Editor’s Note

I am grateful to Victor E. Hanzeli for helping in the preliminary editorial work involved in the
preparation for publication of the three papers which together with his own are herein presented as a
cluster of articles on Error Analysis and Learner’s Language.

Introduction
This issue contains four papers read and discussed at the joint ACTFL/AATSP/AATF
convention in Denver, Colorado, in late November, 1974. The session, chaired by A. Valdman, was
labelled “Error Analysis” but, as it happens so often, it was promptly refocused by Corder upon the
larger topic of Learner’s Language, in terms of both grammaticality and appropriateness. Cohen’s
paper on error correction draws on rich research sources and speaks directly and informatively to
various second-language classroom practices. Valdman and Hanzeli show the relevance of
information derived from the study of Learner’s Language and Error Analysis to various aspects
of second-language methodology.
The work of this group is further witness to the trend, inaugurated at the 1969 Second Inter-
national Congress of Applied Linguistics in Cambridge, that i s taking applied linguistics toward
the pragmatics of communication, “notional” organization of learning sequences, and probabilistic
approaches to pedagogical grammar. In these developments, Error Analysis has played a crucial role
by (a) pruning the excessive claims of Contrastive Analysis and (b) providing data for the emerging
theory of Learner’s Language in all its sociolinguistic and psycholinguistic complexity.
VICTORE. HANZELI
University of Washington

The Language of Second-Language

DON’T WANT to speak immediately about priate to use the term “error analysis” to refer to
I the problem of whether “error analysis” is a
suitable name for what we are doing, because I
the pedagogical objectives of the sort of study
that we are doing, to distinguish it from the per-
think we shall have plenty of opportunity to talk haps more theoretically oriented study which we
about that later. What we are interested in has call the study of interlanguage. What we are
developed from a long history of people attempt- really interested in, as a long-term objective, is to
ing to account for the errors of language learn- gain an insight into the processes of second lan-
ers; in this respect it is the grandfather of our guage learning and to understand something
rather young science. It still is, I think, appro- about the various strategies of language learners.

