Download as pdf or txt
Download as pdf or txt
You are on page 1of 16

System, Vol. 13. No. 1, pp. 9-24, 1985. 0346-251X185 $3.00+0.

00
Printed in Great Britain. 0 1985 Pergamon Press Ltd.

THE Ll = L2 HYPOTHESIS: A RECONSIDERATION

ROD ELLIS

School of Language Studies, Ealing College of Higher Education, London

The purpose of this article is to examine the arguments and the evidence in favour
and against the Ll = L2 hypothesis. First, however, it will be necessary to define
the nature of the hypothesis more narrowly. Then evidence for and against the
hypothesis will be reviewed in terms of the “product” and the “process” of
acquisition. The hypothesis will be explored in terms of current developments
in both FLA and SLA research, which have focused on the cognitive and socio-
cognitive basis of language acquisition. Finally the discussion will be widened to
consider the general distinction between formal and informal learning.

INTRODUCTION

These are several reasons why second language acquisition (SLA) research has paid so much
attention to investigating the similarities and differences between SLA and first language
acquisition (FLA). One reason is historical. Early SLA research was motivated by the need
to disprove the Contrastive Analysis Hypothesis, which stated that the errors that second
language (L2) learners produced could be predicted or explained by negative transfer from
their first language (Ll). If it could be shown that the process and product of SLA were
similar to FLA, then the Contrastive Analysis Hypothesis could be discredited. A second
reason is methodological. SLA researchers have borrowed many of the methods used in
FLA research, so a comparison was bound to occur. A third reason is the applied nature
of much SLA research. Because there was so little direct evidence about SLA to guide L2
teachers, extrapolation from theories of FLA took place (e.g. Cook, 1971).* It is not sur-
prising, therefore, that initial interest in SLA was directed to establishing to what extent
this extrapolation was justified. There were important theoretical, methodological and prac-
tical reasons for examining the Ll = L2 hypothesis.

DEFINING THE HYPOTHESIS

In its pure form the hypothesis states that, all other things except knowledge of language
being equal, FLA is the same as SLA. In practice, however, all other things are not equal.
In particular there is an age difference between Ll and L2 learners, which is marked when
the comparison is between child FLA and adult SLA. Age is a confounding factor, and
as Brown (1980) points out, this has potential implications across the physical, cognitive
and affective domains. If it is discovered that an adult learns a L2 differently from a child
learning a Ll this may be a reflection of differences in mental capacity, which changes
with age and which interacts with language learning processes.

9
10 ROD ELLIS

A study by Cook (1980) gives some idea of how cognitive-as opposed to linguistic-
processes produce differences between FLA and SLA. Cook distinguishes “speech pro-
cessing memory” and “primary (or short-term) memory”. He found that when the load
on memory was too great in the processing of relative clauses, native-speaking children,
native-speaking adults and foreign language learners used the same strategy of treating
the first noun phrase in the sentence as the subject and the first noun phrase after the verb
as the object, irrespective of whether this was in fact the case. In another experiment, design-
ed to establish the maximum number of digits that foreign language learners could repeat
in a L2, Cook was able to show that an adult learner behaves like a native-speaking adult
rather than like a child. On the basis of these experiments Cook concludes:

where the memory process depends on features of syntax the foreign adult is subject to the same
type of restrictions as the child; where the memory process is minimally dependent on language the
adult can transfer his memory capacity to the task substantially unimpaired. (p.39)

The general point that Cook makes is that the Ll = L2 hypothesis is justified “to the
extent that other attributes of the mind are not involved” (p.47). Thus where “speech pro-
cessing memory” is involved, SLA and FLA are the same, but where “primary memory”
is involved differences will arise as a result of general cognitive differences between children
and adults.

Another important point to consider in defining the hypothesis is the level at which it is
to be explored. Does the hypothesis mean that the language produced by Ll and L2 learners
will be the same or does it mean that the internal processes are the same? In other words
it is possible to approach the issue from the point of view of the product of learning or
the process of learning. In the case of the former an answer will be sought by examining
the nature of the errors produced in both types of learning or by comparing the orders
of acquisition of the same set of features. In the case of the latter an answer will be sought
by examining the strategies that learners use.

At first sight this distinction may not seem very useful, given that processes can only be
inferred from product. Similar products should represent similar processes, different pro-
ducts represent different processes. This need not be the case, however. A child learning
Ll English produces early negative utterances of the kind:

No like daddy.

A Spanish adult learning L2 English produces identical negative structures (Cazden et al.
1975). It does not follow, however, that the process is the same, for the adult may be
transferring the negative rule from Spanish (i.e. no + V). It could similarly be argued
that differences in language-learner language between FLA and SLA need not reflect dif-
ferences in processes. McLaughlin (1978) has argued that the real similarity between the
two types of acquisition lies in the fact that all learners use the strategy of “employing
what is known”. This raises the discussion to such a high level of generality that many
important differences can be overlooked. Nevertheless, the distinction between process and
product is a useful one and will be followed on the subsequent sections.

