Second Dialect and Second Language in the Composition Class

Author(s): James R. Nattinger

Source: TESOL Quarterly , Mar., 1978, Vol. 12, No. 1 (Mar., 1978), pp. 77-84
Published by: Teachers of English to Speakers of Other Languages, Inc. (TESOL)

Stable URL: https://www.jstor.org/stable/3585793

REFERENCES
Linked references are available on JSTOR for this article:
https://www.jstor.org/stable/3585793?seq=1&cid=pdf-
reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

is collaborating with JSTOR to digitize, preserve and extend access to TESOL Quarterly

This content downloaded from

61.92.158.55 on Sun, 28 Apr 2024 09:10:22 +00:00
All use subject to https://about.jstor.org/terms
 TESOL QUARTERLY
Vol. 12, No. 1
March 1978

Second Dialect and Second Language in the

Composition Class
James R. Nattinger

In linguistics and in other social sciences there are two kinds of facts for
describing data-"hard" facts, the measurements of external events, and
"soft" facts, the internal responses to these external events. Although these
facts exist in all kinds and on all levels of linguistic description, distinguishing
between them and defining their boundaries often lead to much confusion.
An interesting problem in metatheory would be to take some of the current
linguistic oppositions, such as etic-emic, performance-competence, empiricism-
rationalism, speaker-hearer, and analyze them in terms of this hard-soft dis-
tinction. One such pair is treated here: second dialect and second language.
When teachers speak of problems with the former they are usually speaking
about "hard" facts, or facts of (external) variation; when speaking of second
language, though, they often mean "soft" facts, or those that have to do with
(internal) categories and stereotypes. The two kinds of speakers will thus
benefit from different classroom techniques. The rest of the paper reviews some
current pedagogical methods for teaching and testing composition-among
these, sentence combining, discourse analysis, and cloze testing-and sug-
gests how these can be differently used for the two kinds of speakers.

Lately there has been a lot written about teaching English to foreign spea
ers and teaching it to non-standard English speakers. Some claim that th
should be very little difference between the two approaches and that these
proaches should, in fact, have the same goals and follow the same meth
Others, though, believe that these two groups of speakers share very li
either in attitude or in linguistic background that would profit from the sa
pedagogical techniques. Below I will look at non-standard dialects, whi
call second dialects, and foreign languages, which I call second language
terms of a distinction that may distinguish them in a useful way.
In linguistics and in other social sciences there are two different kinds
facts, both of which are real and necessary in any description of human beha
yet which exist in different senses. The first have sometimes been called "h
facts." These are externally perceivable events which are operationally definab
they usually involve measuring and counting, and are independently verifia
by the replication of some sort of procedure; they are the facts of what hum
do. The other kind of facts are "soft facts." These are more internal respon
to the external data; they are only internally verifiable and are perceptions

Mr. Nattinger, Associate Professor of English and Linguistics at Portland State University,
Fulbright lecturer in Buenos Aires the last half of 1977. He has published in Language Lea
ing, Creativity (Yazigi), and English Teaching Forum.
77

This content downloaded from

61.92.158.55 on Sun, 28 Apr 2024 09:10:22 +00:00
All use subject to https://about.jstor.org/terms
78 TESOL Quarterly

what humans think they do; they involve, in other words, the mental concepts
with which we categorize the world. Charles Pyle (1976: 159-180), who is the
first linguist I am aware of to treat the distinction in these terms, says,
Language users don't simply hear what someone says; they project idealized cate-
gories onto the stream of noise. They selectively perceive and evaluate behavior by
unconsciously interpreting marginal, deviant, exceptional pronunciations as instances
of their discrete concepts. Then people think their evaluations and categorical inter-
pretations of what happens is what really happens. They respond to their projections,
not to the actual behavior of others. In other words, they convert hard facts into
soft facts and think these are hard.