409
410 S. PIT CORDER

In this sense it is clearly a learner-centered inves- behaviour of the language learner is, in that
tigation we are concerned with and only secon- sense, rule-governed and we are able to infer,
darily would it be regarded as concerned with dimly sometimes, the nature of the rules the
teaching. We can only adopt appropriate teach- learner is apparently using when he attempts to
ing strategies, materials and the like, to the communicate by means of this language of his.
degree that we understand what is going on We do this indirectly, of course, through the sort
when people learn. of utterances that he is producing; that is, after
Now, I want to list a number of theoretical all, the way that we get at the structure of a lan-
notions or fundamental assumptions which, I guage anyway - through the actual performance
think, are shared by my co-panelists. They will of native speakers of it. I would go so far as to say
quickly put me straight if I don’t, in fact, that the learner is a “native speaker” of his pe-
represent their views. The statements I am going culiar language. In fact, he is probably the only
to make are of such generality, however, that it native speaker of it, though his language may
would be very difficult to find fault with them. share interesting properties with that of other
They will however give you the general orienta- people who are learning the same target lan-
tion of all our work. guage, particularly if they have the same lan-
First, 1 would regard language learning as guage background as he has. This language of
fundamentally a form of cognitive learning; the his is clearly neither his mother tongue nor is it
language learner is using any human being’s the target language, though it may share some
cognitive learning capacities in the process of properties with these two languages: mother
language learning. These cognitive learning ca- tongue and target. We would expect to find
pacities are, I take it, genetically built-in but similarities under certain conditions with both of
without prejudice as to whether they are lan- these, and this is the reason why this language
guage specific or not. In other words, I don’t that the learner has and is developing all the
propose, and I don’t know whether my col- time, has been given a number of names which
leagues are going to propose, to adopt either a indicate this relationship. The term “interlan-
strong nativist or strong interactionist point of guage” which has achieved wide acceptance, as a
view about language learning. My own inclina- result of Larry Selinker’sproposal some years ago
tion is toward the more general interactionist (Selinker, 1972), emphasizes one feature of this
point of view: that language learning is no dif- language of the learner: that it is intermediate,
ferent in kind, from any other sort of cognitive in some way, between two languages. It is that
learning. aspect of it (when we find it has such features)
The second point is that language learning is a that is stressed by the term “interlanguage.”
creative activity- it is a process of discovering Then we have an alternative name, for exactly
some sort of regularity in the language data pre- the same object: “approximative system” (Nem-
sented to the learner. The learner creates a lan- ser, 1971), which stresses the changing, the
guage system for himself and what he has dynamic characteristics of a learner’s interlan-
created, or what he is continually creating, is, guage, the notion of approximation to some
both formally speaking, in terms of its structural goal, the notion of movement and change, but
properties, and functionally speaking, in terms particularly movement and change in a certain
of what it can be used for, a language. The direction, we hope. Finally, we have my own
learner possesses, then, something we can prop- term, which has not gained very wide currency,
erly call a language at any point during the “transitional competence” (Corder, 1967). This
course of his acquisition of the target language, similarly stresses the notion of change without
and this language can be used for a range of making any specific reference to the relation-
communicative purposes just as any other lan- ships that the learner’s language may have with
guage can. Since it has the structural properties any other languages he may know or be learning.
that we understand language to have, it is, in We can regard the learner’s language, his
principle, describable in the same way and by interlanguage as we may conveniently call it
means of the same theoretical models as any now, as a sort of hypothesis: It is the hypothesis
other language-unknown or known. The verbal which he has developed -creatively developed -
 THE LANGUAGE OF SECOND-LANGUAGE LEARNERS 411

as a result of interaction with linguistic data of children. He may be exposed merely to written
the language which he is learning-the target data in the classroom. He may be exposed to
language data. It is a hypothesis about the struc- other interlanguage speakers using this language
ture of the target language. If the learner is amongst themselves for communicative purposes.
asked what language he is speaking, he will, of In other words, the linguistic environment in
course, say that he is speaking, or at least trying which a learner is setting up his hypotheses about
to speak, the target language. We have a slightly the target language will-crucially, I think-
different point of view about what he is actually determine the nature of the hypotheses which he
doing. is going to come up with. Therefore, we would
In order to test this hypotheses of his- this is in expect to find quite different sorts of interlan-
no sense, of course, a conscious process-he guages developing in differing linguistic environ-
makes utterances which are generated by his ments. The third point is that the actual strat-
particular interlanguage grammar at a particu- egies that are employed by learners may be
lar moment. The behaviour of the teacher or various. I don’t think we can assume that there is
other speakers of the target language enables him only one strategy of language learning nor that
to decide whether any particular hypothesis he any learner has command of all possible strat-
has developed is valid or not. This does not egies. On the contrary, strategies of learning are
mean, of course, that everyone learning a lan- something that the learner has also learned.
guage will necessarily come up with the same hy- They are part of the equipment of cognitive
pothesis about a target language, because there structures he brings with him to the task. In
are a number of variable factors which deter- other words, there is something which we can
mine the nature of the hypothesis the learner will call a “personality factor” in language learning,
arrive at any particular moment. which determines how the learner sets about
First of all, following Piaget and other cogni- interacting with his linguistic environment. For
tive learning theorists, the sense we make of our example, the age of the learner will play a sig-
environment depends upon what we already nificant part in determining the sort of strategies
know about it; we see the world in the light of the that he is going to employ to discover the nature
knowledge we already have about it. In other of the data he is exposed to.
words, the existing cognitive structures which we Now, given these three factors, it would not be
possess -existing hypotheses we have about re- at all surprising, of course, if we found quite a
ality-condition the way we perceive and process fairly large range of different hypotheses about
new experience. In this particular instance, the the nature of the target language that learners
relevant existing cognitive structures may be invent. The range of possible hypotheses is, in
those of the mother tongue, any known dialects fact, very large and I think research does indi-
of the mother tongue or any other partially cate there is a good deal of variability in the
known languages. There are a whole host of pos- interlanguages which people learning a second
sible cognitive structures possessed by a learner language do, in fact, develop. The job, of
and clearly their nature will be related to his age: course, of the teacher is to be aware of the cogni-
the older a person is, the more languages he is tive structures which their learners have, and
likely to know, either partially, or as a native or also to find out the sort of strategies which they
near-native speaker. So that the range of possible adopt and then to adapt the environment; that
cognitive structures by means of which he assimi- is, to adapt the linguistic material to what he
lates this new experience will be different for knows about the preexisting knowledge and pre-
each individual learner and therefore we may ferred strategies of his pupils. The linguistic en-
expect that the hypothesis he develops about the vironment is the only factor in the learning situa-
target language will be conditioned by this fact. tion which is largely within the control of the
Another factor which will determine the na- teacher. Now, interlanguages are unstable; both
ture of his hypotheses is, of course, the sort of the term “transitional competence” and the term
linguistic data to which he is exposed. A learner “approximative system” stress the notion of in-
might be exposed to native speakers using their stability, and indeed instability is something that
language for communicative purposes, adults or we should wish to see in interlanguages, but not
412 S. PIT CORDER