A third point relevant to defining the hypothesis lies in distinguishing a strong and a weak
form. The strong form claims complete identity in the product and process of FLA and
THE LI = 12 HYPOTHESIS: A RECONSIDERATION 11

SLA, while the weak form claims that they are similar. Thus, simply identifying differences
is not sufficient to negate the hypothesis entirely.

COMPARING THE PRODUCT

In this section the product of FLA and SLA will be compared in terms of evidence provid-
ed by Error Analysis, experiments and cross-sectional and longitudinal studies of the order
of acquisition.

1. Error Analysis
One of the major aims of Error Analysis has been to compare the rival claims of the Con-
trastive Analysis Hypothesis and the Ll = L2 hypothesis. The procedure has involved
identifying errors in a corpus of language-learner language and then classifying them in
terms of whether they are interlingual in source or traceable to other sources. Dulay and
Burt (1974a), for instance, identify four error sources:

1. Interference-like goofs-those that reflect native language structure and are not found in Ll acquisition
data of the TL.
2. Ll developmental goofs-those that do not reflect native language structure, but are found in Ll ac-
quisition data of the TL
3. Ambiguous goofs-those that can be categorized as either interference-like goofs or Ll developmental goofs
4. Unique goofs-those that do not reflect Ll structure, and also are not found on Ll acquisition data
of the TL

They provide examples of each type, some of which are reproduced below:

Source Error type (Spanish-English) Example

Interference Omission of obligatory ‘how’ in English, optional in Spanish I know to do all that.
Developmental Irregular plural treated as regular. He took her teeths off.
Ambiguous Wrong ‘no’ placement; ‘no/not’ distinguished
Unique Overuse of ‘do’ We do get no more book.

Dulay and Burt (1973) attempted to calculate the frequencies of the different types of er-
ror. They collected samples of speech from Spanish-speaking children and concluded that
85% of the errors were developmental, 3% interlingual and 12% unique.2 In other words,
they found little evidence of the learners’ Ll influencing the type of errors produced.

Not all researchers have concluded that the Ll plays such a minimal role. George (1972)
notes that one third of the errors that occurred in his foreign student’s theses could be
traced to their Ll. Ervin-Tripp (1970) attributes between one third and two thirds of French-
English errors in her study to the Ll. James (1980) reviews the various attempts to deter-
mine the proportion of interlingual errors and concludes: “It seems then that between a
third.and a half of learner errors may be caused by the Ll:L2 misfit.” (p.146).

Dulay and Burt’s (1973) estimate, therefore, is not compatible with the general finding.
There is considerable disagreement among researchers. This may be the result of differences
in the Ll:L2 relationships under investigation (e.g. the greater the linguistic magnitude
between languages, the greater the proportion of interlingual errors (Corder, 1981)3) or of
12 ROD ELLIS

differences in the way the data were collected (e.g. written data may show more inter-
lingual errors than spoken data) or of differences in the subjects (e.g. adult learners may
produce more interlingual errors than children) or of differences in the way in which the
data were processed.

The evidence from Error Analysis is far from conclusive. Although most researchers agree
that there are large numbers of errors which cannot be attributed to the Ll, they disagree
markedly in precisely what proportion. The problems of coding errors in terms of the
psycholinguistic source are substantial. It is doubtful whether Error Analysis is capable
of providing a clear answer.

2. Experimental evidence
There have been relatively few attempts to investigate the Ll = L2 hypothesis by experimen-
tal manipulation. One interesting attempt was made by Cook (1973). He replicated the
research undertaken in FLA by Carol Chomsky (1969). This consisted of examining learners’
ability to produce a number of “advanced” syntactical patterns- e.g. relative clauses and
patterns involving verb complements (“The wolf is happy to bite” vs. “The duck is hard
to bite”).4 Cook found that adult L2 learners made the same kind of mistakes in the use
of these patterns as did Ll learners and concluded that for these structures adult SLA did
not appear to be facilitated by the Ll. He pointed out that the adults must have known
from their Ll that sentences can have other sentences embedded in them (as in relative
clauses) but seemed unable to use this knowledge in processing such structures in a L2.

Cook’s study lends some support to the Ll = L2 hypothesis. It is a pity there have not
been more such studies.

3. Order of acquisition comparisons


There have been several attempts to compare the order of acquisition of grammatical
features in FLA and SLA. The data used have been obtained both cross-sectionally and
longitudinally.

