightly more difficult to illustrate than the others, since it
functions more often at the higher levels (which I have not yet
described adequately enough to use as examples) than the lower ones.
The principle is this: In primary exponence, a slot is normally filled
by a unit from the next lower level; if this level is skipped, and the
unit which fills the slot comes from an even lower level that normal
level, we then have 'level-skipping' exponence, for example, 'If you
can join me, we'll go together. Otherwise, I'll look for someone else'.
In this case, we have two hypothetical sentences; in the first one, both
the protasis and the conclusion are expounded by clauses; in the
second, however, only the conclusion is expounded by a clausethe
protasis ('otherwise') is expounded by a word (skipping both the
clause and the phrase levels). Elliptical units fall under this category,
and I will occasionally refer to them by that term; at this stage,
however, it is to our advantage to underline that the same mechanism
(that is, 'exponence') is taking place, regardless of the unit that fills
the slotwhether large or small.
2.3.1.6 Embedding
A shorthand term for both 'recursive' exponence and 'back-looping'
exponence is 'embedding'. One could say that 'a growing distrust of
their frequent prevaricating' is a Noun Phrase in which the Modifier
of the Noun Phrase 'a growing distrust' is expounded recursively by
another Noun Phrase'of their frequent prevaricating'. This specifies
the exact relationship of the filler to the slot. In this volume, as is done
in many others, I will sacrifice some of this precision, and will bypass
explicit statement of the nature of this exponence (except where an
93
explicit statement is necessary); the result is that I will describe the

above phrase as 'a Noun Phrase that contains an embedded Noun
Phrase'. In this enquiry I will refer to 'a Narrative text embedded in
an Expository text', 'a Predictive paragraph embedded in a Narrative
History speech formula', and so on. The concept of embedding is the
most important to my later discussion of those which have been
described so far; the material preceding this has served primarily as
background so that this feature may be the more readily understood.
By way of further clarification of this concept, we can draw connections between it and other concepts we have looked at. The most
cogent of these is the twinned concepts of 'Slot' and 'Filler'. The
mechanism of exponence (and, therefore, of embedding) is the direct
connection between the slot and the filler.60 By referring to the
relationship between two units, of 'embedding', I am highlighting the
fact that 'paragraph X' is functionally related to 'clause Y' (e.g. where
a paragraph of Reported Speech fills the Direct Object slot in a
Transitive Clause). The same is true of this concept's connection to the
'Tagmeme-Syntagmeme' pair of concepts'paragraph X' is a manifestation of the 'tagmeme P', in and of itself, yet it functions with
other tagmemes to create a permissible manifestation of the
'syntagmeme Q', which results in 'clause Y [the Reported Speech
paragraph embedded in the Transitive Speech Formula Clause]'.
This concept of 'embedding' is a considerable help in elucidating the
grammatical role of Reported Speech in, say, Narrative History, since
it highlights the interconnectedness between Reported Speech and the
framework that supports it. This is essential for the integrity of the
framework, and also permits us to examine Reported Speech as individual, fully self-contained, units, which also happen to function as part
of something else. The benefit of this is that we are able to ask questions
of individual Reported Speech texts, and to compare them selectively
(for example, to other units of the same text-type), so as to gain a
greater understanding of the features of distinctive text-types. Likewise,
we gain insights into the nature of Reported Speech as a grammatical/
syntactic mechanism by approaching it in this way. I will, as I have
intimated above, return to this in fuller detail in Chapter 5.
60. Although the terms 'embedding' and 'embedded' could be used to described
'level-skipping exponence' (see immediately above), it serves primarily to indicate
the insertion of a larger unit into a slot normally filled by a smaller one. Both situations are manifestations of the 'Slot-Filler' interrelationship.
94
2.2 Specific Features of the 'Text' Level

I will now look at the highest level of the linguistic hierarchy
'Text'from a variety of vantage points. That such a level exists is
indicated both by behavioural evidence (people usually know when an
oral report, or a conversation, or a joke, is finished, for example),
and by linguistic evidence (languages tend to demonstrate linguistic
patterning at the text level, for example, languages tend to mark the
'peak' event of a story).61
2.2.1 The Identity of the Text
Our first approach to the text will be to examine the varieties of texts,
or 'text-types'. I will look first at several binary parameters that will
be used to build a matrix for the identification of text-types.62
2.2.1.1 The First Parameter: 'Agent Orientation'
'Agent Orientation' (abbreviated 'AO') is a feature of texts that
highlight the participants. For example, a story generally highlights
the activities (or lack thereof) of a participant, or several participants;
a theoretical paper about, say, a matrix for distinguishing text-types,
does not focus on participantsif any are mentioned (for example,
the reader, or the author) it will only be obliquely, and their activities
as related in the text will have little to do with the flow of the text.
Again, if a builder is explaining the day's schedule to a team of
workers, reference will have to be made to who is to do what, where
a manual on how to rewire a house will prioritize the activities performed rather than the people doing them (it makes no difference who
does them, as long as they are done in the manner required by safety,
by the housing code, etc.).
2.2.1.2 The Second Parameter: 'Contingent Temporal Succession'
Of the three parameters that will define our matrix, this may be the
most difficult to grasp. 'Contingent Temporal Succession' (abbreviated
'CTS') refers to whether or not events or doings in a text are related
to (or 'contingent upon') prior events or doings.
61. See below, this chapter.
62. This material is derived from Longacre's Grammar, pp. 3-6; the vast majority
of the terminology is his own, yet I will only reference extended quoted material; I
am too greatly in his debt for this material to reference every allusion to his publications. Having said that, I must add that his own explanations of these concepts,
except in the classroom, achieve minimalist standards.
95
Events in a story, or instructions on how to assemble a model airplane, typically exhibit a certain temporal development: each activity
develops in some way out ofor is at least sequentially related to
previous activity. This is not as true of a lecture on the nature of
metamorphic rock in southern Scotland, or an ideological pamphlet
about how life would be better if humanity truly put Marxist
Socialism into action.
A good example of this contrast can be seen in the first chapter of
the Gospel of John. In the first five verses (in particular, but this is
also true of vv. 6-18), the text describes the nature of 'the Word', and
there is only minimal reference to sequences of eventsmost of the
clauses are stative. In v. 19, however, the treatment of events changes,
and one event leads to another, and the emphasis is on sequence rather
than state.
Longacre points out63 that 'Narrative and procedural discourse have
chronological linkage, while expository and hortatory discourse have
logical linkage', which may be a useful confirmatory criterion in texttype identification.
2.2.1.3 The Matrix with Two Parameters
When we put these two parameters together, their combinations define
four broad categoriesNARRATIVE, PROCEDURAL, BEHAVIOURAL and EXPOSITORY.
+ Agent Orientation
- Agent Orientation
+ CTS
NARRATIVE
PROCEDURAL
-CTS
BEHAVIOURAL
EXPOSITORY
I will come back to this matrix shortly, to add another parameter, and
will interact with its categories more fully at that time; until then,
Longacre's brief summary of these classifications will suffice:
Narrative discourse (broadly conceived) is plus in respect to both parameters. Procedural discourse (how to do it, how it was done, how it takes
place) is plus in respect to contingent succession (the steps of a procedure
are ordered) but minus in respect to the agent orientation (attention is on
63. Longacre, 'Discourse', pp. 19-20.
96

what is done or made, not on who does it). Behavioral discourse (a broad
category including exhortation, eulogy and political speeches of candidates) is minus in regard to contingent succession but plus in regard to
agent orientation (it deals with how people did or should behave).
Expository discourse is minus in respect to both parameters.
A certain care has been taken in positing these initial parameters, for if
they are defined too broadly we get into difficulty in classifying some discourses. Thus, Hebrews chapter 11 is really an expository discourse on
faith. On first inspection, however, there are difficulties in classifying this
discourse. If, for example, we were to define the first parameter simply as
chronological succession (as in Longacre 1976),64 then Hebrews 11
would be plus in respect to this parameter since it orders its examples of
believing men and women according to the chronological framework of
the Old Testament. It is plain, however, that the chapter does not present
the actions of any person of faith as dependent on those of a person previously mentioned. There is chronological succession, but not contingent
succession. The writer is exemplifying faith and simply mentions his
various examples in the order in which they are mentioned in the Old
Testament. Likewise, while there is a great deal of action and many agents
mentioned in the chapter, the chapter is oriented towards those agents who
act as examples of faith. Furthermore, there is disparity instead of identity
of reference. Actually, then, if we define our two parameters carefully
enough, Hebrews 11unlike true narrativesis minus with respect to
both parameters.65
I am not content to leave this section without clarifying one thing

further. Longacre's discussion above of Hebrews 11 with reference to
Agent Orientation is unclear. He writes that, 'the chapter is oriented
towards those agents who act as examples of faith'; this, I believe is
slightly misleading, since he is trying to explain why the text is to be
considered '-Agent Orientation'. One might more clearly say, 'the
chapter is oriented towards faith itself, and because faith is a human
response to circumstances, this requires the text to deal with agents and
events. Thus, though there is to some degree a highlighting of agents
and actions, this is only to achieve the goal of speaking about faith.'
64. His exact reference here is opaque; he lists three works in his bibliography
that appeared in 1976it is unlikely, however, that they did not share the same
definition of these categories.
65. Longacre, Grammar, pp. 3-4; although his contribution to Brend and Pike,
Tagmemics ('Discourse,' pp. 1-44), is concerned largely with the developmental
history of this area of analysis, he provides here as well a good summary of these
parameters and their resultant text-types (pp. 18ff.).
97
2.2.1.4 The Third Parameter: 'Projection'

Waterhouse writes, concerning the above-described four 'types of
discourse':
At workshops of the Summer Institute of Linguistics in the Philippines in
1967-68, Longacre began work on a broad front in a variety of languages.
This resulted in a three-volume work (Longacre 1968-96): Discourse,
Paragraph, and Sentence Structure in Selected Philippine Languages.66
He found four major types of discourse: narrative, procedural, expository,
and hortatory. These occurred in most languages investigated. Narrative
discourse was that which told some type of story. Procedural discourse
told how something was done or made. Hortatory discourse attempted to
influence or change conduct or outlook. Each type of discourse was
described in terms of its chronological and person orientation; its
initiating, closing and nuclear tagmemes; and the types of linkage of units
within it.67
A matrix with only two binary oppositions does not allow us to differentiate texts that are clearly different in make-up. For example,
one Procedural text may record how something was made or done,
but another Procedural text may instruct the reader for the making or
the doing. 'Narrative discourse' includes prophetic, as well as historical, narratives. To represent these in the matrix we require another
parameter: Projection.
This parameter is, I think, the easiest to grasp. If a text is 'plus
Projection', it looks toward the future in some way; if it is 'minus
Projection', it does not. For example, a set of instructions about how
to turn lead into gold, will be 'plus projection'; a lab report about how
lead was turned into gold, will be 'minus projection'.68 A prophecy
about the fall of an empire, and the historical report of its fall, differ
in terms of 'Projection'.
2.2.1.5 The Matrix with Three Parameters
The addition of Projection to the matrix divides it into eight sections,
and gives us a means of defining eight notional text-types:
66. (Summer Institute of Linguistics Publications in Linguistics and Related

Fields, 21; Santa Ana, CA: Summer Institute of Linguistics, 1968).
67. Waterhouse, History, pp. 48-49.
68. And fictional. These parameters do not have to do with such things as truth
content, injection of humour, and other such features.
98

+ Agent Orientation
- Agent Orientation
NARRATIVE
PROCEDURAL
Prediction
How-to-do-it
+ Prqj.
Story
How-it-was-done
- Proj.
BEHAVIOURAL
EXPOSITORY
Exhortation
Promisory Speech
Budget Proposal
Futuristic Essay
+ Proj.
Eulogy
Scientific Paper
- Proj.
+ CTS
-CTS
These categories, though descriptive of the 'deep structure' of language, are commonly seen in the surface structures of languages as
well, and therefore provide labels, and rationale, for handling them
independently of one another. This is the greatest value of such a
matrix: it enhances our perception of distinctions that are marked
(perhaps only subtly) in real language data.
These eight categories69 are distinctive, and, in those which are
distinctive for a given language, they are marked by the surface
structure of that language.70 It would be beneficial, perhaps, to
illustrate these notional categories, giving an example from each, on
the same topic. For the sake of simplicity we will take 'tying one's
shoes' as the subject:
In the Narrative category (+ AO, + CTS), we could have a prediction
69. My own 'technical labels' for these categories (by which I will refer to them
in the analyses that follow) are slightly different from those given above; they are:
Narrative Prediction, and Narrative History; Procedural/Instructional, and
Procedural/Lab Report; Hortatory (which is the only type of Behavioural text we
encounter in the texts examined in this study); and Expository (when a strict distinction between ' Projection' texts is required, I refer to the former as Expository/
'What-it-will-be', and the latter as Expository/'What-it-was'). I will always capitalize
these terms when they refer to text-types.
70. Not all of these categories are distinguished in the surface structure of each
language, and in some languages the 'encoding' of one notional structure may be
very similar to, or may partially or completely overlap, the 'encoding' of another.
99
about a child tying its shoes'at the age of four, he will tie his shoes
without help; he will reach down, take a lace in each hand, and cross
them; he will then...' and so on ( + Proj.); likewise, the same thing
could be looked on in the past, as history'and at the age of four, he
suddenly knew how to tie his shoes: he just reached down, took a lace
in each hand...' and so forth ( - Proj.).
If the topic of tying shoes is found in the Procedural category
( - AO, + CTS), the agent will be mentioned only because the activity
requires one; so, a set of clinical instructions will write, 'when the
laces have been crossed, one lace is tucked under the crossing, and the
two are pulled reasonably tight; a bow is formed next, by...'
( + Proj.); this can be contrasted with a lab report on how this action
is accomplished: 'the wearer crossed the laces, and tucked one under
the other; the laces were then drawn tight enough for comfort, and a
bow was then formed. This was done by folding one of the laces...'
( - Proj.).
In the Behavioural category ( + AO, - CTS), an exhortation by a
judgmental peer might run like this: 'If you want your laces to stay
tied, you've got to tie them tighter. Pull the laces harder...when
you're doing the bow, hold onto it longer...and make sure before you
start that your laces are even...don't cross the laces the wrong way
either, when you're making the first knot...' and so on ( + Proj.); a
' - Proj.' example of this might be imagined as a reflection by a
widow on the precision with which her husband used to tie his shoes:
'he always made sure his bows were the same length, and he would
make sure he had pulled the laces securely tight before he began the
process. He didn't mind if he had to do it twice, but it had to be right.
He had a habit of patting each side of the shoe when he had finished,
and he usually brushed them off with his hand just before starting to
tie them...'
In the Expository category ( - AO, - CTS), we are dealing with
primarily descriptive material. A ' + Proj.' Expository text might
read, 'When the child learns to tie a shoe properly, the bow will be
neat and secure, the laces even. The foot will be comfortable in the
shoe, and the laces will not be too tight'; if such a text is ' - Proj.', the
time reference will not be futureit may be present or past (though
the addition of further parameters could differentiate past from present): 'The shoe was beautifully tied; the laces were not too tight, and
the bows were even and...'
100
This rather exhaustive run on the permutations of 'shoe-tying' texts

should aid the reader to relate these categories to real language data.
2.2.1.6 The Question of other Parameters
This matrix could be complicated by the addition of other parameters;
Longacre suggests the addition of 'Tension'. This is a category which
permits distinction between, say, an account of history, such as the
book of Judges, where similarity of events, and repetition, feature
more highly than does the building of tension through a series of
scenes until a climax of tension is achieved and released, such as is
seen in the individual histories in Judges, or in the Gospel of John.
This same parameter is decisive in the distinction between a theological essay that is merely 'exploratory' and one with a polemic thrust
(e.g. one of Martin Luther's treatises against the theology of the
Roman Catholic Church of his day).
Sociolinguistic factors enter the picture as well, and complicate it to
a certain degree, especially with regard to Hortatory texts.71
In addition, I have alluded to other factors that could be included;
however, for each step toward greater specificity we sacrifice
simplicity and clarity. For the present work, I will limit differentiation to the three parameters given in the matrix above.
2.2.2 The Internal Structure of Texts
At this juncture, we shift our attention to the internal structure of
these text-level units. This structure is marked by features that can be
collected into two loose groups: the first includes those features which
tend to extend throughout the length of the textthey are roughly
similar to the 'warp' in woven cloth; the second includes those
features which tend to break up those in the first categorythey are
comparable to the 'weft' in woven cloth. As with woven cloth, it is the
working together of these features that results in a completed product.
2.2.2.1 'Longitudinal' Features of Texts
These are those features that extend throughout the text; they include
two pairs of categories that are in opposition to one another, as well as
several other features that tend to be examined in terms of the way
they thread their way through the text.
71. As we have already noted with regard to Longacre's treatment of these texts
in Joseph.
101
'Main-Line' versus 'Off-Line'

Many have remarked on the predominance of the we + Prefix conjugation in storytelling/historical sections of the Hebrew Bible; it has been
dubbed 'the narrative form'. It forms the backbone of the story, but is
occasionally interrupted with clauses of different types. It is comparable to the English 'simple past', the French 'passe simple', the German
'Imperfekt'. All these forms characterize Narrative History in the same
way that command forms characterize Hortatory texts. These are the
forms that are responsible for moving the 'story'/'exhortation' forward
towards its end. Longacre terms these 'on-the-line' or 'mainline'.72 It
has been found that every distinctive text-type in a language has a
clause-type that it prefers; I have already mentioned command forms
for HortatoryI might also point to the tendency for English
Procedural texts to employ the passive forms (particularly Procedural/
Lab Report texts, using the passive forms of past tenses), and the
tendency world-wide for Expository texts to rely on stative clauses.
In addition, there occur other clause-types, which do not have the
job of moving the text inexorably forward; these usually contain
material that sets the scene, or in some other way departs from the
main task assigned to a clause in their particular text-type. In
Narrative History, a Verbless clause is an anomalyit adds detail to
the context, but does not advance the story; conversely such a clausetype is the main-line form of Expository texts73where a clause with
a Suffix form of (e.g.) ntoy is inserted, then, it provides an event-setting which sets the scene for observations that relate to that context.
This type of clause, too, can (rather, must) be analysed according to
its text-type. Consider the following example:
Mistress Madwin had a third bed brought in and hastily assembled and
made up. One of the beds already there stretched nearly from wall to wall
in length, and had obviously been meant for Loial from the start. There was
barely room to walk between the beds. As soon as the innkeeper was gone,
Rand turned to the others. Loial had pushed the still-covered chest under
his bed and was trying the mattress. Hurin was setting out the saddlebags.
'Do either of you know why that captain was so suspicious of us? He
was, I'm sure of it.' He shook his head. 'I almost think he thought we
might steal that statue, the way he was talking.'74
72. Longacre, Joseph, pp. 64ff.
73. Or so I propose; we will return to this at several points in Chapters 4 and 5.
74. R. Jordan, The Great Hunt (New York: Tom Doherty Associates, 1991),
p. 316.
102
Who is the speaker in the second paragraphwho is 'he' who shakes

his head? How can we tell? At first blush, the identity of the speaker is
left ambiguous; however, if one traces through the previous clauses to
the last clause in simple past (the main-line form in English for storytelling), we find that all ambiguity is resolved. Mistress Madwin has
left the room; and the other two characters are occupied with things
that are recorded only as off-line clauses (a sequence of three passive
causative clauses are followed by several descriptive clauses, two
imperfect clauses, and one pluperfect clause). Only Rand is the subject
of main-line clauses.75
'Foreground' versus 'Background'
The opposition between main-line and off-line is a syntactic question;
the opposition 'foreground' versus 'background' is similar, but is
more a 'notional' distinctionin some ways, it is a deep structure distinction that is encoded by a surface structure opposition of mainline
clauses versus off-line clauses. Foreground material is that which
moves the story/exhortation/instructions/etc, towjard its essential goal
(whether that be the highlighting and resolution of a peak event, or
some other text-type-appropriate goal). Background material is that
which does not significantly advance the story/etc. Both 'off-line' and
'background' material can be categorized in terms of 'distance from
the main-line' or 'degree of backgrounding'; the more unlike the
main-line clause-type an off-line clause-type can be shown to be (in
terms of its tense-aspect-mood values, for example), the further offline it can be said to be. This is the principle behind Longacre's
dines.76
Payne's contribution to Longacre's Theory and Application in
Processing Texts in Non-Indoeuropean Languages,11 though borne of
75. Here we have an example of the interplay between several kinds of text-level
phenomena, the one (in this case, the analysis of main-line and off-line forms) helping resolve difficulties in another (the question of participant reference, which will be
discussed below, in this chapter). This is a case where either the writer was conscious of these features in writing this section, or his innate grasp of language style
has led him to craft this section in such a way as to maintain tension (through ambiguity) without abandoning clarity altogether.
76. I have already commented briefly on these in our examination of Joseph; they
will receive in-depth attention later in this chapter.
77. D. Payne, 'Activity as the Encoding of Foregrounding in Narrative: A Case
Study of an Asheninca [Central Peru] Legend', in Longacre (ed.), Theory and
103
research using a somewhat different model, provides a helpful tool

for assessing degree of foregrounding. The researcher assesses clauses
in terms of high versus low activity, with various factors being
important:
Mode
Realis (e.g. Affirmative, NonFuture, Indicative)

Aspect Telic (e.g. Perfective, Punctual)
Agent Agentive (e.g. Volitional,
Intentional)
Object Transitive (e.g. Individuated
Object)
Irrealis (e.g. Negative, Future,

Conditional)
Atelic (e.g. Stative or Durative)
Non-Agentive (e.g. Involuntary,
Accidental)
Intransitive (e.g. NonIndividuated Object)78
The oppositions 'main-line' versus 'off-line' and 'foreground' versus

'background' are more readily apprehended when seen in their
setting; and when we are working through texts, we will have ample
occasion to cite examples and fine-tune these concepts.
Participant Reference
This term refers to where and how participants enter, are referred to,
and exit, their texts. It is found that repetition of a person's name, for
example (in a story), when the context identifies the person without
him or her being named, can signal an important moment in the
storyline.
This is a topic that itself warrants a volume this size; yet, analysis of
this feature of texts will not figure highly in the material which
followsthe scope of this study will be sufficiently limited that I will
only make occasional reference to participant tracking (when, for
example, it provides significant confirmation of a doubtful hypothesis).
Application in Processing Texts in Non-Indoeuropean Languages (Papiere zur

Textlinguistik, 43; Hamburg: Helmut Buske Verlag, 1984).
78. Payne, 'Activity', pp. 63-64; this Transitive-Intransitive opposition is
defined differently from what we are accustomed to, and the Object category is
evaluated on a 'sliding scale' of sorts, so this approach would need to be examined in
closer detail than the above were it to be considered for application to our data; but
this much gives a good understanding of the issues involved in differentiating
foreground and background; even the title of this article suggests the 'deep structure'
nature of foregrounding, with its reference to 'encoding' (the process whereby deep
structure language propositions, etc., are rendered according to the surface structure
rules of any given language).
104
Topical Cohesion
Topical cohesion is a feature of texts similar to that of Participant reference; it refers to the continuity, or lack thereof, through a text of
the topic(s) referred to in that text. Tracking their introduction,
maintenance, relinquishment and possible reintroduction feeds information into an analysis of a text's internal structure. By and large, this
approach to the text will, like the above, serve no more than a
confirmatory role in the analyses that follow this chapter.
Other 'Longitudinal' Features of Texts
Other elements of a text may be examined for their clues to the
overall construction of the text. An example of such an element is
temporal referencewhere does it occur? and in what kinds of
clauses? Is all passage of time explicitly marked?and so on. Lexical
cohesion is another example: how many times do words having to do
with 'X' occur in this stretch of text? Is that consistent with the rest of
this text? and so forth. A good theoretical base will suggest likely
features to watch out for, and while any feature of a text can be
examined to see if it has a text-level role, some leads will be more
productive than others.
'Profile' Features of Texts
Features such as topical or lexical cohesion are examined in terms of
their continuity through the text. Other features are not. These are
'one off features, and the ensemble of these solitary occurrences
usually tell us much about the text as a whole. We will examine these
features in terms of their deep structure concepts and then in terms of
how they are realized in surface structure.
Plot
It may seem odd to run across this term here; is it not a literary analyst's term? Have we wandered outside our own domain? This may in
fact be true, but it merely underlines the interdependency of these
various ways of approaching texts. Each is seeking to understand how
the human mind creates and communicates meaning; many of the
underlying truths will surface through application of widely differing
procedures. This simply confirms the existence of these features as
real, rather than theoretical.
One such feature is 'plot'. Literary analysts (not to mention secondary school teachers) have long considered this one of the structural
105
givens of Narrative History. That stories tend to have a beginning,

development, high point and conclusion is fairly well acknowledged.
As a result of studying a wide range of languages, which have a wide
range of storytelling techniques, a hypothetical underlying structure
has been proposed.
Longacre proposes seven elements underlying the typical Narrative
History text:
1.
2.
3.
4.
5.
6.
7.
Exposition
Inciting Moment
Developing Conflict
Climax
Denouement
Final Suspense
Conclusion
'Lay something out'

'Get something going'
'Keep the heat on'
'Knot it all up proper'
'Loosen it'
'Keep untangling'
'Wrap it up'79
These elements underlie and inform the surface structure of not only
Narrative History texts, but also the other text-types. Perhaps some of
the terms may seem a bit alien for such a text-type as Procedural/
Instructional, or Hortatory, but the concepts of beginning, sustaining,
coming to the 'point', settling everything out and concluding, can be
found to have their place in texts of any variety of text-type.
Constituent Structure
As with any underlying, deep-structure 'notion', realization of these
concepts as surface structure features is not always straightforward,
but there is enough consistency to propose a set of [roughly] corresponding features:
I
II
III
IV
V
VI
VH
Vm
IX
Title
Aperture
Stage
(Pre-Peak) Episodes
Peak
'Peak Prime'80
(Post-Peak) Episodes
Closure
Finis
79. Longacre, Grammar, pp. 20ff.

80. This term is used where the deep structure 'climax' is encoded as more than
one surface structure peak; in languages that employ specific peak-marking devices (a
specific particle, for example), when such a device is used to mark two different
episodes, the second will be referred to by this term.
106
I, II and IX are identified by Longacre as 'formulaic',81 and are

surface structure features only. As is always the case in mapping relations between deep and surface structure features, one cannot always
map one-to-one correlationsin this case, for example, the deep
structure features '2. Inciting Moment' and '3. Developing Conflict'
both realize as 'IV. (Pre-Peak) Episodes'. The deep structure features
do not require us to seek exact replications in the surface structure,
rather they permit us to look for a range of potential structures in the
texts we examine. If we sought a specific system of surface features
the underlying organization of the text might elude us.82 Suffice it to
say, we are engaged in looking at texts to discover what they show us
in terms of their patterning; the awareness of deep structure features
will alert us to language tendencies that may be reflected in surface
structures in our texts, which might otherwise escape our notice.
Internal features or units that function to build tension or release it,
to confirm a point, to identify the key exhortation, and so on, suggest
an overall pattern to the text, a pattern of ebb and flow, of rise and
fall, which gives a sense of a unified whole. These features suggest
where we will find breaks in the text, and also that we will find
functional sections bracketed by those breaksand that these breaks
and functional sections fit into an overall plan.
These deep- and surface-structure features give to the text a
'profile'. This sort of thing can be illustrated readily from familiar
stories, such as 'The Boy who Cried "Wolf!"'. A boy, sent to watch
the flock of sheep in the hills, became lonely, and twice raised the
alarm when there was in fact no threat, to alleviate his loneliness. The
villagers ceased to trust his alarm-cry, and when the sheep were
indeed threatened by the wolf, the boy could raise no help.
Additionally, this story had a moral, and the listeners were meant to
take to heart the value of integrity and the consequences of sacrificing
it. Now, I haven't rigged the data; this is a well-known story, told in a
81. Longacre, Grammar, p. 20.
82. To go into greater detail on the nature of correlations between deep and surface structure features would muddy the waters here significantly. The purpose of
raising the issue is to demonstrate the fact that these surface-level features are
grounded in basic human functions, that they have a real cognitive existence which is
not derived from, for example, conventions of literary stylein fact, the converse is
true, that conventions of literary style, etc., derive from the normal innate human
tendency to organize texts along these lines.
107
very similar fashion in a wide variety of cultures. But it is the

sequence of episodesthe rising and falling of tension in the storyline, and the like, which breaks it into sectionswhich we recall,
rather than the word-for-word tale that we first heard. At a subconscious/intuitive level, we have internalized the 'profile' of this story,
rather than simply its surface structure encoding. The only way we
have been able to absorb this 'profile', however, is by interpreting the
clues left for us in such things as the introduction of off-line material
to break the flow, and the repetition of key words, phrases and structures, which create for us the patterns of the story.
2.2.3 The Overlapping and Interleaving of Features
In looking at texts, then, we are faced with more than a handful of
different types of features, some of which form part of the continuity
of the text, others of which function to segment it. These features
work together; rarely is there significant disharmony in the indicators
of the structure of a text (e.g. a text in which the participant reference,
lexical cohesion, backgrounding versus foregrounding, and plot
structure, all seemed to indicate radically different text structures).
Rather, all these features tend to point the same direction toward a
harmony of results. Thus, one often finds that the 'Exposition Stage'
section of a story is set apart from the remainder by containing a
much higher proportion of off-line material, and that boundaries,
which are indicated by breaks in the main line, are also confirmed by
topic shifts, introduction of new characters, and the like. Even when
exploring only one surface-level feature (e.g. a particular clause-type,
as does Eskhult),83 therefore, it is wise to keep an eye on other
features of the text, in order to see the single feature within its larger
and more holistic context.84
83. Eskhult, Studies.