an undirected or random instability, obviously. form to do more than one job where the native
The instability, that is, the dynamic nature of speaker uses it to do one job. But I think also-
interlanguages, must be goal-oriented. We do and this has been less thought about and cer-
not wish to see hypotheses being set u p by tainly less investigated- that the language of the
learners about the nature of the target language learner is functionally restricted. In other words,
which are uneconomic; in other words, caused the sort of messages, the sort of communicative
by faulty teaching or faulty materials. Learners functions which he can perform with that re-
unfortunately do set up hypotheses about the duced code are themselves restricted. He cannot
target language which have to be discarded use his interlanguage for all the communicative
later, and which serve no purpose in the learning needs which he has as a native speaker of his
process. We want the development of the inter- mother tongue. This leads, as I think many
language to be as direct as possible towards the teachers must recognize when they persuade
goal. That is, of course, where language teachers their pupils to actually try and use this interlan-
have a very considerable responsibility: to make guage for communicative purposes, to frustra-
sure that learners are not, in fact, developing tion on the part of the learner. What does he do
useless and uneconomic hypotheses about the when he becomes frustrated in his communica-
target language as a result of the teachers’ tech- tive desires? This wants looking into rather care-
niques and their materials. There is, of course, fully. A small piece of work has been done on
the interesting phenomenon which has been this by a Hungarian, Tamzis Vsradi (Vziradi,
called “fossilization”by Larry Selinker. This de- 1973), and I am sure there is a great deal of pos-
scribes the state of affairs in which people actu- sible research to be done here to find out what
ally stop learning and get stuck with the hy- strategies a learner adopts to overcome this func-
potheses they then have; in other words, finish up tional inadequacy when he really does require to
with an interlanguage code which is not the best communicate by means of his interlanguage. I
possible for purposes of interaction with native think it is worth mentioning at this point that an
speakers of the target language. It’s a matter of interlanguage, because of its formal and func-
considerable interest, and a matter for further tional restrictiveness, is a sort of pidgin. It is a
research to find out just under what conditions characteristic of pidgins that they are both func-
learning stops, either in adults or in children. tionally restricted and formally reduced. This is
This stresses the necessity for longitudinal studies the reason why some people have suggested,
of the development of approximative systems, to Richards amongst them (Richards, 1972), that
find out just in what way the interlanguage interlanguages are in that class of languages
develops in learners, and under what sort of con- amongst which pidgins are to be found. Both of
ditions this development does or does not take them develop in situations of language contact.
place. We want to be able to relate, in other This is something also which needs further inves-
words, the nature of the development to the tigation, to see just in what degree the properties
nature of the learning environment. of interlanguages and the properties of pidgins
Not only is the interlanguage unstable; it is are similar.
also a “reduced system,” in the sense that it is a One thing I want to draw attention to, be-
simplified grammatical code when compared cause it opens up a whole new area of investiga-
with either the mother tongue or the target tion which as far as I am aware has not yet been
language or any fully developed code for com- even begun, is that languages are not just beauti-
munication. One result of the reduced nature of fully structured systems, they are also used for
the interlanguage code is that learners tend to communicative purposes and it is becoming evi-
“over-indulge.” By this I mean that, given that dent that the way that we use the code is itself a
they have fewer rules, they tend to overuse these rule-governed type of behaviour. This would
rules in their desire to communicate. This ten- seem to be what is coming out of work in socio-
dency to overuse certain forms, i.e., to make a linguistics. A learner who tries to use his inter-
language form do more work than the native language for communicative purposes must have
speaker makes it do, means to say that there is a rules for doing so. In other words, he must have
semantic loss involved. The learner will use a a set of “rhetorical rules” or a *‘rhetorical sys-
 I -.
- C. %.r”
THE LANGUAGE OF SECOND-LANGUAGE L . E A m p & i @ m 413