Perhaps the most obvious of these mental concepts is the phoneme. A pho-
neme is not completely definable in phonetic terms, as so many post-Bloomfield-
ians have pointed out; neither is it measurable or externally perceivable. Instead,
the phoneme is an abstract idealized conceptualization of pronunciation; it is a
convention that speakers of a particular language share and in terms of which
they categorize what they hear. Like other soft facts, it is a fact about what
speakers think they do, or think they should do; yet it is not something that
can be found in the data. The data may, in fact, contradict this soft fact. In
English, for example, there is clear phonemic contrast between /t/ and /d/, yet
often this contrast is neutralized in the stream of speech, and in such pairs of
words as writer/rider, bitter/bidder they are pronounced the same. On the
other hand, there may be clear phonetic contrast in the data and no correspond-
ing phonemic one. Cat and can't are a minimal pair differentiated by vowel
nasalization phonetically, but the difference we hear phonemically is the pres-
ence or absence of the nasal consonant. What this shows is that there is a gap
between hard and soft facts, that there is certainly no one-to-one relationship
between them, and that soft facts probably cannot be invalidated by hard facts:
soft facts and hard facts, therefore, describe distinct pieces of data.1
The implications of this distinction ought to be explored more than they
have. These facts exist at all levels and in all kinds of linguistic description; yet
they are rarely discussed separately and are often confused. An interesting prob-
lem in metatheory would be to take some of the current linguistic oppositions
and analyze them in terms of the hard and soft distinction. For example, (a)
etic and emic, in the sense in which Pike generalizes these suffixes, treat the
distinction rather differently than I have here. The etic view concerns hard facts
and the emic, soft facts; the difference from the above is that these describe
the same piece of data, not distinct pieces, and are thus two views of the same
thing; (b) performance and competence are again different. They describe one
kind of fact, the hard facts of performance, as being generated from another,
the soft facts of competence. In this distinction, soft facts are primary and prior
to hard facts, which they'produce; (c) empiricism and rationalism differ in the

1I am aware that these conclusions do not easily follow from the scanty evidence here; how-
ever, I have intended only to sketch Pyles' argument, not recreate it. The reader should refer
to Pyles' important article for the details.

This content downloaded from

61.92.158.55 on Sun, 28 Apr 2024 09:10:22 +00:00
All use subject to https://about.jstor.org/terms
 Second Dialect 79

kinds of facts they value. Empiricism looks for hard facts, those that can be
measured, and derived from inductive procedures. Rationalism, using more d
ductive methods, attempts to discover the mental projections by which
categorize the world, things which are not, of course, measurable or countab
(d) speaker and hearer are terms that have often been used synonymously in
linguistic descriptions. But if we look at them in light of the soft-hard distin
tion, it appears that they denote quite separate behaviors. Sociolinguistics has
been showing us for some time now that when we speak we do so with great
variation, and when we hear or perceive what is spoken we do so stereotypicall
-the nature of perception is quite different from the nature of production. Th
the analysis of speech has much to do with hard facts, the analysis of percep
tion has much to do with soft facts.
Second dialect and second language is another distinction that can be
analyzed in these terms. Definitions may have to be stretched a bit to do it,
for second dialect and second language have more to do with pedagogical con-
cepts than with theoretical ones as I treat them here, but the hard-soft differ-
ence is a pervasive one and can give us some insight as to the nature of this
distinction, also. Second dialect facts are essentially hard facts in that they
concern the variable ways certain mental concepts are projected onto the ex-
ternal world. The learning problems of second dialect speakers are problems
of learning the various ways that basic language categories can be manifested.
Second language facts, on the other hand, are usually soft facts. They are prob-
lems of recognizing and learning cultural stereotypes, thus they are problems
of establishing these mental categories themselves. This difference has important
implications for a college writing program.

Implications for a college writing program

Portland State University, like any public urban university, has many sec-
ond dialect and second language speakers, and like most, assigns these students
to composition classes without regard for linguistic background. I believe this
is the correct thing to do for second dialect speakers, but I do not think it is
for second language speakers, who, I feel, should study English composition
apart, with a teacher trained in ESL methods.
Second dialect speakers share the same soft linguistic facts with the teacher.
The teachers' chore therefore is to tap this underlying knowledge and get the
students to understand and perform variable ways of writing the language they
already know. To do this, teachers must be aware of some of the current work
in linguistics, or at least be acquainted with some of the basic principles of
language variation, register, and social dialects. Sophisticated teachers will in-
dicate the places where the students' writing differs from standard English
and will discuss these differences in terms of language variability and prestige
patterns. Most importantly, however, these teachers will point out where there
is true ambiguity and unclarity in the students' work and not just where there
are infractions of social values. Second language speakers, on the other hand,

