Professional Documents
Culture Documents
of Interlanguage Discourse
Gabriele Kasper
University of Aaihus
The paper examines the influence of the formal classroom as a learning environment on the
discourse behaviour of advanced German students of English in conversations with English
native speakers. It is proposed that FL teaching as a causal factor in the formation of
IL-specific rules can operate either directly, in presenting the learner with FL material which
deviates from target norms, or indirectly, by triggering off psycholinguistic processes which
in turn lead to IL-specific rule formation. These are referred to as primary and secondary
teaching induction, respectively. On a more concrete level, the impact of two constituents of
FL teaching on IL discourse is discussed:
In conclusion of the data analysis, a classroom specific pidgin will be hypothesized which,
when transferred to non-classroom settings, leads to pragmatically inappropriate com-
municative behaviour.
On a more general level, it will be postulated that second language acquisition hypotheses
should be formulated with reference to specific types of acquisition/learning contexts.
1. Introduction
One of the many issues which interlanguage (IL) researchers disagree about is the
extent to which the development and properties of ILs depend on particular acqui-
sition contexts1. Quite a few researchers subscribing to the identity hypothesis of
L2 acquisition—i.e. to the assumption that all language acquisition is guided by the
same (linguo-) cognitive2 principles—have in their search for such processes ig-
nored the potential influence of the acquisition environment (see in particular
Dulay and Burt's studies on morpheme acquisition orders, summarized in Dulay
and Burt 1980). Others compared the acquisition of particular L2 phenomena—
again English morphemes—in informal acquisition contexts and in the L2 class-
room and found no significant differences in acquisition order, from which they
concluded that the contextual variable has no decisive impact on L2 acquisition
(e.g., Fathman 1975, 1977; Larsen-Freeman 1976; Perkins and Larsen-Freeman
1975). It seems rather questionable, however, that findings on certain highly
99
100 Studies in Second Language Acquisition, 4, No. 2, 1982
A third group of researchers in the "identity tradition" adopts the view that while
all language acquisition is based on the same acquisition system, the activation of
this system is modified by different acquisition contexts. An exponent of this
assumption is the Kiel project which compares the acquisition of some formal
aspects of English and German by children under various contextual conditions
(see in particular Wode 1981).
Likewise, one should be careful not to generalize too quickly from acquisition
regularities observed in One particular L2 subsystem as to how L2 as "(a subset of) a
language" is acquired. As was mentioned in connection with the morpheme
studies above, the acquisition procedures applied to formal systems like phonology
and morphology might well differ from those activated in the acquisition of func-
tional aspects like speech act and discourse rules. Again, the integration of the
regularities observed in the acquisition of different L2 components into a com-
prehensive theory of L2 acquisition presupposes the analytical step of describing
these components as acquisition objects in their own right.
The above remarks are intended to indicate the general presuppositions which
underlie the more specific purpose of this article, viz. to show how certain features
observable in IL discourse can be traced back to a particular learning environment,
i.e. the PL classroom. As this study is part of a larger investigation into pragmatic
aspects of IL use (Kasper 1981), it will only report on learner-specific behaviour at
the pragmatic level. Morphological, syntactic, and lexical IL features which are
likely to be brought about through teaching have been well documented in the
literature (e.g. George 1972, Selinker 1972; Corder 1973: 291f; Jain 1974:199;
Richards 1974a; Stenson 1975; Swain 1977; Holmes 1978; Raabe 1979: 82f;
Schmidt and Richards 1980). Such aspects will only be included here if they are of
pragmatic relevance.
Teaching-Induced Aspects of Interlanguage Discourse 101
2. Data
The data on which this study is based was collected in a project on communicative
competence in FL teaching, conducted at Ruhr—Universitat Bochum, Federal
Republic of Germany, from 1976 to 1980 (for descriptions of the project, see
Edmondson, House, Kasper and McKeown 1979; Edmondson, House, Kasper and
Stemmer 1980,1982). It comprises three batteries of 48 face-to-face dialogues each:
dialogues between pairs of native speakers of English and German first-year stu-
dents of English, between native speakers of English, and between native speakers
of German. The dialogues were all elicited through role play and were audiotaped.