84. In my comments on Eskhult's Studies I have criticised his narrowness, for
this very reason; likewise I have affirmed Khan's endeavour to identify features other
than Ex and PAR which could perform the same function(s). Khan stands a better
chance of understanding his chosen grammatical feature than Eskhult does, because
he sees it in a wider context.
108

3. Methodology
85
With these comments on the tagmemic model of text analysis, I have

endeavoured to present a sufficient theoretical base so that we can test
a few hypotheses raised by the model, against data from the Hebrew
Bible. The intermediate step, however, is to discuss the kind of
approach to the text this theoretical base will lead us to employ. There
will be two sections to this: (1) a set of general principles; and
(2) those general principles, given specific characteristics to fit our
specific context, and to meet our specific needs.
3.1 General Principles
Before going into the specifics of how to approach a text for analysis
using the Tagmemic' model, I must lay out a few basicsthese form,
as it were, the Credo:
3.1.1 The Primacy of the Data
Our first principle of methodology must be that the data are worthy of
analysis. If we approach a text (or any other research data, for that
matter) with the conscious or unconscious presupposition that it can
only be as we expect it to be before our contact with it, then we are
suggesting that the system we impose upon the text has greater value
than the system which produced it (i.e. that if the two are not
identical, the former is above suspicion). This is rarely the case.
Moreover, such a presupposition forgets that we go to the data to be
informed about the world that produced them, and that if such an
attitude had prevailed among previous researchers, the very
presuppositions with which we approach the data would never have
evolved.
We must consider the possibility that the data can correct our
presuppositions, else we might as well give up 'scientific endeavour'
altogether. The language had a life of its own before we came to it
85. In an earlier version of this study, I used the phrase 'these few comments'. It
has been suggested that this might annoy the reader, as thirty pages of tagmemic
theory will hardly seem like 'a few comments'. Yet, the reader will surely be aware
that this has been a rather sketchy introduction to the model; a much more thorough
introduction would be needed to do justice to the security with which this model
allows us to describe linguistic nuances.
109
with our 'theory'. We must discover how it came to be the way it is,86
and our only means to this end is to let it speak to us with its own voice.
3.1.2 Selecting a Manageable Task
With a healthy respect for the integrity of the data, we must then
decide what we would like to ask. We must use wise restraint in this.
Our selection of a topic must be informed by such concerns as (1) the
size and characteristics of the database, and (2) the amount of time the
researcher intends to give to the studyother factors, such as the
needs of the academic community at the time of research, will enter
the debate as well. There is little point in asking of Hebrew data, 'what
is the text-linguistic structure of the Pentateuch?', or 'what is the
function of syntax in poetic texts?', unless we are prepared for a long
and exhausting project, for example.87 The choice of which topic to
pursue will determine, to a surprising extent, the value, as well as the
accuracy, of the results.
3.1.3 The Theoretical Starting Point
Using the theoretical base introduced above for discovering text-types,
it would be foolish to embark on a phonological enquiry like 'which
syllable structures are permissible?'. This is because the model one
works from will be predisposed to function better for some kinds of
enquiries than for others. In addition, when one examines a language
from the perspective of a certain model, one will find that the state of
description, to date, of that language, will inevitably show certain gaps
when compared to what might surface if description were carried out
to its fullest on that language, using the model in question; that is,
certain questions, for which this model would be a particularly helpful
tool, will already have answers, and others will not. Therefore, we
must allow the model we choose to help us determine the enquiry in
which it will function.
86. Those who are more favorable to text-critical work will respond here that this
is also their goal; the difference I would cite between these two goals lies in the
underlying assumption of the one that the text is largely 'in a form intelligible to the
original users, if not to us' (my own view), while that of the other is that the text has
departed substantially from its original integrity, and that we have the knowledge and
tools to restore it (it is this latter assumption of which I am skeptical).
87. The additional fact that the groundwork for such a study has not yet been
adequately laid is another consideration.
110
3.1.4 Working with a Hypothesis

Once we have determined the scope, and starting point, of our topic,
we are embarked on a process that will be governed largely by the
question we ask.
Selecting a Working Hypothesis and a Data Sample
I have enumerated several factors above which will influence the
choice of question(s) we will ask of the data. The question must also
be formulated in such a way to lead us to a productive answer. 'Do
verbs have a macro-syntactic significance?' would be next to useless as
working hypothesis; we need a starting point that will result in a
concrete observation about the language. 'Does the distribution of
verbs with rrn in non-'Reported Speech' sections of Narrative History,
in comparison with other clause-types, indicate possible macro-syntactic significance?' is a much more functional hypothesis.88
Once we have chosen a topic, a theoretical starting point, and a
working hypothesis, we must also decide what will be included in our
data sample. In most language studies, the database is more or less
theoretically, if perhaps not functionally, unlimitedsince living
speakers will continually be creating further dataand therefore the
researcher must select appropriate material for a study (that is,
material particularly suited to the study undertaken). Thus, if one is
wanting to discover the text-level features of hortatory texts, it is
unhelpful to include prophetic narrative, expository and procedural
texts in the database, since these will clutter the study, making seeing
patterns that much more difficult. On the other hand, if one is wanting
to determine the similarities and dissimilarities between expository
texts and hortatory texts, it is wise to have several of both on hand to
allow suspicions to be confirmed or denied. In short, choices must be
made about the type of data examined, and the quantities thereof.
Other factors may influence the choice as well; ifas is true in the
case of Classical Hebrewthe integrity of the data may be called into
88. In fact, this is a kind of sub-hypothesis, since it would be correlative to other
questions we must ask at the same time. Perhaps the 'umbrella' hypothesis might be:
'Do suspected clause-level macro-syntactic devices for non-Reported Speech of
Narrative History converge to frame a complete picture of the constituent structure of
the text?'. Were we to be strictly legalistic about this, hypotheses could not be framed
as questions, but would have to take the form of 'if-then' sentences. This last
scruple is perhaps helpful on occasion, but I tend to bypass it, since the question
form accomplishes the task sufficiently well, and without any great ambiguity.
111
question on the grounds of textual variants and other, similar difficulties, it may be wise to look first at less suspect texts, so that the results
achieved can be evaluated without undue complications.
These choices, like any others in the process, add an element of
subjectivity to the endeavour, and this, of course, must be kept to a
minimum. Restrictions on the database are going to limit the scope of
the results to an equal degree, and so should be made wisely. On the
other hand, however, language contains such variety that simply
taking a handful of data at random rarely allows patterns to surface
(unless one is, for example, studying the frequency of a particular
feature in a specifically random sample of data).
Working to Disprove the Hypothesis
Now comes the difficult part: we have to disprove the hypothesis. If
we cannot do that, only then have we substantiated it. This is a
difficult discipline for a mind trained in 'the Arts'; 'pure' scientists are
more familiar with this principle. And yet all of us are aware that if a
theory about a feature is to become accepted as 'true', it must be able
to withstand all challenges; therefore, if the researcher proposing a
hypothesis throws as many of these challenges at the data as possible,
he or she stands the better chance of ushering the hypothesis into the
realms of accepted truth.
But this is tremendously difficult to do; once an hypothesis has led
the researcher to propose a solution, it is difficult to get it out of one's
head, and to ask of it if it might, in fact, be better explained another
way. Yet this is the only way of securing the proposal.
3.1.5 Charting
This is, strictly speaking, not a 'general principle', in the same way
the previous five sections have been, yet it nonetheless deserves attention at this point. When one examines 'texts', the very size of the units
of data imposes some difficulty; therefore, how we handle our data is
at least as important as in other types of research. When studying
whales, one's 'field' methods will be slightly different from those used
to study goldfish. Beavon writes,
The analysis of texts depends on the use of charts. The better the chart,
the more readily one sees the structure of the discourse.89
89. Beavon, 'A Partial Typology of Konzime (Bantu [Cameroon]) Discourse',
112
Choice of charting techniques must eventually rest with the individual

using them; and, of course, the more creative the person, the wider
variety of charting techniques available. The principal issues which
must be resolved here are (1) the construction of a chart that actually
allows patterns to surface, and (2) achieving a balance between clarity
and detail. If too much detail is highlighted, clarity is obscured, and
patterns may be overlooked; if too little detail is included, patterns
may be overlooked.
3.1.6 Drawing Conclusions and 'Feeding' the Theoretical Base
This stage involves determining what, exactly, the data have brought
forth during the enquiry. There are two facets to this: (1) description
of the results in and of themselves; and (2) comparison of these results
with those which the theoretical base led the researcher to expect at
the outset. The first requires little comment, other than that it is
wisefor the sake of one's reputation as a scholar, if for none
otherto be circumspect in one's write-up, and that one's goal in this
description is (a) to be true to the data, and (b) to communicate
clearly one's findings as economically as possiblein a word,
'elegantly'.
The second, on the other hand, is less transparent. The point is this:
if we start by assuming the primacy of the data, and something goes
awry, then we must ask questions of (1) our theoretical base, and
(2) our methodology. If, for example, we look for a strongly
suspected pattern, and find none, we must ask if our charting
technique, or perhaps our choice of data-sample, has been at fault.
Yet, perhaps the charting technique and the data-sample were beyond
reproachthen it is time to ask the same question of the theoretical
base, since it suggested that such a pattern would exist.
Returning to the example we used earlier, a phonological survey of
possible syllable-types: if our theory says that certain patterns are
never permitted, and yet we find words in which these patterns exist,
then we must ask if perhaps we have erred when dividing our words
into syllables, and have included phonemes in certain syllables that
actually belong in others. If this does not seem to be the case, then we
must amend our theory to permit such syllable-types.
p. 250 (in Longacre [ed.], Theory, pp. 210-55); here 'discourse' is used in the
sense of 'text'.
113
3.2 Specific Methodological Plan for Chapters 4 and 5

Having now become somewhat acquainted with some of the factors
that inform one's approach to the text, the reader will now be able to
see how these qualify my own approach to the data in the analyses that
follow.
3.2.1 Yet again, 'The Primacy of the Data'90
I have argued that the data are the only independently existing, and in
that sense the only truly real, element of language description. The
models and procedures with which we approach the data serve merely
to aid us to determine what the data really are. This means, then, that
textual emendation of the book of Exodus, for example, will be a last
resort during this enquiry; it means as well that questions of scribal
additions to the text of Ruth will not deter us from examining the text
as a unit, for we presume that scribes were the literati of their day,
and if they 'contributed' to a text, they will have done so in way that
did not violate the kinds of patterns we wish to examine. In short, it
means that, in the pages to follow, I will be prepared to call into
question theory and methodology before calling into question the data
pool.
3.2.2 Selecting a Manageable Task
This is a slightly different matter. Given that I have chosen (a) to
evaluate several other works at the outset of this volume, and (b) to
introduce the tagmemic model of linguistic analysis, and the principles
of its methodology (in rather a greater degree of detail than is usually
90. I may be accused of whipping a dead horse, here. The overstatement of this
principle, however, is for polemic reasons; it is clear from even a cursory
examination of the history of biblical studies, that unchallenged presuppositions have
controlled the field for extended periods, resulting in what might be termed a rather
insular attitude among scholars. Modern linguistics, with its greater openmindedness
to new discoveries from the data, has much to offer the study of Classical Hebrew,
but for the most part, hebraists are less than enthusiastic about its benefits. The
perspective of 'if it ain't broke, don't fix it' appears to be in the mind of the reticent
biblical shcolars here, yet the fact remains that the 'old ways' have programmed the
researcher to find only what has already been found in the data, rather than anything
new. It 'ain't broke', but the data have been pointing in directions that are obscured
by the presuppositions of the 'old ways'; we must allow our presuppositions to be
corrected by new insights from the data.
114
seen in such works), I am then left with less space in which to present
substantiation of my 'claims', than I would like. I have chosen to
compensate for this factor in two ways: (1)1 have selected a narrowly
circumscribed topic for examination; and (2) I have decided to
address this topic with a lessened concern for thorough substantiation
than would otherwise have been the case, in order that a greater
number of seeming 'loose ends' from the previous chapters may be
tied together for the reader.91 And since the theoretical base I have
elucidated here has focused to a fair degree on the identification of
text-types, and on questions of main-line versus off-lineand since
these were singled out in Longacre's work on Genesis 37, 39-48, as
his most significant insight for contemporary text-linguistic description
of Classical Hebrewit is most apposite that we turn our attention to
identification of Classical Hebrew text-types, and their attendant mainline and off-line clause-types.
3.2.3 The Theoretical Starting Point
We have looked at Longacre's matrix of 'notional' text-types, and I
have cited his 'verb rank clines' of main-line and off-line forms as
being particularly productive for Classical Hebrew; we have also
looked at constituent structure of texts as something that may be
marked by off-line features. I will therefore take as my starting point
these theoretical concepts, and examine the data to see whether they are,
in fact, viable for describing our language. I have also dealt bluntly
with Niccacci and Eskhult, in terms of their treatment of 'discourse'
(i.e. 'Reported Speech'), and have suggested that their analyses are
deficient because they do not deal well with this feature of the text.
Therefore, the enquiry about text-types, main-line versus off-line
clause-types, and constituent structure of texts, will do well to include
reference to the Reported Speech feature as well.
The sections that follow present the verb-rank clines for four texttypes as a hypothetical representation of what this theoretical base
leads us to suspect.
91. All this is based on my firm conviction that treatments of broader topics, with
fuller substantiation, will appear in the not too distant future to offset the shortcomings of this particular study.
115
The Narrative History Cline

Although this cline has already been presented above in the discussion
of Longacre's Joseph, I include it here for the sake of completeness.92
NARRATIVE DISCOURSE TEXT-TYPE VERB-RANK CLINE
Bandl:
1. Preterite
Storyline
Band 2:
2.1 Perfect
2.2 Noun + Perfect (with noun in focus)
Backgrounded Actions
Band 3:
3.2 Participle
Band 4:
4.1 Preterite of hayd, 'be'
Setting
4.2 Perfect of hayd, 'be'
Band 5:
5. Negation of verb clause: irrealis (any band)
'momentous negation' promotes
5. to 2.1/2.2.
The Narrative Prediction Cline93

NARRATIVE PREDICTION TEXT-TYPE VERB-RANK CLINE
1. we + Suffix
Band 1:
Line of Prediction
Band 2:
2.1 Prefix
Backgrounded
2.2 Noun + Suffix (with noun in focus)
Predictions
Band 3:
3.2 Participle
Band 4:
4.1 we + Suffix of hayd
4.2 Prefix of hayd
Setting
92. Longacre, Joseph, p. 81; I have stated that I believe that the degree of
detail in these 'clines' is difficult to accept with unreserved confidence, and these
various, finely tuned, layers will not be contrasted with one another in the analyses to
follow.
116
The Hortatory Cline94

HORTATORY
Band 1:
Primary Line of
Exhortation
Band 2:
Secondary Line of
Exhortation
Band 3:
Results / Consequences
(Motivation)
Band 4:
Setting
(Problem)
TEXT-TYPE VERB-RANK CLINE

1.1 Imperative (2nd person)
1.2 Cohortative (1 st person)
1.3 Jussive (3rd person)
2.1 7 + Jussive / Prefix
2.2 'Modal' Prefix
3.1 we + Suffix
3.2 7 / pn + Prefix
3.3 Suffix (with future reference)
4.1 Suffix (with past reference)
4.2 Participles
The Expository Cline

Longacre does not posit a cline for Expository texts. Sufficient
research has not yet been done to permit such a specific hypothesis,
but I will hazard here a tentative cline:
EXPOSITORY TEXT-TYPE VERB-RANK CLINE
Band 1:
1.1 Nominal clause (Verbless)
Primary Line of
1.2 Existential clauses (with 'en or yes)
Exposition
Band 2:
[Clauses with hayd]
Secondary Line of
Exposition
Band 3:
Band 4:
Contest/Setting
[Here we would expect to find those clausetypes which have the greatest amount of
action and transitivity]95
3.2.4 A Working Hypothesis and Data Sample

The theoretical details are coming together enough now that we
can formulate a working hypothesis. So far, we have affirmed (a) the
95. I refer the reader again to Longacre, Joseph, pp. 11 Iff.: 'Elements at the
bottom of the clines for Narrative, Procedural, and Hortatory discourse have the
highest ranking in Expository' [p. 111]conversely, those clause-types which are at
the top of the other clines will be at the bottom of this one. I will examine this more
closely in the next two chapters, but my results even then will remain tentative.
111
primacy and integrity of the data, (b) the need to select a fairly
restricted topic, which (c) ought to be related to the material presented
in Longacre's matrix and clines, and to 'Reported Speech' versus 'nonReported Speech'. Therefore, I have chosen to ask, as a preparation to
shaping a working hypothesis, 'Can we substantiate the existence of
main-line clause-types for more than just the Narrative History texttype? and do these have relevance for Reported Speech as well?'
If we frame these in terms of a testable hypothesis, we might say the
following:
Main-line clause-types are text-type specific, and can be described as
such; they will predominate in the text, and text-types can be identified by
the predominant clause-type.
The main-line clause-type for the Narrative History text-type is the we +
Prefix clause.
The main-line for the (Narrative) Predictive96 text-type is the we + Suffix
clause.
The main-line clause-type for the Hortatory text-type is built on a
'command' form.
The main-line clause-type for the Expository text-types is the Verbless
clause.
Other text-types, whose clines have not been described nor intimated by
Longacre in Joseph will be identified first by features other than 'mainline' clause-types (as these have not yet been proposed), and then clause
distribution within those texts for which we have able to posit a text-type
identity, will be examined with a view toward proposing their main-line
forms.
The constituent structure of texts will be marked by divergences from the
main-line form in all text-types; off-line marking of constituent structure
will be confirmed by other types of marking devices, and will reflect a
comprehensible underlying notional structure;
The results of the above analyses will be expected to hold true for
Reported Speech as well.
This sequence of hypotheses is simply a specific outworking of the

question we have posed above, though it may look like a series of
rather too loosely related studies. In addition, one may be tempted to
96. I will occasionally refer to this using both labels, to reaffirm the close
connection between this text-type and Narrative History.
118
object that I have proposed to keep this enquiry rather severely

circumscribed, and that this list does not seem in keeping with that
goal. However, my charting techniques and so on will enable me to
kill several birds with one stone; therefore I will have occasion to test
more than one hypothesis with it.
In spite of claims at several earlier points that the corpus of data is
small, I will, nonetheless, need to restrict further the material I
examine. Given the issues with which I am concerned, and the space
limitations of this particular study, I will work with no more than,
say, a dozen texts, and not many of these in great depth.
It will be advantageous to employ texts with clear boundaries, for
this grants a ready assurance that we are dealing with a structural unit
that is natural, rather than artificial and perhaps misleading. Shorter
texts will be easier to work with than long ones, but they will need to
be of such a length as to show at least some internal structurefewer
than 50 clauses in a text, to pick a number somewhat at random,
would mean that the text will not be likely to show us much about the
language's mechanisms for marking paragraph divisions, and so on.
As I mentioned earlier in this chapter, textual concerns with the
Hebrew Bible recommend to us that we avoid, where possible, texts
that contain seriously suspect sections. It will be seen that our choice is
not severely restricted by this caution.
Since I am taking as a starting point the more readily accepted view
that the we + Prefix forms are the main-line forms for Narrative
History texts, it will be wise to examine this hypothesis first, both to
confirm its veracity and to demonstrate the effectiveness of the charting techniques before moving on to texts-types whose results may be
received with less confidence.
Following this, a text from another text-type will be required. Since
I cannot hope to substantiate the existence and nature of every texttype in the matrixowing primarily to the limitations of this study
some will have to be left behind. Hortatory texts are particularly
familiar, and that they employ command forms is not difficult to grasp;
therefore, I will take for granted, in this study, Longacre's Hortatory
Verb-Rank Cline. In addition, Narrative Prophecy has received
similar attention in the literature to that received by Narrative
History, and its characteristics are likewise more readily accepted
without extensive prooffor the sake of this preliminary study, that
isand therefore I will 'borrow' Longacre's cline for this text-type as
119
well. Because there is considerable Procedural/ Instructional material

in the Hebrew Bible, this is a logical choice for the next text-type to
be examined. Several of these suit the needs of this study admirably
(being well-defined units of a reasonable length, with few textual
difficulties). In addition (and perhaps most importantly), we are
fortunate to have a pair of textsthose concerning the building of the
Tabernaclethat differ from one another solely by text-type: one is a
set of instructions, the second is a report about the completed process.
This pair of texts will allow me to compare these two text-types
without the complication of differing subject matter.
These are the major factors involved in the choice of texts I will
look at in this study; there should be little cause for complaint with
regard to the decisions made (particularly in that, at the end of this
study, I will not make any hard and fast claims to accuracy, but rather
claim only to have pointed the way forward in this discipline).
3.2.5 The Charting Methodology
Because my first hypothesis centres on the question of whether the we
+ Prefix clause-type can be demonstrated to be the main-line clausetype for Narrative History, my first decision regarding the analysis
that follows was to display the text 'one clause per line'; I have also
chosen to separate out those factors which were irrelevant to my
quest. This means that syntactically subordinated material is placed in
a separate column, because clause-initial particles preclude the use of
the we + Prefix form, and therefore the absence of such a form is
easily predictable. For the same reason, Reported Speech material is
separated from non-Reported Speech material (because the former
clearly contains other text-types than Narrative History, and would
clutter up the Narrative History material if included with it, to the
point that no patterns in Narrative History would be discernible)
but, because Reported Speech might, nonetheless, contain some
Narrative History material, I have decided that subordinated material
should likewise be placed in a column separate from non-subordinated
Reported Speech material, in order that similar processes of
evaluation may be carried out on Reported Speech sections as on nonReported Speech sections.97
97. And because it was found that subordinated clauses more than occasionally
contained, on the one hand, clauses coordinated within that subordination, and on the
other hand, clauses subordinated within that subordination, it was decided to allow
120
The distribution of clauses in non-subordinated non-Reported

Speech material can thus be examined to see if patterns regarding this
distribution can be found, drawing from the three theoretical starting
points considered to be primary: (1) text-type identification according
to the matrix; (2) main-line clause-type identification, in comparison
with Longacre's clines; and (3) the reliability of off-line clause-types
as constituent structure / macro-syntactic markers.
Because these patterns are more visible if the clause-types are
immediately discernible, it was decided to use a colour-coding
scheme.98
3.2.6 Drawing Conclusions and Feeding the Theoretical Base
I have talked about the primacy of the data, and that the results of the
research must be allowed to inform one's approach to the data, and
one's theoretical base. It might, therefore, be instructive for the
reader to have a look behind the scenes, so to speak, at earlier,
preliminary studies, with regard to some of the abandoned hypotheses
which at one point accompanied or preceded the ones given above.
At the outset, I examined Hosea, hoping to propose something with
regard to its constituent structure based on clause-type distribution.
the distinction to be marked in the charting: where subordinationor other break in
the flow from one clause to the nextoccurred, a space was inserted between lines;
otherwise, whether within subordination, Reported Speech, or neither, no space was
inserted, implying continuity.
98. In the dissertation that lies behind this study, 16 colours were originally used
to mark by circling all manner of possible macro-syntactic indicators; this rather
robust spectrum was found confusing by all but myself, so it was pared down to
eight by coalescing such categories as hinneh, lo', and yes (among others), into a
single category, and by eliminating others from the marking scheme altogether (e.g.
infinitives, which were found to have no substantial relation to the hypotheses under
examination). This color-coding was perhaps the most striking feature of the original
charting procedure, and enormously helpful. Owing to printing restrictions, this
color-coding scheme cannot be reproduced for this volume. An alternative scheme
would be to underline those items considered to be the principal marking device for
the main-line, and circle any feature of interest that was not the main-line marker
(e.g. the we + Prefix forms in a Narrative History text would be underlined, and
such features as hinneh, Suffix forms, and Verbless clauses would be circled; this
scheme is less satisfactory, however, for it fails to provide a means for identifying at
a glance different text-types, marking items according to function rather than form,
and the function of a form will necessarily vary according to the text-type in which it
is found.
121
This was not successful, since I had not yet investigated the Narrative
text-types, nor had I begun to suspect that 'poetic' style might not
permit as simple an approach as I was hoping to adopt; I found this
material difficult to work with for these reasons, and no patterns surfaced. I next looked at Jonah, then Ruth, and these Narrative History
texts proved more fruitful.
I had initially presupposed that infinitives were clause-level features, and I attempted to examine them as macro-syntactically significant; this was relatively fruitless, and I dropped it from my 'theory'.
In addition, it looked as though participles had a rather interesting
distributionthat they were used attributively or substantively only in
subordinated material, and predicatively only in non-subordinated
material (i.e. that they had a non-overlapping distribution)this,
however, was not borne out by examinations of other data, and turned
out to have been a result of faulty analysis of a few occurrences of the
participle. My methodology was adjusted to prevent further mistakes.
Further, I had taken strong exception to Niccacci's claim that
'narrative discourse' was describably distinct from Narrative History
found outside of Reported Speech. I challenged this in my work with
the data, and was both justified, and corrected, by the data: I found
that, yes, what Niccacci calls 'narrative discourse' is indeed distinctive, but that this is traceable to factors related to the embedding of a
unit of Reported Speech in the Speech Formula clause, rather than it
being a distinctive type of text in its own right. In this case, I had
distrusted Niccacci's observations, while they were accurate, and yet
found a different explanation which, I believe, describes the data more
accurately and elegantly than his.
4. Final Comments: On the Relationship of this Chapter to the
Following Two Chapters
It will be clear to the reader, both from the nature of this volume, and
from the way the preceding sections have been written sometimes in
future tense, sometimes in present, and sometimes in past tense, that
the analysis that follows this section was undertaken prior to it. Yet, as
data returns us to theory and methodology, which then inspires us to
return to the data, this is not unfitting. The approach which we will
undertake with regard to the following material has been adequately
laid out in this chapter; further explanations, and reiterations, of
122
the 'hows and whys' of this analysis will accompany the procedures
themselves. The conclusions drawn from the analysis will be interspersed in the accompanying comments, and will be summarized in
the final chapter. The reader should be in no doubt, however: the
material that follows this chapter is not intended to be a thoroughgoing text-linguistic analysis, but serves, rather, as an illustration of
the application of Tagmemic text-linguistic theory and methodologies
to Classical Hebrew data
Chapter 4
TEXT-LINGUISTIC ANALYSIS OF NARRATIVE
AND NON-NARRATIVE TEXT-TYPES, WITH DATA SAMPLES

FROM JUDGES, LEVITICUS, AND EXODUS
1. Introduction
Now that we have examined the theory and methodology that will
instruct our approach to the data, we can look at a few texts. I propose
first to analyze briefly a short section from Judges which will give us
a chance to test the water with that linguistic text-type which is best
known in biblical studiesNarrative History. I will then turn to some
Procedural/Instructional material, from Leviticus; also from Leviticus
is a text that will illustrate the concept of 'embedding'. The final
section in this chapter will look at material from Exodus on the
building of the Tabernacle, another Narrative History text, and its
parallel, a Procedural/Instructional text. This will give us a chance to
look at a new text-type in contrast with a nearly identical version in
the more familiar Narrative History text-type.1
1. In phonological analysis (analysis of the sound system of a language), for
example, one looks for 'minimal pairs' of words to confirm that a certain pair of
sounds are contrastive at the 'emic' level (significant for the language, able to carry
distinctions of meaning, etc.), rather than contrastive at the 'etic' level (below the
level of awareness of the native speaker, etc.)the words top, 'call', and jrip,
'tear', demonstrate that the contrast between t* and s is significant for Classical
Hebrew. On the other hand, the English /k/ sound may have as many as nine 'etic'
variants (distinguished by where the consonant is articulated, and how it is released);
yet it is impossible to find a 'minimal pair' where the difference in meaning between
two words can be attributed to a contrast between any of these variants (for example
an aspirated /k/ and an unaspirated /k/). Finding a minimal pair where one word had
an aspirated /k/, and another of a different meaning, whose only phonetic difference
was a lack of aspiration after the /k/, would be sure evidence of aspiration as a significant feature of the language in question. In the same way, the existence of these two
nearly identical Exodus texts is a substantial confirmation of the existence of different
124
I have noted the frequency with which Hebrew narrative text level
patterns have been acknowledged by contemporary (as well as earlier)
writers. This being the most common text-type in the Hebrew Bible,
its features are more readily perceived. My treatment of Judges 2, a
Narrative History text, will be brief, for two reasons: (1) it is assumed
that the reader will not be surprised at the basic text-level features of
Narrative History (its predominating forms, etc.), these features often
having been discussed in the literature; and (2) we will be returning to
another Narrative History text in the following chapter, where we will
examine in much greater detail those and other text-level features.
2. Judges 2
The choice of Judges 2 for our examination is somewhat arbitrary, yet
because it is not very complex in structure, does not present any serious textual difficulties, and does not rely heavily on reported speech as
a means of carrying the story-line forward, it is a good starting point.2
A glance at the role of Judges 2 in the book of Judges will be helpful. Judges 2 is part of the introduction to the series of stories which
follow and which focus on specific leaders. In this chapter, however, a
synopsis is given, highlighting the cyclical nature of the reported history and giving a moral assessment of this cycle. It opens with a message from the angel of YHWHan indictment of the people of Israel
for their failure to obey YHWH's commands, and a statement of his
refusal to intervene any further for them. After this opening section
of reported speech mechanism, the story-line proceeds by simple
narration. Israel's apostasy, and YHWH's rejection of Israel,
announced by YHWH's messenger, is followed by Israel's repentance,
YHWH's response and Israel's salvation; this is detailed as the cycle
that will be replayed throughout the rest of the book of Judges. The
simple narration that forms the body of the chapter is bracketed by
reported speech, ending the same way it began, with another message
of judgment from YHWH.
This section is set off from the preceding material by two clauses
that are irregular for the Narrative History text-type.3 The syntactic
text-types, for the significant differences between the two texts are explainable by
'text-type' alone.
2. The full text of Judg. 2, in four-column format, is found in Appendix 1.
3. The first, 1.35.3, is a we + Prefix < irn clause, the second is a verbless
4. Text-Linguistic Analysis
125
marker of the second episode boundary (2.23.2) is what Longacre

calls a 'momentous negation' clause:4 i;oin"T3 D3n]"î 'and he did not
give them into the hand of Joshua'. This boundary is also marked by a
topic shift in the following episode (episode boundaries need not
always be syntactically marked; it is usually the combination of features, syntactic and otherwise, rather than a specific mechanism, that
conveys the desired signals), and, I propose, by a change to the
expository text-type.5
It is the material between the two judgment oracles that will require
our attention at this stage. Of the 80 clauses in this episode,6 22 are
found in reported speech; I will observe here only that, of these 22, 9
are subordinated. The remaining 58 clauses are simple narration; of
these, 13 are subordinated, with the rest (45 clauses) non-subordinated.
Now, these 45 clauses include a variety of types:
1
1
1
2
5
35
Ellipsis (2.18.1)
Asyndetic Suffix clause (2.17.4)
Suffix clause with -i copula (2.19.2)7
Suffix clauses with a preposed element; one with rrn
Negated suffix clauses
we + Prefix clauses; 3 with rrn
45
non-subordinated clauses
A full 82% of these clauses are we + Prefix clauses. This is clearly the
form of the verb that is preferred for conveying Narrative History
information. We can separate out, provisionally, the we + Prefix
clauses with rrn, since they tend to indicate states rather than events
clause; both are common devices for indicating a break in the flow of the narrative, as
we will see in the following chapter.
4. That is, a clause in which an important event is indicated by the lack or failure
of an action.
5. I will discuss the Expository text-type more fully in the following pages.
6. The consideration as 'clauses' of at least two of these (2.18.1 and 2.22.3,
both of which are elliptical) may cause some consternation, but I have nevertheless
chosen to include them in the count as full clauses; it will be seen that this approach
does not greatly influence the analysis.
7. In this, and in the preceding category, we have two examples of what
Niccacci says cannot happen (The QATAL which has first position in the sentence is
distinct from a second position QATAL. The first kind occurs in discourse [my
'Reported Speech'] but never in narrative' [Syntax, p. 30]). My analysis will have
less difficulty explaining this feature.
126
(and thus do not advance the main line of narration), and are often
used to signal paragraph and other macro-syntactic boundaries. Even
with this taken into consideration, we + Prefix clauses (with verbs
other than rrn) account for 76% of the non-subordinated clauses.
It is this kind of distribution which numerous authors have noted,
and which justifies reference to this form of the verb as the 'narrative'
form, and the like.8 The identification of this clause-type as the mainline Narrative History clause-type is the starting point for our
research into text-types and their uses of clause-types.
Yet we are faced with approximately a quarter of the clauses in
non-subordinated narration being what is termed 'off-line'. What are
these doing? If the we + Prefix clauses are main-line, advancing the
narrative by consecutive events, and so onif these are the bones of
the narration, what then are these other clauses?
In brief, if the former are the 'bones', the latter are the 'joints'. We
have seen in Longacre, in a variety of works on other language
groups, and in his work on Hebrew (and his findings are confirmed
independently by others such as Niccacci),9 that material in non-mainline clauses adds to the narrative, not by moving it forward, but by
contributing background information and creating a setting for the
narrative. Theses studies have shown that non-main-line information
impedes the flow of narration, and therefore serves the purpose of
arresting the reader's progress, either to highlight a particularly significant moment in the narration, or to provide means of distinguishing one sub-section of a narrative from another which follows it.10
It is important to note that these non-main-line clauses are evenly
interspersed through the text; they tend to occur in groups, breaking
the flow of main-line clause-types. In Hebrew, as in many other
8. We have already referred to GKC's comments on this form: 'imperfect with
waw consecutive... serves to express actions, events, or states, which are to be
regarded as the temporal or logical sequel of actions, events, or states mentioned
immediately before. The imperfect consecutive is used in this way most frequently as
the narrative tense...' (11 la). While GKC's description of this form is inadequate
at some levels, it must be admitted that the quote above has captured almost exactly
the sense of Longacre's concept of ' + Contingent Temporal Succession'.
9. See Niccacci, Syntax, p. 35, et passim.
10. See, in particular, Longacre's work, but I have also pointed out Niccacci's
comments about main-line and off-line formssee esp. my p. 29 (Niccacci, Syntax,
pp. 71 and 107); these are but a sample of what is becoming a widely substantiated
understanding of text-level features.
127
languages, the more off-line clauses included at a specific point in a

text, the greater the focus on the division being signalled, or on the
event being highlighted. It is worth graphically examining this
interruption of main-line 'flow'.
In the chart on the next page, the clause numbers are indicated in
the left-hand column of each section;11 moving to the right from that,
we have main-line non-subordinated clauses; in the next two columns
we have off-line non-subordinated clauses, and subordinated clauses,12
respectively.
Clauses 2.1.3-2.3.4 and 2.20.3-2.22.3 are the reported speech sections.13 In this chart, they can be seen clearly to bracket the remaining
material. Each clause is marked on the chart with a dot, indicating the
kind of clause (with reference to the main line)with the exception of
the elliptical 2.18.1, which is marked by an empty circle.
The features that become most visible in this chart, aside from the
two reported speech sections, are the groupings of main-line clauses
(we + Prefix clauses with verbs other than rrn), and the clustering of
off-line material between 2.16.2 and 2.20.2.14
Imagine that this chart is a musical score: each line (each clause)
stands for one measure, and all are the same length; the first column is
what the drum plays, the second is the tambourine, the third is the
triangle. Now, if we were to hear this 'played', we would find that the
steady, pulsing 'thump' of the drum, which predominates in the earlier
part of the 'piece', is progressively more and more often interrupted
by the tambourine and the triangle, as time goes on, till we hear very
little of it near the endeventually the drum-beat is nearly swallowed
up by the tambourine and the triangle. If we propose that the
11. Throughout this study, I employ a Chapter/Verse/Clause number system, as

introduced earlier: hence 'Judg. 2.18.1' refers to the first clause of v. 18 in Judg. 2;
there are four other clauses in this verse (2.18.2-2.18.5).
12. I have not made a thorough study of the role played by subordinated clauses
in the inter-play of 'on-line' versus 'off-line' clauses. The work I have so far done
suggests strongly that subordination may have macro-syntactic significance, but I
will make further comments on this at a later point.
13. Since I am not going to address issues of Reported Speech in this particular
text, these sections have not been given specific clause-type treatment in the charted
display.
14. The reasons for assigning 2.17.3 to the 'subordinated' column will not necessarily be clear at first glance. Further comments will follow in the text.
128
reported speech sections be scored for flute, the 'musical' bracketing

becomes even more strongly apparent.
ML
2.1.1
1.2
1.3
1.4
1.5
1.6
1.7
2.1
2.2
2.3
2.4
3.1
3.2
3.3
3.4
4.1
4.2
4.3
5.1
5.2
6.1
6.2
7.1
7.2
7.3
7.4
8.1
9.1
10.1
10.2
10.3
10.4
11.1
11.2
12.1
12.2
12.3
12.4
12.5
13.1
OL
Sub
ML
13.2
14.1
14.2
14.3
14.4
14.5
15.1
15.2
15.3
15.4
15.5
16.1
16.2
17.1
17.2
17.3
17.4
17.5
17.6
18.1
18.2
18.3
18.4
18.5
19.1
19.2
19.3
20.1
20.2
20.3
20.4
20.5
21.1
21.2
21.3
22.1
22.2
22.3
23.1
23.2
OL
Sub
129
There are four short sections of non-subordinated off-line material