This content downloaded from

61.92.158.55 on Sun, 28 Apr 2024 09:10:22 +00:00
All use subject to https://about.jstor.org/terms
80 TESOL Quarterly

often do not share many of the soft facts of the language, let alone the hard
facts. Thus, their errors are likely to be more fundamental, and they will likely
violate more basic categories of the language. They need to become aware of
the facts of English language variation, too, just as the second dialect speakers,
but usually they need practice with more basic rules first.
It follows from this that the techniques used to teach and test composition
should be different for these different kinds of speakers. But before I list some
current teaching methods I must pause for a matter of definition. Up to this
point I have neglected to define what I mean by "second dialect" and for a
good reason-it is extremely difficult to do. Structurally, second dialects are
not discrete systems that are qualitatively different from standard dialects, nor
do they designate clear social or ethnic boundaries. Yet most of us are willing
to agree they exist and will seriously listen to research on such subjects as Ap-
palachian speech, Black speech, Ozark speech and others. And most of us will
agree, through some sort of symbolic interaction, that there are kinds of lan-
guage that can be described as standard also, although we are always a bit
embarrassed by that impression and hasten to qualify it. This impression we
have of qualitatively different categories coupled with our ability to find only
quantitative differences in the data is the result, once again, of the soft-hard
distinction: standard and non-standard speech are soft categories that are mani-
fested by hard data, data that often contradict these categories.
Below are some methods currently in use for teaching composition and
some suggestions about how they might be used differently for the two groups
of speakers. One current and popular method of teaching composition is by a
method called "sentence combining." This is an approach to composition that
developed from the work of Bateman and Zidonis (1964) at Ohio State, John
Mellon (1969) at Harvard, and Frank O'Hare (1973) at Florida State. Essen-
tially it is a technique for teaching the complexities of syntactic embedding that
is based, loosely, on the early Syntactic Structures model of transformational
grammar and makes use of such notions as "kernel sentence" and certain com-
bining transformations. There are several variations of this technique around,
from very controlled applications in which kinds of possible embeddings are
introduced slowly and systematically, to very free applications in which stu-
dents are at once given many types of embeddings to use. In all of these varieties,
though, the object-to expand the options that students have in putting a sen-
tence together-and the basic method-to provide students with kernel sen-
tences and to have them practice different ways of combining these-are the
same.

Second language speakers will, I feel, be better served by

guided approaches than will second dialect speakers. Cha
has developed one such controlled approach. In it, students lea
of embedding gradually, moving to the next rule only after
vious, and, since they take the kernel sentences for these emb
the text, they will not have to invent the content of their senten

This content downloaded from

61.92.158.55 on Sun, 28 Apr 2024 09:10:22 +00:00
All use subject to https://about.jstor.org/terms
 Second Dialect 81

dings for second language speakers would probably concern very basic opera-
tions that would only be tedious and unnecessary drills for second dialect speak
ers. For example, there might be drills for the proper placement of articles,
single adjectives, participles, appositives; other attributive phrases of time, lo
cation, and manner; prepositional phrases of most types; particle movement in
two-part verbs; rules of contraction; subordinate clauses with relative pronoun
such as who, which, where, that, when these intervene between subject an
predicate. All of these obviously concern operations of a very basic sort i
English, many of which William Labov (1969: 24-26) would call Type I Rule
("automatic, deep-seated patterns of behavior which [for native speakers] are
not consciously recognized and never violated"); yet they are the sort that mo
second language speakers have learned imperfectly and that require furth
practice. Second dialect speakers, as I mentioned above, would find most of thes
drills unnecessary; these students require a less guided approach to senten
combining.
William Strong (1973) gives his students a series of kernels, just as Cooper
does, but he presents them with several possibilities for combining all at once
and lets the students discover the options available to the writer. His point is
that even a series of kernels offers several possibilities, as Marilyn Sternglass
(1976) says, and after students combine some of these they can make a choice
as to which they prefer. Thus the idea here is not only to build syntactic ma-
turity, but also to teach students how to vary the language they already know
in acceptable ways. Useful operations for second dialect speakers to practice
would include multiple embeddings, absolutes, sentence compoundings, any but
the simplest relative clause constructions, and, perhaps most importantly, dif-
ferent kinds of nominalizations, particularly clauses used as subjects or objects
of the sentence. These drills certainly go beyond fundamental operations of
sentence construction; they deal much more with the options we all have of
combining these elementary structures.
Another method for teaching composition which so far has not become
popular, unfortunately, is one based on Willis Pitkin's (1973) model for dis-
course analysis. This model follows the idea that information in writing, as
elsewhere for that matter, is chunked into increasingly larger units or "discourse
blocs" which themselves are regrouped into even larger units by combining two
blocs at a time. Pitkin's model is a hierarchy that is built by successively group-
ing binary functions, and, in this, it is very similar to an Immediate Constituent
display in establishing a structure through successive binary groupings (or cuts,
looked at from the other direction). However, there is an important difference.
For Pitkin, structure has very little to do with determining where the cuts
should come; instead he proposes the cuts be made entirely on a functional or
conceptual basis and suggests that there is no necessary correlation between
the grammatical "weight" of structural elements-their shape, length, or punc-
tuation, for example,-and their position in the functional hierarchy; that is,
there is no reason for saying, a priori, that the Sentence or the Paragraph is a