Cross-sectional comparisons
The basis for the cross-sectional comparisons is the morpheme studies investigated by Dulay
and Burt. Dulay and Burt (1974b) collected data from Chinese and Spanish speaking children
aged 6 to 8 years. They computed the accuracy orders for 11 morphemes and used this
to make claims about the order of acquisition by aquating the “order of difficulty” with
the “order of acquisition”. They found that the order was the same for the Chinese and
Spanish children, suggesting that the children’s Lls did not play a major role. However,
when they compared the “acquisition order” obtained in their study with that obtained in
both longitudinal studies of Ll English (e.g. Brown, 1973) and cross-sectional studies of
Ll English (e.g. de Villiers and de Villiers, 1973), they found that they were different. Dulay
and Burt explained this difference in terms of a theory of “creative construction”. They
suggested that because there were cognitive differences between Ll and L2 learners, the
language-learner language they produced would not be the same. However, the L2 learner
does not fall back onto his L 1, but instead uses his linguistic and cognitive faculties to
reconstruct the L2 in a way not dissimilar to that observed in FLA.
THE Ll = L2 HYPOTHESIS: A RECONSIDERATION 13

Dulay and Burt’s study has been repeated many times with learners from different Ll
backgrounds, with adults instead of children and using alternative means of collecting data.
The results are mixed and require very careful interpretation, but the general findings are
more or less the same as those reported in Dulay and Burt (1974). If Dulay and Burt's
“creative constructive” explanation is accepted, then the view expressed earlier that the
Ll = L2 hypothesis is maintainable where purely linguistic processing is involved but not
where general cognitive factors become involved is lent some support.

The “morpheme studies” have recently come under considerable attack. Doubts have been
expressed about the way in which the data have been collected and in particular about
equating “accuracy” and “acquisition” orders (see Hatch, 1983 p. 48ff for a review of
the criticisms that have been made). It is not possible to accept the evidence provided by
these studies with confidence.

Longitudinal comparisons
Longitudinal research into the order of acquisition in SLA has largely been restricted to
negatives, interrogatives and sentence patterns. These will be considered separately and
comparisons with the order of development reported in FLA made.

a. Negatives. Table 1 below contrasts the order of development for negatives reported by
Klima and Bellugi (1966) for FLA with that reported for SLA by Hatch (1974). The
similarities outweigh the differences. The general direction in both FLA and SLA is from
external to internal negation and then towards the integration of the negative particle with
the auxiliary system. One difference is that Klima and Bellugi do not report any instances
of the use of the negative particle before a modal verb (e.g. “I no can swim”). Both FLA
and SLA researchers emphasize that the stages of development are not water-tight; there
is a carry-over of “old” rules into the next stage.

Table 1. The acauisition of neeatives in FLA and SLA comuared


FLA (based on Klima and Bellugi, 1966) SLA (based on Hatch, 1974)

External negation (i.e. declarative nucleus preceded 1. External negation (i.e. negative particle is
or followed by the negative particle) preposed)
e.g. No the sun shining e.g. No very good.
Wear mitten no. No you playing here.
Internal negation (i.e. the particles “no” and 2. Internal negation (i.e. the negative particle is
“not” are incorporated into the structure of the moved inside the utterance; “no” and “not” used
utterance; “can’t” and “don’t” are used as before auxiliary and main verb)
unanalysed units) e.g. Mariana not coming today
e.g. Don’t bite me. I no can swim.
1 not crying. Negative attachment to modals (i.e. the negative
Analysed aux + not (i.e. the particle “not” is used particle is attached to modal verbs)
with a range of auxiliaries that are marked for e.g. 1 can’t play this one.
tense and number; “no” no longer substitutes for I won’t go.
“not”) I don’t like.
e.g. Paul didn’t laugh. Do-support with negation (i.e. the particle “not”
He won’t cry. is attached to do-aux that is marked for tense and
number)
e.g. I didn’t said it.
He doesn’t know anything.
14 ROD ELLIS

b. Interrogatives. Table 2 contrasts the order of development for interrogatives in FLA


and SLA. Once again the similarities outweigh the differences. Initially learners make do
with intonation questions. A little later WH interrogatives appear without subject-verb
inversion (except in formulas5) and then inversion slowly spreads through both WH and
yes/no questions. Embedded questions appear first with inversion and later without. One
possible difference is that in FLA inversion occurs in yes/no questions before WH ques-
tions, while it occurs in both types simultaneously in SLA. As with negatives, each stage
is not clearly marked. Development consists of a gradual revision of one set of rules in
favour of another.