(2.4.1; 2.10.1; 2.14.5-2.15.1 [2 clauses]; and 2.23.2), and one long
section (2.17.1-2.19.3 [14 clauses]). Remembering our hypothesis that
the macro-syntactic purpose of off-line material is to signify with reference to main-line material, we begin to look for their significance:
every time we hear the tambourine we must ask why it is being played
at that time.15 What are these devices signalling?
I propose that this background material functions in two ways:
(1) to signal the divisions in the text, and (2) to highlight the key
point of the history. The first two off-line clauses (2.4.1 and 2.10.1)
signal the beginning and end, respectively, of the first section of this
introduction. I would begin a new paragraph at 2.10.2; there is no
reason to connect 2.10.2 to 2.10.1, syntactically, and no objection can
be raised at any other level (save perhaps at the lexical level, as both
clauses refer to dwr, but this is very weak objectionmany examples
can be found in the commentaries citing just such cross-boundary
lexical cohesion as the height of authorial craft.16
I would suggest as well that 2.14.5 likewise signals the end of a section.17 The final clause of the chapter (2.23.2), as well, signals closure. On the other hand, I would propose that the material in the
longer section, flanking as it does the single main-line clause 'and he
saved them from the hand of their enemies all the days of the judge'
(2.18.4), serves to identify the peak event of the section. This, in fact,
is the message of the book of Judges as a whole; it is the truth that is
to be learned from all these historiesand it is syntactically marked
as such by 5 preceding off-line non-subordinated clauses (with
15. To return for a moment to the question of subordination as a possible macrosyntactic device: this particular chart shows a remarkable frequency of alignment
between [other?] off-line clauses and subordination. The fact that both slow down
the forward movement of the main line by contributing background information
suggests that they may function similarly. I will comment again on this in more detail
in the treatment of Ruth in the following chapter.
16. Boundaries are often 'transition zones' rather than strongly delineative borders:
]rm // fpun BTK ^mrr -,i;oin"T3 D3n]"î~n~m'And all
that generation were also gathered to their fathers // And there arose another generation after them, who did not know YHWH. .. (Judg. 2.10.1-2). Compare Exod. 1.8,
where a new paragraph is introduced with nearly identical wording (minus the lexical
cohesion, excepting that of 'Joseph'):]rm // fpun BTK ^mrr -,i;oin"T3 D3n]"î~n~m
17. With v. 15 being connected closely to 2.14.5, as a sort of amplification
thereofsee my further comments below on 2.17.3.
130
4 subordinated clauses thrown in), and by 3 following off-line nonsubordinated clauses (with 1 subordinated clause).
The interaction of vv. 14-15, and v. 17, with the rest of the text
requires comment. Verse 17 contains five clauses:
2.17.1
2
3
4
5
It is 2.17.3 which is most intriguinghere we have an occurrence of

a main-line Narrative clause assigned to a subordinated section.
Though this may seem alarmingly out of keeping with my proposed
assignment of we + Prefix forms to the main line, this clause is clearly
to be considered a continuation of the previous clause, which in itself
is subordinated. It is a history within a history or, to put it more
technically, we have here Narrative History material, encoded as such,
but set into a sentence to serve as an explanation; the fact that it is an
explanation has cast it in the form of subordinated material; this in
turn forces certain syntactic adaptations on the first of these two
clauses: it is forced by the preceding ky to forsake the normal [clauseinitial] we + Prefix form, and uses the Suffix form instead; the second
of the two clauses in the explanatory Narrative History continues the
history, in the normal way, with the main-line form.
Although it is beyond the scope of this study to pursue identification
of paragraph-types, it may be said that these three clauses (2.17.12.17.3) form what Longacre calls a 'narrative amplification' paragraph, about which he writes, 'Amplification paragraphs are paragraphs that consist of a Text and an Amplification, with the latter
adding new information not contained in the former, while at the
same time essentially incorporating the material found in it'.18
The second section requiring comment is 2.14.5-2. 15. 5.19 The six
clauses together (three non-subordinated clausesthe first a negated
suffix clause, the second a clause with a suffix form of rrn and a
proposed subject and temporal phrase, and the third a we + Prefix
18. Longacre, Joseph, p. 97; Here Longacre uses Text' in much the same way
as school grammars tend to use 'Topic Sentence'. His 'Discourse' provides a skeletal
introduction to this level of text-linguistic analysis (see pp. 20ff.).
19. See Appendix 1 for full text.
131
clauseand three ISDN clauses) articulate one thought: that Israel

suffered without YHWH's help.20 This unit serves to conclude the
second of the simple narration sections. The following section begins
with a shift of topic, with the emphasis being on the salvation of
Israel, rather than their humiliation.
So, then, this chapter is in five sections (I-V). It is bracketed by
reported speech sections (the first of which [I] is introduced by two
main-line clauses, and the second of which [V] is concluded by two
clauses, one of which is non-main-line). The first of the simple narrative sections (11) is set into motion by a we + Prefix clause with rrn;21
it deals with the people's faithfulness under Joshua, and is concluded
with a Suffix clause with preposed subject. The second simple
narrative section (111) is concluded by the 'amplification paragraph'
discussed above. The final simple narration section (IV) is the peak
episode of this historical overview; it opens with a topic shift. It is
bracketed with two main-line clauses at either end (2.16.1-2 and
2.20.1-2), and contains only one other main-line clause (2.18.4); this
is set off from the surrounding main-line material by 13 non-mainline clauses on either side of it. After the reported speech section
(V), the following chapter begins with a shift in topic, and possibly a
shift from the Narrative to the Expository text-type.22 Judges 2, then,
exhibits the following structure:
I
2.1.1-2.3.4
FIRST JUDGMENT SPEECHboundaries: 2

initial main-line clauses (2.1.1-2), Reported Speech
II
2.4.1-2.10.1
ISRAEL UNDER JOSHUAboundaries: 1 initial

off-line clause, and 1 final off-line clause
III
2.10.2-2.15.5 ISRAEL WITHOUT LEADERSHIPboundaries:

initial topic shift; final 'amplification paragraph'
(sequence of 7 clauses, 5 of which are off-line)
20. This is another amplification paragraph; v. 15 provides explanatory detail to

v. 14.
21. A temporal clause, which is a frequently used device for initiating new
sections (see notes on Ruth in the following chapter)other mechanisms exist for
non-macro-syntactic introduction of temporal material (e.g. nm1? m-nrrn,-nrrn,
nm1? m-nrrn, from v. 15).
22. I, personally, am convinced of this. However, until I can present the whole
of the book of Judges as an analysed text (as an introduction to the body of which
this putative expository section serves, in the same way that we + Prefix < rrn
clauses and verbless clauses may be introductory to shorter sections of narrative), I
will not make this claim.
132

IV
2.16.1-2.19.3 ISRAELUNDER JUDGESinitial topic shift; peak

event marked with 13 off-line clauses
2.20.1-2.23.2 SECOND JUDGMENT SPEECHboundaries:

initial topic shift, final off-line clause followed by
topic- and text-type (?) shift
This brief overview of the macro-syntactic features of Narrative

History is the starting point for looking at other text-types. We have
seen that the 'we + Prefix' form is the main-line clause-type for
Narrative History, and that non-main-line clause-types function as
'break' markers or peak markers. Next we will look at what I call the
'Procedural/Instructional' text-type, with a text sample from Leviticus
14. This will not be an exhaustive analysis, but instead will build on
the observations just made on the Judges 2 text, and will present to us
our first non-Narrative text-type.
3. Leviticus 14.1-32
We now look at a text of a different text-type. Lev. 14.1-32 is a text
(on more or less the same topic) of sufficient length for us to see some
internal structure; other texts offer certain features that we don't find
in this one, but are generally shorter than we would like for a starting
text. It is relatively free of textual difficulties.
This text presents instructions for the ritual cleansing of a 'leper'. It
is set, as is the overwhelming majority of legal material in the Hebrew
Bible, into a Narrative History framework.23 The unit is introduced by
a typical speech formula;24 the following unit also begins with a
speech formula, identical but for the inclusion of Aaron. Within these
boundaries, the text is further bracketed by an Expository introduction and summary: immediately following the introductory speech
formula is an Expository introductory clause (...mm rrnn nT, 14.2.1);
the final sentence in the section echoes this (...~itf mm nT, 14.32.12), recapitulating the entire chapter. In addition, the central division
(14.21.1-2) consists of two existential clauses (one verbless [with DK],
the other with p*); this, however, is not necessarily a reflection of
23. It is also, it would seem, presented as an Expository text within this Narrative
History framework. This need not delay us here; my purpose is to examine the
Procedural material; see the notes on Lev. 6.1.1(Heb)-7.37.2, following the treatment of Lev. 14.
24. noK1? noD^N mrr trm, 14.1.1.
133
some attempt to tie this material into an 'Expository framework', but

is instead likely to be a syntactic marking of the paragraph division.
ML OL Sub
14.2.1
2.2
3.1
3.2
3.3
4.1
4.2
5.1
5.2
6.1
6.2
7.1
7.2
7.3
8.1
8.2
8.3
8.4
8.5
8.6
9.1
9.2
9.3
9.4
9.5
9.6
10.1
11.1
12.1
12.2
12.3
13.1
13.2
13.3
13.4
13.5
14.1
14.2
15.1
15.2
ML
16.1
16.2
16.3
17.1
17.2
18.1
18.2
18.3
19.1
19.2
19.3
20.1
20.2
20.3
21.1
21.2
21.3a-22.le
22.2
22.3
22.4
23.1
24.1
24.2
25.1
25.2
25.3
26.1
27.1
27.2
28.1
28.2
29.1
29.2
30.la-31.ld
30.2
31.2
31.3
32.1
32.2
32.3
OL
Sub
134
This chapter contains 81 clauses;25 the first is the introductory speech

formula (a we + Prefix clause), which presents the material that
follows as part of what YHWH said to Moses on the mountain.26 Of
the remaining 80 clauses, 15 are subordinated clauses, 18 are nonsubordinated off-line clauses, and 47 are we + Suffix clauses (not
< rrntwo clauses with we + Suffix < rrn occur, at 14.9.1 and at
14.22.3). Clearly, we have here the same kind of reliance on a mainline clause-type (in this case, we + Suffix < non-rrn verbs) as we saw
in the Narrative History, Judges 2, text. In addition, we see the same
kind of distribution pattern as we recognized in Judges 2.27
The features that I find significant in this text are relatively straightforward. In the first place, there is a predominance of we + Suffix
forms (the main-line clause-type [left-hand column of the above
chart]); these 47 clauses comprise 59% of the clause totals.28 They
tend to occur in strings. Where these strings are broken by nonsubordinated off-line clauses, we can propose paragraph divisions, for
example:
14.3.3
14.6.1
25. The reader will notice an anomaly in the above chart: the sequence of clauses
runs 14.29.2 / 14.30.la-31.Id / 14.30.2. This displacement of 14.30.2 has its root
in the charting methodology that I have employed, where every clause is treated as a
unit. Where a subordinated clause is embedded inside the clause to which it relates, it
is removed from that clause, and follows it in the chart on a separate line.
26. It is not surprising that religious systems that derive a substantial part of their
authority from direct connection with a founder (as is the case in Judaism,
Christianity, Islam, Mormonism, etc.), should seek to tie their non-narrative (nonhistorical) material to a historically validating setting, by passing that material on in a
Narrative History format. The ' + Agent Orientation' feature is as important to the
authority of the text as is the ' - Projection' feature.
27. See chart of Judg. 2, p. 128.
28. The percentage of these main-line clauses may not seem like a vast majority;
however, other clause-types fall significantly below this percentage: there are 15 subordinated clauses of a variety of types, 19%; there are 10 non-subordinated prefix
clauses with verbs other than rrn (one of these is clause-initial, 14.9.2), 12.5%; there
are 3 verbless clauses, 4%; there are 2 we + Suffix clauses with n'n, 2.5%, and one
each of Prefix < n'n, existential clause with ]', and nan + Suffix form (a conditional
protasis).
135
or peak moments in the text, as in the following sequence:
14.8.5
14.8.6
14.9.1
14.9.2
14.9.3
14.9.4
14.9.5
14.9.6
The proposal that these 'off-line' clauses mark the peak events of the
episode offers a reasonable explanation for the fact that the first
shaving of hair is described with a we + Suffix clause ("^DTIK n"7:n
nuto, 14.8.1), while the second is encoded with the off-line 'direct
object + Prefix' clause (14.9.3). The suppliant is brought back into the
camp, but first only as an outsider, and finally is accepted back into
society; the text marks these events as the goal of the entire procedure
(cf. also 14.16.1-19.3; 14.25.2-31.3).
The off-line clauses used to mark paragraph division tend to occur
singly; those which mark peak sections tend to occur in collections,
and form clusters around single main-line clauses, or short strings
thereof.29
This profile is so similar in nature to that of the Judges 2, Narrative
History, text that it is difficult to understand how the existence of a
Procedural/Instructional text-type has been overlooked by contemporary Hebrew text-linguists.30
4. Leviticus 6.1 (Hebrew)-?.37
We will turn our attention very briefly to another text from Leviticus.
In this data sample I will present an example of complex embedding,
with an overview of any other significant text-level features.
This text deals, in general, with laws of sacrifice and priestly function, and is broken up into six shorter units of differing topics. It
29. This is true of such a large number and variety of texts, that it can serve
almost as a rule of thumb.
30. One could deny them both, of course, but would be hard pressed to explain
these features in a more satisfying manner. The data demand an even more streamlined description than what we propose here, if the present explanation be rejected.
136
belongs to the first section of Leviticus which deals with sacrificial

lawsit is difficult to say whether this section extends from the
beginning of the book through to the beginning of the narrative section in chs. 8-10, or from the beginning of the book to the beginning
of the 'Holiness Code' in ch. 17, with chs. 8-10 being an interpolation.31 The initial boundary of our data sample is marked by a
Narrative History speech formula, accompanied by a topic shift; it
follows a summary statement of the material covered in the preceding
paragraph. The terminal boundary of our sample is marked in a
similar fashion, and will be discussed in greater detail below.
The material is set into a Narrative History framework, where it is
recounted that Moses was commanded by God to command the people
(and here we have the Hortatory text-type, which uses command forms,
and we + Suffix clauses for the main line) to do certain things. Each of
these units is introduced by a verbless clause beginning with ni:
The Narrative History and Hortatory introduction:
6.1.1 (Heb.)
6.2.1
the topical 'paragraph headings':
6.2.2
6.7.1
6.13.1
6.18.2
7.1.1
7.11.1
the topical summary:
7.37.1
and the Narrative History and Hortatory closure:

7.38.1
31. For our purposes this mention of the larger structure of Leviticus will suffice;
none of the details presented in the material which follows hinges on concerns of this
sort.
137
These features make it clear that we should consider this section a single
larger unit, for the final sentence (7.37.1-7.38.1) summarizes each of
these verbless clauses, then recapitulates the Hortatory, and the
Narrative History, settings into which this material is set. The narrative
summary is, in fact, a conclusion to the first part of Leviticus, for it
looks back to the opening statements of the book, giving a specific time
reference and topic content, which can be traced only to 1.1, and refers
to the material contained in the intervening chapters as 'bringing offerings', that is, summary of the whole by lowest common denominator.
Like an onion, made up of concentric layers, these Procedural/
Instructional and Hortatory texts are wrapped in Expository (the ni
clauses) material, then in Hortatory, then Narrative History, layers.
The hearer's 'way into' this text is by way of story-telling, which
recounts a situation in which Moses was commanded to explain to the
Israelites, what they must do, this latter material being the bulk of the
text and taking the form of instruction and exhortation. This should
not appear exaggerated, since the 'lexical cohesion' of the bracketing
of this text substantiates fairly well these hypotheses.
5. Parallel Pericopes from Exodus
Our next texts for consideration are found in the parallel accounts of
the building of the Tabernacle. The first account, Exodus 25-31, is set
in the context of Moses receiving instructions on how the Tabernacle
is to be built;32 the second account, chs. 35-40, is given the form of a
historical account of how the building of the Tabernacle was
accomplished. We will look at only a sampling of these texts on the
building of the tabernacle; a brief analysis will be sufficient for our
purpose, which is to mount a comparison between the Narrative History
text-type and the Procedural/Instructional text-type. In addition, we will
take another look at the question of nature of the Expository text-type.
I have organized the material into 'pericopes' (sections exhibiting a
certain semantic cohesion), in the following manner:
32. For a synoptic presentation of these two sets of texts, see 'Appendix Two:
Exodus Texts Compared in Columnar Format', together with a colour-coded sample
of this material (pericopes B and C), in my PhD dissertation, which is the basis of
this book (University of Edinburgh, 1993). Owing to the limitations of printing these
are not included here. An alternative, though in my opinion inferior, methodology for
marking different clause-types has been suggested in this volume (Chapter 3).
138

Pericope
A
B
C
D
E
F
G
Narrative
History Texlt
37.1-9
37.10-16
37.17-24
36.8-34
36.35-38
38.1-8
38.9-20
Instruction
Text
25.10-22
25.23-30
25.31-40
26.1-30
26.31-37
27.1-8
27.9-19
Semantic Contents
The Ark
The Table
The Lampstand
The Structure of the Tabernacle
The Veil
The Altar of the Burnt Offering
The Court
The sequential order of the first set of texts (Exod. 25-27) will be
followed here; this will displace some pericopes of the second set
from their own literary sequence.331 will move somewhat freely back
and forth between the texts and the Historical Account Procedural/
Instructional texts, commenting from time to time on the two as a set
of 'parallel texts'.
My goal is to test the hypotheses constructed from our first examinations of data on these texts from Exodus. I will not, in this section,
go much beyond reconfirming the main-line clause-type of these texttypes; there will be occasions where the text throws up difficult
passages, but for the most part we will have to leave these sections
unattended to here (though I will comment on how the theory we are
working with here would attempt to resolve these problems, or will
point the way toward other possible solutions).34
33. I make no claims for the chronological precedence of one or the other set of
pericopes (although the first [Instructional] set does appear from certain indications to
be the earlier version), these concerns being largely irrelevant to the present inquiry.
My decision to follow the order of the first set is more or less arbitrary, conditioned
in part by the temporal relationship between the contents of the two sets, and the realtime construction of the Tabernacle.
34. The reader may well challenge me on this pointam I not just avoiding unaccomodating data? My response to this challenge is that I am proposing a detour
around concerns that require more information than this text has yet been able to
present. This study has never had as its goal a thorough-going description of Hebrew
macro-syntax, but hopes rather merely to initiate the process by illustrating some
very basic concepts; we will not be able to return during the present study to solve all
these problems. This, too, may sound escapist. However, the purposes of this study
are sufficiently broad that I will not be able to fine tune each of my identifiable texttypes, in addition to presenting the assessments of other works, and the presentation
of linguistic principles, theory and methodology. Such a finetuning would require
further substantiationespecially as regards the finer details of each text-typethan
139
Our analysis of the text begins with the question of initial and
terminal boundaries of the pericopes. This series of pericopes shifts
abruptly from one topic to the next, and tends to mark pericope
boundaries semantically rather than macro-syntactically; it does not
exhibit much macro-syntactic paragraph indication (of the sort that we
observed in the text of Judg. 2), that kind of indicator being largely
unnecessary here, owing to explicit topic-shifts.
Since the subject matter (rather than the text-type itself), requires
the inclusion of a fair number of 'measurements', we find that the
Expository text-type (which focuses on 'state' rather than 'action') is
often embedded into the Procedural/Instructional material and the
Historical Account, as a mechanism to incorporate the measurements
into the main text.35
We will look first at Pericope D,36 since it is the longest of the pericopes in our selection. The Historical text has 43 clauses; the Procedural
text has 51. Three clauses of the Procedural text are represented by
non-clausal elements in the Historical text (26.11.3 and 26.11.4 >
36.18.Ib; 26.24.4 -36.29.3b), while one clause in the Historical text
is represented by a non-clausal element in the Procedural text
(36.25.la-26.1-2 -> 26.20.lg-21.11). The remaining additional clauses
in the Procedural text are not represented at all in the Narrative text.
The Historical text is relatively unexceptional, apart from a higher
proportion of x + Suffix clauses37 than we might normally expect
(exactly one third of the non-subordinated clauses, and only one fewer
than the we + Prefix clauses without rrn). We will return to this
feature later to examine its possible significance.
There is a significant break in the pericope between vv. 13 and 14,
which is marked by this pericope's only we + Prefix clause with rrn
(36.13.3"TITO puran TTI); this clause is followed by a clear topic shift
space permits if I am to present a taste of the linguistic systems under examination.
At the end of the day, the reader will, I hope, concur with me: the material omitted
from discussion will be minimalthough, I grant, not inconsequentialin contrast
with that with which we will actually engage.
35. As was pointed out in the preceding chapter, the concept of embedding is
central to this approach to the language; it is advisable that the reader understand this
principle well before going further into the data, for it will feature highly in this
chapter and the next.
36. 36.8.1-36.34.3; 26.1.1-26.30.2.
37. Where 'x' represents any clause element(s) coming before the verb; cf.
Niccacci, Syntax, p. 13, etfreq.
140
(from the curtains of linen to the curtains of goats' hair), and can
therefore be said clearly to mark a paragraph boundary. This is
consistent with the function of such rrn clauses in other Narrative
History texts (cf. Judg. 2).
There is another interruption of the more common forms at verses
29 and 30, where three non-subordinated Prefix < rrn clauses occur
this distribution is remarkably similar to the peak-marking devices we
saw in our earlier, Narrative History, text-samples.38
36.29.1
2a
b
3a
b
SO.la
b
c
Verbless clauses are 9, in total (21% of non-subordinated clauses), and
occur in three clusters of three clauses each. Each cluster comes after
the introduction of a new topic (i.e. an item is reported as made, then
described by measurement; the details of manufacture and/or installation follow),39 and, in fact, since there are only the three topic shifts
in this pericope, the verbless clauses add directly to the macrosyntactic identification of new paragraphs. This format is repeated in a
large number of the pericopes in our material.40
38. I follow the proposed emendation in 36.29.2 from rrr to rn; the 'ketib' is
easily explained as a borrowing from the form in the parallel text (26.24.2), and that
rrr resembles closely the preceding, clause-initial, nrn. One might wonder why the
making of the two frames for the rear corners of the Tabernacle should be considered
the peak event of the episode, yet a very plausible case could be made for exactly
thisfor this is the last set of instructions concerning the building of the Tabernacle
itself, rather than its furnishings.
39. 36.8.1, topic shift from the contributions for building, to the construction of
linen curtains, followed by three verbless clauses (36.9.1-3);
36.14.1, topic shift to goat's hair curtains, followed by three verbless clauses
(36.15.1-3);
36.20.1, topic shift to the boards for the Tabernacle, followed by three verbless clauses (36.21.1-2, 36.22.1).
40. I concede that it is 'logical' to find such descriptive material where we do,
and that we are not required to posit macro-syntactic significance in order to justify
141
The Procedural text shows similarity to other texts of the same texttype. It, too, shows the same slightly higher percentage of clauses with
Prefix forms of verbs other than rrn as did the Historical text just
examined.41 This is not surprising, for where there is clause-to-clause
correspondence between the two pericopes,42 only the following 4 (or
5) clauses render their material with a change of syntax as well as
form:
26.1.1
36.8.1
26.3.1
36.10.1
26.3.2
36.10.2
and,
26.24.3-4
36.29.3
Verbless clauses appear to have the same function of marking topic
shifts as do those in the parallel, Narrative History, text.
Clauses with rrn appear to mark boundaries (26.6.3; and 26.11.4
and 26.13.1), and may mark peak events (26.24.1-4again it would
appear that the completion of the structure of the Tabernacle is
marked as the peak event).43
its presence; but equally good logic can be summoned for their distribution elsewhere
in the text, where they might have no macro-syntactic function. The fact that their
presence in the text at this particular juncture can be tied to other facets of the text
than its macro-structure does not negate the possibility of macro-syntactic significance, and merely underscores my presupposition of the interrelatedness of many
different ways of approaching text-level features.
41. Non-rrn clauses: we + Suffix clauses occur 17 times and comprise 35% of
the non-subordinated clauses; x + Prefix clauses, 13, and 26%; verbless clauses, 10,
and 20%.
42. 42 clauses97% of the Historical text, and 82% of the Procedural text.
43. A possible exception to this is 26.3.1 (noted in the preceding table), which
may mark the beginning of action in the pericope, or may not. It is an odd clause: its
142
Looking at other pericopes in this material continues to confirm our

hypotheses, and we will see in our analysis of Pericopes A through C,
below, the same general patterns we have seen elsewhere. We look first
at Exod. 37.1.1-9.3; 37.10.1-16.3; and 37.17.1-24.1. These exhibit
macro-syntactic patterns similar to those discovered in the Judges 2
text and in Exod. 36.8-34 (Pericope D).
Pericope A (37.1.1-9.3), the description of the building of the Ark,
has 22 clauses, which are distributed as follows:
10 of these are we + Prefix clauses (of which one is a clause with rrn);
9 are verbless clauses;44
the remaining three are Suffix clauses (of which one is with rrn).
The pericope ends with a sequence of 6 clauses off-line (37.7.237.9.3); the only other off-line clauses are three groups of verbless
clauses (37.1.2-4; 37.3.2-3 and 37.6.2-3). The layout of this pericope
is as follows: the opening clause (we + Prefix) introduces the topic,
the building of the Ark; this is followed by the measurements of the
Ark (3 verbless clauses) and eight clauses describing the various processes involved in the building of the Ark, interrupted by two putative
verbless clauses, which may serve to separate the construction of the
Ark itself from the construction of the various secondary features of
it; the section concludes with seven off-line clauses which have to do
with the placement of the cherubim. The initial and terminal boundaries of this pericope are clearly marked by off-line clauses, and offline clauses may also serve to articulate an internal structural division.
rrn existential statement is followed immediately by another clause, a verbatim repetition of the first, minus the n'n form. The same thing occurs in the parallel text
(36.10.1-2), where the we + Prefix form of nnn is transmuted into the Suffix form
(both finite forms replace the participles found in the Procedural text); this is less
strange, in that it does not require comment as a macro-syntactic feature at the basic
level with which we are now working, but it, too, is curious.
44. 37.3.2 and 37.3.3 contains material that could be considered appositional
phrases, and which do not stand on their own as full clauses, but which I have
chosen to consider full 'verbless clauses', largely because they are introduced by the
copula (which suggests that they be taken as on a par with other sequential clauses).
If we accept them as full clauses their presence here may mark a division between the
construction of the Ark itself and the construction of the items associated with it. If
we accept them as appositional phrases the structure of the pericope is simplified, and
verbless clauses appear only at the beginning and end of the pericope. From the
standpoint of macro-syntax, I find the latter preferable, but owing to microsyntactic
considerations, I am more comfortable with the former.
143
The distribution of the 17 clauses in Pericope B (37.10.1-16.3) is as

follows:
10 we + Prefix clauses (none with rrn);
3 are non-subordinated verbless clauses;
one is a Suffix clause with rrn;
three more are subordinated, two of these being verbless, and one
containing a Prefix form.
The pericope ends with two subordinated clauses (37.16.2-3). The ten
we + Prefix clauses are broken in only two placeswith three verbless clauses (37.10.2-4) at the initial boundary of the pericope, and
with a subordinated verbless clause followed by a Suffix < rrn clause
(37.13.3-14.1); here again, the off-line clauses appear to mark a
'paragraph division'.
In Pericope C (37.17.1-24.1), we find 16 clauses. Here, however,
we find a very different distribution of clause-types:
there are 10 verbless clauses (two of which include a participle);
there are four Suffix clauses (two of these with rrn);
and there are two we + Prefix clauses.
The chiastic structure here is unmistakable. One might be inclined to

see this pericope as artistically structured to highlight the central
clause,45 but there is little logicand I, to date, have seen no precedentfor such an intensive highlighting. It is no doubt cleverly
devised, and no doubt the central clause is being highlighted, but this
explanation alone is unsatisfactory. The placement of these clauses is
fascinating:
17.1
17.2
17.3
18.1
18.2-21.2
21.3
22.1
23.1
24.1
1
1
1
1
8
1
1
1
1
we + Prefix clause ntoa

Suffix clause < nto
Suffix clause < nvi
Verbless clause, with Ptc. < sir
Verbless clauses
Verbless clause, with Ptc. < RT
Suffix clause < n-n
we + Prefix clause < rroi?
Suffix clause < no^
A
B
C
D
E
D'
C
A'
B'
45. 37.19.2: minn-p crrcrn n'3pn rroo'? p, 'thus [it was] for the six branches
going out of the lampstand'.
144
Our theoretical base suggests the possibility that the group of 'static'
clauses (the two rrn clauses, and the ten verbless clauses,46 37.17.337.22.1)as an embedded Expository unit, with its own aperture,
body of material, peak statement, and closure, inserted into an otherwise unexceptional narrative history text. This also produces a much
more streamlined ('elegant') description.
Returning our attention to the larger text, we will find it helpful to
look at these counts from a different angle, to get a bird's-eye view of
the clause distribution. The clause distribution chart opposite shows, in
a condensed form of my previous charting technique, the clause count
of Pericopes A through C; as in the earlier chart, the main-line clauses
are in the left-hand column, non-subordinated off-line clauses are in
the center, and subordinated clauses are in the right-hand column.47
There are 85 clauses in these five pericopes, of which 37 are we +
Prefix clauses with verbs other than rrn. This may seem like a severely
weakened 'Narrative History', with so few main-line clauses, but an
alternative view of this distribution gives results more in line with
expectations of this text-type. If we regard clusters of verbless clauses
and n-n clauses as embedded 'Expository' material, the clause counts
regain some equilibrium.48 From our tally of Narrative History, nonsubordinated off-line material, we may then exclude those sections
where two or more of these 'Expository-type' clauses occur in
sequence,49 and find this distribution:
37
13
46. The latter of which Longacre proposes to be the main-line clause-type for
expository texts (Longacre, Joseph, pp. 11 Iff.; 'Perspective', pp. 88-89).
47. I have included in this chart, for the sake of comparison, the clause counts of
37.25.1-38.8.2 (two pericopes, 31 clauses; the first pericope has 12 clauses, the
second, 19). Although I will not examine them in depth, this glance at their structure
gives an idea of the kinds of forms that predominate in these contiguous texts.
48. This may seem dangerously close to 'doctoring the data', but the reader must
permit me, at least temporarily, this hypothesis. Other material will be presented, in
due course, which will help to substantiate this position.
49. A moment's consideration of these sections of the data (31 clauses) will show
that they tend to be corroborative descriptive material, and are only loosely tied to the
main-line material. The exclusion of the data from these clause-counts is intended
merely to show more clearly the similarities between these pericopes and other
'Narrative History' texts; we still recognize the text-level function of intrusive material (it serves to highlight a break in the text, or a peak), even if we seek to identify
the text-type of that material.
wc + P
(ML)
1
Pericope A
Other
(OL)
145
Sub
3
3
2
4
2
1
6
1
Pericope B
3
6
1
1
3
2
1
1
Pericope C
1
1
10
1
1
1
(Exod. 37.25.1-37.29.1)
1
4
1
2
1
3
(Exod. 38.1.1-38.8.2)
1
4
1
1
2
1
5
1
1
1
Clause Totals
37
Clause Distribution Chart
44
146
This is what we would expect to find in such a text, and the

explanation given of embedding satisfies both logic and theory.
However, we find considerable mixing of clause-types in these
pericopes; and although there is sufficient presence of we + Prefix
clauses to warrant proposing that these texts are related in text-type to
the Judges 2 'Narrative History' that we examined earlier, we need to
consider another option as well. Longacre writes that
procedural/instructional discourse looks very similar to predictive. Both,
for instance, have a mainline which consists of waw-consecutive perfects.
But while in predictive discourse imperfects can occur both in VSO
clauses to mark a secondary storyline and in NV clauses [Noun-Verb
clauses] to mark an action/event relative to a noun, in procedural discourse
the imperfect occurs only in NV clauses.50
What does he mean? Another look at the matrix of text-types will

elucidate the question:
+ Agent Orientation
- Agent Orientation
NARRATIVE
PROCEDURAL
Prediction
How-to-do-it
+ Proj.
Story
How-it-was-done
-Proj.
BEHAVIOURAL
EXPOSITORY
Exhortation
Promisory Speech
Budget Proposal
Futuristic Essay
+ Proj.
Eulogy
Scientific Paper
-Proj.
+ CTS
-CIS
If 'Predictive discourse' (which I prefer to call the 'Narrative