This content downloaded from

61.92.158.55 on Sun, 28 Apr 2024 09:10:22 +00:00
All use subject to https://about.jstor.org/terms
82 TESOL Quarterly

necessary functional unit of the hierarchy. Thus, this is not a description that
treats the sentence as basic; rather, sentences here are sub-units of texts and
can vary greatly in their functional value. For example, a single sentence can
be included in one of these discourse blocs or it could itself include one or
more of these blocs. Some of the "blocs" that Pitkin suggests are functional
shaping discourse are: problem:solution, assertion:reassertion, premise:conclu
sion, and concession:response. These blocs can describe a text in terms o
tree diagram that branches by splitting a higher discourse bloc into two low
ones-much the same as an Immediate Constituent display, as I said befo
only here the constituents are conceptual ones and not structural ones.
The major advantage of this model for a composition class is that it
discourse analysis and not a sentential one, that it, therefore, treats sentences
paragraphs, for that matter) not as basic units but simply as sub-units of tex
which is certainly closer to a proper notion of how discourse is put togethe
This is something that most composition methods do not do, from tradition
sentence diagramming, to the kind of sentence combining discussed above.
There are many problems with this model, though, particularly when it
pulled into the classroom. Identifying the discourse blocs themselves is diffic
since they are not tied to any particular form; also there is always some of
text that cannot be explained in terms of binary features-at least one quart
of a typical, expository text, I have found; finally, the diagrammatic techn
Pitkin gives for displaying the relationships of these blocs is, to say the lea
unwieldy. But in spite of these problems, Pitkin can be used valuably in
limited way, and, to get back to the theme of this paper, it should be
differently for second language and second dialect speakers. The latter h
found it valuable to look at texts in terms of these polarities, to analyze
construct their own essays in terms of them, and to discover what some of
structural reflexes of these discourse blocs are. All of this was impossible fo
most of the second language speakers. They simply could not find the binar
oppositions we were discussing. I suspect this is due to differences in genera
language patterning, even on this highly abstract, functional level, and that f
such a method to be successful, there would have to be some sort of contrasti
analysis of functional discourse patterns much as is done in phonology
morphology to determine what might be some of the differences in overall p
terning. Then students could see the ways their language structured discour
and the ways that was different from English patterning. After that they wo
be more able to analyze and produce in terms of the English patterns.
Francis Christensen (1967), in an earlier model for discourse analysis, als
treated the sentence as a sub-unit of the text. However, unlike Pitkin, Christen
treats the sentence as an integral functional unit of discourse and says that
sentence must be seen as either coordinate or subordinate to the sentence that
precedes it. Paragraphs are patterns of this sentence sequencing that usu
move in a general to specific direction. The typical paragraph, he feels, begi
with a general topic sentence, one that presents the controlling idea of