Table 2. The acquisition of interrogatives in FLA and SLA compared

FLA (based on Cazden, 1972) SLA (based on Ellis, 1982a)

1. One word utterances are used as questions i.e. a 1. An early ‘non-communicative’ stage when the
holophrastic stage. learner repeats the question he has been asked.
2. Intonation questions and some formulaic WH 2. Intonation questions and some formulaic WH
questions5 appear. questions5 appear.
3. Intonation questions become more complicated WH questions consisting of WH pronouns with
and productive WH questions without inversion a declarative nucleus appear.
occur. Inversion occurs in yes/no and WH questions. In
4. Auxiliary verbs are developed. Inversion occurs in some studies there is main verb inversion
yes/no questions but not in WH questions. e.g. Like you icecream?
5. Tag questions develop (‘huh’ initially) and ‘Be’ inversion precedes ‘do’ inversion.
inversion occurs in positive WH questions. Embedded questions first with inversion
6. Tag questions with an auxiliary verb appear and e.g. He asked me how do I do it.
there is inversion in negative WH questions. and later without
7. Embedded WH questions develop. e.g. He asked me how I do it.
appear.

c. Sentence patterns. FLA is characterized by a gradual growth in the overall complexity


of utterances. This has been traditionally measured in mean length of utterance (MLU).
Initially utterances are one or two morphemes in length and then are gradually expanded
to three and four morphemes and longer (Brown, 1973; Crystal, 1976). SLA is also
characterized by a gradual extension in utterance length, although the greater short-term
memory of L2 learners leads to a fuller use of formulaic utterances. It is for this reason that
MLU is not used as a measure of development in SLA (Larsen-Freeman, 1978). In child
L2 learners, however, a very similar pattern of development to that seen in FLA is evi-
dent. Pienemann (1980) illustrates the incremental nature of three Italian girls’ acquisition
of L2 German. Initially, utterances consist of single constituents, while two and then three
constituent utterances follow later. Pienemann argues, however, that the one word stage
is not analogous to FLA as it reflects a strategy of communication rather than lack of
cognitive development. The children economize by omitting those constituents of the stan-
dard sentence which are contextually or cotextually given. Greenfield and Smith (1976),
however, have argued that an identical process takes place in FLA (see p. 00).

Felix (1978) provides a comparison of the acquisition of sentence types in SLA and FLA.
He reports on the acquisition of L2 German by two English speaking children aged 4 and
7 years, He found that their early utterances were constrained by processing limitations
THE Ll = L2 HYPOTHESIS: A RECONSIDERATION 15

in the same way as those of Ll learners. During the first 2 to 3 months the children pro-
duced only three different multi-word utterance types; (1) sentence imitations, (2) noun
phrases, usually of the adjective + noun pattern and (3) copular sentences. Felix claimed
that the children acquired different sentence structures “in a distinct chronological order”.
He argued that this contrasts with FLA, where learners have been shown to produce a
multitude of different structures as early as the initial two-word stage. Thus, in comparison
with FLA the sentence patterns of the two L2 learners were “amazingly restricted”. This
might be because, whereas the Ll learner’s utterances express conceptual rather than gram-
matical relations, the L2 learner’s utterances are syntactically constructed because he is
already familiar with the syntactic organisation of language. According to Felix, therefore,
the initial stages of FLA and SLA are very different as a result of the previous knowledge
of language which the L2 learner brings to the task.

The longitudinal studies of SLA have provided much direct evidence for evaluating the
Ll = L2 hypothesis. In general the evidence from the study of negatives and interrogatives
gives support to the hypothesis, although the equivalence is not total. Wode (1976, 1978)
has shown that L2 English negatives and interrogatives can be influenced by the Ll when
there is a crucial similarity measure between the two languages.6 In other words, the
learner’s Ll influences the development route when the L2 learner feels that the L2 system
is sufficiently like the Ll system to justify transfer. The evidence from the study of sentence
patterns likewise suggests that, although there are strong similarities between FLA and
SLA, there are also differences which are the direct result of prior language learning. The
strong version of the Ll = L2 hypothesis is not tenable, but the weak version is supported
by a comparison of the product of the two acquisitional types.

COMPARING THE PROCESS

The “process” of language acquisition can be accounted for in a number of different ways.
In FLA research explanations for how children learn their mother tongue have been sought
in “operating principles” (Slobin, 1973), the pre-linguistic or cognitive basis of acquisi-
tion (e.g. Sinclair de Zwart, 1973), in functional accounts of development (e.g. Halliday,
1976) and the conversational origins of linguistics knowledge (e.g. Wells et al., 1981). These
explanations have only just begun to be explored in SLA research, however, so the follow-
ing comparisons must necessarily be tentative.

1. ‘‘Operating principles ”
Slobin (1973) suggests that the ways in which children process language are determined
largely by the fact that human language is produced and received in rapid temporal
sequence. Children need strategies for quickly programming and receiving messages and
also for storing and organising the linguistic system. He suggests the now well-known
“operating principles”:

A.Pay attention to the ends of words.