Prediction' text-type, to preserve its close ties with Narrative History)
is so similar to the Procedural/Instructional text-type that they share,
for example, main-line clause-types, then it is possible that our
'Historical' text is not an example of the 'Narrative History' text-type,
but rather of a related one. It stands to reason that, if Narrative
Prediction and Procedural/Instructional text-types are similar, then
50. Longacre, 'Perspective', p. 183.
147
Narrative History and 'Procedural/Lab Report' (as I shall call it) may
likewise differ in no more than subtle ways.
It is worth juggling the idea, then, that both of these Exodus texts
may be, in fact, Proceduralthe one being the Procedural/
Instructional, the other being Procedural/Lab Report, text-type.
The set of parameters from Longacre's theoretical base provide us
with a means to analyse the 'deep-structure' differences between the
two texts. The most obvious of these is Projection: one text puts the
doing of these things into the future, and, therefore, according to
Longacre's terminology, is 'plus', with reference to 'Projection',
whereas the other text places the doing of these events in the past,
and is therefore 'minus,' with reference to 'Projection' (cf. chart on
p. 138):
Parameter
Exod. 35-^40
Exod. 25-31
Projection
The texts of instructions are clearly 'minus Agent Orientation' and

therefore Procedural, or Expository; the second set of texts be 'plus
Agent Orientation', as the agent/actor is mentioned by name on more
than one occasion (Exod. 35.30-35, 36.1, 37.1, etc.), or it, too, may
be 'minus Agent Orientation', as the agent/actor references are very
few, and the sense of the text does not hang in any way on interplay
between the agent/actor and event (as it so clearly does in 'pure'
Narrative History).
Parameter
Projection
Agent Orientation
Exod. 35-40
+ (-)
Exod. 25-31
+
-
On the other hand, it can be said with confidence that both sets of texts
are 'plus' with reference to 'Contingent Temporal Succession'; that is
to say, both texts emphasize a certain 'following on' from one event
to the next: one thing leads to another, and each event tends to be
connected in some way with its (immediate) predecessor. Thus, we
have:
Parameter
Projection
Agent Orientation
Contingent Temporal Succession
or, to return to Longacre's matrix:
Exod. 35-40
Exod. 25-31
+ (-)
+
+
+
148
+ CTS
+ Agent Orientation
- Agent Orientation
NARRATIVE
PROCEDURAL
+ Proj.
Exod. 25-31
Procedural/Instructional
Exod. 35-40
Story or Procedural/Lab Report
- Proj.
So then, we have two sets of texts, the texts from Exodus 25-31 are
clearly
+ CTS / - Agent Orientation / + Projection, or 'Procedural/Instructional',
and the texts from Exodus 35-40 are either
+ CTS / + Agent Orientation / - Projection, or 'Narrative History',
or they are
+ CTS / - Agent Orientation / - Projection, or 'Procedural/Lab Report'
The jargon of the preceding paragraphs is a technical way of saying

thatin addition to being nearly identical at the micro-syntactic, lexical and semantic levelsExodus 25-27 and Exodus 35-38 are a
'minimal' pair of texts51 which belong to different (yet very similar)
linguistic text-types, and that we see here two sets of texts whose very
few differences are largely traceable to their difference of text-type.
This pair of texts allows us to examine text-type differences without
the confusions of working with unrelated texts.
Yet another possibility that we must consider is that Hebrew does
not significantly distinguish the Narrative History text-type from the
Procedural/Lab Report text-type. In a majority of languages studied to
date by linguists, these two text-types are distinguished from one
another by language-specific features52and it may be that the same is
true of Classical Hebrew. Likewise, it may be that in Classical Hebrew
these text-types form a 'porte-manteau' category (two halves folded
into one), where both receive the same encoding treatment.53
51. Or perhaps, what is termed an 'analogous pair', in which the differences
between the two compared items are not so great as to throw doubt on the analysis,
but are slightly greater than the single difference required for the term 'minimal pair'.
52. In English, for example, the Procedural/Lab Report text-type relies heavily on
passive forms, which are rare in Narrative History.
53. There is ample precedent for the coalescing of two logical categories into one
149
Our data here present no conclusive evidence with regard either to

the text-type identification of these 'Historical' texts or to the question
of whether the two text-types under discussion are differentiated in
the surface-structure of Classical Hebrew. (The differences observable
between other Narrative History texts and the texts under discussion
here may be occasioned by stylistic factors, for example; but we may
yet find that these two different text-types54 encode their information
with subtle differences in choice of clause-types, as Longacre proposes
with reference to Narrative Prediction and Procedural/Instructional
texts.)
With the current state of research, however, we must content ourselves simply to note the similarity between these 'Historical' pericopes
with the material we examined from Judges 2, and to note that if these
texts are evidence that Hebrew does distinguish these two text-types,
then we must conclude that it does so at a fairly subtle level. Further
research on this question will permit us to comment on this issue with
greater confidence. We may proceed in spite of this insecurity, however, to gain as much ground as we can at this early stage in our
description, for little in the way of further text-linguistic description
in this volume will be hindered by the relative insecurity of these particular observations.
The consistent similarity to each other of the two columns requires
little further comment. The texts in the two columns are very nearly
identical. Apart from certain lacunae (at both the word- and the
clause-levels), there are few changes in vocabulary, and only minimal
changes in syntax. In the samples given (pericopes A through G), the
Exodus 25-31 account contains 191 clauses (of which 46 are omitted
from, or replaced by non-clausal elements in, the other), and the
Exodus 35-40 account contains 161 clauses (of which 16 do not occur
in the Exod. 25-31 account).
'less logical' oneone has only to look at the verbal 'themes' of Classical Hebrew,
where the nine-cell chart has two 'holes', whose functions are adopted by other
forms. Despite these holes, the language continued to function without apparent difficulty. In English the Hortatory/Exhortation text-type often overlaps with the Procedural/Instructional text-type, due to socio-linguistic patterns of mitigationwe
exhibit a reluctance to frame exhortations to bluntly, and often modify our presentation of them to 'tone down' their bluntness.
54. That is, different at a deep-structure level.
150
Historical material
alone
16
both accounts
145
Procedural/Instructional
material alone
46
These clause totals do not require closer attention; they represent the
same kind of shifts and omissions that were commented upon above,
in the notes on Pericope D.
Before we leave the Exodus texts, we will examine one more pericope (the others included in the appendix provide no new surprises).
Pericope G55 begins as one would expect, given the data we have so
far surveyed, with a 'we + the appropriate conjugation' clause, and is
marked as well by a topic shift (to 'the court of the Tabernacle'). The
terminal boundaries of the two texts are marked by topic shifts.
The exceptional feature of this pericope is that these are the only
clauses with finite verbs; all others are verbless clauses. In the futureoriented text the only distinctive feature of the remaining clauses is
that two of them contain participles (27.16.1 and 27.17.1); the same is
true of the historically-oriented text (here the corresponding clauses
occur in the opposite order38.17.4 and 38.18.1). These texts are
clearly neither Narrative History (or Procedural/Lab Report) nor
Procedural/Instructional; following clues from their semantic content
and their macro-structure, we are led to conclude that we have here
another instance of an embedded Expository text (cf. Longacre's
matrix).
+ Agent Orientation
- Agent Orientation
BEHAVIOURAL
EXPOSITORY
Budget Proposal
Futuristic Essay
+ Prqj.
Scientific Paper
- Prqj.
-CIS
The proposed 'embedded Expository texts' from Pericope C can be

added to these data, and some conclusions can be ventured, based on
the kinds of things we have seen in other text-types.56 These texts are,
55. 27.9.1-27.19.1 and 38.9.1-38.20.1.
56. Owing to the small amount of data currently reviewed, I offer these only
as a tentative working hypothesis, although I do not believe they will require much
151
in my view, clear representatives of the Expository text-type. Although

long stretches of this text-type are not frequent in the Hebrew Bible,
the text-type is seen frequently as smaller pieces embedded in other
text-types.
The main-line clause-type of this text-type appears to be the verbless clause, and its principal off-line clause-types, x + Participle
clauses, and clauses with finite forms of rrrt (which identify, where
necessary, the temporal reference ['plus- ' or 'minus- Projection'] of
the material)our data so far have shown us only x + finite forms of
rrn, and I would suggest that clauses with we + finite forms of rrn will
be even less common, for they move further away from the stative
sense of the main-line clauses.57
In the preceding pages we have examined a variety of texts, and
alteration as more data is processed. The existence of this as a logical, 'notional'
(deep structure) category, attested often enough in other languages to qualify as a
language universal, and the fact that this text-type is encoded in relatively consistent
ways in other languages, permits us to 'borrow in' the category for the description of
Classical Hebrew, despite the limited amount of data available for analysis. We
accept this 'borrowing in,' as a working hypothesis, looking to language universals
for possible characteristics, and we rely on the data 1) to deny us this hypothesis; 2)
to force us to admit that our data are insufficient to the task of proving or disproving
the hypothesis; or 3) to confirm the existence of the category proposed. Leaving
out this text-type considerably complicates our description, where including it as a
feature of the language enables us to formulate a more elegant description.
57. On the clause-types one can expect in Expository texts, Longacre writes:
'... it is sufficient to note that as the inverse of narrative discourse (and to some
degree of predictive as well), expository discourse can be defined as discourse in
which the most static verb forms of a language predominate and have the highest
ranking. For this reason, elements at the bottom of the clines for Narrative,
Procedural, and Hortatory discourse (cf. Chapters 3 and 4) have the highest ranking
in Exposition.
'Thus the nominal (verbless) clause is the static clause par excellence. Clauses
with yes, 'there is', and 'en 'there isn't', (also the negative of nominal clauses) have
about the same ranking as nominal clauses. Clauses with copulative uses of hdyd,
'be', rank a step lower; by virtue of having any sort of verb at all they are not as
completely static as verbless clauses. Possibly, clauses with stative/denominative
perfects rank next; these verbs are essentially adjectival in function. Below all of
these rank clauses with participials [i.e., participial clauses]since these encode
activities in whatever discourse type they are found. Finite verbs rank lowest... Not
that the sort of clauses that predominate in expository discourse are the typical stuff,
for example, of Setting in narrative paragraphs. Clearly, what is off-the-line in
narrative is on-the-line in exposition' [Longacre, Joseph, pp. 111-12].
152
these have shown us at least three text-types: Narrative History,

Procedural/Instructional, and Expository. Other text-types which have
been hinted at, but which we have not yet examined are Narrative
Prediction and Hortatory; in addition, a possible Procedural/Lab
Report text-type may come into play in the 'minus Projection' texts of
the Exodus pericopes we looked at.
In the next chapter, we will continue our examination of macrosyntactic features, and will apply our theory and methodology in the
analysis of a complete text to test our initial working hypotheses, and
the hypotheses derived from this chapter's analyses. In brief, these
latter are:
1.
2.
3.
4.
5.
The main-line clause-type of the Narrative History text-type

is the we + Prefix clause-type; its off-line forms are we +
Prefix clauses with rrn, Suffix clauses, and verbless clauses,
with or without Participles.
The main-line clause-type of the Procedural/Instructional
text-type is the we + Suffix clause-type; its off-line forms are
we + Suffix clauses with rrn, Prefix clauses, and verbless
clauses, with or without Participles.
The main-line form of the Expository text-type (whether
'+ Projection' or '- Projection') is the Verbless clause; '+ or
- Projection' will be indicated in this text-type by some of its
off-line clauses (those with finite verbs will show the temporal orientation of the texts); its off-line clause-types are
verbless clauses with Participles, clauses with irn, and clauses
with other finite verbs.
A shift from the main-line clause-type to an off-line clausetype is indicative of a break in the flow of the text; this can
serve three functions (though usually not all three at the same
time): (a) to indicate the introduction of background information; (b) to signal a change of scene; and (c) to indicate the
peak event, most important fact, and so on, of the text.
Texts may be embedded in other texts; just as a Participle can
'fill the slot' or 'perform the function' of Subject in a clause
(e.g. 'Typing is fun'), so also can a Verb Phrase, or Sentence
(e.g. To be or not to be is not the only question'), that is,
items from the next lower structural level are the expected
constituents of a construction, but items from other levels
may in fact be the constituents.
6.
153
Material of one text-type may be imported into another by

means of embedding; the embedded material becomes part of
the incorporating unit in the same way as does any other
functional part of that unit. The identity (e.g. the text-type)
of the embedded material is not altered by the embedding
process.
Chapter 5
JEPHTHAH AND RUTHREFINING
AND TESTING THE HYPOTHESES
1. Introduction
In this chapter, I will present two texts in detail (Judg. 10.6-12.7, 'the
Jephthah story', and the book of Ruth), building on observations made
in the last chapter, with the intent of working our way towards a
balanced presentation of Hebrew macro-syntax, as descriptive of what
is seen in the data.
My main purpose, as I turn to the Jephthah story, is to underline a
principle which will be received with scepticism by some hebraists,
and which, therefore, will require more thorough explanation. The
principle is this: Features that are characteristic of specific text-types
will be found in material of that text-type, whether in Reported
Speech sections or not.1
To this end we will look at the non-subordinated narrative, and
compare it to five Reported Speech sections of the Jephthah story.
Here I will be confronting directly Niccacci's thoughts on this passage,
for he comes to very different conclusions from my own. We will
need to till the soil fairly deeply in this sectionmoving slowly in
order to prepare the ground upon which to work in the analysis which
follows.
We will also need to wrestle once again with the principle of
embedding; many of the examples of individual text-types that I
identify in the analyses that follow occur in combination with other
text-types, and unless the concept of embedding is well-understood,
the reader will not be able to judge accurately my results and
proposals. Here again we will move at a fairly deliberate pace.
When we have completed our examination of the Jephthah story's
1.
As is asserted by hypothesis 6 (see previous page).
5. Jephthah and Ruth
155
Narrative History sections, which will serve to illustrate and clarify

certain theoretical points, we will turn to the book of Ruth. Here I
hope to leave aside, as much as possible, apologetic and explanatory
material about theoretical issues, in order that we may examine a
single text with all our tools at hand, to see what results.
2. The Jephthah Story: Judges 10.6-12.7
2.1 The Main Narrative
In the Jephthah story, we will look first at the patterns found in nonsubordinated narrative (chiefly those created by the shifts away from
the main line), and will then focus our attention on the selected
sections of Reported Speech. This material will permit us to examine
the Narrative History text-type in Reported Speech, to see whether, in
fact, it does obey different rules, as Niccacci claims.2
I have found it most helpful thus far to arrange these clauses in the
format described for Judges 2 in ch. 4, with subordinated material set
apart from un-subordinated, and 'narration' set apart from Reported
Speech. This gives us the following distribution of the clauses3 in this
story (see next page).
A large proportion of the text is Reported Speech (54%);4 this is, in
fact, fairly standard in Hebrew Narrative History.5 It is common to
accomplish a large proportion of the story-telling by recording the
interaction of the participants.
In early parts of a Narrative History, however, the scene is being
set for what follows, and Reported Speech is not used6we will look
at this again in the 'Jephthah Story' at a later point, and will examine
the role of Reported Speech in much greater detail in the book of Ruth.
2. 28 (p. 49) and 75 (pp. 102ff.), esp. p. 107.
3. A total of 228 clauses, not including the elliptical clause proposed for
11.13.2; this 'empty' clause is proposed on the basis of the fact that the clause that
'follows it' is syntactically subordinated, and requires something to be understood:
...]rm // fpun BTK ^mrr [...] T3 D3n]"î~n~m11.13.1, [2], 3.
This empty clause does not figure in any of the clause counts presented.
4. See chart, following page.
5. Cf. Ruth, where 56% of the clauses are Reported Speech; see chart below in
this chapter (p. 176).
6. This is the most fitting explanation for the relative lack of reported speech in
Judg. 1, and especially, 2it is introductory to the remainder of the book, and as
such its job is not so much to describe events as to establish setting.
156

D
C
(subordinated (un-subordinated
rep'd speech)
rep'd speech)
42 clauses
80 clauses
19%
35%
Reported Speech
54%
B
A
(subordinated (un-subordinated
narration)
narration)
7 clauses
99 clauses
3%
43%
Narration
46%
As with most of the Judges stories, the boundaries of this pericope are
marked primarily with semantic features: most often the report of the
death of a particular judge marks the end of one pericope, as is the
case in this instance (12.7.2-3); and the inception of a new section is
marked either by the introduction straightaway of a new judge (if the
judge's reign is dealt with only cursorily [e.g. Shamgar, 3.31]), or by
the formula, 'And the people [again] did evil in the eyes of YHWH'
(e.g. Gideon: mrr Tin mn 'arifcr-n TfcHD [6.1]), as in this case (10.6.1,
with ^DK). There is no heavy macro-syntactic marking, since the
pericopes are clearly distinguishable on the basis of semantic content.7
With the borders of our passage thus secured, we will first examine
so-called 'narration' (that material which is not related as Reported
Speech). Distribution of the clause-types in this material is shown below:
(Column B)
Subordinated
Clauses
7
(Column A) (Column A)
Non-main-line
Main-line
Clauses
Clauses
18
81
The total of these clauses is 106, the percentages being 6.6%, 17% and
76.4%, respectively. The clause-type distribution for nonsubordinated, non-Reported-Speech clauses (column A in the
'columnar' text) is:
we + Prefix
we + Pref of rrn
Suffix
Suffix of rrn
Verbless Clauses
Clause with )'N
Prefix
81
6
5 (3 negated)
2
3
1
I8
7 . It may be that this identification by contents rather than by syntax is merely a

matter of the author's choice, and therefore 'stylistic'; on the other hand Classical
Hebrew may betray a preference for one or the other of these across the board, or in
certain contexts. No strong pattern has emerged from the data so far evaluated.
8 . Since these figures total 99 clauses, the percentages are self-evident.
157
This indicates once again a decided preference for we + Prefix clauses

to tell the story. The thread of we + Pref forms in column A is
broken by eighteen clauses (not counting subordination and Reported
Speech), producing 12 breaks in the main line of the storymost of
these clauses occur singly, but at three places the off-line clauses occur
in 'clumps',9 forming the major hinges in the story.
Following macro-syntactic clues (with minimal reference to other
[e.g. semantic] clues), we can make the following outline of the story,
as it is told by the clauses in Column A.
1
1.1
1.2
2
2.1
2.2
2.2.1
2.2.2
2.2.3
2.3
2.3.1
2.3.2
2.4
2.4.1
2.4.1
3
Introduction
General Introduction:
Israel's apostasy
Specific Introduction:
10.7.1-18.4
YHWH's response, and
Israel's oppression
11.1.1-12.7.3 Jephthah'sLife
11.1.1-3.4
The Introduction of
Jephthah
11.4.1-33.2
Jephthah and the
Ammonites
The Ammonites Wage
11.4.1-2
War on Israel
11.5.1-28.2
Jephthah's Defence
11.29.1-33.2 Jephthah's Offence
11.34.1-40.1 Jephthah's Daughter
the Victory Sacrifice
Returns
12.1.1-7.3
Jephthah and the
Ephraimites
The Conflict
12.1-5.1
12.5.2-6.7
The Aftermath
Concluding Formula
12.7.1-3
10.6.1-18.4
10.6.1-4
Outline of the Jephthah Story
9.
These occur at (1) 11.1.1-2 (2): nso-p im //Vn TO: rrn 'isfain nrsn
... -ufn "i"pn // nut.
(2) 11.34.2-35.1 (4): p imn VrpR //rim* 'rt pm // ...ry inn mm
... - jnp'i // nm vnuro vn // m~; and
(3) 11.39.5-40.1 (3): na'D" mro //^RiiD'3 pn-nm //D- run"*1? rn
...&* pa^'i //...btoiD' ni]3 nn^n.
158
The asterisks in the chart mark non-subordinated off-line clauses: '*/'

indicates one off-line clause at the beginning of a section, 7**' indicates two off-line clauses at the end of a section, and '*/**' indicates
one off-line clause at the beginning, and two at the end, of a section).
There is nothing very new about this division of the text, based on
its macro-syntax and its semantic content; it is a fairly standard outline
of the unit. The significant point is that the macro-syntax points to the
same structure as do other features of the text.101 will comment only
briefly on the more important features of this text.
I have mentioned above breaks in the stream of main-line clauses; I
have specifically mentioned, so far, only those places where the main
line is broken by off-line, non-subordinated clauses.111 will shortly be
looking as well at breaks in the main line created by subordination and
by Reported Speech. First, however, we return to the question of offline clauses in column A.
These clauses have a distribution similar to those we looked at in
Judges 2; that is, those which occur singly appear to indicate minor
paragraph breaks; those that occur in larger blocks appear to indicate
a major break, or a peak in the episode. Applying this observation to
the text at hand, we find that there are 9 such 'minor breaks', and (as
previously mentioned) 3 'clumps'.
These are found in the text as follows:
10.6.4
neg'd Suffix cl.
11.1.1-2
11.4.1
Subj. + rrn cl.

verbless cl.
TH + N Phr
11.5.1
'm + sub'd cl.
11.28.1
neg'd Suffix cl.
11.29.1
Tim + Subj.
+ PrepPhr.
end of intro. to the Jephthah

history
introduces Jephthah episode
introduces episode of Jephthah
and Ammon
introduces (embedded) episode
of the Elders of Gilead and
Jephthah
concludes episode of Jephthah's
messages to Ammon
introduces episode of Jephthah's
battle with Ammon
10. Scholars have long followed both guidelines in constructing outlines of their
texts; I am not claiming anything that has never before been noticedrather I am
trying to collate it into a new form of description.
11. I.e. remaining within the confines of 'Column A' in our formatted text.

11.29.4
Suffix cl.
11.34.2-35.1 mm + Ptc. cl.

Verbless cl.
159
concludes 'aperture' section of

episode of Jephthah's battle
with Ammon
marks peak event (the appearance
of Jephthah's daughter)
1'K Cl.
11.39.1
11.39.4-5
12.5.2
12.6.4
TH + InfPhr.
TPI + PrepPhr.
Subj. + Neg'd
Suffix cl.
-rim + NPhr
rrm + sub'd
speech fmula
neg'd Prefix cl.
introduces closure section of

(embedded) episode of
Jephthah's daughter
concludes closure section of
episode of Jephthah's daughter
introduces 'Shibboleth' episode
concludes 'Shibboleth' episode
The deployment of these clauses is not uniform; for example, there is

no off-line clause (or series of them) at the end of ch. 10, where we
might expect one, to indicate that the 'introduction' has ended and the
'main episode is about to begin. This should not be perceived as a
failure of the system; such features as lexical and participant strings
suggest such a division even where no off-line clauses are employed.12
Yet this is not the final word on the matter; other features (such as
subordination, Reported Speech, and longer stretches of main-line
clauses) also contribute to the 'text-map' given to the reader/hearer.
Subordination from the main-line (or rather from 'Column A')
occurs only 7 times in this story. In the previous chapter, I examined
the distribution of subordinated clauses and strings of clauses, in the
texts under study, and commented on the possibility of a text-level
function for subordinated material.13 Here again, the distribution of
subordinated clauses arouses suspicion, for they occur only at
boundary or peak sections:
12. Cf. my comments on Judges in the previous chapter, and on the Exodus
pericopes, where episodes are not distinguished so much macro-syntactically as
semantically.
13. This is not surprising at a logical levelsubordinated clauses constitute, by
their very nature, a break in the main line of the text-type. It is not inconceivable that
they serve here, in Narrative History, as do rrn clauses, for example, to alter the
rhythm of the text to indicate a 'high point' or a juncture in that text.
160

10.8.3-4
11.5.2
11.28.2
11.39.4
12.4.4
12.5.3
'... all the people of Israel, ~m were beyond the

Jordan in the land of the Amorites, ~n is in Gilead'
'And when ("Koto // "m) the Ammonites made war',
'the message of Jephthah TBfc he sent to him'
'according to the vow iefa he had made'
'for CD) they had said, "...'"
'And when (... "3 rrm) any of the fugitives of
Ephraimsaid,"...'"
The first two clauses (10.8.3-4), and the next (11.5.2), form part of
the aperture sections of the episodes in which they occur. 10.8.3-4
may seem a bit too distant from the first clauses of the section to be
considered 'aperture', but examination of the flow of this episode
indicates that the action of the section is contained in the Reported
Speech (YHWH's message to Israel10.11.1-14.4is the peak
moment of the episode); the clauses in question occur in the build-up
to that action, rather than being part of it. 11.5.2 likewise is part of
the build-up section of its episode.
The next two clauses occur at the end of their respective episodes:
11.28.2 concludes the episode of Jephthah sending and receiving messages from the King of Ammon; 11.39.4 is part of the closure of the
episodes about Jephthah's daughter.
The remaining two subordinated clauses bracket what appears to be
the peak clause of the final (and post-peak) episode of the Jephthah
story, which clause reads nns6 ]"rvn ni-owrnK i^ D'TI 'And Gilead
took the fords of the Jordan against Ephraim' (12.5.1). The peak (in
terms of the modern reader's interest is usually considered to be the
Ephraimites' dialect betraying their identity, but syntactically, the
marked peak is the geographical detail.14 It is understandable that our
attention be drawn more vividly to the material that follows, with its
fascinating revelation of contemporary Hebrew dialectal phonology,
than to this sectionbut the warfare and victory against Ephraim (in
particular, the taking of the fords of the Jordan), may well have been
the more significant detail for the contemporary reader.15
Another reading of this would be to consider this sequence of
clauses (12.4.1-5.3) a rather complicated initial boundary marker for
14. 12.5.1.
15. I grant that, had the phonological detail no interest value for the original
reader, it would not have been included. I maintain, however, that it is more on the
line of 'added colour' than 'peak event'.
161
this episode. This appears to me less likely, as this section, and other
Judges material studied to date, tends toward simplicity, rather than
complexity, for boundary marking.
A case may be made for seeing this section as focusing on the detail
of the pronunciation of n^nc; this is done more by default than by
intention, it would seem, since it is accomplished by discounting the
significance of the taking of the fords. Most commentators write at
length on the linguistic security procedures employed at the fords of
the Jordan, but little mention is made of the actual taking of the fords,
except to remark that the taking of the fords was a measure taken to
cut off the retreating Ephraimites.16 Boling's comments, however, are
instructive:
The account of Ephraim's expedition against Jephthah sustains the implicit
comparison of Jephthah and Gideon (7.24-7.3). Gideon's problem with
the Ephraimites stemmed from his being a west bank judge who had
become an east bank feudalist. Jephthah's problem with the Ephraimites
no doubt stemmed from his east bank prominence and the consequent
threat to Ephraim's prior west bank influence within the confederation.
Given the widespread devastation and power vacuum which Abimelech
created in a few years at Shechem, it is not surprising that the center of
early resistance to the Ammonite challenge shifted to Gilead, with tribal
politics taking on a whole new configuration.17
This underlines the geographical importance of this victory, which

is hinted at in the macro-syntax of the text. Webb writes:
The diplomacy of vv. 1-3 was superficially like that between Jephthah
and the Ammonites, but with significant differences. The same applies to
the report of the battle, only here the contrasts are more striking than the
similarities. This time there is no mention of divine charisma, nor does
Yahweh give the victory. Indeed, as far as we know Yahweh is not
involved in any way. This battle is not presented as a holy war, but with
wry humour as a rather squalid tit-for-tat feud. The Gileadites answer the
taunt of the Ephraimites ('you are fugitives of Ephraim', 4e) by putting
them to inglorious rout and thereby making them the true 'fugitives of
Ephraim', (v. 5)D'nsK 'D^s again. The pronunciation test of vv. 5-6
adds a further sardonic touch to the scene. The tactic of seizing the fords
of the Jordan (v. 5) previously used to such effect by Israelites against
Moabites (3.27-30) and by Israelites against Midianites (7.24-28) is now
16. Cf. Martin, The Book of Judges, p. 148.

17. Boling, Judges, pp. 213-14.
162

used by Israelites against Israelites. The slaughter is prodigious (v. 6).
The intertribal feud under Jephthah is part of a thematic development
(progressive internal disintegration) which reaches its climax in the civil
war involving the whole of Israel at the end of the book (chs. 19-21).18
I find both hypotheses convincingon the one hand, that these off-line
clauses bracket the peak clause; and on the other, that they merely
introduce the whole episodebut not equally so. I favour the former,
for the reason (in addition to that cited above) that, while both
hypotheses are justifiable on the basis of plot structure, the former has
the added merit of conforming more closely to expected macrosyntactic behaviour: we have seen that single main-line clauses flanked
by off-line clauses tend to stand out as the peak clause of a unit.19
Reported speech also appears to serve a sort of text-level function.
The developing and releasing of tension in a story, for example,
can be described as a curve like that of a camel's hump. At the
aperture of an episode, the tension has not been introduced; at the
closure, the tension has (in theory, at least) been resolvedthus, at
both ends of the curve there is little tension. At the peak, the tension is
at its highest.
The techniques for creating, maintaining, or resolving tension in a
story are language-specific. A feature which is commonly used for
increasing tension in the story would, very likely, be out of place at
the end of a story (and therefore might be placed there intentionally
by the word-smith, or avoided altogether). So, those things which
have a function in developing the story-line in a particular way are
generally distributed according to a discernible plan.
In Classical Hebrew, Reported Speech tends to function as a tensionmaintaining device. So, for example, in Judg. 10.6-15, once the stage
is set, and a certain amount of tension has been created,20 Israel and
YHWH have a conversation.21 The end result of this conversation is
18. Webb, The Book of Judges, p. 72.

19. For example, at Judg. 2.18.4; we have seen similar features in other texttypes as well, in our examinations of Leviticus and Exodus texts; see ch. 4 above.
20. The apostasy of Israel and their subjugation (10.6.1-9.2), though a formulaic
theme in the book of Judges, nevertheless introduces a significant tension to the
story-line.
21. Verses 10-15; we will return shortly to look at the Reported Speech material
of this section in greater depth.
163
recorded in the seven main-line clauses that follow: Israel puts away
its foreign gods, and YHWH's heart is turned again to them. But they
are still under oppressionand the Reported Speech device is
employed again to maintain the tension during the transition to the
next episode: 'Who is the man who will fight for us? He shall be our
chief. This maintaining of tension is reiterated by the use of a
question introducing a new topicthe quest for a leader.
Reported Speech serves this purpose of maintaining tension
throughout the text.22 It is used extensively to detail the several
obstacles Jephthah must overcome in his life: the rejection by his
people is highlighted in 11.1-10; the intractability of the King of
Ammon in 11.12-27; the stark exigency of his vow to YHWH, and its
outcome in 11.30-31 and 11.35-38; the aggression of the Ephraimites
in 12.1-3 and 12.4-6. In most of these cases little is added which has
not already been recounted on the main line of the story, or which
will not be retold on the main line at a later point.
Conversation exhibits a tendency to bracket peak events.231 cannot
yet afford to make a solid proposal about this, as the data offer too
little evidence for such conclusions. We will, however, encounter the
same concerns as we look at more data. The reader may find this
hypothesis less suspect after our examination of Ruth.
Before moving away from the text we have been examining, I
would like to make another tentative observation: it would appear that
'aperture' and 'closure' sections of text have an affinity for unbroken
chains of main-line clauses. The opening 'Israel and YHWH' section,
10.7.1-18.4 (1.2 in the outline above), opens and closes with stretches
of these main-line clauses, as do 11.1.1-3.4 (2.1.1 'Jephthah's Early
Life'), 11.29.1-33.2 (2.1.4, 'Jephthah's Offense') and 12.1.1-6.7 (2.3,
22. We will look at this feature only in passing here, but will return to it again in
our analysis of the book of Ruth. It is worth noting in passing, at this point, that the
book of Jonah ends with Reported Speech, which seems to vitiate my hypothesis that
it has a text-level function. Yet, if we hold to our idea that it maintains tension over
transitional sections of text, we can propose that the unusual occurrence of Reported
Speech at the end of Jonah actually serves to strengthen the purpose of the bookit
leaves the reader hangingthe final resolution is, in a sense, the reader's/hearer's
own responsibility. The point of the tale is hammered home all the more firmly by
this lack of resolution of carried-over tension.
23. Cf. the treatment, above, of the subordinated clauses 12.4.4 and 12.5.3, both
of which, in fact, introduce Reported Speech sections.
164
'Jephthah and the Ephraimites'). The story as a complete unit in itself

likewise opens, and is concluded, with sequences of main-line clauses
(10.6.1-3 and 12.7.1-3). In addition, other episodes conclude with
stretches of main-line clauses although their apertures are handled
differently: the subsections (embedded narratives) of 'Jephthah and the
Elders' (11.11.1-3), 'Jephthah's Daughter Goes to the Hills' (11.38.35), and 'Jephthah's Daughter' (11.39.2-3).
One section of text does not fit this pattern: 12.4.1-3. The sequence
of clauses here is followed immediately by the two subordinated
clauses (which may mark the peak of this episode), and therefore is
not likely to be episode-marginal. This entire section raises difficult
questions, and I hesitate to draw any conclusions from it.
2.2 Narrative History in Reported Speech
In this section I will present five samples of Reported Speech. For the
sake of simplicity I have restricted the data we will examine to
Narrative History texts in Reported Speech. The conclusions I reach in
this brief section will be applied to other text-types in Reported
Speech when we turn to the analysis of Ruth.
Judges 12.2.1-3.6
This text is actually the last, sequentially within Judges, of the five
we will examine; it is, however, also the least difficult. The text
is presented, below, like most of the other sample texts, in a format
that permits comparison between narration and Reported Speech, on
the one hand, and between subordinated material and nonsubordinated material, on the other. Columns A and B, on the right
hand side of the page, contain narration, that is, material that is not
Reported Speech; the left half of the page contains columns C and
D: Reported Speech material. In each of these two halves, the righthand column contains non-subordinated material, and the left,
subordinated.24
24. Owing to formatting restrictions of this volume, Column B is not fully

independent of Column A, nor is Column D of Column C. This should, however,
occasion no difficulties.