As the learners had already left school when they participated in the project, direct
observation of their FL classroom activities was not possible. In order to compen-
sate for this information gap to some extent, the syllabus according to which they
had been taught and the course materials they had used were analyzed. Information
on their FL training and experiences in using English outside the classroom was
obtained by questionnaire. Informal interviews, which were conducted with each
learner after the role playing exercise in order to obtain "authoritative interpreta-
tions" (Corder 1972) by the learners of ambiguous or incomprehensible utterances,
also frequently shed light on the learning process. Furthermore, video recordings of
various English lessons at German secondary schools were analyzed (see Kasper
1981 for a detailed account).
3. Analysis
Pragmatic errors and appropriate but learner-specific IL features were identified by
native speakers of British English, by comparison of the learners' performance with
the English native speaker dialogues and by reference to available linguistic de-
scriptions. A pragmatic error was defined as a pragmatic or discourse feature which
is neither used in the given context nor accepted as appropriate by native speakers.
faced with incorrect rules or texts but rather with inappropriately arranged exer-
cises, leading to wrong rule generalization.
RU: yes I've done everything now thanks and I just wanted to say Mrs Bell
how lovely it's been living here and to thank you for everything
This inappropriate formal register use might be due to the intensive work with—
especially non-contemporary—literary texts, which is the main activity during the
last three years of English in German secondary schools. If one follows this in-
terpretation, the learners can be said to generalize a particular register from the
appropriate contexts to contexts calling for a different register ("pragmatic gen-
eralization"); the errors in question can then be accounted for as being secondarily
teaching induced.
However, one also finds quite a few instances of inappropriate register use in our
learners' course books, of which the following exercise is an example:
(i) "Stella and Barbara want to go to the theatre to see a play. They are
dressing.
Stella: Mother, I can't find my comb.
Mother: It may be in the cupboard in the bathroom, dear.
Stella: No, it isn't. May I have your comb, Mother?"
"On the way to the theatre.
Barbara: Shall we be allowed to keep our coats on in the theatre?
Stella: Yes, we shall."
(Learning English A2: 122).
The implicit rules underlying the use of may, be allowed to, and shall in these
dialogues are likely to lead directly to non-native IL rule formation, which can thus
be characterized as being primarily teaching induced.
Inappropriate use of modal verbs. Example (2) and (i) lead to a connection with the
second area of pragmatic errors due to teaching materials, viz. the area of modal
verbs. Practicing epistemic and non-epistemic may in close proximity and without
previous explanation and practice of these two functions, as is the case in the
104 Studies in Second Language Acquisition, 4, No. 2, 1982
course book which (i) is taken from, is likely to result in IL specific rule formation.
On the product level, such rules materialize in errors like the double-marking for
probability in (3).
(3) (Landlady reproaching student)
E: well you did make a lot of noise last night when you were coming in
L: oh well I can't remember so well but perhaps you may be right yes
RU: oh I don't remember last night very well but if you say so you're
probably right
Another modal verb which our subjects often use inappropriately ismust, as can be
seen from examples (4) and (5).
(4) (E and L witnessing a car crash)
E: what shall we do then what shall we 'do
L: oh ermjjerhaps we must stop other cars^coming
RU: perhaps we'd better stop other cars
IM: ja vielleicht mussten wir andere Autos anhalten
(5) (as 4) ^ ^
E: .. . I think there's a farmhouse on the other side of the road there
L: ye\er so one of us must go there and ask for help
RU: yes er so one of us had better go there and ask for help (I suppose)
Although these pragmatic errors will mainly be a result of transfer from German
miissen, it seems likely that the IL rule which functionally equates English must
and German miissen was reinforced by inappropriate grading of the rules express-
ing obligation in the learners' course book. Must is introduced as the first verb of
obligation at a very early learning stage, followed by have to a year later, whenmust
is already a stable and well-automatized element of the learner's IL system (see also
Speight 1977 for this point). Even then, however, the functional difference between
these two modal verbs is neither explicitly stated nor practiced; have to is merely
presented as a substitute for the morphologically defective must. Had better,
which is frequently used for expressing suggestions by the English native speakers
in our data, is not presented for productive use in this course book at all.