B. The phonological forms of words can be systematically modified.
C. Pay attention to the order of words and morphemes.
D. Avoid interruption and rearrangement of linguistic units.
16 ROD ELLIS

E. Underlying semantic relations should be marked overtly and clearly.


F. Avoid exceptions.
G. The use of grammatical markers should make semantic sense.

In so far as the short-term processing constraints are common to children and adults, ac-
cording to Slobin, it is not unreasonable to suggest that they are also applicable to SLA,
where learners are also faced with similar problems of storing and organising linguistic
knowledge. There have been only sporadic attempts to demonstrate this applicability,
however (e.g. Andersen, 1983). The discussion below will attempt to show how at least
some of Slobin’s operating principles can be used to account for the process of SLA.
The acquisition of negatives and interrogatives in FLA and SLA demonstrates the opera-
tion of several of the principles. Operating Principle D is evident in external negation, where
the learner attaches the negative particle to the beginning or the end of a declarative nucleus.
It is also evident in non-inverted WH questions, where a WH pronoun is attached to the
declarative nucleus. Operating Principle B is also evident in negation in the free use learners
make of negative particles-no, not, don’t-early on. Principle C is evident in the early
attention the SLA learner pays to word order. Indeed, if Felix (1978) is right and L2 learners
are syntactically sensitive right from the start, C is more important in SLA than in FLA.
Ervin-Tripp (1974) in an investigation of English children’s acquisition of L2 French
suggests that the importance the children attached to word order is a feature that is common
to FLA. She observes that the children interpreted utterances consisting of noun + verb +
noun as subject-verb-object sentences irrespective of morphological information denoting
the passive. Here is further evidence of Operating Principle C in SLA. However, it is
Principle F that is lent the strongest support in SLA research. Over-generalization of TL
rules is widely commented on in the Error Analysis literature. The avoidance of exceptions
is one of the main devices which the learner uses to simplify the learning task without
impairing his communicative effectiveness.
At the level of Operating Principles, then, there is some evidence to suggest that the Ll
= L2 hypothesis is correct. It is not clear, however, whether all of Slobin’s Principles can
be found in SLA. Wode’s (1980) observation that free morphemes such as articles and
prepositions tend to be learnt before bound morphemes such as tense inflections casts doubt
on Principle A. Also it is not clear how the L2 learner’s knowledge of his Ll affects the
nature and the working of the Operating Principles. The attractiveness of Slobin’s
framework rests in the possibility of being able to explain why some structures are learnt
before others in terms of general cognitive operations, which can also be observed in FLA.

2. The cognitive basis of language acquisition


Cognitive psychologists working in the Piagetian tradition have strongly challenged Chom-
sky’s (1966) assertion that language constitutes a faculty that is independent of other
cognitive systems. The cognitive position is clearly put by Sinclair de Zwart (1973):
.... it seems easier, and much more helpful, to suppose that the child brings to the task of acquiring
his mother tongue a set of universal cognitive structures which have been built up during the first year
of life and which provide enough assumptions about the nature of human language to enable the child
to begin to join the talking community about the age of 1’/2.

Here is the explanation for why the child’s first use of language reflects conceptual rather
than linguistic organisation.
THE Ll = L.2 HYPOTHESIS: A RECONSIDERATION 17

It follows from this that the child’s early language can be better handled within a semantic
than a grammatical framework. A number of researchers (e.g. Greenfield and Smith, 1976;
Wells, 1974) have turned to Fillmore’s (1968) case grammar to account for the semantic
organisation of early language in FLA. Below is a modified version of a description system
developed by Greenfield and Dent (1980) to describe the various semantic roles and rela-
tionships which appear in children’s speech:

A. Arguments
1. Agent (Ag): the animate instigator of an action-process
2. Object (0): the inanimate entity affected by the action process
3. Dative (D): the animate entity affected by the action-process. This can refer
to a recipient, beneficiary, possessor or experiencer of a
perception
4. Locative the location or direction of the action or state-process
(L):
B. Predicates
5. Action- the action which is performed by the agent and affects the object
Process (AP):
6. State- the state or condition that influences an object or dative
Process (SP):
7. Property state or condition of the object
of Object (PO):
8. Locative state of object which results from the locative action-process
State (LS):
9. Property state or condition of locative
of Locative (PL):
Example: I kicked the yellow ball into the swirling river.
Ag AP PO 0 LS PL L