D
165
B
A
12.2.1
2a25
b
3
4
3.1
2
3
4
5
6a
b
The speech formulae, in this as in all other Reported Speech units

examined in this book, are included as part of those units. In an
instance such as we have above, the text reported to have been spoken
is an embedded direct object of the verb of speech; the direct object
'slot' may be filled by anything from a single word or particle (e.g.
12.5.7-8 ? -on) to a fully developed, complex text (e.g. 11.15.127.3 43 clauseswe will examine this in depth after dealing with the
shorter texts). Therefore the above quotation from Judges is but a
single, complex sentence. (The same could be said of vast sections of
the Pentateuch, where whole series of chapters are introduced by a
single speech formula [thus making the series of chapters merely an
embedded Direct Object of the speech formula]; however, beyond
acknowledgement that this is, indeed, the case, there is little to be
gained from considering such sections as single clauses.) This text
contains, apart from its speech formula, 9 clauses:
1 Subordinate clause with pt
3 Suffix clauses
12.3.2
12.2.2 (with nnn), 12.2.4 (negated);
12.3.6 (in an interrogative clause)
5 we + Prefix clauses
In examining Reported Speech sections, we must be on the lookout for

clues as to why the text was spoken; this gives us our principal clues
25. Where a single clause spans more than one line in these charted texts, the
device 'a', 'b', and so on will be employed to so indicate; this will be true as well of
clauses which span more than one verse (cf. Judg. 11.39.6-11.40.1, one clause
which spans four lines).
166
for analysing the 'constituent structure' of the text. In this case, we

may say that Jephthah is repudiating the right of the Ephraimites to
challenge his behaviourthe final clause reads: //^RiiD'3 pn-nm //D-
na'D" mro //^'Why then have you come up against me this day, to fight
against me?' The earlier material is his justification for his choice not
to request Ephraim's help; he accomplishes this by explaining the
historical background of that decision. This historical background is
presented in an embedded Narrative History text.
Internally, the Narrative History text is composed of two episodes:
the first is a description of the circumstances in which Jephthah found
himself (12.2.2-4), the second, his response to those circumstances
(12.3.1-5). The boundary between the two episodes is marked by the
negated clause (12.2.4), and the repetition of that material in the
sentence // jrtfra -promr-wi (12.3.1-2).
The most significant point here is that there is nothing out of the
ordinary about this Narrative History text. It opens, as do many
Narrative History texts, with the stage-setting device of a rrn clause;26
its boundaries are marked by the same sort of features as we have
noted in other Narrative History texts. This begins to call into doubt
the conclusions voiced by Niccacci on Judges 11: that somehow 'narrative discourse' has a different shape to it from narrative proper.27
Judges 10.10.1-4
We move now from the last sequentially, of our sample texts, to the
first. This one is brief, but still permits analysis:
10.10.la
b
2
3
4
26. We will see the same device employed in the book of Ruth on more than one
occasion. Logic as well gives us a rationale for such a function and device: do we not
say that a story has to 'start somewhere?' We expect a story to start with some kind
of anchor into space and timethis kind of anchor is provided by such things as
Verbless, and rrn, clausesand if such a setting slot were filled by an embedded text
of more than one clause in length, we would expect it to be an Expository one, which
has as its main line just such clauses.
27. Niccacci, Syntax, pp. 102ff.
167
The formulaic opening is one of a sequence of main-line, we + Prefix

forms, and requires no particular comment. The content of the
Reported Speech is a text of three clauses.
Let us, for the record, describe this unit in terms of its constituents;
we have:
an Intransitive Clause:28
DirObj
with the following construction:

PredComplement < IndObj < Subj < Predicate,
where
Pred = Finite Intransitive Verb
Subj = NounPhr
IndObj = PrepPhr
PredComplement = InfPhr,
and
InfPhr = Infinitive + DirObj,
and, finally,
DirObj = Embedded Narrative History Text.
10.10.3-4 is an embedded Narrative History text, filling the Direct

Object slot in the speech formula. The second clause in this embedded
text begins with a we + Prefix form, but the first begins with -3, which
must be clause-initial; so the verb form in this first clause (and therefore the clause-type) must accommodate itthis is the only reason we
need to seek for the Suffix clause replacing a we + Prefix clause.29
Here we have an embedded three-clause historical unit, in which the
first clause gives the general material, and the two remaining give the
28. It may seem odd, at first blush, that I consider a verb of speech 'Intransitive',
for I am arguing that the speech material is an embedded Direct Object of the speech
verb; yet in this case the speech material is the Direct Object of the infinitive ~\Q$b,
not ofifwi; the latter is the defining feature of the main clause, and its syntax defines
it here as an Intransitive verb.
29. We are, of course, speaking of non-entities; there is no 'we + Prefix clause',
and therefore we can only posit that it 'would otherwise have been there'. It is clear,
however, that if it 'wanted' to be there, it nevertheless could not be there, owing to
the restrictions placed on the clause by the subordinating conjunction.
168
specific reason for the statement of the first. The latter two give, first,
the specific reason with regard to YHWH, and second, that reason with
regard to other gods.
Judges 10.11.1-14.4
YHWH's response to the speech we have examined above comprises 12
clauses (including the speech formula). The speech formula is a
transitive clause (in this case, without the Predicate Complement =
Imperative), where the Direct Object slot is filled by an embedded
text. The text itself is rather complex:
11.1
2a
b
12.1
2
3
13.1
2
3
14.1
2
3
4
YHWH's response begins with a question (an elliptical one, at that).

Questions are features only of Reported Speech in Classical Hebrew
narrative, and seem to play a part in the establishment of participant
roles,and the like. There seems to be an element of protest, scorn,
condescension, and/or anger, to rhetorical questions in our data; their
purpose appears to be to establish the order of precedence and deference between speaker and addressee, as well as to highlight to the
address information that the addressee should take (or should have
taken) into account.30
30. Each instance of rhetorical questions in the Hebrew Bible will need to be
analysed in terms of its sociolinguistic setting (inferior superior, equal > equal,
superior -> inferior, to name but a few of the possible settings we might have to deal
with in this sort of description), the degree of intensity of communication, the texttype of the material (if any) which follows the rhetorical question, the position of the
rhetorical question in the Reported Speech, and many other issues. The nature of the
topic precludes us from examining it even briefly during this study, though we are
169
The purpose of our text is essentially that of exhortation:

31
The reason for this exhortation is that YHWH will no longer deliver
them (DDnK pmi? ppttrtf? p1?, Therefore I will deliver you no more',
10.13.3); and in turn the reason for this state of affairs is that YHWH
had delivered them but they have turned away from him to serve
other gods. This latter section is expounded by a Narrative History
text of five clauses' length.
This Narrative History text is introduced by a Suffix clause
(providing, along with the rhetorical question preceding it, a setting
for the text that follows), and a second episode of it is signalled by
another of the same. This text presents us with no surprises. We can,
therefore, divide it into its constituent parts:
Judges 10.11.1-14.4:
Speech Formula
(a main-line NH cl.)
NH text Aperture?
11.1
2a
b
12.1
2
3
13.1
2
NH text
Narr. Prediction Text

Hortatory Text
14.1
2
3
4
We have not so far looked at Hortatory text typesfor the time being,
however, they will occasion little comment, for two reasons: (1) their
impoverished for our lack of understanding of its functions and purposes.
3 1 . 'Go // Cry to the gods // whom you have chosen // Let them deliver you in the
time of your distress' .
32. This section of text poses some textual difficulties, but they need not delay us
here; I have already stated my intention to bypass rhetorical questions in the current
study.
170
simplest features are relatively obviousthey are composed of

cohortatives, imperatives and jussives, on the main line, with Prefix
forms running close to the main line; their off-line clauses would
require greater attention, but since we are not looking yet at any
longer texts, this can be postponed; and (2) this text-type is
considerably influenced and shaped by the sort of sociolinguistic
factors that I mentioned above in reference to rhetorical questions
deference to one of higher rank, for example, precludes the use of
command-forms, and such mitigation can have a variety of nuances.33
This will be yet another area that we must leave relatively unexplored,
although where the data require us to give this issue our attention, we
will do so. The lack of in-depth attention to this text-type will not
seriously hinder us, however; I will rely on Longacre's 'Verb Rank
Cline' for this text-type34 to provide us with working hypotheses, and
will deal only cursorily with questions of mitigation.
Judges 11.7.1-5
This short text also contains an embedded Narrative History text.
11.7.1
2
3
4
Its introductory formula is unexceptional and requires no extended

comment. The opening clause is again a rhetorical question about
historical fact; it is followed by a we + Prefix clause. Another rhetorical question with a subordinated clause concludes the Reported Speech.
Despite my disclaimers above about rhetorical questions, I will toy
with them a little here. If we were to render the first rhetorical
question as a statement, the result would be 'you hated me'or
33. Longacre writes, 'The presence of mitigation (making a command more
mollified and socially acceptable) and deference (e.g., of commoner to monarch)
make the analysis of hortatory discourse the more difficult' (Joseph, pp. xif.); he
devotes 21 densely reasoned pages to an examination of this concern (ch. 5,
pp. 119-40). It is well beyond the limits of this study to 'unpack' and render that
material here. I will content myself with the occasional elucidation from other data.
34. I will do the same for the Narrative Prediction text-type (which Longacre calls
Predictive).
171
perhaps, more emphatically (acknowledging the nuance of the

negative structure of the rhetorical question), 'you hated me greatly'
(perhaps with an infinitive absolute construction in Hebrew?)in any
case, a simple statement of a past event.
The second rhetorical question, however, would require a more
complex structure: the resulting unit would be either a Narrative
Reason Paragraph (to use Longacre's terminology: 'You come to me
now, because...'), or a Narrative Result Paragraph ('You come to me
now, in order to/that...').35
Thus, I consider 11.7.2 to be a Narrative History clause, turned
inside out to express annoyance, or superiority, or some other such
nuance. It serves the purpose of introducing the historical setting (and
first event) of a brief Narrative History text.
Judges 11.15.1-27.3
This text-sample contains 43 clauses (with a short and unexceptional
introductory speech formula); Niccacci has examined this in Syntax,
and makes a number of observationsI will take exception to some of
these, and concur with others.
First, however, we will examine the data for ourselves; later, we
will return to compare our findings with Niccacci's.
11.15.1
2
3a
b
16.1
2
3
17.1a
b
2
3
4
5
6
35. This sort of 'back-formation' gives us a clearer idea of the underlying

structure of the unit; however, due to the tentative nature of such analyses at the
present time, these results will not be used as the basis for further speculation.
172

18.1
2a
b
3
4
5
6
19.1a
b
2
3
20. la
b
2
3
4
21.la
b
c
2
3a
b
22. la
b
c
23.1a
b
2
24.1
2a
b
3
4a
b
25.1a
b
2
3
26.1a
b
173
c
2
3
27.1
2
3a
b
This unit constitutes Jephthah's final message to the King of Ammon,

who had ordered Jephthah's people to restore his land to him, accusing
them (unjustly, it would seem) of having taken it, on their coming up
from Egypt.36 Jephthah responds in the negative. After the Narrative
History speech formula, Jephthah's speech employs another speech
formula (a formal one11.15.2), and is followed by a denial of the
accusation (11.15.3) and a justification for that denial (a Narrative
History text, embedded within a subordinated clause11.16.1-22.1);
this denial and justification is followed by a summary of the historical
justification, and a series of questions and statements on the subject of
'take what your god has given you'. It concludes with a jussive: //rim*
'i /nm vnuro vn // m~na'D" mro //'Let YHWH, the Judge, judge this
day between the Israelites and the Ammonites' (11.27.3).37
There are two negated indicative clauses in this unit, both of which
deny the King of Ammon's claims of injustice:
jn$p'i // nm vnuro vn // m~na'D" mro //'Israel did not take the land
of Moab and the land of Ammonites' (11.15.3)
,i;oin"T3& D3n'I have not sinned against you' (11.27.1)
These clauses form the aperture and closure of the speech unit; the
second of which initiating a closure paragraph, composed of 3 clauses
(neg'd Suffix cl., Verbless clause with Ptc. and Jussive clause).
The boundary between the first section (11.15.2-11.22.1) and the
second section (11.23.1-11.27.3) is clearly marked: not only is there
a shift from we + Prefix forms to Prefix forms, there is a shift to a
36. 11.13.1-4.
37. This section combines all the difficulties of rhetorical questioning with those
of mitigated Exhortation; in short, the body of the second section is awkward. We
can analyse the bracketing of this section with the opening of the speech, but any
actual identification of the text-type(s) of the second section would be, at best,
provisional.
174
predominant use of the verb root czrr, which occurs in this next section
six times (in six contiguous clauses).38 Although other features of the
second section of this text could be noted here, my real interest in this
passagefor this studyis the first, Narrative History, section of the
Reported Speech material (11.16.1-22.1).39
The text commences with a quote from the King's message to
Jephthah: onsan imîo, 'In his coming up from Egypt' (11.13.3),
becomes onsoo nrrbsn, 'In their coming up from Egypt' (11.16.1). It
is followed by 17 we + Prefix clauses, which are from time to time
interrupted by eight off-line subordinated, or Reported Speech
clauses. These eight clauses occur in three groups: 11.17.2-17.5,
11.18.5-6 and 11.19.3-4.
The middle group of these three (11.18.5-6a neg'd Suffix clause,
and a subordinated Verbless clause) marks a text division; this division
is emphasized by the use of similar events at the initial boundaries of
both sections. The first section commences with Israel's arrival at the
borders of Edom, and their sending a request to Edom's king, which
is refused (11.16.3-17.3); the second takes place at the borders of the
Amorites, and commences with their sending a request to their king,
which is likewise refused (11.19.1-20.1). These features serve to
create two distinct episodes: 11.16.1-18.6, with 15 clauses, and
11.19.1-22.1, with 11 clauses.
The other two groups of off-line clauses each contain a Reported
Speech 'request for right of passage' to a king, followed by negated
Suffix clause(s). I identify these as the syntactically marked peaks of
the two Narrative History episodes.
2.3 Conclusions from Analysis of the Jephthah Story
At no point in my examination of the material in the Jephthah Story
have I seen evidence that would lead me to conclude that Narrative
History in Reported Speech behaves differently from that in nonReported Speech, as Niccacci concludes. What little evidence Niccacci
38. This latter feature, combined with a shift from, primarily, Prefix forms, to
Verbless clauses and Suffix forms, leads me to suspect a subdivision at 11.25.1.
39. Admittedly, I am inclined to see 11.23.1...rm // fpun B577TK ^mrr 'And now, YHWH, God of Israel, has given...'as a hinge that functions both to
conclude the first section and to introduce the second, but we will deal with it here as
extraneous to the embedded Narrative History text, since it is clearly not to be
considered subordinated under the o in 11.16.2, as is the Narrative History text.
175
cites in favour of his proposal is more elegantly explained by two,

somewhat overlapping, principles.
First, Narrative History texts I have so far examined (and here I am
referring not only to those texts in Classical Hebrew) tend to begin
with a stage-setting section, in which clause-types that are off-line for
Narrative History predominate. This is a normal, and logically
justifiable, state of affairsa story requires background and setting.
The only consistent exception to this trend is where the text in
question is too short to employ such a device, or where previously
introduced material obviates the necessity for such a device.
Second, even where a text might conceivably employ a clauseinitial, main-line for Narrative History, we + Prefix clause, this does
not happen. The language prefers to underscore the relationship
between the embedded Reported Speech material, and the speech
formula into which it is embedded as Direct Object.
This may sound very like what Niccacci has proposed; it is, in fact,
radically different. Niccacci proposes that 'narrative discourse' is a
different 'type of narrative';40 I propose that Narrative History texts
are formulated according to the same principles whether in Reported
Speech or non-Reported Speechthere is no 'different type of
narrative' for Reported Speech. Material in Reported Speech is,
however, subject to restrictions with regard to its initial clausein
order to call attention to the fact of its being embeddedas if to say
'this is not actually the beginning of the clause; rather, it is the Direct
Object of another clause'.
3. The Book of Ruth
3.1 Basic Facts and Statistics
The boundaries of this text are clearly marked. Aside from the
obvious onesthe beginning and end of a self-contained bookthere
are clear macro-syntactic markers. The opening boundary is marked
by two Tn clauses;41 the terminal one is marked by two verbless
clauses and a genealogical appendix (an embedded text of a different
type).42
41. treason BSO 'D'a "rr\ 'In the days when the judges were judging' (1.1.1),
and pa am TH, 'there was a famine in the land' (1.1.2).
42. 4.17.4-22.2.
176
The book of Ruth contains 399 clauses, by my accounting; ten of

these form the genealogy at the end of the book. For the percentages
given in our discussion, however, I omit these genealogical clauses,
giving a working total of 389 clauses.43 The distribution of these
clauses in the four columns is as follows:
D
(Subordinated
(Un-subordinated
Reported Speech) Reported Speech)
74 clauses
(19%)
146 clauses
(38%)
(Subordinated
Narration)
(Un-subordinated
Narration)
13 clauses
(3%)
156 clauses
(40%)**
Reported Speech
Narration
220 clauses
(57%)
169 clauses
(43%)
We will first examine the non-Reported Speech material. After I have

commented on the structure and constituents of that portion of the
text, I will analyse the Reported Speech sections.
3.2 RuthThe Non-Reported Speech Text
I will be examining here the interplay between main-line clauses (by
which I mean 'main-line for the Narrative History text-type', for
which I propose the we + Prefix clause), off-line clauses, subordinated
clauses, and Reported Speech sections.45 As we have come to expect,
these interruptions of the main line are not spread haphazardly
through the text, but function as indicators of episode divisions and as
peak markers.
Where concentrations of off-line clauses occur, the significance of
the break, or of the peak event, is the greater. There are six places in
43. For some statistical purposes, the genealogical clauses would weight the proportions too much in one direction or another; for example, the ratio of main-line
clauses to off-line clauses in the main Narrative History would look significantly different if these clauses were allowed to influence itfor example, if we figure the
percentage of Suffix forms in the main Narrative History (non-subordinated, nonReported Speech clauses), the difference is noticeable (4%, calculated with the lesser
number; 11%, when calculated with the greater number).
44. Percentages are given in round figures.
45. These last will be dealt with as individual units on pp. 187ff.; here they will
be examined only as deviations from the main line of the narration.
111
the text where concentrations of more than one off-line clause occur
(1.1.1-2; 1.2.1-3; 1.4.2-3; 1.22.2-3.1.2, 4 clauses; 4.8.1-4; and
4.17.4-18.1, 2 clauses). If we allow for the possibility that
subordination may have some macro-syntactic significance, 3 more
clause-groups present themselves (2.3.4-2.4.1, 2 clauses; 2.17.3-4;
and 4.1.3-4); in addition, several of the features we are looking at can
be found separated by only one or two main-line clauses (e.g. 3.6.1
and 3.6.4).46
3.2.1 Ruth 1.1.1-5.2
I have already spoken briefly about the initial and terminal boundaries
of this text; we now return to look more closely at the initial boundary.
As was said above, there can be no question of this not being the
initial boundary; however, there is much yet to be said on the nature
of this boundary, and its extent.
The text opens with two Tn clauses (1.1.1-1.1.2). Two sm clauses in
succession is in itself very unusual. One search of the Hebrew Bible
using a CD-ROM data-base indicated 918 occurrences of vn.47 I
found, among these 918, only 17 pairs of -m clauses in succession
(e.g. Job 1.3 and Exod. 12.41), and two triplets (Gen. 39.2 and
Josh. 17.1-2). This gives a percentage of less than 4%.48 This is a
very rare feature.
Some of these occurrences are clearly paragraph-initial if not
episode-initial (e.g. Ruth 1.1, Gen. 39.2 and Job 1.3); others may
either be episode-initial, or they may require to be divided, thus
marking one boundary each (initial/terminal). Gen. 27.30 is an
example of a 'maybe'this pair may initiate the section wherein Esau
seeks a blessing from Isaac (which Jacob has just 'stolen'), or the first
of the two Tn clauses may serve to conclude the previous section,
while the second clause opens the following section.49
46. Rather than list these occurrences here in some kind of indecipherable shorthand form, I will identify them as we come upon them in the text.
47. I make no claim to absolute accuracy with these figures, but I estimate the
margin of error to be no greater than 1-2%.
48. There are six TH pairs of Gen. 1the formulaic.. .nrrm // fpun BTK
which I have excluded from my examination, rightly or wrongly; if these are added
into the count, the percentage is raised to just over 4%.
49. This function appeared fairly frequently among the Tn pairs which were
separated by one or two clauses not containing 'm (e.g. Gen. 39.10-11), and may be
normative, or at least common, for contiguous pairs as well.
178
In any case, the Ruth passage is unquestionably text-initial, and

cannot be a bridging mechanism, as may be suspected of other occurrences of multiple TH clauses.50
And yet, not only do we have two Tn clauses here, but they are also
separated from six other off-line clauses by only four main-line
clauses. 1.6.1 introduces the journey back to Bethlehem, and begins a
new section; we can extend this opening section to the end of v. 5;51
this gives us an opening section of 16 clauses. The ratio of main-line
clauses to off-line clauses in 'Column A' of this opening section of
Ruth is 1 : 1, whereas the ratio for the remainder of the text is
roughly 15 : 2. This opening section clearly does not reflect the
standard pattern of clause distribution for Narrative History.
Yet, on a larger scale, this is fairly normal, as has been pointed out
above; story-telling usually starts from nothing, and the most basic
information must be given at the outset, in order that the
hearer/reader can make sense of the text. All but the very shortest of
our sample texts have had a sort of 'setting the stage' section at their
beginning.52
3.2.2 Ruth 1.6.1-19.1
The next sectionthe return to Bethlehembegins with four mainline clauses, interspersed by three subordinated clauses (1.6.1-7.4),
which are followed by a sequence of 58 clauses, interrupted in three
places by 31 clauses of Reported Speech, and again by an off-line
clause (1.14.5), and by one subordinated clause (1.18.2).
A more detailed look shows the introductory section (1.6.1-7.4) to
be followed by the first Reported Speech (Naomi advising her
daughters-in-law to return to their own culture and families, 1.8.19.2). Three more main-line clauses follow this (the reaction of the
daughters-in-law, 1.9.3-5), and two speech sections (their intention to
50. I did find, however, one other feature of these clauses which is of interest: in
several of the clause pairs I have examined, the first provides general background
information, which the second narrows into a specific context (often with a temporal
function; cf. 1 Kgs 17.12)these two clauses are a textbook case of this pattern.
How frequently this is the case, however, I cannot yet say. This feature, however,
strengthens the case for reading these as inseparable pairs.
51. 1.5.2the conclusion of the 'death' motif, the first motif of the story.
52. Judg. 2.4.1:...mrr JR'PQ ~mD TTI; Lev. 14.2.1:...mm rrnn nr; etc.; the
Jephthah story, on the other hand, does not have a syntactically marked 'stage', but
nevertheless begins with onesee commentary above on Judg. 10.6.1-5.
179
accompany Naomi, 1.10.1-3, and Naomi's response, 1.11.1-13.6).

Another set of four53 main-line clauses (Orpah's decision to leave
Naomi, 1.14.1-4 [or 1.14.1-3]) comes to an end with a Suffix clause:
%nb bpqr nrm'and [but] Ruth clung to her' (1.14.5). This off-line
clause marks the peak, in my opinion, of this first episode; either the
clause itself is the peak, or it signals that the following interchange
between Naomi and Ruth (two speech sections, 1.15.1-3, Naomi
reiterating her advice that Ruth leave her, and Ruth's refusal, 1.16.117.6) is the peakthis latter is the more likely, I believe.54 The
remainder of the episode contains three main-line clauses, broken by
one subordinated clause, which describe the journey itself (1.18.119.1, 1.18.2 being the subordinated clause).
The string of main-line clauses, in addition to being interrupted by
Reported Speech and a subordinated clause, is broken by a second offline clause. The first off-line clause, the peak-marking 1.14.5,
occurred within the text, the second occurs as the initial boundary of
the following section (1.19.2).55
Thus the first episode is bounded, initially and terminally, by strings
of main-line clauses interrupted by subordination, initially by a clear
topic shift, and terminally by the off-line clause marking the initial
boundary of the following section. The peak is marked by a Suffix
clause.
This is the first instance of Reported Speech in this text, and it will
be helpful to explore here its role in Narrative History texts.
Without exception, Reported Speech material can be 'translated'
into main-line Narrative History we + Prefix clauses, but something is
lost in so doing. For example, Naomi's first exchange with her
daughters-in-law could be re-cast as, 'Ruth told her daughters-in-law
to return to their families, and blessed them; and she kissed them and
they raised their voices and wept, and refused to go'. Yet to reduce
this and the remaining interchanges between these characters to such a
53. Or three, if one does not accept the reintroduction from the LXX of 1.14.4
bgve569090whose omission by haplography is easily explained on the basis of
14.3nmonb ns~iu pcmand whose sense and syntax fit easily into the text.
Neither their inclusion in, nor their deletion from, the text alters its macro-syntactic
structure.
54. I have not come across any obvious Narrative History peaks encoded as offline clauses; off-line clauses generally serve to point to something else.
55. See below.
180
Narrative History text would shear them of much of their power and
nuance. We would fail to capture a sense of Naomi's stoic attitude and
bitterness, and of Ruth's deep love and unyielding commitment. Such
nuances could not be conveyed in simple Narrative without a laboured
proliferation of clauses and relative materialother text-types,
however, convey these things better, and can be introduced by way of
Reported Speech.
Such a proliferation of clauses in the narration would have the
additional effect of 'distancing' these events from the reader/hearer.
Reported Speech has the opposite effectthe reader/hearer is
deposited, as it were, into the very place and time of the interaction,
heightening the 'presence' of the events.
This movement into and out of the interchanges is part of the
'breathing' of the textpart of the increasing and decreasing plottension. These moments of immersion in the actual event are vital to
the development of characters and plot.
3.2.3 Ruth 1.19.2-22.3
This episode, dealing with the arrival of the two women at Bethlehem,
begins with the aforementioned off-line clause: nnb rra rritns TPI, 'As
they entered Bethlehem', 1.19.2. This is a short episode; after the
initial clause, there are one main-line clause (1.19.3), two speech
sections (1.19.4-5, and 1.20-21.5Naomi's interchange with the
women of Bethlehem56), another, single main-line clause, and the
terminal boundary. The terminal boundary is marked by two off-line
clauses (a verbless clause, 1.22.2, and a Suffix clause, 1.22.3). The
final three clauses summarize the return to Bethlehem, and the last, in
addition, points forward toward the next section by indicating that
they returned 'at the beginning of the barley harvest'. The summary
material indicates that this break concludes not only the most recent
sub-section but also the general section which began with 1.6.1, if not
the whole of this first chapter.57
56. The function of this section of Reported Speech appears to be to carry the
tension of the 'widowhood' (with all its attendant insecurities in that culture) section
into the following section, where Ruth meets Boazthis is no joyous homecoming,
but a retreat toward at least marginal security.
57. The mention of 'Ruth, the Moabite, her daughter-in-law' points to an
intentional inclusion of the material in 1.1.1-5.2.
181
3.2.4 Ruth 2.1.1-17.3

This next section of the text likewise has a syntactically marked
boundary of two clauses, which, when taken alongside the previous
two clauses, form one of the most significant 'hinges' in the story. No
other boundary in Ruth (apart from the opening of the book as a
whole, and the terminal genealogy) is marked by so many off-line
clauses.
These boundary clauses introduce the other main participant of the
story, Boaz. The episode is composed of 15 main-line clauses, with 12
Reported Speech sections (in five groups), two subordinated clauses,
and one off-line clause.
This episode contains a great deal of Reported Speech material,
maintaining the tension created in the main-line clauses (and, of
course, contributing to it through the speech contents themselves)
through this section and into the next. This gets underway immediately, with two short speech sections on the subject of Ruth going
out to glean (2.2.1-4 and 2.2.5-6); it is followed by a section of nonSpeech clauses reporting her finding a place to glean. Here, we have
four main-line clauses (2.3.1-4), followed by a subordinated clause
(2.3.5), which in turn is followed by a Suffix clause beginning with mm
(2.4.1). This mm + Suffix clause identifies the peak of this episode.58
It is followed by three speech sections (2.4.2-3, 2.4.4-5 and 2.5.12), one main-line clause (2.5.3), and two more longish speech sections
(2.6.2-7.6 and 2.8.1-9.8)these cover Boaz's exchange of greetings
with his workers, his question to his foreman about Ruth, the
foreman's response, and Boaz's first words to Ruth. Ruth's response
contains two main-line clauses and one speech (2.10.1-2 and 2.10.3-5).
Boaz addresses himself again to Ruth (one main-line clause, 2.11.1,
and a Reported Speech section, 2.11.2-12.3); this is followed by two
more speech sections (Ruth's gratitude to Boaz, 2.13.1-4, and Boaz's
invitation to the meal, 2.14.1-4). A series of six main-line clauses
(2.14.5-15.1) describes Ruth's meal.
58. This claim is likely to be met with some scepticism. However, the clause
marking the peak is not to be equated with the peak itself (the meeting of Boaz and
Ruth), which follows, and is expounded in conversation. That this comes so early in
the episode is no less understandable: this is a peak without complete resolution, and
the tension thereof is carried through the remainder of the story, and is built upon,
until the main peak of the episode is reached (the agreement at the city gate between
Boaz and the go 'el).
182
The episode comes to an end with a speech section (Boaz's

instructions to his young men, 2.15.2-16.4), two main-line clauses and
a single subordinated clause (2.17.1-2 and 2.17.3which describe
Ruth's gleaning). The Reported Speech material in this episode is
concerned mainly with character development. The terminal boundary
is marked by a 'm clause, indicating the measure of her gleaning:
* 'rt pm // ...ry'And there was about an ephah of barley' (2.17.4).
3.2.5 Ruth 2.17.4-3.7.7
The initial boundary of this episode is not syntactically marked, but
rather is identifiable on the basis of the previous, off-line, clause, and
on the basis of the topic shift from 'gleaning' to 'returning home'. The
episode contains 17 main-line clauses, eight speech sections (in three
blocks) and four subordinated clauses.
The first seven clauses contain five main-line clauses, broken by
two subordinated clauses; these describe Ruth's return home to
her mother-in-law. Naomi questions her about her day (Reported
Speech: 2.19.1-4); Ruth's response is introduced by a single main-line
clause, with a single subordinated clause (2.19.5-6), then the speech
formula; the five speech sections that follow (2.19.7-9; 2.20.1.3;
2.20.4-6; 2.21.1-7; 2.22.1-4) recount Ruth's day, and Naomi's
amazement. Then follow two main-line clauses describing Ruth's
gleaning throughout the harvest, and then two speech sections where
Naomi instructs Ruth on the right way to approach Boaz (2.23.1-2;
3.1.1-4.8; 3.5.1-3).
Ruth 3.6.1-7.8 contains nine main-line clauses, and one subordinated clause, and describes Ruth's actions in fulfilling her mother-inlaw's instructions.
The Reported Speech material in this section reflects the tension
developed in the previous material and hints at a possible resolution.
I am intrigued by the presence of the four subordinated clauses. I
am tempted to propose that their presence here serves to underscore
the concern with stability and security in this story (a decisive
majority of the subordinated clauses in this text occur in sections
occurring at home, or in some way connected with economic or
domestic security). A wider sampling of material will be necessary
however, before any more secure hypotheses may be proposed.
There appears to be no syntactic marking of the peak in this
episode, unless we consider the repeated speech formula without
183
change of speaker, in 20.20.4, to be performing this function.59 Some

would make an episode division at 3.1.1, owing to the semantic content,
and owing to the fact that the preceding two verses summarize an
elapsed period of time. I am not averse to this; the preponderant
majority of texts that I have so far examined mark episode boundaries
and the like syntactically, but not all. It must be pointed out, however,
that the lack of syntactic markingin a text which has demonstrated a
tendency to mark boundaries syntactically does indicate that the
break, if any, is fairly understated.
The end of the episode is marked once again by the following
episode-initial TH clause.
3.2.6 Ruth 3.8.1-18.6
This episode, introduced by another Tn clause, deals with the nightscene at the threshing floor, and Ruth's return afterward to her
mother-in-law. After the opening (temporal) clause, we find two
main-line clauses, then a verbless clause with a participle, introduced
by mm. This, I would say, introduces the peak event of the episode.
This is followed by a conversation between Ruth and Boaz (three
sections: 3.9.1-2; 3.9.3-6 and 3.10.1-13.9) concerning her redemption; this is followed by two main-line clauses and a subordinated
clause, and two Reported Speech instructions to Ruth about her return
home (3.14.1-2; 3.14.3; and 3.1.4.4-15.4, both by Boaz). This is
followed by Ruth's return, and report, to Naomi (five main-line
clauses, 3.15.5-16.1, and three speech sections, the middle one of
which is introduced with a main-line clause + subordinated clause +
speech formula, 13.16.2-3; 3.17.1-4; and 3.18.1.6). The terminal
boundary is once again indicated by the initial-boundary-marking
devices of the next episode.
3.2.S Ruth 4.1.1-13.1
The initial boundary here is marked by a Suffix clause. The episode
contains seven main-line clauses, interrupted by nine Reported Speech
sections, five off-line clauses and two subordinated clauses.
The Suffix clause is followed by a single main-line clause (4.1.2),
then a verbless clause with participle (which is introduced by mm
4.1.3), a subordinated clause (4.1.4), and a speech section (4.1.5-7).
The action continues with three main-line clauses, a speech section and
59. This would appear to be the logical peak of the episode, as well. See below.
184
another main-line clause (4.1.8-2.1; 4.2.2-3; 4.2.4). The participants