This content downloaded from

61.92.158.55 on Sun, 28 Apr 2024 09:10:22 +00:00
All use subject to https://about.jstor.org/terms
 Second Dialect 83

paragraph, and follows with sentences that fill in detail, either giving parallel
examples in coordinate sentences or further specifying an idea in subordinate
sentences. Second dialect and second language speakers both find this model
very useful, but, as with the others, in different ways. Second dialect speakers
begin writing paragraphs that conform to the typical paragraph described above,
but soon begin to practice permutations of this general to specific style and dis-
cuss the stylistic effects of these changes. Second language speakers, since they
are still learning basic patterns in the language, do better to avoid unpredictable
material here as in other areas of their language learning. They should practice
the typical paragraph as Christensen presents it, a general topic sentence first,
followed by more specific supportive statements in subordinate or coordinate
relationships, and then, only when they are at ease with that, go on to the varia-
tions of that form.

Testing
Obviously the different uses of the methods described above will call for
different kinds of testing: the specific choices students use in combining sen-
tences will vary, of course, as will the binary polarities they locate, and the
kinds of paragraphs they can put together. Tests will have to be constructed in
terms of these various expectations.
There are other tests, not connected with any specific teaching technique,
that could also be used to test writing competence. I will mention just one of
these here. Cloze tests have been around for a long time and have gained a
certain amount of deserved popularity. Cloze tests measure something that no
one understands very well yet that has to do with what John Oiler (1973) calls
the "grammar of expectancy,"10 the ability to anticipate certain features of dis-
course. Many feel that this ability is closely related to both reception and pro-
duction skills, that is to a "global proficiency" in the language. Classical cloze
deletes every fifth word of a text and asks students to fill in the blanks with
something that fits. Second dialect speakers have no trouble doing this, and
second language speakers usually do. The native speaker is able to predict from
many more features in the text than is the second language speaker, and is able
to fill in the blanks quite easily. In several newer versions of the cloze, these
deletions are not as random but are more planned. In some, a deletion occurs
among every seventh to twelfth word; in some, all inflectional affixes are de-
leted; in others, all function words (Bondaruk, Child, and Tetrault: 1975). As
more and more of the redundant signals are systematically removed through
deletion, the more trouble even a native speaker will have in filling in the blanks.
This kind of "reduced redundancy" testing can obviously be very useful in test-
ing students' recognition of written patterns. A test that deleted many transition
signals from sentences in a text would test students' perception of the organiza-
tion of that text, since their ability to supply these items would indicate their
ability to see the systematic relationship there. But, again, this is very difficult
for second language speakers; any reduction of redundancy decreases their func-

This content downloaded from

61.92.158.55 on Sun, 28 Apr 2024 09:10:22 +00:00
All use subject to https://about.jstor.org/terms
84 TESOL Quarterly

tioning in that language far more than it does native speakers'. In testing as well
as teaching college composition, then, there are advantages in distinguishing
second dialect from second language speakers.
REFERENCES

Bateman, D. R. and F. J. Zidonis. 1964. The effect of a knowledge of g

mar upon the growth of language complexity. U.S. Office of Educa
tive Project, 1746. Columbus, The Ohio State University.
Bondaruk, John, James Child and E. Tetrault. 1975. Contextual test
guage proficiency. Arlington, Center for Applied Linguistics, 89-1
Christensen, Francis. 1967. Notes towards a new rhetoric. New York, H
Cooper, Charles. 1973. An outline for writing sentence-combining pro
Journal 62, January.
Labov, William. 1969. The study of non-standard English. Washington,
Mellon, J. C. 1969. Transformational sentence-combining: a metho
the development of syntactic fluency in English composition. Urba
O'Hare, Frank. 1973. Sentence-combining: Improving student writing w
grammar instruction. Urbana, NCTE.
Oller, John W. 1973. Pragmatic language testing. Language Sciences
Pitkin, Willie. 1973. A pedagogical model of discourse structure. U
sertation, University of Southern California. Ann Arbor, University
Pyle, Charles. 1976. Soft facts. Problems in Linguistic Metatheory.
ceedings, Michigan State University, Department of Linguistics
159-180.
Stemglass, Marilyn. 1976. Composition teacher as a reading teacher. College Com
position and Communication 27, 378-382.
Strong, William. 1973. Sentence combining. New York, Random House.

This content downloaded from

61.92.158.55 on Sun, 28 Apr 2024 09:10:22 +00:00
All use subject to https://about.jstor.org/terms