Whereas a convincing case can be made out for the cognitive basis of FLA, it is difficult
to see how SLA can be explained in this way. In FLA the child is cognitively immature,
but in non-simultaneous SLA the learner has developed well past the sensori-motor stage.
It is still possible, however, that early SLA speech, or at least much of it, is semantically
organized. The L2 learner shares with the Ll learner the need to articulate meaning inten-
tions of a greater complexity than he is capable of, either because he simply lacks the
necessary language or because he cannot construct and execute plans sufficiently quickly.
All language learners need to communicate beyond their linguistic means. The solution
to this problem lies in reducing the complexity of propositions by omitting semantic roles
that can be inferred from the context. Thus Ellis (1982b) hypothesizes:
The L2 learner utilizes his knowledge of the conceptual organization of events and simplifies their
representation in the L2 according to the principles of informativeness. He operates a strategy of semantic
simplification. (p. 214/5)

Table 3 analyses a number of utterances produced by an Ll and three L2 learners7 in


terms of the semantic categories described on p. 11. It will be noticed that the Ll utterances
do not always adhere to the word order of the adult code, whereas the L2 utterances do,
but, this apart, there is a striking similarity in the kinds of propositional reduction evident
in both.
18 ROD ELLIS

Table 3. A semantic analysis of selected Ll and L2 learner utterances

Ll utterances L2 utterances

Utterance Context Analysis Utterance Context Analysis

Milk Moving towards a 0 Gun A boy is pointing 0


cup of milk a gun at him
Daddy Daddy was writing Ag AP You Describing action Ag AP
write down what she was writing of another pupil
saying
Finish Telling teacher his AP 0
Cutting Her mother was AP 0 book exercise book was
meat bringing a knife to full
cut some meat
Mariana Teacher had asked Ag AP
Lie down She was just about AP Ag no where Mariana was
Lwindi to lie down coming Someone had said AP LS L
Sit down Daddy had told her AP L Eating at she didn’t eat any
room not to leave room School meat
Want Chasing the cat SP 0 D I want Requesting a D SP 0
pussy colour coloured pencil
Lwindi
Blue Requesting a blue PO 0
Big chair Pointing at a chair PO 0 colour pencil
Daddy Watching daddy Ag 0 Berinda Describing what Ag 0
bread eat some bread book another pupil was
doing

This account of semantic simplification suggests that both FLA and SLA derive from the
learners’ knowledge of how reality can be segmented into perceptual categories which match
the semantic organization of language. It suggests that human cognition is at the root of
how language is used and learnt and provides strong grounds for supporting the Ll =
L2 hypothesis. However, the support is for the weak rather than the strong form of the
hypothesis. It would be counter-intuitive to argue that ail L2 learner utterances are
characterized by semantic simplification. Adults, in particular, have the capacity to pro-
duce fully formed and relative complex propositions almost from the onset of aquisition.
They too, however, are likely to resort to utterances that are semantically simplified when
under communicative pressure.

3. Functional accounts of language development


Since the middle of the 70s FLA research has been increasingly interested in how a child
discovers the meaning potential of language, in particular how he uses it to perform a variety
of interpersonal functions such as promising, asking permission, requesting help etc. Halli-
day (1978) has described this new perspective as an “interorganism” one and contrasts
it with the “intraorganism” perspective of the Chomskian era. Halliday argues that the
nature of language is determined by the uses to which it is put. It follows that in order
to understand the process of FLA it is necessary to examine how the child uses language
and, in so doing, internalizes a knowledge of the grammatical system.

Halliday (1975) offers his own account of one child’s (Nigel’s) FLA. Nigel began by develop-
ing a proto-language consisting of a limited number of sounds that were used to express
specific functions. The second stage of development is marked by two general language
THE Ll = LZ HYPOTHESIS: A RECONSIDERATION 19

functions, the Mathic (which is self-oriented) and the Pragmatic (which is other oriented).
This distinction is reflected in Nigel’s use of intonation- a falling intonation for the Mathic
function and a rising intonation for the Pragmatic. Eventually these macrofunctions develop
into the metafunctions of the adult system-the ideational, interpersonal and textual func-
tions. When Nigel reaches this stage he is able to map grammatical structures onto each
other, integrating categories belonging to all three metafunctions by means of appropriate
selections from the lexical and grammatical systems of the language. In this way Halliday
tries to show how the development of language is linked to the development of meaning
potential.

In SLA research there have been few studies of the relationship between the development
of pragmatic competence and the development of linguistic competence. The L2 learner
already has a fully developed pragmatic competence, and, it has been argued (James, 1981),
needs only to transfer this to the L2. There is, however, some evidence to suggest that
the L2 learner requires time to sort out the form-function relationships of the L2 (Wagner-
Gough, 1975; Eisenstein et al., 1982). An “interorganism” perspective may, therefore,
prove as insightful for understanding the process of SLA as it has been for FLA. It is
unlikely that L2 learners will pass through the same three stages as Nigel, but it is quite
likely that the acquisition of specific linguistic features is dependent on which interper-
sonal meanings the learner feels a need to encode.