are all now present at the city gate: Boaz, the go 'el, and the ten elders
of the city.
I do not find it surprising that a clause which I have consistently
considered to be a peak-marking device (the rum + ptc, verbless
clause4.1.3) is found so early in this episode. First, we have seen it
used fairly early on other occasions (see commentary above), and
secondly, this is the Peak Episode of the entire story; and thirdly, little
additional material needs to be added to this episode (additional
material which would normally form the 'stage-setting' section of the
episode)any such stage-setting has been accomplished in earlier
episodes. The author is free to launch immediately into the peak event.
If we propose that this clause-type serves to mark the opening of the
peak scene, then we can find ample logic for its use where we have
found it.60
Then follow four Reported Speech sections alternating between
Boaz and the go'el, fine-tuning the details of the redemption of Naomi
and of Ruth (4.1.3-4.10; 4.4.11-12; 4.5.1-2; and 4.6.1-6). The next
four clauses are intriguing: a verbless clause (4.7.1); two verb-initial
Suffix clauses (4.7.2-3); and another verbless clause (4.7.4, a resume
of the first). These clauses occur at the precise moment of Boaz's
receiving the right of redemption.61
The mechanisms of peak-marking devices are language specific, as
we have noted before. The language may have a preference for
saturating the text with verbs (thus increasing the speech of the
events), or with adverbials and relative material (thus increasing the
detail of the events); it may work to slow down the forward march of
events, giving a kind of 'slow-motion' effect to the text. Classical
Hebrew appears to have a preference for the latter. Texts marked
clearly for peak events show the tendency to interrupt the flow of
events, just at the point where the tension is at its peak. It is not
surprising to me, therefore, that we find this set of four clauses where
we find it.
60. In 2.4.1, it begins the scene of Ruth's first encounter with Boaz. In 3.8.1, it
begins the scene of Ruth's second (recorded) encounter with Boaz. Both of these
scenes are the keys to the building up toward this, the final, key scene.
61. E.F. Campbell, Ruth (AB; Garden City, NY: Doubleday, 2nd edn, 1975
p. 148. On the question of whether or not this verse is intrusivea later addition
which does not blend into the surrounding materialsee my own notes below.
185
The question of its integrity in the text could be argued either direction, in my opinion, without impact on our assessment of this section's
macro-syntactic function. If it is intrusive, it was added in such a way
as to serve the purposes of the text admirably well (taking into
consideration its effect as a peak marker); if it is part of the original
story, its contribution is the same. The question will have to be
addressed without reference to its having 'broken up the flow of the
text', for that is what it would have been intended to do, whether as an
original part, or as an added part, of the text.
The speech of the go'el resumes (4.8.1-2), and he performs the
ceremony mentioned in v. 7 (4.8.3). The story continues with two
speech sections (Boaz addresses the witnesses, 4.9.1-10.3, and the
witnesses respond, 4.11.1-12.3); the latter of these two is introduced
by a speech formula containing a subordinated clause. The episode
concludes with a single main-line clause (rm~nK Tin npn, 'And Boaz
took Ruth', 4.13.1), and a Tim clause ('and she became his wife',
4.13.2).
3.2.8 Ruth 4.13.2-17.4
What follows now in the text is a pair of what may be called 'postpeak episodes'62 (4.13.2-16.2 and 4.16.3-17.4), whose role is to
unravel the tension of the story; they perform the function of a
denouement. The first has no syntactically marked initial boundary,
but is identifiable on the basis of the preceding Tim clause, and on the
basis of a topic-shift, from the marriage, to the marriage bed. The
second follows another Tim clause (terminating the first alleged postpeak episode), in which Naomi becomes nurse to her grandson. While
both do contain speech sections, they contain a lesser proportion to
narrative than we have seen in the major episodes of the story; nor is
this surprising if we maintain our hypothesis that conversational
material in Hebrew has a tendency to maintain or strengthen tension in
the textthe opposite effect of what is desired here.
In the first of these episodes, we find three main-line clauses, then a
speech of blessing on Naomi by the townswomen; this is followed by
two main-line clauses, in which Naomi takes up the child (4.13.2;
62. We need to examine the option that these are not actually separate episodes;
this question will be addressed in depth after the present hypothesis has been
commented on.
186
4.13.3-5; 4.14.1-15.4; and 4.16.1-2), and is concluded by the 'Naomi

as nurse' Tim clause (4.16.3).
The second of these contains a speech section, and another main-line
clause, and concludes with a verbless clause (4.17.1-2; 4.17.3; and
4.17.4these acclaim Naomi, name the child, and indicate his significance in history).
It is worth noting that the speech materials in these last two sections
are fairly static in natureeven though the first is '+ Projection' and
the second is historical, both are concerned more with state than with
action. Further work will enable hypotheses to be constructed with
regard to specific tension-building capacities of different text-types.
The question arises as to whether rrn clauses that translate into
English as 'become' clauses have different macro-syntactic significance. I have strong doubts about this, for the real issue seems to
have more to do with the semantic domains of the English words 'be'
and 'become' than with the function of the Hebrew clause-type(s).
Nevertheless, the question presents itself, and the more so here
because the semantic content of the passages in which they are found
does not admit altogether readily to their identification as episodeboundary features. In 4.13.1-3n VrpR //rim*'rm // ...ry inn mm
iikn n}]', //, 'And Boaz took Ruth, and she became his wife, and he
went in to her...'), for example, the clauses all deal with the same
subject, and flow on from one to another without looking much like a
break of any sort.
And in 4.16.1-17.1- jnp'i // nm vnuro vn // m~na'D" mro //^RiiD'
...&* pa^'i //...btoiD' ni]3 nn^n'And Naomi took the child, and laid him
in her lap, and became to him a nurse; and the neighbour women gave
him a name, saying...) the same thing is the case, although the case
becomes stronger here for assigning the Tim clause the role of episodeterminal boundary indicator. Further work on this verb will give us
more freedom to draw conclusions about its various functions.
3.2.9 Ruth 4.18.1-22.2
The remaining ten clauses do not require much comment. They
conclude the book, as is obvious; they are not of what we have
described as the Narrative History text-type. They are introduced by a
single verbless clause (4.18.1), which also ties the section (loosely) to
the preceding material (by mention of the name 'Perez', also found in
4.12.1).
187
3.3 RuthThe Reported Speech Material

In this section, I will examine the individual texts embedded, as
Reported Speech, into other clauses. I cannot hope to do this justice
each text deserves, if not requires, the same kind of attention we have
paid to the other portions of Ruth examined above, and both time and
space impose rather strict demands. In addition, not much has been
contributed towards description of non-Narrative History text-types in
Classical Hebrew. Therefore the scope of the following analyses will
be restricted to (1) general identification of text-type(s) involved in
each text, and (2) comment on internal structure, where this is
possible without undue speculation.
I will go through Ruth from start to finish63 and will comment on
every speech unit, even if only to remark on why I cannot pursue
analysis of it at this point. Those sections where analysis is limited by
the size of the text, or where the text admits no description without a
far greater commitment of energy, will be relegated to footnotes.64
I will not hesitate in this section to be succinct in, and to forego
extended justification of, my analyses of these texts, once the reader
has been introduced to my approach to the data.
In each case, clause counts of the texts will include the speech
formula, to reiterate the embedded nature of the Reported Speech
matieral in the speech formula.
3.3.1 Ruth 1.8.1-9.2
Naomi's advice to her daughters-in-law to return to their homes. This
is a text of seven clauses: 1 Speech Formula (1.8.1), 3 Imperative
(1.8.2-3, 1.9.2), 2 Jussive (1.8.4, 1.9.1) and 1 Subordinated (Suffix)
(1.8.5).
1.8.1
2
3
4
5
9.1
2
63. Though it is strongly tempting to analyse these texts by grouping them first
according to text-type, this approach, and its attendant benefits, must await a less
restricted study.
64. Eleven such sections will be addressed only in footnotes.
188
As can be seen from the clause-types, this text is readily identifiable as

a Hortatory text. We can propose another Hortatory text embedded in
the first (1.8.4-9.1)a blessing, since there is a shift to third
personbracketed by the two, and one, imperative clauses.65
Ruth 1.1L1-13.666
This text allows us much more freedom to explore. It contains 18
clauses: a speech formula, and an embedded text of 17 clausesthe
last eleven of these are subordinatedI will outline this section more
fully below; the first six contain: 3 Imperative clauses (1.11.2, 1.12.12), 1 Prefix (question) (1.11.3), 1 Verbless (question) (1.11.4) and 1
we + Suffix < rrn (1.11.5), followed by a subordinated clause (1.12.3),
into which the remaining material in this Reported Speech section is
embedded by another subordinated clause, a speech formula (1.12.4).
11.1
2
3
4
5
12.1
2
3
4
5
6
7
13.1
65. As I intimated while working on Judg. 10.6.1-12.7.3, I am adopting

Longacre's proposed verb-rank 'clines' for Narrative Predictive (see his Joseph,
p. 107, and above) and Hortatory (ibid., p. 121, and above).
66. (1.10.1-3) This unit is composed of a speech formula, and a subordinated
clause for which we must supply an ellipsis: fiai?1? 3ie?3 "[HK '3 n*?'n3"inm. I can
offer no further comment on this regarding text-typeits subordinating conjunction
enforces a non-initial verb form, so we cannot know its true text-type. As it is in
response to Naomi's advice, and as the verb is a Prefix form, we may guess it to be
from the Narrative Predictive text-type (allowing for the fact that the context, and the
content, would suggest that this response would be '+ Agent Orientation').
189

2
3
4
5
6
The structure of this unit, and the text-types involved here, are
complex. It is clear that there is a break at 1.12.3, at the subordinating
conjunction; the section preceding the subordination is clearly another
Hortatory text, with another embedded text within it (the embedded
material being a sort of sub-text within the paragraph, bracketed by
the repeated Tin nno clauses). I will not at this time comment further
on this text, owing to the difficulties of sifting through the implications of the questions. (The second and third clause, being a Verbless
clause and a 'we + Suffix < rrn' clause, point us in the direction of an
Expository/'What-it-will-be' text-type.)
The remaining section (1.12.3-13.6) likewise leaves us few clues to
text-types. The subordinating conjunctions, adverbials and question
formats make identification of text-types nearly impossible (we can
identify 1.12.5, of course, as belonging to one of the Expository texttypes, but this is hardly an improvement). On the other hand, we can
explore the internal structure of this second section with more success.
We can assign the two questions and their intervening subordinated
clause (1.13.1-3) to a single subsection; and I propose that the -maN o
of 1.12.4-7, which immediately precedes it, is a sort of protasis, to
those questions. The questions are answered by the speaker (they are
obviously rhetorical) in the negative and reasons are given (1.13.4-6).
3.3.3 Ruth 1.15.1-3
A simple speech formula introduces this unit of two clauses: the first
clause (a Suffix clause) is introduced by run; the second is an
Imperative clause.
15.1
2
3
This is a Hortatory text, where the reason for the command is

given; this arrangement is called by Longacre a Hortatory Reason
Paragraph.67
67. See Longacre, Joseph, p. 92, also p. 125 et passim.
190
Ruth 1.16.1-17.6
This masterful section, introduced by a simple speech formula,
contains 13 clauses, only one of which (a negated Imperative clause
[1.16.2], occurring first in the sequence of Reported Speech clauses)
does not occur as subordinated text. The subordinated clauses (1.16.31.17.6) include 8 Prefix clauses, broken by 3 Verbless (1.16.7-8,
1.17.5 [w/ Ptc.]) and 1 Jussive (1.17.4).
16.1
2
3
4
5
6
7
8
17.1
2
4
5
6
The same configuration prevails here as in the last unit analysed: this
is another Hortatory Reason Paragraph. The reason in this case is an
extended list of Narrative Predicative clauses, many of which are
paired by subordination. I am tempted to propose that two of the
Verbless clauses (1.16.7-8) are peak-markers. This, of course, is
speculative, since we cannot be certain of the nature of the text for
which we are proposing the peak.
The text concludes with an oath (1.17.4-6the oath formula itself
[a Prefix clause, and Verbless clause w/ Participle], and the attendant
restriction [a subordinated Prefix clause]).
191
Ruth 1.20.1-21.5
The contents of this section are: 1 Speech formula (1.20.1), 2 Nonsubordinated Suffix clauses (1.21.4-5), 3 Subordinated Suffix clauses
(1.20.4-21.2), 4 Prefix (1.20.2 [negated], 1.21.3 [a question]) and 1
Imperative (1.20.3).
20.1
2
3
4
21.1
2
3
4
5
The first two clauses are clearly Hortatory; the remainder is less easy
to place. This is poetic in style, and we have very little to go on in
terms of studies of poetic syntax.69
3.3.7 Ruth 2.2.1-4
This unit comprises four clauses; the speech formula contains one, the
remainder are two Cohortative clauses, and one Prefix clause (2.2.4,
embedded by subordination). The text is clearly Hortatory.
2.2.1
2
3
68. (1.19.4-5) This is a very brief speech section, containing only three words: a
speech formula, and a two-word Verbless clause in question formatpR //rim* 'rt pm /
'QJJ3). It is a non-rhetorical 'yes-or-no' question, which allows us to speculate on the
kind of form it would take if it were not a question; it is likely that this is a
representative of the Expository text-type.
69. My feeling is that poetic concerns displace text-type features sufficiently that
text-type identification of highly poetic passages is nearly impossible, or at best,
irrelevant. Dr Longacre tells me (personal communication, 1992) that very little
application of his theory and methodology to poetic texts has been made to date.
192
Ruth 2.4.2-310
This unit contains one speech formula clause, and one Verbless clause.
4.2
3
Although the embedded text contains Verbless clause, it is Hortatory

(it contains a blessing, not a command) rather than Expository. Here
is a case where semantics, rather than syntax, identifies for us a text's
type. As I have repeatedly noted, evaluation by means of one rarely
excludes the other.
Ruth 2.5.1-211
This unit contains one speech formula clause, and one Verbless clause
(a question).
5.1
2
72
This embedded text is Expository.
Ruth 2.6.1-7.6
This speech unit contains nine clauses: 2 Speech formula clauses
(2.6.1-2, both we + Prefix), 3 we + Prefix (2.7.1, a speech formula,
introducing the next two, embedded clauses, and 2.7.4-5), 1 Verbless
clause (2.6.3), 1 Cohortative (2.7.2) and 1 we + Suffix (2.7.3) and
the final, 'badly disrupted',73 clause, whose precise syntax will,
unfortunately, very likely remain a mystery.
6.1
2
3
7.1
70. (2.2.5-6) This unit contains one speech formula clause and one Hortatory
(Imperative) clause: TD 'D1? n1? -own.
71. (2.4.4-5) This unit contains one speech formula clause and one Jussive
clause. The latter is a Hortatory textpR //rim* 'rt pm //
72. The question is not rhetorical, which simplifies understanding its purpose,
and enables us to determine more easily its nature.
73. Campbell, Ruth, p. 96.
193
2
3
4
5
6
This is the first instance in Ruth of a coordinated speech introduction;

this phenomenon intrigues me greatly but I will not comment on it
here. It is a feature which surely imports unique detail into the text,
yet without study of all such occurrences I cannot draw any
conclusions.74
Despite the great variety of clause-types, and the textual difficulties
with the final verse,75 this text is fairly easy to sort out. The initial
clue is the 3 we + Prefix clauses; these, and the semantic content,
secure the identification of this text as an embedded Narrative History
text. The first of these clauses is another speech formula, which
introduces an embedded Hortatory text (2.7.2-3). The first and last
clauses in the embedded Narrative History text look very like the sort
of things we have begun to expect at the initial, and terminal,
boundaries of Narrative History texts.76
Ruth 2.8.1-9.8
This speech sections contains 13 clauses: 1 Speech formula clause
(2.8.1), 5 Prefix clauses (2.8.3-4 [negated], 2.9.2 and 2.9.8
[subordinated]), 4 we + Suffix (2.9.3, 2.9.5-7), 2 Suffix (negated)
(2.8.2, 2.9.4) and 1 Verbless (2.9.1).
74. The exhaustive treatise of Meier (Speaking of Speaking: Marking Direct

Discourse in the Hebrew Bible) is the best starting place for a fuller treatment of this
phenomenon. Although he does not devote a specific section to the question of
multiple-clause introductions to Reported Speech, he makes several comments on
sections containing such speech formulae.
75. For further comments on these textual difficulties, see Campbell, Ruth,
pp. 94ff.
76. Both the difficult MT reading (in essence, a Verbless clause with a substantival Infinitive), and the apparently derivative LXX reading (a negated Suffix clause)
could fit easily into the role expected of a terminal clause.
194

8.1
2
3
4
5
9.1
2
3
4
5
6
7
8
This section, though it, too, looks somewhat difficult at first glance, is
relatively straightforward. The questions (2.8.2 and 2.9.4) are difficult, but do not impede us from analysis of the rest of the text. 2.8.39.2 are clearly Hortatory clauses; 2.9.3 and 2.9.5-8 are clearly
Procedural/Instructional. Both units conclude with a subordinated
clause.
Ruth 2.11.1-12.311
This text begins with a two-clause introductory speech formula; it
continues with 4 Suffix clauses (2.11.7 [negated], 2.11.3-4, 2.12.3), 2
we + Prefix (2.11.5-6) and 2 Jussives (2.12.1-2, the second with [[ ).
11.1
2
3
4a
b
5
6
7
77. (2.10.3-5) This unit is comprised of 3 clausesa'D" mro //^RiiD'3 piiD'3 p

&*^'i //...bta speech formula, a Suffix clause (a question) and a Verbless clause.
Once again, we have a non-rhetorical question, whichin combination with the
Verbless clausepermits us to determine this as an embedded Expository text.
195
12.1
2a
b
3
78
This is the second instance of a coordinated speech formula.

Leaving aside 2.11.3, whose text-type affiliation is elusive, I turn to
the next four clauses (2.11.4-7), which are clearly Narrative History
(note once again the concluding subordinate clause); the following
three clauses (2.12.1-3) are another blessing text (Hortatory, with a
yet another concluding subordinate clause).
Ruth 2.13.1-5
This speech unit contains a simple speech formula, and four other
clausestwo Suffix clauses (subordinated), bracketed by two Prefix
clauses (the latter negated and of II']])
13.1
2
3
4
5
The first section of this unit remarks on the unexpected kindness of

Boaz (which in this text is followed by two subordinated clauses); the
second is a statement of identity. This is similar in structure to 2.10.35;79 the specific paragraph- and text-types of these units are difficult to
ascertain.
Ruth 2.14.1-4
This speech unit contains a speech formula and one Imperative clause,
followed by two we + Suffix clauses.
14.1
2
3
4
78. See comments on 2.6.1-7.6 above.

79. See note 77 above.
196
The embedded text appears to be a command, and its result; this may
fall under the category of Hortatory, or it may be a combined Hortatory and Narrative Predictive paragraph.
Ruth 2.15.2-26.4
The unit contains one speech formula clause, and 4 Prefix clauses
(2.15.4, 2.16.4 [negated]; and 2.15.3, 2.16.1) and 2 we + Suffix
clauses (2.16.2-3).
15.2
3
4
16.1
2
3
4
The embedded text appears to be a solid stretch of Procedural/

Instructional material, much of it in secondary (off-line) forms owing
to negation or fronting of emphasized clausal elements. The repetition
of Q3 at 2.15.3 and 2.16.1 may indicate the onset of paragraphs; this
suggests two units 2.15.3-4 and 2.16.1-4both of which are terminated by negated clauses. That this material is intentionally structured
as the two units proposed here is further substantiated by a shift from,
on the one hand, forbidding the workers to harass Ruth, to, on the
other hand, requiring them to be intentionally generous with the grain
left for her.
Ruth 2.19.5-9*
The response to Naomi's questions (2.19.1-4) comes in the form of a
Narrative History text, and of a Reported Speech section. We can see
the former as part of a composite introductory speech formula. The
remaining clauses are two: a Verbless clause, and a subordinated
Suffix clause. The speech text is Expository.
80. (2.19.1-4) The unit consists of 1 speech formula, 2 Suffix clauses (in
question format), and 1 Jussive (blessing) clause. I cannot comment any further at
this point, since the difficulties of question texts, and of Suffix clauses unaccompanied by contextual material to help with identification, preclude greater precision
/ nm vnuro vn // m~na'D" mro //^RiiD'3 pn-nm
197
19.5
6
7
8
9
Ruth 2.20.1-3
The following text is composed of a simple speech formula, a
Verbless clause, and a subordinated negated Suffix clause.
20.1
2
3
This embedded text can be described as a Hortatory unit, despite its

similarity of clause-types to the preceding text; the Verbless clause, in
this case, presupposes a Jussive form of rrn.
Ruth 2.20.4-6
This unit consists of one speech formula and two Verbless clauses.
The embedded text is Expository.
20.4
5
6
This text is unique in Ruth in that the speech formula is repeated,

without a change of speaker and/or addressee. This phenomenon
occurs at the point where connections are being made (i.e. that Boaz is
a near relative who can serve as a redeemer) which will be integral to
the resolution of the storywhether this is coincidence or artifice is
unclear at this point.
Ruth 2.21.1-7
This section contains 1 Speech formula clause (2.21.1), 2 Suffix
clauses (2.21.3 [subordinated], 2.21.6), 3 Verbless clauses (2.21.5,
2.21.7 [both subordinated]) and 1 Prefix clause (2.21.4).
198

21.1
2
3
4
5
6
7
The rather disjointed embedded text commences with an ellipticalor

rather, defectiveclause (does this reflect excited speech?), to which
is subordinated a speech formula; the remainder amounts to an
embedded Procedural/Instructional or Hortatory text, with an
appended temporal clause.
Ruth 2.22.1-4
This section contains four clauses, in order: one speech formula, one
Verbless clause and two Prefix clauses (one negated).
22.1
2
3
4
The embedded text in this unit is Expository, with the reason for the
Expository statement being expounded by an embedded (by
subordination) Narrative Predictive text.
Ruth 3.1.1-4.8
This unit contains 19 clauses: 1 Speech formula (3.1.1), 8 we + Suffix
clauses (3.3.1-4, 3.4.2, 3.4.4-6), 6 Prefix (3.1.2-3, 3.3.5 [both
negated]; 3.4.3, 3.4.7-8), 2 Verbless (3.2.1 [a negated question], 3.2.3
[w/ mn + Ptc]), 1 Suffix < rrn (subordinated) (3.2.2) and 1 Prefix
(<II, IInegated) (3.3.5).
199
3.1.1
2
3
2.1
2
3
3.1
2
3
4
5
4.1
2
3
4
5
6
7
8
The first two clauses are difficult, again owing to the 'question of
questions'; they appear to be introductory to the remainder of
Naomi's speech. The two Verbless clauses, and their intervening
subordinated rrn clause, provide the setting for the following
Procedural-Instructional text (3.3.1^.6); this text is divided into two
episodes, the first terminating with a negated Prefix clause, the second
beginning with a Prefix clause < rrn. The subordinated clause (3.4.3)
may function to introduce the peak.81 The final two clauses are
Narrative Predictive (3.4.7-8), forming a Procedural Result sort of
paragraph.
Ruth 3.9.3-tf2
This speech section is composed of a speech formula clause, 2
Verbless clauses (1 subordinated), and 1 we + Suffix clause.
81. 'Peak' is perhaps an inappropriate term for Procedural/Instructional material;
what is being referred to here is the clause or sequence of clauses which form the
focal set of instructions.
82. (3.5.1-3) This unit, composed of a speech formula clause, and 2 Prefix
200

9.3
4
5
The embedded text I would identify as Procedural-Instructionalor

an Expository unit and a Procedural-Instructional unit. The former
explanation considers the Verbless clause at the commencement of the
speech unit to be a background 'setting' for the Instruction section; the
latter explanation sees the two sections as more on an equal footing, or
even that the first is of greater 'weight' than the second. I lean slightly
toward the latter explanation, but consider the two sections more or
less equal.
Ruth 3.10.1-13.9
This unit contains 20 clauses: 1 Speech formula (3.10.1), 7 Verbless
clauses (3.10.2 and 3.11.4 [w/ Ptc];3.11.5, 3.12.2-3), 6 Prefix (3.11.1
[negated], 3.11.2-3, 3.13.4-6 [the last functioning as a Jussive]), 2
Imperative (3.13.1, 3.13.9), 2 we + Suffix (3.13.3 [< rrn], 3.13.7), 1
Suffix (3.10.3), 1 Clause with & (3.12.4) and 1 Elliptical clause
(omitted from the clause count) (3.12.1).
10.1
2
3a
b
11.1
2
3
4
5
12.1
2
clauses (one of which is subordinated), contains a Narrative Predictive text (Town

ntoR <Q> 'w 'TBumitfK "53 rrt>R).
(3.9.1-2) This unit contains 2 clauses: a speech formula, and a Verbless
clause. The embedded text (a question), appears to be Expository (nsr'a nntn).
201

3
4
13.1
2
3
4
5
6
7
8
9
The embedded text opens with a Hortatory Reason Paragraph (a blessing text 3.10.2-3); it is followed by another of the same, but this
time the second part is expounded by a Predictive text, rather than a
historical one. This latter contains a reason for the reason, as it were
(two Verbless clauses).
The third section (a complex Hortatory text, 3.12.1-13.9) is divided
from the second by nnin, as was the second from the first. The initial
clause of the third section is defective; an Expository text is
subordinated to it (3.12.2-4); this is followed by the body of the
Hortatory text, which itself has several layers of embedding.83 Suffice
it to say that there is Predictive material, and there is Hortatory
material, as well as Expository material, embedded in this unit.
Ruth 3.14.4-5
This speech section consists of three clauses: the speech formula, a
negated Jussive clause and a subordinated Suffix clause. There is no
difficulty in identifying the embedded text as Hortatory, on the basis
of the single main clause.
4.4
5
6
Ruth 3.15.1-4
This unit comprises a speech formula and two Imperative clauses separated by a subordinated Verbless clause. This text is likewise Hortatory.
83. I would risk going too far with this analysis, were I to attempt to describe
closely each clause in this section; any suggestions would be no more than
provisional. This, too, I will leave unattended.
202

15.1
2
3
4
Ruth 3.16.4-17.4
This section contains 6 clauses: 3 speech formula clauses and 2 Suffix
clauses (one of which is subordinated), followed by a negated Prefix
clause.
16.4
5
17.1
2
3
4
This composite speech formula is very similar to the one we saw at
2.19.5-7, which occursas does this onein the context of Ruth
recounting to Naomi an important encounter with Boaz. The 'past
tense' parts of the embedded text hint at Narrative History, but the
evidence is not sufficient to secure an identification; the subordinated
clause, however, is a speech formula clause, and introduces an
embedded Hortatory text.
Ruth 3.18.1-6
The six clauses of this speech unit include 1 Speech formula (3.18.1),
1 Imperative clause (3.18.2), 3 Prefix (all subordinated) (3.18.3-5 [the
last is negated]) and 1 Suffix (3.18.6).
18.1
2
3
84. (3.16.2-3) This section is identical in structure to 3.9.1-2 (see note above;
the same comments apply), apart from the vocative added to the end of the
question.
203
4
5
6
The text is an embedded Hortatory Reason Paragraph, the second half

(3.18.5-6) being a Predictive text with a temporal (Suffix) clause.
Ruth 4.1.5-7
This unit contains a speech formula clause and two Imperative clauses;
the embedded text is Hortatory.
5
6
7
Ruth 4.3.1-4.10*5
This unit contains 13 clauses: 1 Speech formula clause (4.3.1), 4

Prefix (3.4.5 and 3.4.7 [subordinated; the latter is negated], 3.4.2,
3.4.8), 3 Imperative (3.4.3-4, 3.4.5), 2 Suffix (3.3.2, 3.4.1), 2
Verbless (3.3.3, 3.4.10), 1 Clause with |< (3.4.9).
3.1
2
3
4.1
2
3
4
5
6
7
8
9
10
85. (4.2.2-3) This unit contains 2 clausesa speech formula and an Imperative
clause (ns~aci intn); its embedded text, like that in the previous one, is Hortatory.
204
The embedded text contains a speech formula (4.4.1-2), which introduces an embedded Hortatory text (4.4.3-10)the latter part of this
(4.4.8-10) is an amplification of 4.4.7; the earlier section (4.3.2-4.2)
is unclear, though I take it to be a stage-setting device for the
Hortatory text that follows it.
Ruth 4.4.11-12
This unit contains a speech formula, and a Prefix clause; the embedded
text is Predictive.
11
12
Ruth 4.6.1-5*6
This speech unit contains 5 clauses: a speech formula clause, 3 Prefix
clauses (2 negated; 2 subordinated) and 1 Imperative clause (3.6.5).
6.1
2
3
4
5
The embedded text is composite: the opening clause is Predictive, and

is followed by a result statement (Predictive?); the second section
(4.6.4-5) is a Hortatory Reason Paragraph, the second half being a
summary of the first (Predictive) clause.
Ruth 4.9.1-10.3"
There are 8 clauses in this unit; after a single clause speech formula,
we find: 4 Verbless (4.9.4-5 [subordinated], 4.9.2, 4.10.3), 2 Suffix
(4.9.3, 4.10.1), 1 Prefix (negated) (4.10.2).
86. (4.5.1-2) This section is composed of a speech formula and a Suffix clause.
Identification of the text-type of the embedded clause is not possible ('DT3 TO nan
irftrtrttf rcrrntS D'pn1? <Q> nmp narrneto rroKiori rrn num 'DCJ rro mton inup).
87. (4.8.1-2) This unit is composed of a speech formula and an Imperative
clausefmjp tra1? *an ion; the embedded text is Hortatory.
205
9.1
2
3
4
5
10.1
2
3
The embedded text is bracketed by two Verbless clauses (identical);

these identify the incorporating text as Expository; the material contained in the subordinated clauses is some sort of historical/expository
material.88
Ruth 4.11.1-12.3
This speech section is introduced by a composite speech formula containing 2 clauses, and includes a further 8 clauses. These are: 3 Prefix
clauses (4.11.4 [Jussive], 4.12.1 [< rrn], 4.12.3), 2 Suffix (subordinated)
(4.11.5, 4.12.2), 2 Imperative (4.11.6-7), 1 Verbless (4.11.3).
11.1
2
3
4
6
7
12.1
2
3
88. This appears to be formal speech; we simply do not have enough data
processed to be able to venture conclusions about its text-linguistic features.
206
In response to Boaz's formal speech to them as witness of the transaction, the ten elders respond with an Expository 'we are witness', and
continue on with a blessing of the couple; this divides into two units
(semantically, the first section dealing with Boaz, the second with
Ruth), which both conclude with subordinated clauses.
Ruth 4.14.1-15.4
This section contains 8 clauses: 1 Speech formula (4.14.1), 3 Suffix
clauses (subordinated) (4.14.3 [negated], 4.15.2-3), 2 Verbless
(4.14.2, 4.15.4), 1 we + Suffix < rrn (4.15.1), 1 Prefix (possibly as
jussive) (4.14.4).
14.1
2
3
4
15.1
2
3
4
This embedded text is Hortatory (a blessing)it has a historical

section (4.14.2-4) giving the reason for the praise.
Ruth 4.17.1-2
The final Reported Speech section of the book is composed of two
clauses: a speech formula and a Suffix clause. The embedded text does
not admit to any more precise description than 'historical'.
17.1
2
3.4 Summary and Conclusions