It should be added that in the role play which the opening move in (7) is taken from
the learner is supposed to know the native speaker well and is expected by the
latter, so that it is difficult to see why he should want to express doubt about the
other person's identity or availability for talk, as is conveyed by the rising intona-
tion. I assume that the generalization of rising intonation to contexts which require
falling intonation is induced by the initiation—response—feedback sequences
which are typical for all traditional classroom discourse (e.g. Sinclair and Coulthard
1975; Griffin and Mehan 1981). The functional difference between requests for
information in non-classroom situations and teachers' initiations is that in the
former case, the speaker's intention is to fill a gap in his knowledge by eliciting
information which is (supposedly) available to the hearer (see Maas 1972: 213ff). In
the latter case, however, the speaker does possess the relevant information and
wants to find out, for reasons deriving from the institutional properties of the
formal learning/teaching context, if this information is part of the hearer's knowl-
edge (see Rehbein 1979 for an analysis of instructional elicitation processes). As
this function of teachers' initiations is part of learners' institutional knowledge,
they very often perform their responses by not only providing the requested infor-
mation but at the same time appealing for feedback from the teacher as to whether
their response is correct (cf. Edmondson 1979, 1980, 198 la ; Kahrmann 1979). A
typical "pedagogic exchange" of this kind has been documented by Edmondson
(1979:207):
I11) T: no^look at the words,, check through^ is that all rights I'll ask a few
questions^who wants to go to Germany^Conny
Con: Tom wants to go to Germany
T: ja a^what family^Tom belongs to what family^Ute
Ute: it is^erm (studies book) the family Bartlett
T: yes it's the Bartlett family^rigrit
In the second exchange, the learner uses rising intonation in responding to the
teacher's initiation, and the teacher confirms Ute's response in the following turn.
Lack of marking for speech act modality. The predominance of the referential over
other speech functions is furthermore evidenced in the insufficient degree to which
relational and expressive functions are explicitly expressed in the learners' con-
tributions. Native speakers normally take care of such functions by formulating
their speech acts more or less directly, and by intensifying or downtoning their
illocutionary force, a phenomenon which has been termed speech act modality (see
House and Kasper 1981; Kasper 1981). In six of the 48 learner-native speaker
dialogues of our corpus, the learners do not mark their speech acts for modality at
all. On the process level, this has been tentatively explained in terms of the learners
using the communication strategy of global modality reduction (Faerch and Kasper
1980 and forthcoming). In other dialogues, the learners mark some, but not all, of
their utterances for speech act modality. In psycholinguistic terms, one might
describe the process leading to such individually unmarked utterances as local
modality reduction (Kasper 1979) or modality generalization, depending on
whether it seems more likely that the learner has difficulties in finding the appro-
priate linguistic means for marking for that function and therefore decides to
reduce on it, or whether he is not aware of the necessity for expressing it at all. In
the later case, his communicative behaviour is generalized from contexts with a
predominantly referential function to those where relational functions play a more
important role. Examples of learners' contributions whose lack of markedness for
speech act modality can be accounted for by modality generalization or modality
reduction as primary causal factors are (12) and (13).
something for a while and I mean it's not often that this happens
Whereas in the cases mentioned so far, the hypothesis is that the learners planned
their utterances on the basis of their IL knowledge, there are also speech act
realizations with a lack of markedness for modality which seem to be planned on
the basis of the learner's LI. However, while the learner preserves the illocution
and propositional content of his L1 plan in his IL realization, the modal component
is reduced. I have referred to this psycholinguistic process as LI transfer with
reduction. It can be detected if one compares the learner's intended meaning as
expressed in his L1 with his IL realization. Examples (4) above and (14) can illustrate
this point.
108 Studies in Second Language Acquisition, 4, No. 2, 1982
At the level of individual linguistic elements which can serve as "modality mark-
ers" (House and Kasper 1981), the total learners' performance shows a considerable
underrepresentation olappealers (e.g. question tags) and cajolers (e.g. you know, I
mean, actually); the English native speakers use appealers twice as often, cajolers
10 times as often as the learners (for a more detailed account see Kasper 1981:
246ff.|.
4. FL "pidgin"
Looking at the principles governing social interaction, one finds that many types of
interaction arenot governed by the conversational principles set up by Grice (1975),
which emphasize propositional, logical and rational aspects of communication.