Ellis (1982a) gives several examples of this. In the spontaneous speech of three classroom
learners he observed that the first instances of “can” occurred in utterances requesting
goods (e.g. “Can I have one yellow book, please?“). He suggests that the learners were
not so much mastering “can” as learning how to ask for objects they need in their class
work. Another example is the use of “don’t”. This was first used in all three learners’
speech to regulate the behaviour of other pupils. That is, it was likewise tied initially to
a specific function. SLA may be an outwards-inwards process in a way not entirely
different to that in FLA described by Halliday.

4. Learning through interactions


In the last few years strong claims have been made (e.g. by Scollon, 1976; Wells et al.,
1981) that the best account of how a child comes to master his mother tongue lies in the
conversational work that takes place between himself and a caretaker to maintain inter-
subjectivity in face to face communication. Scollon, for instance, has suggested that “ver-
tical construction” (i.e. constructions that are built up over two or more discourse turns)
is the developmental basis for horizontal constructions. The child is able to build more
complex utterances by imitating part of another speaker’s utterance and then adding to
it. The typical topic-comment structure of many early two-word utterances is the result
of exploiting the adult’s contribution to the conversation. Wells (1980) has been able to
show that adults’ speech aids FLA, but that it is not the syntactic properties that are
facilitative so much as the interactional features, in particular the preparedness of the adult
to allow the child to propose topics for joint attention and to help to sustain a topic by
comprehension checks and requests for clarification etc. Wells emphasises that FLA is an
“apprenticeship in meaning” which takes place through negotiating conversations.

SLA research has also recently begun to explore the role that conversational interaction
plays in language development. Hatch (1978a,b) has illustrated how linguistic competence
20 ROD ELLIS

can grow out of learning to participate in conversations. She has demonstrated how the
conversational roles in learner-native-speaker interaction constrain the types of discourse
structure that occur and that these, in turn, determine the frequency with which particular
grammatical rules are modelled for the learner in the input. According to Hatch it is the
special nature of the conversations in which learners participate which governs the course
of development. The role played by native-speaker speech adaptations has also been con-
sidered by Long (e.g. 1983). Long, like Wells, emphasizes the importance of interactional
adjustments rather than formal modifications. He, too, discusses the need for “negotia-
tion” to ensure that the L2 learner receives sufficient “comprehensible input”.

The Ll = L2 hypothesis is, perhaps, most acceptable from an interactional perspective.


Spontaneous conversations involving L2 learners do not look so very different from those
involving Ll learners, so if learning does occur through interaction, it is reasonable to ex-
pect the pattern of development in both to be very similar. However, the L2 learner-
especially the adult-has the ability to plan ahead what he is going to say, so that not all
conversations he takes part in will be spontaneous. When this occurs a different pattern
of use-and therefore of development- is to be expected. In this respect there is no com-
plete identity in the process of FLA and SLA. The distinction between “unplanned” and
“planned” discourse is part of a wider distinction between “informal” and “formal” lear-
ning, which is the central issue for understanding the extent to which SLA and FLA can
be equated.

INFORMAL AND FORMAL LEARNING

The view that has been built up in the foregoing discussion is that SLA strongly resembles
FLA in a number of important ways, but that there are other aspects in which the two
types of acquisition differ. The similarities and differences can be explained in terms of
a general distinction between “informal” and “formal” styles of learning.

The main characteristics of informal and formal learning are summarized in Table 4. The
essential difference lies in the fact that whereas informal learning is embedded within the
context of socially relevant action, formal learning occurs within a context of language
and symbolic activity (Bruner, 1966).
Table 4. Characteristics of formal and informal learning contrasted

Aspect Informal learning Formal learning

Type of Implicit knowledge (i.e. learner Explicit knowledge (i.e. learner is able
knowledge cannot articulate what he knows) to articulate what he knows)
Manner of Knowledge to be aquired is Knowledge to be acquired is present
presentation demonstrated to the learner rather through verbalization of rules and
than its nature articulated principles
Feedback Incidental opportunities occur for Feedback is often presented in terms of
the learner to receive feedback about evaluation of the leaner’s performance
the appropriateness of his/her as correct or incorrect
performance
Affective roles There is empathy and identification The emphasis is placed on what is
with persons and social roles rather taught rather than who is doing the
than appeal to subject matter teaching
THE LI = L2 HYPOTHESIS: A RECONSIDERATION 21

FLA almost invariably involves informal learning. The characteristics of adult-child


interaction are in tune with the manner of presentation, feedback and affective roles of
informal learning. Also attempts to teach children correct grammatical forms (e.g. Cazden,
1965) have not been successful.