The conclusions we may draw from this examination of Reported
Speech material in Ruth are hindered by three factors:
1.
2.
3.
207
Although I have, throughout this volume, cited the need to

test our conclusions against further data, here the need is the
greater; in short, the data-sample has been too small to make
any but the most obvious, and the most tentative observations,
for a variety of reasons. Nonetheless, here more than
elsewhere we need to process more texts.
The first of these reasons for needing a larger data-base is
that text-types within Reported Speech material shift rapidly,
and it is not common to find long stretches of material in a
single text-type. Broadening our data-base would bring to us
more texts of a greater length, which are a better starting
point for research than shorter ones.
It is clear that subordination (which is more common in
Reported Speech than in non-Reported Speech), specifically,
and embedding, more generallyby reason of their cohesion
with other units within their contextboth limit the kinds of
clauses that can occur at the outset of any text unit in such
a section. This immediately means that we have a greater
number of clauses than we would like whose surfacestructure signals as to text-type have been obscured by such
permutations.
On the other hand, however, there have been encouraging results as

well. One of my working hypotheses89 was that the constituent
structure of a text would be marked by divergences from the mainline form in all text-types, and that the 'off-line marking of
constituent structure will be confirmed by other types of marking
devices'. Where the text has been ample enough to examine both
main-line and off-line clauses in a single text-type, we have seen this
hypothesis substantiated: Ruth 2.15.2-16.4 is a good example of this,
where syntactically marked divisions are confirmed by a shift in
topical focus.
It was also found that the data I have examined follow the patterns
established for their individual text-types; any disruptions of the
expected pattern can be seen to be consistent with the hypothesis that
these were conditioned by syntactic relationships between the
embedded (Reported Speech) text and the clause into which it is
embedded.
89. Seep. 117.
208
We have seen in our examination of Ruth strongly consistent

tendencies, both within the book itself and in comparison with other
texts. In particular, the Narrative History text-type and the
Procedural-Instructional text-type (being the two text-types with
which the reader will now be most familiar) were shown to have
consistent boundary, main-line and peak-marking features throughout
the data, whether in simple narration or in Reported Speech. The
treatment of Reported Speech in Ruth has shown that the kinds of
features observed in other texts examined occur equally when the
material is conveyed as Speech. I anticipate that this hypothesis will be
found the more solidly substantiated as more data are processed.
Chapter 6
SUMMARY, CONCLUSIONS AND IMPLICATIONS
1. Summary
In this study, I have examined five currently influential works on the
text-linguistic description of Classical Hebrew, a theoretical base and
methodology for such description was presented, and several texts
were worked according to this theoretical model. My goals have been
to underline the need in such undertakings for good theory and
methodology, and for clear and direct communication of findings.
To this end, in the first two chapters, I surveyed Niccacci's Syntax,
Eskhult's Studies, Andersen's Sentence, Khan's Studies, and
Longacre's Joseph. Each of these contributes to our growing understanding of text-level features in Classical Hebrew. Each of them also
fails to achieve our ideal standards of theoretical-methodological
integrity and clarity of presentation.1 It is claimed (1) that the Hebrew
language can be described elegantly and helpfully at the level of 'text',
and (2) that this cannot be accomplished if the researcher's theoretical
starting point does not allow for the possibility of a variety of texttypes, or if the write-up does not explain itself so that linguistically
astute, but non-linguistically trained, hebraists can both trace the
procedures and comprehend the results.
Of the five works we examined, it was claimed that Joseph offered
the greatest steps forward in the description of the languagethat is,
its description of text-types by a matrix with three distinctive parameters, and the description of each text-type in terms of its own
specific scale of clause-type distribution (which Longacre terms
'clines'); and since Longacre does not offer much theoretical
explanation, the third chapter attempted this task. Since space was
limited, it was decided that attention should be given primarily to that
1. With the exception of Khan, whose topic is so restricted, and is so alone in
its class, that its usefulness is limited, even if it is a model of scholarly work.
210
portion of the theoretical basis which would permit the reader

quickest access to the most significant contributions of Joseph; this has
meant that we worked toward an understanding of certain basic
features of 'tagmemics' which are particularly important for an
understanding of the matrix and the clines.
This presentation of the theoretical base also entailed discussing
methodological principles, and in the end led to the proposal of some
working hypotheses with which we could give the theoretical base a
'road-test'.
This road-test consisted in asking of several texts whether text-types
and main-line forms did in fact appear to be linked, and whether the
patterns created by the alternation between main-line and off-line
forms coincided with other features to reveal the internal structure of
the texts. In addition, we looked at 'Reported Speech' to determine, if
possible, whether this kind of text had the same text-type and cline
characteristics as non-Reported Speech. The final analysis attempted to
step away from self-conscious theoretical explanation and to apply the
theory and methodology, more freely, to a single, unified text.
Thus, the organization of this volume is as follows:
Chapter 1: Introduction, and Examination of Three Texts: Niccacci,
Eskhult and Andersen
Chapter 2: Examination of Khan and Longacre
Chapter 3: A Summary Introduction of the Tagmemic Linguistic Model
of Text Analysis, and Resultant Principles of Methodology,
Including a Set of Working Hypotheses with which to Test
the Model
Chapter 4: Analysis of Hebrew TextsTaking Text Samples from
Judges, Leviticus and Exodus
Chapter 5: Analysis of More TextsSamples from Judges and the Book
of Ruth
Chapter 6: Summary, Conclusions and Implications
It is merely happenstance that the central chapter of this study focuses

on theory and methodology, yet it is nonetheless significant. In fact, I
might loosely apply the term 'chiastic' to the structure of this work
the central chapter lays out for the reader the details of the theory and
methodology with which I have approached the data; those preceding
it examine modern works with reference to their theory and
methodology (at a fairly basic level)and, generally, find them less
than ideal; those following it apply the proposed theory and
methodology to a set of biblical texts.
6. Summary, Conclusions and Implications
211
2. Conclusions
In our survey of other works (chs. 1 and 2), we found each of these to
contain solid contributions to our field of interest, but also to exhibit
specific weaknesses. In Niccacci and Eskhult, we found that the binary
opposition between 'Narrative', on the one hand, and 'Discourse', on
the other, which both relied upon, was inadequate to describe what is
happening in Hebrew at the text level, particularly within their
category of 'Discourse'. In Andersen, the difficulty encountered was
one of readability; simply, valuable data is obscured by 'jargon' and
idiosyncratic abbreviations, to the extent that few hebraists really
tackle the work at all.
Khan's topic is very narrow, and therefore is limited in its applicability; yet it stands out as a model of controlled, balanced
scholarship, and clarity of communication; little can be said by way of
complaint about this work.
Longacre suffers somewhat from the same 'jargonal' deficiencies as
Andersen; this is not simply a characteristic of this model of analysis,
because Khan, for all his lucidity, cites Longacre as one of his
strongest influences; rather, it is merely a feature of scientific studies
that 'process' is highlighted over communicability, with the result that
some obfuscation of the results is not uncommon. On the other hand,
Longacre imports unique insight to the treatment of Hebrew texts, and
his contributions are extremely valuable.
I would like to stress here that chs. 1 and 2 are not simply introductory, providing background for the study at hand; rather, these
chapters contribute a major part of the substance of this work. I have
provided an evaluation of theory and, more importantly, methodology, in a variety of text-linguistic undertakings, and a recommendation for a particular theory and methodology for further analysis of
biblical Hebrew. Therefore, examination of these works is not merely
required as a setting for my own project; it is my project.
Many authors give some attention to text-linguistic research without
entering fully into it, and many others supply confirmation of the
reality of text-level features by approaching texts from a completely
different angle, yet arriving in the end at very similar conclusions.
Moreover, there are those authors who have consciously eschewed the
realm of text-linguistic description, even while acknowledging, on
occasion, its value. Waltke and O'Connor's Introduction to Biblical
212
Hebrew Syntax was singled out as a particularly significant example of

this, and was addressed in an excursus; for though they process an
enormous amount of theoretical and otherwise difficult material, for
the benefit of the less well-read user of their book, yet they bluntly
refuse to do this same kind of processing when it comes to textlinguistics.
Since Longacre's insights are so significant, I have introduced some
necessary complications of theoretical linguistics in order to make
these contributions more readily accessible, and to equip the reader to
a small degree to engage in his or her own text-linguistic research
using this model. I have claimed strongly that this theory is a
significantly better starting point for analysis than many others
because it has been immersed, since its inception, in the problems and
intricacies of real language data; few other models have had this kind
of exposure to the world's languages, and fewer still are as innately
responsive to the nuances of the data, and as respectful of the primacy
thereof, as the tagmemic model.
So, at the risk of oversimplification, I have offered a brand of
'tagmemic' theory that will equip the reader to work more effectively
with the contributions of Longacre, Andersen and Khan (et al.), and
in particular, to appreciate the integrity of this kind of scholarship. It
serves, in addition, the purpose of laying the groundwork for the
methodology that I have applied to the texts examined here. The
theoretical groundwork is closely linked with the methodology, which
was presented with explanation of the formats employed for
examining the data. Chapter 3 is the core of this study, and it is hoped
that the reader will thus have available enough theory to do first hand
evaluating of the data.
In the presentation of the methodological principles in this chapter,
I proposed a general working hypothesis, and several 'sub-hypotheses',
which space limitations would permit us to take to the data. These
were:
Main-line clause-types are text-type specific, and can be described as
such; they will predominate in the text, and text-types can be identified by
the predominant clause-type.
The main-line clause-type for the Narrative History text-type is the we +
Prefix clause.
The main-line for the (Narrative) Predictive text-type is the we + Suffix
clause.
213
The main-line clause-type for the Hortatory text-type is built on a

'command' form.
The main-line clause-type for the Expository text-types is the Verbless
clause.
Other text-types, whose clines have not been described nor intimated by
Longacre in Joseph, will be identified first by features other than 'mainline' clause-types (since these have not yet been proposed), and then
clause distribution within those texts for which we have been able to posit
a text-type identity, will be examined with a view toward proposing their
main-line forms.
The constituent structure of texts will be marked by divergences from the
main-line form in all text-types; off-line marking of constituent structure
will be confirmed by other types of marking devices, and will reflect a
comprehensible underlying notional structure.
The results of the above analyses will be expected to hold true for
Reported Speech as well.
I applied these hypotheses first to Judges 2, with the assumption that

the reader would find the results easiest to assimilate; since many have
already undertaken studies of the text-level features of Narrative
History prose, these features will be more familiar to the reader. I
examined this text's clause distribution statistics and proposed that
they fit with our hypothesis that the we + Prefix clause is the main-line
form for this text-type, and that other clauses fit the pattern as off-line
forms. Some sort of text-level function in the distribution of subordinated clauses began to look plausible.
The next text examined was a section of Leviticus 14. In this sample
we found that the we + Suffix form predominated, and that this
appeared to be the main-line form for Procedural-Instructional texts;
off-line clause types were posited as well. We looked briefly at
another text from Leviticus (from chs. 6-7), which encouraged us
with regard to these identifications. Here we addressed the question of
texts being embedded in other texts. Here it was seen that the
Procedural-Instructional material (wrapped, albeit minimally, in
Expository material) was set into a Hortatory framework, which was
itself set into the over-arching Narrative History structure of the
Pentateuch.
We moved from these texts to a comparison of portions of the two
series of pericopes concerning the building of the Tabernacle. These
two were shown to be, on the one hand, Procedural-Instructional
214
with a similar distribution of features to that which we saw in the

Leviticus samplesand on the other, Narrative History, or ProceduralLab Report. The second set was difficult to identify precisely; it
exhibited many of the same features that we had seen in the Judges
material, but there were enough ambiguities to withhold assignment to
a particular text-type identity. Longacre has stated that ProceduralInstructional texts look very similar to Narrative Predictive texts,
sharing the same main-line clause types, and so on. It stands to reason,
then, that Procedural-Lab Report texts might likewise share main-line
clause-types with Narrative History texts. Our data at this stage are
inconclusive.
The Exodus pericopes provided opportunities to try our hand at
identifying the Expository text-types, with their clause distributions,
as well. Longacre's suspicions about the verb ranking which one
would expect to find in these text-types (stative clause-types occupying
the main-line slots, etc.) were found to be good guidelines.
With these data under our belt, so to speak, we were able to add
some more working hypotheses (rather, 'sub-hypotheses') to those we
started with; these were not so much different from the former, as
simply more specific.
Chapter 5 introduced two new texts: the story of Jephthah (Judg.
10.6-12.7), and the book of Ruth. In the first of these, I focused on
the Narrative History text-type. Here, however, the intent was to
examine Niccacci's view that historical texts in Reported Speech are a
different text-type from those in non-Reported Speech. To this end, I
examined first the non-Reported Narrative History text, then compared it with five Narrative History texts found in Reported Speech.
These were found to conform significantly to patterns we had seen in
earlier Narrative History material, with one slight exception: the first
clause in the embedded Narrative History material never took a we +
Prefix form. However, rather than following Niccacci's lead in
defining this as a different type of Narrative, I proposed that the first
clause in any Reported Speech unit always indicates its status as an
element in the (speech introduction) clause in which it is embedded,
and therefore it is never, in terms of surface structure, clause initial.
This allows us to maintain the symmetry of the matrix, and yet
accounts for the distinctive features noticed in the texts so far examined, while remaining consistent with the theoretical base, and
permitting a less cluttered description of the data.
215
Our final engagement with the data involved the book of Ruth,
where we attempted to work through the text with all our tools in
concert. Allowing for the development of hypotheses as we went
along, we examined Ruth, first, in terms of the main-line and off-line
clause-types, subordinated clauses, and embedded speech texts. These
were seen to work remarkably closely with one another to show forth
a 'plot' structure for the book as a whole and for individual episodes
within it.
When our examination of the constituent structure of Ruth was
completed, we turned our attention to analysing the internal structure
of the individual speech units. Each Reported Speech section was
addressed, even where the text was too short, or too complex to
permit conclusions. It was found, howeverthese short, or difficult
passages asidethat what we proposed as a result of our study of the
Narrative History sections in Reported Speech in the Jephthah story
was true as well here. That is to say, once the hypothesis had been
permitted that opening clauses of Reported Speech units reflect their
non-initial status as embedded units, we found no reason to suggest
that text-types encode any differently in Reported Speech than in nonReported Speech.
In the long run, this study has shown that text-types other than
Narrative History have features as particular to themselves as those
which have come to be recognised as features of Narrative History. In
addition, these text-types show a strong preference for a particular
clause-type, and this serves as the backbone of the text. Off-line
clauses are also identifiable, and the twomain-line and off-line
serve with other features to mark the constituent structure (breaks in
the flow, and peaks) of the text; each text-type deploys clause-types in
a characteristic fashion. In addition, apart from requiring that the
embedded material shows its cohesion with the text in which it is
embedded, Reported Speech does not appear to challenge any of these
hypotheses significantly. As a by-product of this research, strong hints
surfaced that subordination plays a role in the constituent structure
marking of Narrative History at least (principally occurring at the
ends of episodes), and that Reported Speech does likewise (though its
distribution leads us to the conclusion that its function is to sustain
tension in the story or to develop it slightly).
216

3. Implications for Progress
I believe that it is in turning to fully productive linguistic models that

progress will be made in assessing and describing features of Classical
Hebrew. What has been presented here is merely a start along the
way. What is needed is solid linguistic study of the language, written
in such a way that it makes a difference to those who use the language
on a day-to-day basis, that is, rabbis and pastors, teachers and
students, and so on.2
I do not want to imply that it is linguistic scientists who should be
doing this work; true, their contributions are singularly welcome
(write in reasonable lucid English, or other modern language, please),
but to exclude linguistically astute hebraists with no formal linguistic
training would be a grave mistake, for these are often the people with
the most intimate understanding of the Hebrew text. Though they lack
formal tools, their knowledge is irreplaceable. What is needed is a
bridge between abstruse scientific theory and the reader who would
like to work on his or her own text-level analyses, that is, something
that allows those who cannotor do not want toundertake study of
linguistic theory in a university setting (and even were they to do so,
what is the likelihood of finding someone to teach grammar discovery
procedures, and the like?), to bring their considerable skills to the text
with a high level of scientific integrity. Hebraists need to be provided
with the tools of the trade, without requiring them to give up the trade
to acquire them. It has been my goal, in this volume, to justify, and to
make a beginning at, this effort; it has also been my goal to show that
this task is not unmanageable, and thateven with a minimum of
good text-linguistic theorymuch can be discovered without undue
mental gymnastics.
This, of course, will be easily recognizable as my idee fixe, my
hobby-horse. The principles have been underlined time and again in
this book: good theorygood methodologygood communication.
Anything less does not constitute effective scholarship.
That said for the last time, what can we say about specific areas of
progress?
I have made much of Longacre's observations. His matrix of texttypes and his verb-rank clines are significant improvements. Other
2.
Not to exclude linguists, and other students of language, of course.
211
scholars are also contributing. Much new material is coming out,

which has the potential of changing radically the way Hebrew is taught
and studied. This will require careful monitoring, of course; much can
seem helpful that is not. The most important measure of a description
is how well it deals with all the language, especially the difficult data.3
Loose ends in this study (and Longacre's as well) include: further
work in identifying text-types and their characteristics (especially
Expository), further analysis of Reported Speech, examination of
more texts with a view toward understanding specific functions of
specific features,4 and so on. Certain other endeavours should very
likely be undertaken only by those with an intimate knowledge of the
language, and at least some awareness of linguistic methodology; these
include: the question of text-linguistic features in poetic material, and
the question of text-linguistic features in formal language. Further to
these, little has been done to examine syntax (micro- or macro-) in
poetic materialO'Connor's Hebrew Verse Structure does this at the
level of micro-syntax, but suffers from the same plague as do
Andersen and Longacre in varying degrees and different waysthat
is, it is so difficult to make sense of that it is off-putting; Watson's
Classical Hebrew Poetry is easier to work with, but includes little real
syntax.5
It may be taken nearly for granted that greater precision in textlinguistic description of Classical Hebrew is to be desired because such
description is of value in and of itself. The relevance of this increase
in knowledge for other facets of biblical studies may not be so
obvious.
3. I am aware that the final part of this work does not measure up well on this
front; yet it does not purport to be a 'description'; rather, it is very brief introduction
to the application of one promising linguistic model. It should be evaluated as such.
4. I must stress, however, that studies like Khan's and Eskhult's will never be
able to give us full answers. For this we need more studies of full texts, looking at
stories, etc., as units, and describing the patterns that emerge as we compare hundreds
of these to each other. Only then will we be able to see whether subordination (for
example) has a specific text-level function, or whether Tn has a strong preference for
episode-initial contexts.
5. Professor J.C.L. Gibson has observed to me that poetry appears to be
governed syntactically as well as 'stylistically' by parallelism (personal communication, 1992); this feature may be a governing factor like the 'embedding/cohesion'
one which I allege controls first clauses in Reported Speech. Such hypotheses as
these will require long and data-intensive study.
218
Text-level analysis can contribute significantly to text exegesis; for

example, the results of my analysis of off-line clauses led me to
propose that Judges 12.5.1 is marked as the peak of the last episode of
the story. It is easily overlooked that our interest in the pronunciation
of the word Shibboleth would not necessarily have been shared by the
earliest hearers of the account; text-linguistic research helps objectify
such elements of the text-to-reader and reader-to-text relationships. In
a different vein, a more thorough examination of the book of Ruth
than I have been able to present in this work would allow us to trace
the themes and purposes of the book, as reflected in the peak marking,
topic continuity and shifts, participant reference, and the deployment
of tension-maintaining devices in the text.
Text-linguistics may also contribute to text-critical discussions;
assessment of emendations according to text-linguistic, in addition to
other, criteria makes enormous sense.
However, it is clear that the most immediateif not the greatest
benefit from text-linguistic research will be for students and teachers
of the language. In the same way that checking the cards in a deck to
determine which (if any) are missing is far easier if the cards are
arranged in numerical order by suits, so also learning (and therefore
teaching) any language is greatly simplified if its forms are systematizedall the more so if it is a dead language.6 If the system of texttypes were presented to students (I do not mean the theoretical
parameters, but rather the simple existence of these text-types), and
their associated main-line forms, then this much, in one stroke, would
give the learner a handle to begin sorting through the various
distributions and functions of the Hebrew verb.7
I have endeavoured to convince the reader of this volume that textlinguistic analysis and description offers significant insights into the
structure of the language and the function of its forms. I have
attempted as well to elucidate one model of text-linguistics that is at
the same time relatively approachable and exceptionally acute in the
results it leads to. This was illustrated by applying the model in seven
6. I am not a supporter of the inductive method of learning a dead language,
since inductive learning requires a deep saturation in that language in order to bring
lasting success, i.e., one must hear it without ceasing, and in all manner of contexts;
this is, of course, denied students of Classical Hebrew, Ugaritic, Sanskrit, Latin, etc.
7. I recognize that this approach alone would not cover the material which a
student needs to cover; however, it is a very good starting point.
219
medium- to large-scale analyses of biblical texts. Although results

from these analyses could not hope to be precise, they do in fact
demonstrate the distinctive features of specific text-types, and confirm
several of the hypotheses that were suggested by the model.
Text-linguistic analysis of Classical Hebrew, and other ancient
languages, is still in its early years; we can look forward to significant
developments in the understanding of our texts as more of this
research takes place.
Appendix 1
THE TEXT OF JUDGES 2 (BHS) IN COLUMNAR FORMAT
2.1.1
2
3
4
5
6
7
2.1 a
b
2
3
4
3.1
2
3
4
4.1 a
b
2
3
5.1
2
6.1
2a
b
7.1
2
Appendix 1. The Text of Judges 2 (BHS) in Columnar Format
221
3
4
5
8.1 a
b
9.1 a
b
10.1
2
3a
b
4
11.1
2
12.1 a
b
2
3
4
5
13.1
2
14.1
2
3
4
15.1
2
3
4
5
16.1
2
17.1
222

2
3
4
5
6
18.1
2
3
4
5
19.1
2a
b
3
20.1
2
21.la
b
20.33
b
20.4
20.5
21.2
21.3
22. la
b
c
2
3
23.1a
b
2
Appendix 2
THE TEXT OF RUTH (BHS) IN COLUMNAR FORMAT
1.1.la
2
3a
b
2.1
2
3
4
5
6
3.1
2
4.1
2
3
4
5.1
2
6.1
2
3
4a
b
7.1
2
3
4
8.1
224

2
3
4
5a
b
9.1
2
3
4
5
10.1
2
3
11.1
2
3
4
5
12.1
2
3
4
5
6
7
13.1
2
3a
b
4
5
Appendix 2. The Text of Ruth (BHS) in Columnar Format
225
6
14.1
2
3
4
15.1
2
3
16.1
2
3
4
5
6
7
8
17.1
2
3
4
5
6
18.1
2
3
19.1
2
3
4
5
226

20.1
2
3
4
21.1
2
3
4
5
22.1
2a
b
3
2.1.la
b
2
2.1
2
3
4
5
6
3.1
2
3
4
5
4.1
2
3
4
227
5
5.1
2
6.1
2
3
4
5
7.1
2
3
4
5
8.1
2
3
4
5
9.1
2
3
4
5
6
7
g
10.1
2
3
4
5
228
11.1
2
3
4a
b
5a
b
6
7
12.1
2a
b
3
13.1
2
3
4
5
14.1
2
3
4
5
6
7
g
9
15.1
2
3
4
16.1
2
3
229
4
17.1
2
3
4
18.1
2
3
4
5
6
19.1
2
3
4
5
6
7
8
9
20.1
2
3a
b
4
5
6
230

21.1
2
3
4
5
6
7
22.1
2
3
4
23. la
b
2
3.1.1
2
3
2.1
2
3
3.1
2
3
4
5
4.1
2
3
231
4
5
6
7
8
5.1
2
3
6.1
2
3
7.1
2
3
4
5
6
7
8.1
2
3
4
9.1
2
3
4
5
6
10.1
2
3a
b
232

c
11.1
2
3
4
5
12.1
2
3
4
13.1
2
3
4
5
6
7
8
9
14.1
2
3
4
5
6
15.1
2
3
4
5
233
6
7
8
16.1
2
3
4
5
17.1
2
3
4
18.1
2
3
4
5
6
4.1.1
2
3
4
5
6
7
8
9
2.1
234

2
3
4
3.1
2a
b
3
4.1
2
3
4
5
6
7
8
9
10
11
12
5.1
2a
b
c
6.1
2
3
4
5
235
7.1 a
b
2
3
4
8.1
2
3
9.1
2
3a
b
4
5a
b
10.1 a
b
2a
b
11.1
2
3
4a
b
5
6
7
12.1
2
3a
236

b
13.1
2
3
4
s
H.I
2
3
4
is.u
2
3
4
16.1
2
3
17.1
2
3
4
is.i
2
19.1
2
20.1
2
21.1
2
22.1
2
BIBLIOGRAPHY
Aejmelaeas, A., 'What Can we Know about the Hebrew Vorlage of the Septuagint?',
ZAW 99 (1987), pp. 58-89.
Alter, R., The An of Biblical Narrative (London: George Alien & Unwin, 1981).
Andersen, F.I., The Hebrew Verbless Clause in the Pentateuch (JBL Monograph Series,
vol. XIV; Nashville: Abingdon, 1970).
The Sentence in Biblical Hebrew (The Hague: Mouton, 1974).
Andersen, F.I., and A.D. Forbes, A Linguistic Concordance of Ruth and Jonah: Hebrew
Vocabulary and Idiom (The Computer Bible, 9; Biblical Research Associates,
1976),
Andersen, F.I,, and D.N. Freedman, Hosea (AB, 24; Garden City, New York:
Doubleday, 1980).
Auld, A.G., Joshua, Judges and Ruth (The Daily Study Bible; Edinburgh: St Andrew
Press; Philadelphia: Westminster Press, 1984),
Bandstra, B.L., 'Word Order and Emphasis in Biblical Hebrew Narrative: Syntactic
Observations on Genesis 22 from a Discourse Perspective [1986]*, in
W.R. Bodine (ed.), Linguistics and Biblical Hebrew.
Bar-Efrat, S., Narrative Art in the Bible (Sheffield: Almond Press, 1989).
Ban, J., Comparative Philology and the Text of the Old Testament (London: SCM Press,
1983).
The Semantics of Biblical Language (London: SCM Press, 1983).
Beattie, D.R.G., Jewish Exegesis of the Book of Ruth (JSOTSup, 2; Sheffield: JSOT
Press, 1977).
Berlin, A., 'Lexical Cohesion and Biblical Interpretation', Hebrew Studies 30 (1989).
'Point of View in Biblical Narrative', in S.A. Gelier (ed.), A Sense of Text (1983).
Bernstein, M.J., 'Two Multivalent Readings in the Ruth Narrative', JSOT 50 (1991),
Bodine, W.R, (ed.), Linguistics and Biblical Hebrew (Winona Lake, IN: Eisenbrauns,
1992).
'How Linguists Study Syntax [1987]', in W.R, Bodine (ed.), Linguistics and Biblical
Hebrew.
'The Study of Linguistics and Biblical Hebrew1, in W.R. Bodine (ed,), Linguistics
and Biblical Hebrew.
Boling, R.G., Judges (AB, 6; Garden City, NY: Doubleday, 1975).
Brend, R.M. (ed,), Kenneth L. Pike: Selected Writings (Janua Linguarum, Series Major,
55; The Hague: Mouton).
Brend, R.M. and K.L. Pike (eds.), Tagmemics. I. Aspects of the Field (Trends in
Linguistics, Studies and Monographs, 1; The Hague: Mouton, 1976).
Brenner, A., 'A Triangle and a Rhombus in Narrative Structure: A Proposed Integrative
Reading of Judges IV and V, VT 40.2 (1990), pp. 129-38.
238
Brownlee, W.H., Ezekiel 1-19 (WBC, 28; Waco, TX: Word Books, 1986).
Burney, C.F., The Book of Judges (London: Rivingtons, 1918).
Buth, R.J., Word Order in Aramaic from the Perspective of Functional Grammar and
Discourse Analysis (PhD Dissertation, UCLA: United Microfilms International,
1987).
Calloud, J., Structural Analysis of Narrative (trans. D. Patte; Philadelphia: Fortress Press;
Missoula, MT: Scholars Press, 1976).
Campbell, E.F., Ruth (AB, 7; Garden City, NY: Doubleday, 2nd edn, 1975)
Cook, E.M., Word Order in the Aramaic of Daniel (Monographic Journals of the Near
East: Afroasiatic Linguistics, 9, issue 3 [Dec. 1986]; Malibu, CA: Undena
Publications).
Cotterell, P., and M. Turner, Linguistics and Biblical Interpretation (London: SPCK,
1989).
Dawson, D.A., 'Text Linguistics and Biblical Hebrew: An Examination of
Methodologies' (PhD dissertation, University of Edinburgh, 1993).
Dearman, A. (ed.), Studies in the Mesha' Inscription and Moab (Archeology and
Biblical Studies, 2; Atlanta: Scholars Press, 1989).
Eisenman, R., and M. Wise, The Dead Sea Scrolls Uncovered (Rockport, MA: Element,
1992).
Elson, B.F., and V.P. Pickett., An Introduction to Morphology and Syntax (Santa Ana,
CA: Summer Institute of Linguistics, 2nd edn, 1964).
Eskhult, M., Studies in Verbal Aspect and Narrative Technique in Biblical Hebrew
Prose (Uppsala: Uppsala University, 1990).
Exter Blokland, A.F. den, 'Clause Analysis in Biblical Hebrew NarrativeAn
Explanation and a Manual for Compilation', Trinity Journal 11 NS (1990),
pp.73-102.
Fewell, D.N., and D.M. Gunn, Compromising Redemption: Relating Characters in the
Book of Ruth (Literary Currents in Biblical Interpretation; Louisville, KY: John
Knox Press, 1990).
Fokkelman, J.P., Narrative Art in Genesis: Specimens of Stylistic and Structural
Analysis (Assen: Van Gorcum, 1975).
Freedman, D.N., and K.A. Matthews, The Paleo-Leviticus Scroll (Cambridge, MA:
American Schools of Oriental Research, 1985).
Fuchs, E., 'Marginalization, Ambiguity, Silencing: The Story of Jephthah's Daughter',
Journal of Feminist Studies in Religion 5 (1989), pp. 35-45.
Garr, W.R., 'The Qinah: A Study of Poetic Meter, Syntax and Style', ZAW 95 (1983),
pp.54-75.
Geller, S.A., A Sense of Text: The Art of Language in the Study of Biblical Literature
(JQRMS; Philadelphia: Jewish Quarterly Review, 1983).
'Through Windows and Mirrors into the Bible. History, Literature, and Language in
the Study of the Text', in S.A. Geller (ed.), A Sense of Text (1983).
Gerleman, G., Ruth /Das Hohelied (BKAT, 18; Neukirchen-Vluyn: Neukirchener
Verlag, 1965).
Grant, R., 'Literary Structure in Ruth', BibSac, October-December, 1991, pp. 424-41.
Gray, J., Joshua, Judges and Ruth (The Century Bible; London: Nelson, 1967).
Greenstein, E., 'On the Prefixed Preterite in Biblical Hebrew', Hebrew Studies 29
(1988), pp. 7-17.
Grimes, B. (ed.), Ethnologue (Dallas, Texas: Summer Institute of Linguistics, 1984).
Bibliography
239
Hals, R.M., Ezekiel (FOTL, 19; Grand Rapids: Eerdmans, 1989).

Hoftijzer, J., A Search for Method: A Study in the Syntactic Use of the H-Locale in
Classical Hebrew (Studies in Semitic Languages and Linguistics, 12; Leiden: Brill,
1981).
Hopper, P.J. and S.A. Thompson, 'Transitivity in Grammar and Discourse', Language
56 (1980), pp. 251-99.
Hospers, J.H., 'Polysemy and Homonymy', ZAW 6/1 (1993), pp. 114-23.
Huenergard, J., 'The Early Hebrew Prefix Conjugations', Hebrew Studies 29 (1988).
Hubbard, D.L., 'Ruth IV 17: A New Solution', VT 38 (1988), pp. 19-23.
Hurvitz, A., 'Ruth 2.7"A Midrashic Gloss"?', ZAW 95 (1983), pp. 293-301.
Isaksson, B., Studies in the Language of Qoheleth (Acta Universitatis Upsaliensis:
Studia Semitica Upsaliensia, 10; Uppsala: Uppsala University, 1987).
Jagersma, H., Bijbels Hebreeuws: Basiscursus (Kampen: Uitgeversmaatschappij
J.H. Kok, 1991).
Jenni, E., Lehrbuch der hebrdischen Sprache des alien Testaments (Basel: Helbing &
Lichtenhahn, 2nd edn, 1981).
Johnson, B., Hebrdisches Perfekt und Imperfekt mit vorangehendem we (ConBOT, 13;
Malmo: CWK Gleerup, 1979).
Jones, L.K., 'A Synopsis of Tagmemics', Syntax and Semantics (Current Approaches
to Syntax, 13, 1980).
Joosten, J., 'Biblical Hebrew weqatal and Syriac hwa qatel Expressing Repetition in the
Past', ZAH 3:1 (1992).
Khan, G., Studies in Semitic Syntax (Oxford: Oxford University Press, 1988).
Komer, J., Hebraische Studiengrammatik (Leipzig: Verlag Enzyklopadie, 1988).
Lambdin, T.O., Introduction to Biblical Hebrew (London: Darton, Longman & Todd,
1971).
Levine, E., The Aramaic Version of Ruth (AnBib, 58; Rome: Biblical Institute Press,
1973).
Lilley, J.P.U., 'A Literary Approach to the Book of Judges', TynBul 15 (1967),
pp. 94-102.
Longacre, R.E., An Apparatus for the Identification of Paragraph Types (Notes on
Linguistics, 15; Dallas: Summer Institute of Linguistics, 1980).
'Discourse', in R.M. Brend and K.L. Pike (eds.), Tagmemics.
Discourse, Paragraph, and Sentence Structure in Selected Philippine Languages. I.
Discourse and Paragraph Structure (SIL Publications in Linguistics and Related
Fields, 21; Santa Ana, CA: Summer Institute of Linguistics, 1968).
'Discourse Perspective on the Hebrew Verb: Affirmation and Restatement [1987]', in
Grammar Discovery Procedures (Janua Linguarum, Series Minor, 33; The Hague:
Mouton, 1964).
The Grammar of Discourse (Topics in Language and Linguistics; New York: Plenum
Press, 1983).
Hierarchy and Universality of Discourse Constituents in New Guinea Languages
(2 vols.; Washington, DC: Georgetown University Press, 1972).
Joseph: A Story of Divine Providence: A Text Theoretical and Text-Linguistic
Analysis of Genesis 37 and 3948 (Winona Lake, IN: Eisenbrauns, 1989).
'Tagmemics', Word 36.2 (1985), pp. 137-77.
240
(ed.), Theory and Application in Processing Texts in Non-Indoeuropean Languages

(Papiere zur Textlinguistik, 43; Hamburg: Helmut Buske Verlag, 1984).
Louw, J.P., 'The Function of Discourse in a Sociosemiotic Theory of Translation:
Illustrated by the Translation of zeloute in 1 Corinthians 12.31', The Bible
Translator 39 (1988), pp. 329-35.
MacDonald, P.J., 'Discourse Analysis and Biblical Interpretation [1986]', in
MacGregor, L.J., The Greek Text of Ezekiel: An Examination of its Homogeneity
(Septuagint and Cognate Studies, 18; Atlanta: Scholar's Press, 1985).
Magonet, J., Form and Meaning: Studies in Literary Techniques in the Book of Jonah
(Sheffield: Almond Press, 1983).
Martin, J.D., The Book of Judges (CBC; Cambridge: Cambridge University Press,
1975).
McFall, L., The Enigma of the Hebrew Verbal System (Sheffield: Almond Press, 1982).
Meier, S.A., Speaking of Speaking: Marking Direct Discourse in the Hebrew Bible
(VTSup, 46; Leiden: Brill, 1992).
Meyer, D.R., Hebrdische Grammatik. I. Einleitung, Schrift- und Lautlehre. II.
Formenlehre, Flexionstabellen. III. Satzlehre. IV. Register, Hebrdisches Textbuch
(Berlin: Walter de Gruyter, 1966).
Milne, P.J., 'The Patriarchal Stamp of Scripture: The Implications of Structuralist
Analyses for Feminist Hermeneutics', Journal of Feminist Studies in Religion 5
(1989), pp. 17-34.
Moor, J. de, 'The Poetry of the Book of Ruth: Part I', Or 53 (1984), pp. 262-83; Part
II, Or 55 (1986), pp. 16-46.
Muraoka, T., Emphatic Words and Structures in Biblical Hebrew (Leiden: Brill, 1985).
Murphy, R.E., and O. Carm, Wisdom Literature: Job, Proverbs, Ruth, Canticles,
Ecclesiastes, and Esther (FOTL, 13; Grand Rapids: Eerdmans, 1981).
Myers, J.M., The Linguistic and Literary Form of the Book of Ruth (Leiden: Brill,
1955).
Niccacci, A., The Syntax of the Verb in Classical Hebrew Prose (trans. W.G.E. Watson;
JSOTSup, 86; Sheffield: JSOT Press, 1990).
Noth, M., Leviticus: A Commentary (OTL; London: SCM Press, 1977).
Nyssonen, H., 'Steps to an Analysis of Written Discourse', in H. Nyssonen (ed.), Issues.
(ed.), Issues in Discourse Analysis, Contrastive Phonetics and Literary
Communication (Publications of the Department of English, University of Oulu,
3; Oulu, Finland: University of Oulu Press, 1983).
Nyssonen, H., et al. (eds.), Bibliography of Discourse Analysis and Text Linguistics
(Publications of the Department of English, University of Oulu, 6; Oulu, Finland:
University of Oulu Press, 1985).
Proceedings from the Second Finnish Seminar on Discourse Analysis, Oulu,
September 27-28, 1988 (Publications of the Department of English, University of
Oulu, 9; Oulu, Finland: University of Oulu Press, 1990).
O'Connor, M., Hebrew Verse Structure (Winona Lake, IN: Eisenbrauns, 1980).
Payne, D., 'Activity as the Encoding of Foreground in Narrative: A Case Study of an
Asheninca Legend', in R.E. Longacre (ed.), Theory.
Pike, K.L., Language in Relation to a Unified Theory of the Structure of Human
Behavior (The Hague: Mouton, 2nd edn, 1967).
Bibliography
241
Linguistic Concepts: An Introduction to Tagmemics (Lincoln, NB: University of