Rather, participants organize their conversational behaviour according to princi-
ples which take care of their social needs, and which have been discussed as the
principle of face-saving (Goffman 1967), tact maxim (Leech 1977), politeness rule
(Lakoff 1973), or hearer-support maxim (Edmondson 1981 b). By contrast, the
conversational norms of the FL classroom, as far as using the FL is concerned, do
seem to follow the Gricean maxims, and so do the learners in their communication
outside the classroom. Conversational behaviour is thus orientated to minimal
Teaching-Induced Aspects of Interlanguage Discourse 109
Valdman (1977) establishes three models of such a development, each reflecting the
communicative demands of different acquisition contexts: (1) In an ideal informal
SL context with favourable attitudinal and motivational sets both on the part of the
learner and the native speakers he interacts with, IL develops gradually and sym-
metrically at the functional and formal level. (2) In traditional FL teaching with a
strong emphasis on formal correctness, IL evidences a higher degree of well-
formedness even at the initial stages, whereas its functional use is low ; there is-an
asymmetric development at the two levels with dominance of well-formedness. (3)
In informal SL acquisition contexts with unfavourable attitudinal and motiva-
tional sets on the part of the learner and native speaker interactants, initially strong
emphasis on (a few!) functional aspects, combined with little attention to well-
formedness, will lead to a pidginized L2 variety. Further development towards an
L2 norm is, among other things, impeded by the back-lash effect exerted by a
stigmatized language variety on its user.
If one relates these three models to the present discussion, our learners' FL learning
process can partly be represented by model 2; formal correctness is given high
priority in a German school context, and even though the learners' IL performance
certainly deviates morphologically and lexically from a target norm of English,
these deviations do not suggest pidginization. However, if the concept of well-
formedness is extended beyond the phonological, grammatical and lexical compo-
nent to include formal aspects of the modal level, which I think it should in order to
adequately account for language use in direct interaction, then model 2 is no longer
fully appropriate as representing the learners' IL development. Rather, the de-
ficiencies in the learners' performance at the modal level could be depicted by an
inverse variety of model 1. Given the functional dissociation of LI and FL in the
classroom, expressive, relational and metacommunicative functions are not regu-
larly realized in IL, which therefore does not systematically develop formal features
to match such functions. In analogy to the reductions at the phonological, gram-
matical and lexical level which typically characterize pidgins in informal SL
environments, one might describe these reductions at the modal level as a pidgin-
like variety which typically results from the communicative norms holding in FL
classroom contexts. As LI is available as a means by which non-referential func-
tions can be realized, the restricted use of FL does not result in communication
deficits inside the classroom, and can therefore escape the attention of both
teachers and learners. Outside the classroom, however, it does lead to functionally
110 Studies in Second Language Acquisition, 4, No. 2, 1982
restricted language use, and may well contribute to what Harder (1980) calls the
"reduced personality" of the FL learner.
Notes
1. The following terminology will be observed:
Acquisition: informal acquisition; cover term referring to all types of language acquisi-
tion and learning
Learning: language acquisition under formal classroom conditions
LI: native language
L2: cover term for second or foreign language
SL: second language; non-primary language institutionalized as a regular means of
communication in the learner's speech community
FL: foreign language; non-primary language not institutionalized as a regular means of
communication in the learner's speech community
IL: interlanguage; the linguistic knowledge system the learner activates when trying to
communicate in L2. Like any other language, it comprises pragmatic, semantic, syntac-
tic, morphological and phonological rules. Unlike most other languages, it is typically
developmental and can be permeated by learning and communication strategies.
2. Felix (1981) and Wode (1981) have recently made the valuable point that the explanation
of L2 acquisition in terms of general cognitive principles is insufficient: rather, the
cognitive devices specially geared towards language (LI and/or L2) acquisition should be
in the focus of L2 acquisition research.
3. The "diversity of context" approach to L2 acquisition has its predecessor in the sociolin-
guistic criticism of the homogeneity assumption underlying Chomskyan linguistics: the
description of language variety is a logical precondition for the identification of invariant
properties (see Hymes 1972: 276).
4. This is also implied, though not elaborated on, by Jain 1974 and Richards 1974a.
5. cf. " . . . any inexplicit explanation may be open to misinterpretations by the student. The
teacher must be on guard to monitor drills and explanations to be certain that they do not
lead to false generalizations" (Stenson 1975: 57; my italics).
6. The following conventions are used in the transcripts:
L = German learner of English E = English native speaker
RU = reconstructed utterance IM = intended meaning
^ = fall '-'"*/ = rise-fall
sjl
= rise '-^* = fall-rise
AA
-^ = short pause = longer pause
" = primary stress [ ] = simultaneously spoken passage
References
Bausch, K.R. (ed.). 1979. Beitrage zur Didaktischen Grammatik. Konigstein / Ts.: Scriptor.