tem,” or whatever name we have to give this; he gate what sort of rules learners use in order to
must not only be able to form sentences in his communicate by means of their interlanguage.
language, but actually use these sentences for In other words, if we want to go back to the old
some purpose in a n appropriate way. As far as I notion of “error analysis,” we may now ask what
know, no investigation has even begun into the sort of “errors of appropriacy” do the learners
inappropriateness of learner’s language. In other make besides the “errors of grammar” which
words, nobody has, as far as I know, attempted have been so much investigated in the past?
to describe what we might call the learner’s
“rhetorical system.” Now, it may be unreason-
REFERENCES
able to expect anyone to have done this, in view
Selinker, Larry, “Interlanguage,”IJ2AL 10, (1972), 209-231.
of the fact that we know so little about the nature Nemser, William, “Approximative Systems of Foreign Lan-
of rhetorical functions and how language is used guage Learners,” IRAL 9, (1971), 115-123.
to communicate anyway, but teachers are, of Vbradi. Tambs, “Strategies in Target Language Learner
course, perfectly well aware that learners lack Communication Message-Adjustment,” paper presented
this knowledge. Even when they seem to be able at the VI Conference of the Romanian English Lin-
guistics Project in Timiwara, (May 1973).
to control the target language code reasonably Levenston, E. A., “Over indulgence and Under-representa-
well, learners nevertheless do not use the lan- tion: Aspects of Mother-tongue Interference.” in Gerhard
guage like a native speaker. For want of a more Nickel (editor), Papers in Conlrmfiue Lznguistics, Cam-
explicit name this sort of knowledge has been bridge: Cambridge University Press, 1971.
called “Sprachgefiihl”: the learner knows the Richards, Jack, “Social Factors, Interlanguage and Lan-
guage Learning,” Language Learning, 22.2 (1972). 159-
code, but he can’t speak the language. I would 188.
like to suggest that there is now a new field of in- Corder, S. P., “The Significance of Learners’ Errors,”lRAL,
vestigation just opening up, which is to investi- 5.4 (1967). 161-170.

IN MEMORIAM

FERDINAND F. DI BARTOLO

PRESIDENT,
National Federation of Modern Language Teachers Associations, 1934-1936

BUSINESMANAGER,
The Modern Language Journal, 1938- 1946

Mr. Di Bartolo died on August 25, 1975. An esteemed foreign language teacher he was,
before retirement, associated with the Public Schools of Buffalo, New York.