Nebraska Press, 1982).
Pike, K.L., and E.G. Pike, Grammatical Analysis (Dallas, TX: SIL and University of
Texas at Arlington Press, 1977).
Rainey, A.F., 'Further Remarks on the Hebrew Verbal System', Hebrew Studies 29
(1988), pp. 35-42.
Revell, E.J., 'The System of the Verb in Standard Biblical Prose', HUCA 60 (1989),
pp.1-37.
Schiffrin, D., 'Tense Variation in Narrative', Language 57 (1981).
Schneider, W., Obungsbuch fur den Hebrdische-unterricht (Munich: Claudius Verlag,
1989).
Schoors, A., 'The Pronouns in Qoheleth', Hebrew Studies 30 (1989), pp. 71-90.
Schwartzschild, R., 'The Syntax of ifi in Biblical Hebrew with Special Reference to
Qoheleth', Hebrew Studies 31 (1990), pp. 7-39.
Sihvonen-Hautecoeur, P., 'Discourse Markers and Dialogue in French', in
H. Nyssonen, et al. (eds.), Proceedings, pp. 52-63.
Simpson, C.A., Composition of the Book of Judges (Oxford: Basil Blackwell, 1957).
Smith, M.S., 'The Wow-Consecutive at Qumran', ZAW 4 (1991), pp. 161-64.
Waard, J. de, and E.A. Nida, A Translator's Handbook on the Book of Ruth (New York:
United Bible Societies, 1973).
Waldman, N.M., The Recent Study of Hebrew: A Survey of the Literature with Selected
Bibliography (Bibliographica Judaica, 10; Cincinnati: Hebrew Union College
Press; Winona Lake, IN: Eisenbrauns, 1989).
Waltke, B.E., and M. O'Connor, An Introduction to Biblical Hebrew Syntax (Winona
Lake, IN: Eisenbrauns, 1990).
Waterhouse, V.G., The History and Development of Tagmemics (Janua Linguarum
Series Critica, 16; The Hague: Mouton, 1974).
Watson, W.G.E., Classical Hebrew Poetry: A Guide to its Techniques (JSOTSup, 26;
Sheffield: JSOT Press, 2nd edn, 1986).
Webb, B.C., The Book of Judges: An Integrated Reading (JSOTSup, 46; Sheffield:
JSOT Press, 1987).
Wechter, P., Ibn Barun's Arabic Works on Hebrew Grammar and Lexicography
(Philadelphia: The Dropsie College for Hebrew and Cognate Learning, 1964).
Wickes, W., A Treatise on the Accentuation of the Twenty-One So-Called Prose Books
of the Old Testament (1887).
Wiklander, B., Prophecy as Literature: A Text-Linguistic and Rhetorical Approach to
Isaiah 2-4 (ConBOT 22; Malmo: CWK Gleerup, 1984).
Williams, J.G., 'The Structure of Judges 2.1-16.31', JSOT 49 (1991), pp. 77-85.
Wise, M.R., 'Language and Behavior', in R.M. Brend and K.L. Pike (eds.), Tagmemics.
Witzenrath, H.H., Das Buch Rut (SANT; Munich: Kosel-Verlag, 1975).
Wright, C.H.H. (ed.), The Book of Ruth in Hebrew and Chaldee (London: Williams &
Norgate, 1864).
Zevit, Z., Talking in Biblical Henglish [sic] and Solving a Problem of the YAQTUL
Past Tense' Hebrew Studies 29 (1988), pp. 25-33.
Zuber, B., Das Tempussystem des biblischen Hebrdisch (Berlin: de Gruyter, 1986).
INDEX OF AUTHORS
Anderson, F.I. 13, 23, 46-51, 57, 64,
81, 209-12, 217
Doling, R.G. 161
Bellinger, D. 13
Brend, R.M. 96
Campbell, E.F. 184, 192, 193
Crystal, D. 13, 14, 17, 18, 30, 79, 80
Martin, J.D. 161

Meier, S.A. 193
Muraoka, T. 71
Niccacci, A. 13, 16, 23, 28-39, 42, 43,
45,46,51,54,55,63, 114, 121,
125, 126, 139, 154, 155, 166,
171, 174, 175, 209-11, 214
Dawson, D.A. 137
O'Connor, M. 24-28, 34, 41, 49, 211,

217
Elson, B.F. 85
Eskhult, M. 13, 23, 24, 40-46, 51, 55,
107, 114, 209,217
Exter Blokland, A.F. von 43
Payne, D. 102, 103

Pickett, V.P. 85
Pike, E.G. 74, 86
Pike, K.L. 57, 74, 86, 96
Gibson, J.C.L. 217
Schneider, W. 26
Jones, L.B. 71
Jones, L.K. 11, 71, 73, 74, 80, 81, 86
Waltke, B.E. 24-28, 34, 49, 211

Waterhouse, V.G. 74, 81, 85, 97
Watson, W.G.E. 217
Webb, E.G. 161, 162
Khan, G. 12, 17, 23, 24, 47, 52-56, 59,

69, 107,209,210,211, 217
Longacre, R.E. 11, 19, 23, 24, 44, 47,
52, 53, 56-74, 77-82, 85, 87, 88,
90-92, 94-97, 100-102, 105, 106,
114-18, 120, 126, 130, 144, 14650, 152, 170, 171, 188, 189, 191,
209-13, 216
JOURNAL FOR THE STUDY OF THE OLD TESTAMENT

Supplement Series
1
I, HE, WE AND THEY:

A LITERARY APPROACH TO ISAIAH 53
D.J.A. Clines
JEWISH EXEGESIS OF THE BOOK OF RUTH
THE LITERARY STRUCTURE OF PSALM 2
D.R.G. Beattie
P. Auffret
4
THANKSGIVING FOR A LIBERATED PROPHET:

AN INTERPRETATION OF ISAIAH CHAPTER 53
R.N. Whybray
REDATING THE EXODUS AND CONQUEST

J.J. Bimson
THE STORY OF KING DAVID:

GENRE AND INTERPRETATION
THE SENSE OF BIBLICAL NARRATIVE I:

STRUCTURAL ANALYSES IN THE HEBREW BIBLE (2nd edition)
D.M. Gunn
D. Jobling
8
GENESIS l-ll:
STUDIES IN STRUCTURE AND THEME
P.D. Miller
YAHWEH AS PROSECUTOR AND JUDGE:

AN INVESTIGATION OF THE PROPHETIC LAWSUIT (Rw PATTERN)
K. Nielsen
10
THE THEME OF THE PENTATEUCH

David J.A. Clines
11
STUDIA BIBLICA 19781:

PAPERS ON OLD TESTAMENT AND RELATED THEMES
Edited by E.A. Livingstone
13
ISAIAH AND THE DELIVERANCE OF JERUSALEM:

A STUDY OF THE INTERPRETATION OF PROPHECY
IN THE OLD TESTAMENT
R.E. Clements
14
THE FATE OF KING SAUL:

AN INTERPRETATION OF A BIBLICAL STORY
D.M. Gunn
15
THE DEUTERONOMISTIC HISTORY

Martin Noth
16
PROPHECY AND ETHICS:

ISAIAH AND THE ETHICAL TRADITION OF ISRAEL
Eryl W. Davies
17
THE ROLES OF ISRAEL'S PROPHETS
18
THE DOUBLE REDACTION OF THE DEUTERONOMISTIC HISTORY
David L. Petersen
Richard D. Nelson
19
ART AND MEANING:

RHETORIC IN BIBLICAL LITERATURE
Edited by David J.A. Clines, David M. Gunn & Alan J. Hauser
20
THE PSALMS OF THE SONS OF KORAH

Michael D. Goulder
21
COLOUR TERMS IN THE OLD TESTAMENT

Athalya Brenner
22
AT THE MOUNTAIN OF GOD:

STORY AND THEOLOGY IN EXODUS 32-34
23
THE GLORY OF ISRAEL:

THE THEOLOGY AND PROVENIENCE OF THE ISAIAH TARGUM
R.W.L. Moberly
Bruce D. Chilton
24
MIDIAN, MOAB AND EDOM:

THE HISTORY AND ARCHAEOLOGY OF LATE BRONZE AND
IRON AGE JORDAN AND NORTH-WEST ARABIA
Edited by John F.A. Sawyer & David J.A. Clines
25
THE DAMASCUS COVENANT:

AN INTERPRETATION OF THE 'DAMASCUS DOCUMENT'
26
CLASSICAL HEBREW POETRY:

A GUIDE TO ITS TECHNIQUES
Philip R. Davies
Wilfred G.E. Watson
27
PSALMODY AND PROPHECY

W.H. Bellinger, Jr
28
HOSEA:
AN ISRAELITE PROPHET IN JUDEAN PERSPECTIVE

Grace I. Emmerson
29
EXEGESIS AT QUMRAN:
4QFLORILEGIUM IN ITS JEWISH CONTEXT
30
THE ESTHER SCROLL:

THE STORY OF THE STORY
George J. Brooke
David J.A. Clines
31
[N THE SHELTER OF EL YON:

ESSAYS IN HONOR OF G.W. AHLSTROM
Edited by W. Boyd Barrick & John R. Spencer
32
THE PROPHETIC PERSONA:

IEREMIAH AND THE LANGUAGE OF THE SELF
Timothy Polk
33
LAW AND THEOLOGY IN DEUTERONOMY

J.G. McConville
34
THE TEMPLE SCROLL:

AN INTRODUCTION, TRANSLATION & COMMENTARY
Johann Maier
35
SAGA, LEGEND, TALE, NOVELLA, FABLE:

NARRATIVE FORMS IN OLD TESTAMENT LITERATURE
36
THE SONG OF FOURTEEN SONGS
Edited by George W. Coats

Michael D. Goulder
37
UNDERSTANDING THE WORD:

ESSAYS IN HONOR OF BERNHARD W. ANDERSON
Edited by James T. Butler, Edgar W. Conrad & Ben C. Ollenburger
38
SLEEP, DIVINE AND HUMAN, IN THE OLD TESTAMENT

Thomas H. McAlpine
39
THE SENSE OF BIBLICAL NARRATIVE II:

STRUCTURAL ANALYSES IN THE HEBREW BIBLE
David Jobling
40
DIRECTIONS IN BIBLICAL HEBREW POETRY

Edited by Elaine R. Follis
41
ZION, THE CITY OF THE GREAT KING:
A THEOLOGICAL SYMBOL OF THE JERUSALEM CULT

Ben C. Ollenburger
42
A WORD IN SEASON:
ESSAYS IN HONOUR OF WILLIAM MCKANE
Edited by James D. Martin & Philip R. Davies
43
THE CULT OF MOLEK:

A REASSESSMENT
G.C. Heider
44
THE IDENTITY OF THE INDIVIDUAL IN THE PSALMS

Steven J.L. Croft
45
THE CONFESSIONS OF JEREMIAH IN CONTEXT:

SCENES OF PROPHETIC DRAMA
46
THE BOOK OF JUDGES:

AN INTEGRATED READING
47
THE GREEK TEXT OF JEREMIAH:

A REVISED HYPOTHESIS
48
TEXT AND CONTEXT:

OLD TESTAMENT AND SEMITIC STUDIES FOR F.C. FENSHAM
A.R. Diamond
Barry G. Webb
Sven Soderlund
Edited by W. Claassen
49
THEOPHORIC PERSONAL NAMES IN ANCIENT HEBREW:

A COMPARATIVE STUDY
Jeaneane D. Fowler
50
THE CHRONICLER'S HISTORY

Martin Noth
Translated by H.G.M. Williamson with an Introduction
51
DIVINE INITIATIVE AND HUMAN RESPONSE IN EZEKIEL

Paul Joyce
52
THE CONFLICT OF FAITH AND EXPERIENCE IN THE PSALMS:

A FORM-CRITICAL AND THEOLOGICAL STUDY
53
THE MAKING OF THE PENTATEUCH:

A METHODOLOGICAL STUDY
Craig C. Broyles
R.N. Whybray
54
FROM REPENTANCE TO REDEMPTION:

JEREMIAH'S THOUGHT IN TRANSITION
Jeremiah Unterman
55
THE ORIGIN TRADITION OF ANCIENT ISRAEL:

1. THE LITERARY FORMATION OF GENESIS AND EXODUS 1-23
T.L. Thompson
56
THE PURIFICATION OFFERING IN THE PRIESTLY LITERATURE:

ITS MEANING AND FUNCTION
N. Kiuchi
57
MOSES :
HEROIC MAN, MAN OF GOD

George W. Coats
58
THE LISTENING HEART:

ESSAYS IN WISDOM AND THE PSALMS
IN HONOR OF ROLAND E. MURPHY, O. CARM.
Edited by Kenneth G. Hoglund, Elizabeth F. Huwiler, Jonathan T. Glass
and Roger W. Lee
59
CREATIVE BIBLICAL EXEGESIS:

CHRISTIAN AND JEWISH HERMENEUTICS THROUGH THE CENTURIES
Edited by Benjamin Uffenheimer & Henning Graf Reventlow
60
HER PRICE is BEYOND RUBIES:

THE JEWISH WOMAN EN GRAECO-ROMAN PALESTINE
61
FROM CHAOS TO RESTORATION:

AN INTEGRATTVE READING OF ISAIAH 24-27
Leonie J. Archer
Dan G. Johnson
62
THE OLD TESTAMENT AND FOLKLORE STUDY

Patricia G. Kirkpatrick
63
SHILOH:
A BIBLICAL CITY IN TRADITION AND HISTORY
Donald G. Schley
64
To SEE AND NOT PERCEIVE:

ISAIAH 6.9-10 IN EARLY JEWISH AND CHRISTIAN INTERPRETATION
Craig A. Evans
65
THERE is HOPE FOR A TREE: THE TREE AS METAPHOR IN ISAIAH

Kirsten Nielsen
66
SECRETS OF THE TIMES:

MYTH AND HISTORY IN BIBLICAL CHRONOLOGY
Jeremy Hughes
67
ASCRIBE TO THE LORD:

BIBLICAL AND OTHER STUDIES IN MEMORY OF PETER C. CRAIGIE
Edited by Lyle Eslinger & Glen Taylor
68
THE TRIUMPH OF IRONY IN THE BOOK OF JUDGES

Lillian R. Klein
69
ZEPHANIAH, A PROPHETIC DRAMA
70
NARRATIVE ART IN THE BIBLE
Paul R. House
Shimon Bar-Efrat
71
QOHELET AND HIS CONTRADICTIONS
Michael V. Fox
72
CIRCLE OF SOVEREIGNTY:
A STORY OF STORIES IN DANIEL 1-6
Danna Nolan Fewell
73
DAVID'S SOCIAL DRAMA:

A HOLOGRAM OF THE EARLY IRON AGE
James W. Flanagan
74
THE STRUCTURAL ANALYSIS OF BIBLICAL AND CANAANITE POETRY

Edited by Willem van der Meer & Johannes C. de Moor
75
DAVID IN LOVE AND WAR:

THE PURSUIT OF POWER IN 2 SAMUEL 10-12
Randall C. Bailey
76
GOD is KING:
UNDERSTANDING AN ISRAELITE METAPHOR
Marc Zvi Brettler
77
EDOM AND THE EDOMITES

John R. Bartlett
78
79
SWALLOWING THE SCROLL:

TEXTUALITY AND THE DYNAMICS OF DISCOURSE
IN EZEKIEL'S PROPHECY
Ellen F. Davies
GIBEAH:
THE SEARCH FOR A BIBLICAL CITY

Patrick M. Arnold, S.J.
80
THE NATHAN NARRATIVES

Gwilym H. Jones
81
ANTI-COVENANT:
COUNTER-READING WOMEN'S LIVES IN THE HEBREW BIBLE
Edited by Mieke Bal
82
RHETORIC AND BIBLICAL INTERPRETATION

Dale Patrick & Allen Scult
83
THE EARTH AND THE WATERS IN GENESIS 1 AND 2:

A LINGUISTIC INVESTIGATION
David Toshio Tsumura
84
INTO THE HANDS OF THE LIVING GOD
85
FROM CARMEL TO HOREB:

ELIJAH IN CRISIS
Lyle Eslinger
Alan J. Hauser & Russell Gregory
86
THE SYNTAX OF THE VERB IN CLASSICAL HEBREW PROSE

Alviero Niccacci
Translated by W.G.E. Watson
87
THE BIBLE IN THREE DIMENSIONS:

ESSAYS IN CELEBRATION OF FORTY YEARS OF BIBLICAL STUDIES
IN THE UNIVERSITY OF SHEFFIELD
Edited by David J.A. Clines, Stephen E. Fowl & Stanley E. Porter
88
THE PERSUASIVE APPEAL OF THE CHRONICLER:

A RHETORICAL ANALYSIS
Rodney K. Duke
89
THE PROBLEM OF THE PROCESS OF TRANSMISSION

IN THE PENTATEUCH
RolfRendtorff
Translated by John J. Scullion
90
BIBLICAL HEBREW IN TRANSITION:

THE LANGUAGE OF THE BOOK OF EZEKIEL
Mark F. Rooker
91
THE IDEOLOGY OF RITUAL:

SPACE, TIME AND STATUS IN THE PRIESTLY THEOLOGY
Frank H. Gorman, Jr
92
ON HUMOUR AND THE COMIC IN THE HEBREW BIBLE

Edited by Yehuda T. Radday & Athalya Brenner
93
JOSHUA 24 AS POETIC NARRATIVE

William T. Koopmans
94
WHAT DOES EVE Do TO HELP? AND OTHER READERLY QUESTIONS

TO THE OLD TESTAMENT
David J.A. Clines
95
GOD SAVES:
LESSONS FROM THE ELISHA STORIES
Rick Dale Moore
96
ANNOUNCEMENTS OF PLOT IN GENESIS

Laurence A. Turner
97
THE UNITY OF THE TWELVE

Paul R. House
98
ANCIENT CONQUEST ACCOUNTS:

A STUDY IN ANCIENT NEAR EASTERN AND BIBLICAL HISTORY WRITING
99
WEALTH AND POVERTY IN THE BOOK OF PROVERBS
K. Lawson Younger, Jr
R.N. Whybray
100
A TRIBUTE TO GEZA VERMES:

ESSAYS ON JEWISH AND CHRISTIAN
LITERATURE AND HISTORY
Edited by Philip R. Davies & Richard T. White
101
THE CHRONICLER IN HIS AGE

Peter R. Ackroyd
102
THE PRAYERS OF DAVID (PSALMS 51-72):

STUDIES IN THE PSALTER, II
Michael Goulder
103
THE SOCIOLOGY OF POTTERY IN ANCIENT PALESTINE:

THE CERAMIC INDUSTRY AND THE DIFFUSION OF CERAMIC STYLE
IN THE BRONZE AND IRON AGES
Bryant G. Wood
104
PSALM STRUCTURES:
A STUDY OF PSALMS WITH REFRAINS
105
RE-ESTABLISHING JUSTICE
Paul R. Raabe
Pietro Bovati
106
GRADED HOLINESS:
A KEY TO THE PRIESTLY CONCEPTION OF THE WORLD
107
THE ALIEN IN ISRAELITE LAW
Philip Jenson
Christiana van Houten
108
THE FORGING OF ISRAEL:

IRON TECHNOLOGY, SYMBOLISM AND TRADITION IN ANCIENT SOCIETY
Paula M. McNutt
109
SCRIBES AND SCHOOLS IN MONARCHIC JUDAH:

A SOCIO-ARCHAEOLOGICAL APPROACH
David Jamieson-Drake
110
THE CANAANITES AND THEIR LAND:

THE TRADITION OF THE CANAANITES
Niels Peter Lemche
111
YAHWEH AND THE SUN:

THE BIBLICAL AND ARCHAEOLOGICAL EVIDENCE
J. Glen Taylor
112
WISDOM IN REVOLT:
METAPHORICAL THEOLOGY IN THE BOOK OF JOB

Leo G. Perdue
113
PROPERTY AND THE FAMILY IN BIBLICAL LAW

Raymond Westbrook
114
A TRADITIONAL QUEST:
ESSAYS IN HONOUR OF Louis JACOBS
Edited by Dan Cohn-Sherbok
115
I HAVE BUILT You AN EXALTED HOUSE:

TEMPLE BUILDING IN THE BIBLE IN LIGHT OF MESOPOTAMIAN
AND NORTHWEST SEMITIC WRITINGS
Victor Hurowitz
116
NARRATIVE AND NOVELLA IN SAMUEL:

STUDIES BY HUGO GRESSMANN AND OTHER SCHOLARS 1906-1923
Translated by David E. Orton
Edited by David M. Gunn
117
SECOND TEMPLE STUDIES:

I.PERSIAN PERIOD
Edited by Philip R. Davies
118
SEEING AND HEARING GOD WITH THE PSALMS:

THE PROPHETIC LITURGY FROM THE SECOND TEMPLE IN JERUSALEM
Raymond Jacques Tournay
Translated by J. Edward Crowley
119
TELLING QUEEN MICHAL'S STORY:

AN EXPERIMENT IN COMPARATIVE INTERPRETATION
Edited by David J.A. Clines & Tamara C. Eskenazi
120
THE REFORMING KINGS:

CULT AND SOCIETY IN FIRST TEMPLE JUDAH
Richard H. Lowery
121
KING SAUL IN THE HISTORIOGRAPHY OF JUDAH

Diana Vikander Edelman
122
IMAGES OF EMPIRE
Edited by Loveday Alexander
123
JUDAHITE BURIAL PRACTICES AND BELIEFS ABOUT THE DEAD

Elizabeth Bloch-Smith
124
LAW AND IDEOLOGY IN MONARCHIC ISRAEL

Edited by Baruch Halpern and Deborah W. Hobson
125
PRIESTHOOD AND CULT IN ANCIENT ISRAEL

Edited by Gary A. Anderson and Saul M. Olyan
126
W.M.L.DE WETTE, FOUNDER OF MODERN BIBLICAL CRITICISM:

AN INTELLECTUAL BIOGRAPHY
John W. Rogerson
127
THE FABRIC OF HISTORY:

TEXT, ARTIFACT AND ISRAEL'S PAST
Edited by Diana Vikander Edelman
128
BIBLICAL SOUND AND SENSE:

POETIC SOUND PATTERNS IN PROVERBS 10-29
Thomas P. McCreesh
129
130
131
132
133
135
136
137
138
139
140
141
142
THE ARAMAIC OF DANIEL IN THE LIGHT OF OLD ARAMAIC

Zdravko Stefanovic
STRUCTURE AND THE BOOK OF ZECHARIAH
Michael Butterworth
FORMS OF DEFORMITY :
A MOTIF-INDEX OF ABNORMALITIES, DEFORMITIES AND DISABILITIES
IN TRADITIONAL JEWISH LITERATURE
Lynn Holden
CONTEXTS FOR AMOS :
PROPHETIC POETICS IN LATIN AMERICAN PERSPECTIVE
Mark Daniel Carroll R.
THE FORSAKEN FIRSTBORN:
A STUDY OF A RECURRENT MOTIF IN THE PATRIARCHAL NARRATIVES
Roger Syren
ISRAEL IN EGYPT:
A READING OF EXODUS 1-2
G.F. Davies
A WALK THROUGH THE GARDEN:
BIBLICAL, ICONOGRAPHICAL AND LITERARY IMAGES OF EDEN
Edited by P. Morris and D. Sawyer
JUSTICE AND RIGHTEOUSNESS :
BIBLICAL THEMES AND THEIR INFLUENCE
Edited by H. Graf Reventlow & Y. Hoffman
TEXT AS PRETEXT:
ESSAYS IN HONOUR OF ROBERT DAVIDSON
Edited by R.P. Carroll
PSALM AND STORY:
INSET HYMNS IN HEBREW NARRATIVE
J.W. Watts
PURITY AND MONOTHEISM:
CLEAN AND UNCLEAN ANIMALS IN BIBLICAL LAW
Walter Houston
DEBT SLAVERY IN ISRAEL AND THE ANCIENT NEAR EAST
Gregory C. Chirichigno
DIVINATION IN ANCIENT ISRAEL AND ITS NEAR EASTERN ENVIRONMENT:

A SOCIO-HlSTORICAL INVESTIGATION
Frederick H. Cryer
143
144
145
THE NEW LITERARY CRITICISM AND THE HEBREW BIBLE

David J.A. Clines and J. Cheryl Exum
LANGUAGE, IMAGERY AND STRUCTURE IN THE PROPHETIC WRITINGS
Philip R. Davies and David J.A. Clines
THE SPEECHES OF MICAH:

A RHETORICAL-HISTORICAL ANALYSIS
Charles S. Shaw
146
THE HISTORY OF ANCIENT PALESTINE FROM THE PALAEOLITHIC PERIOD

TO ALEXANDER'S CONQUEST
Gosta W. Ahlstrom
147
148
149
Vows IN THE HEBREW BIBLE AND THE ANCIENT NEAR EAST

Tony W. Cartledge
IN SEARCH OF 'ANCIENT ISRAEL'
Philip R. Davies
PRIESTS, PROPHETS AND SCRIBES: ESSAYS ON THE FORMATION
AND HERITAGE OF SECOND TEMPLE JUDAISM IN HONOUR OF
JOSEPH BLENKINSOPP
Eugene Ulrich, John W. Wright, Robert P. Carroll and Philip R. Davies (eds)
150
151
152
153
TRADITION AND INNOVATION IN HAGGAI AND ZECHARIAH 1-8

Janet A. Tollington
THE CITIZEN-TEMPLE COMMUNITY
J.P. Weinberg
UNDERSTANDING POETS AND PROPHETS : ESSAYS IN HONOUR OF
GEORGE WISHART ANDERSON
A.G. Auld
THE PSALMS AND THEIR READERS : INTERPRETIVE STRATEGIES
FOR PSALM 18
D.K. Berry
154
155
156
MINHAH LE-NAHUM: BIBLICAL AND OTHER STUDIES PRESENTED TO
NAHUM M SARNA IN HONOUR OF HIS VOTH BIRTHDAY

M. Brettler and M. Fishbane (eds)
LAND TENURE AND THE BIBLICAL JUBILEE: DISCOVERING
A MORAL WORLD-VIEW THROUGH THE SOCIOLOGY OF KNOWLEDGE
Jeffrey A. Fager
THE LORD'S SONG: THE BASIS, FUNCTION AND SIGNIFICANCE OF
CHORAL Music IN CHRONICLES
J.E. Kleinig
157
158
159
THE WORD HESED IN THE HEBREW BIBLE

G.R. Clark
IN THE WILDERNESS
Mary Douglas
THE SHAPE AND SHAPING OF THE PSALTER
J.C. McCann
160
161
162
KING AND CULTUS IN CHRONICLES:

WORSHIP AND THE REINTERPRETATION OF HISTORY
William Riley
THE MOSES TRADITION
George W. Coats
OF PROPHET'S VISIONS AND THE WISDOM OF SAGES: ESSAYS IN HONOUR
OF R. NORMAN WHYBRAY ON HIS SEVENTIETH BIRTHDAY
Heather A. McKay and David J.A. Clines
163
164
FRAGMENTED WOMEN:
FEMINIST (SUB)VERSIONS OF BIBLICAL NARRATIVES
J. Cheryl Exum
HOUSE OF GOD OR HOUSE OF DAVID:
THE RHETORIC OF 2 SAMUEL 7
Lyle Eslinger
166
THE ARAMAIC BIBLE

Edited by M. McNamara and D.R.G. Beattie
167
168
169
170
171
SECOND ZECHARIAH AND THE DEUTERONOMIC SCHOOL

Raymond F. Person
THE COMPOSITION OF THE BOOK OF PROVERBS
R.N. Whybray
EDOM, ISRAEL' s BROTHER AND ANTAGONIST:
THE ROLE OF EDOM IN BIBLICAL PROPHECY AND STORY
Bert Dicou
TRADITIONAL TECHNIQUES IN CLASSICAL HEBREW VERSE
W.G.E. Watson
POLITICS AND THEOPOLITICS IN THE BIBLE AND
POSTBIBLICAL LITERATURE
Edited by Y. Hoffman and H. Graf Reventlow
172
174
AN INTRODUCTION TO BIBLICAL ARCHAEOLOGY

V.Fritz
A LITERARY APPROACH TO BIBLICAL LAW:
EXODUS 20.22-23.19
Joe M. Sprinkle
175
SECOND TEMPLE STUDIES: 2

Tamara C. Eskenazi and Kent H. Richards
176
177
STUDIES IN BIBLICAL LAW:

FROM THE HEBREW BIBLE TO THE DEAD SEA SCROLLS
Gershon Brin
TEXT-LINGUISTICS AND BIBLICAL HEBREW
David Allan Dawson

Textlinguistic and Biblical Hebrew Dawson PDF

Uploaded by

Copyright:

Available Formats

Textlinguistic and Biblical Hebrew Dawson PDF

Uploaded by

Document Information

Original Title

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

Textlinguistic and Biblical Hebrew Dawson PDF

Uploaded by

Copyright:

Available Formats

JOURNAL FOR THE STUDY OF THE OLD TESTAMENT

Sheffield Academic Press

This page intentionally left blank

David Allan Dawson

Journal for the Study of the Old Testament

Copyright 1994 Sheffield Academic Press

Typeset by Sheffield Academic Press

British Library Cataloguing in Publication Data

INTRODUCTION, AND EXAMINATION OF SEVERAL

EXAMINATION OF KHAN's 'STUDIES' AND LONGACRE's 'JOSEPH'

AN INTRODUCTION TO ONE MODEL OF TEXT ANALYSIS,

TEXT-LINGUISTIC OBSERVATIONS ON NARRATIVE

JEPHTHAH AND RUTHREFINING AND TESTING

SUMMARY, CONCLUSIONS AND IMPLICATIONS

THE TEXT OF JUDGES 2 (BHS) IN COLUMNAR FORMAT

THE TEXTHE OF RUTH (BHS) IN COLUMN

This page intentionally left blank

Text-Linguistics and Biblical Hebrew

terminology and theory. But its principal focus is on methodology,

This page intentionally left blank

1.1 Introduction to the Content and Goals of this Volume

Text-Linguistics and Biblical Hebrew

the bridge is insecure; likewise, if one's communication of truly valid

Chapter 1 introduces the present research project, and will provide a

analysis of one of the currently most influential works in this field,4

Text-Linguistics and Biblical Hebrew

7. These 'linguistic dependencies' are the focus of 'text-linguistic' analyses; that

darken, for linguistics generates new terminology at a mind-boggling

Text-Linguistics and Biblical Hebrew

It will be clear from the comments above, and to follow, that I

Text-Linguistics and Biblical Hebrew

Another language universal is that a language will employ one of two

Text-Linguistics and Biblical Hebrew

Text-Linguistics and Biblical Hebrew

many applications for it to be free enough of unwanted connotations

The essential theoretical elements that inform my methodology are:

These basic elementslike the dry bones in the valleywill be knit

Text-Linguistics and Biblical Hebrew

an enquiry of exactly this sort, and I acknowledge at the outset that, in

Text-Linguistics and Biblical Hebrew

And yet they also write,

acceptance. And yet, although there is likewise a consensus of opinion

Text-Linguistics and Biblical Hebrew

His general statements with regard to language principles are often

In addition, this book contains some remarkably succinct statements

Essentially, Niccacci's starting point, and conclusion as well, is that

Text-Linguistics and Biblical Hebrew

conjugations, in narrative, and in 'discourse'.51

is, in my opinion, his assessment of the foreground/background

He has lumped all conversational material (reported speech)

The clause types used in the two sentences immediately indicate to us

Text-Linguistics and Biblical Hebrew

description would distinguish the first from the second.52 This is

With reference to the second so-called 'flaw': Niccacci only once

overstatement of 'rules' of syntax, some of which are occasioned by

Text-Linguistics and Biblical Hebrew

here arise out of a partially insufficient theoretical base and a few

main narrative are marked as more or less significant on the basis of

Text-Linguistics and Biblical Hebrew