Borner, W. and K. Vogel. 1976. Zum Verhaltnis von wissenschaftlicher und padagogischer
Grammatik in der Fremdsprachenlehre. Franzosisch lehren und lernen, ed. by W. Borner, B.
Kielhofer and K. Vogel, 7-39, Kronberg: Scriptor.
Brown, P. and S. Levinson. 1978. Universals in language usage: politeness phenomena.
Questions and politeness. Strategies in social interaction, ed. by E.N. Goody, 56-289. Cam-
bridge: Cambridge University Press.
Burt, M.K. and H. C. Dulay. 1980. On acquisition orders. Second language development, ed.
by S. W. Felix, 265-327. Tubingen: Narr.
Teaching-Induced Aspects of Interlanguage Discourse
Leech, G. N. 1977. Language and tact. L.A.U.T. paper No. 46, Series A, Trier: University of
Trier.
Maas, U. 1972. Grammatik und Handlungstheorie. Pragmatik und sprachliches Handeln, ed.
by U. Maas and D. Wunderlich, 189-276. Frankfurt: Athenaum.
Mehan, H. 1979. Learning lessons. Cambridge, Ma.: Harvard University Press.
Perkins, K. and D. Larsen-Freeman. 1975. The Effect of formal language instruction on the
order of morpheme acquisition. Language Learning 25. 237-244.
Raabe, H. 1979. Der Fehler beim Fremdsprachenerwerb und Fremdsprachengebrauch.
Fehlerlinguistik. Beitrage zum Problem der sprachlichen Abweichung, ed. by D. Cherubim,
61-93. Tubingen: Niemeyer.
Rauch, R. 1979. Ein Kapitel Lehrergrammatik: Vorschlage zur Didaktisierung des 'Sub-
jonctif'. In Bausch, 220-237.
Rehbein, J. 1978. Reparative Handlungsmuster und ihre Verwendung im Fremdsprachenun-
terricht. MS, Bochum.
. 1979. Elizitieren. Fragemodifizierung im Unterricht. MS, Bochum.
Richards, J. C. 1974a. A non-contrastive approach to error analysis. In Richards 1974b,
172-188.
. (ed.) 1974b. Error analysis. London: Longman.
|ed.). 1978. Understanding second and foreign language learning. Rowley, Ma.:
Newbury House.
Schmidt, R. and J. C. Richards. 1980. Speech acts and second language learning. Applied
Linguistics 1. 129-157.
Schumann, J. 1975. Implications of pidginization and creolization for the study of adult
second language acquisition. In Schumann and Stenson, 137-152.
Schumann, J. 1978. The pidginization process: a model for second language acquisition.
Rowley, Ma.: Newbury House.
and N. Stenson (eds.). 1975. New frontiers in second language learning. Rowley,
Ma.: Newbury House.
Selinker, L. 1972. Interlanguage. IRAL 10. 209-231.
Sinclair, J. McH. and M. Coulthard. 1975. Towards an analysis of discourse. The English used
by teachers and pupils. London: Oxford University Press.
Speight, S. 1977. German English and English as Esperanto. Kongressdokumentation der 7.
Arbeitstagung der Fremdsprachendidaktiker Giessen 1976, ed. by H. Christ and H. E. Piepho,
156-158. Limburg: Frankonius.
Stemmer, B. 1981. Kohasion im gesprochenen Diskurs deutscher Lerner des Englischen.
Manuskripte zur Sprachlehrforschung Nr. 18. Heidelberg: Groos.
Stenson, N. 1975. Induced errors. In Schumann and Stenson, 54-70.
Strevens, P. 1976. A theoretical model of the language learning/teaching process. Working
Papers on Bilingualism 11. 129-152.
Swain, M. 1977. Future directions in second language research. Proceedings of the Los
Angeles Second Language Research Forum, ed. by C. Henning, Los Angeles: Los Angeles
Second Language Research Forum.
Wode, H. 1981. Learning a second language. Tubingen: Narr.
Valdman, A. 1977. On the relevance of the pidginization-creolization model for second
language learning. Studies in Second Language Acquisition 1 (2). 55-75.
Yorio, A. 1976. Discussion of "Explaining sequence and variation in second language acqui-
sition". In Brown, 59-63.
Zydatiss, W. 1976. Tempus und Aspekt im Englischunterricht. Kronberg/ TS.: Scriptor.