In contrast much SLA involves formal learning. As D’Anglejan (1978) has noted, classroom
SLA consists of explicit teaching through the presentation of rules and careful correction.
D’Anglejan believes that classroom SLA is often less successful than untutored SLA, where
the learning that takes place is informal.

The contrast between FLA as informal learning and classroom SLA as formal learning
is also commented on by MacNamara (1973). He suggests that the success of “street” SLA
compared to classroom SLA is because the natural processes of learning, which are evi-
dent in the former-and in FLA-do not take place in the classroom. The teacher seldom
has anything to say to his pupils that is so important that they are eager to guess his mean-
ing. Likewise the pupil seldom has anything to say which is so urgent that he is prepared
to improvise with whatever communicative skills in the L2 he possesses.

The positions taken by both D’Anglejan and MacNamara are that, although FLA and
SLA are often different processes in that the former takes place through informal learning
and the latter through formal learning, they can be identical processes, when SLA involves
informal learning. D’Anglejan and MacNamara envisage two different processes for SLA,
one relating to informal learning and the other to formal learning. This distinction is also
at the root of Krashen’s (1981) distinction between “acquisition” and “learning”, where
“acquisition” takes place unconsciously through exposure to natural contexts of use and
“learning” is the result of formal study of the L2. Krashen (1976), however, points out
that the correlation between “acquisition” and informal environments and “learning” and
formal environments is not a complete one. He argues that the classroom can function
as an “intake environment” as well as a formal environment, while it is perfectly possible
that an adult learner in a natural setting will study the L2 as well as be exposed to it. The
notions of “informal” and “formal” learning, therefore, cannot be equated with those
of untutored and tutored SLA.

An important question relating to the informal/formal distinction is whether one type of


learning can influence the other, particularly whether formal learning can enhance or alter
the direction of informal learning. D’Anglejan argues that transfer of learning can take
place within the types of context in which the learning occurs but is not possible across
contexts. Thus, formal learning which takes place in one formal setting is available for
use in other formal settings but cannot be accessed in informal settings. Krashen (1981)
also argues that “acquisition” and “learning” are entirely separate processes. This posi-
tion, however, does not meet with general agreement. Strong arguments have been put
forward for believing that explicit learning does facilitate the implicit use of L2. Also the
distinction between “formal” and “informal” learning may not be an absolute one-
much learning may involve combinations of formal and informal characteristics.

From the point of view of the Ll = L2 hypothesis, however, the distinction between “in-
formal” and “formal” learning is a helpful one. When the conditions of learning are close
22 ROD ELLIS

to the informal end of the continuum (as they nearly always are for FLA), SLA and FLA
are likely to be very similar processes and the resulting language-learner language of a univer-
sal kind. When SLA is the result of formal learning-whether inside or outside the
classroom-the process of learning is likely to differ from that observed in FLA and the
resulting language-learner language to differ. The extent to which the Ll = L2 acquisition
hypothesis can be supported, therefore, will depend on the extent to which the L2 learner
engages in informal and formal learning.

NOTES:

1. Cook, for instance, argues that teaching techniques should be modelled on the way a mother speaks to a child.
2. Dulay and Burt’s 1973 study does not give the percentages of errors that were “ambiguous” (these were ex-
cluded from the analysis). Their 1974 study does not give frequencies of the four error types.
3. Although Corder considers interference more likely when the degree of difference between the two languages
is great, others have suggested the interference requires a degree of similarity between the Ll and L2 (see p. 10).
4. Chomsky found that children begin by assuming that all verb complements work in the same way and that
exceptions of the kind:
The duck is hard to bite (where “the duck” is the notional object of the verb “bite”)
are misinterpreted until the child is 8 years old or older
5. “Formulas” are units of language, usually well-formed, that are learnt as wholes and therefore have no con-
stituent structure. Examples of interrogative formulas are “What’s this?” and “How do you do?“.
6. Wode’s examples come from L2 English learnt by German speaking children. He noted examples of full-verb
inversion:
Go you to school?
which parallels the German pattern. Main verb inversion is not characteristic of L2 English when the learners’
Ll is not German, Norwegian, Swedish or Danish.
7. The learner was my own daughter, Lwindi, at the age of 2 years. The L2 learners were three children learning
English in a London Language Unit. They were aged 10, 11 and 13 years. All the utterances listed in Table 3
were produced in the classroom.

REFERENCES

ANDERSEN, R. (1983) How to build your own interlanguage in N easy steps: step 1, Paper given at TESOL
Convention, Toronto, Canada.
BROWN, H. D. (1980) Principles ofLonguuge Leurning ond Teuching, Prentice-Hall.
BROWN, R. (1973) A First Lunguuge: the Early Stages, Harvard University Press.
BRUNER, J. (1966) in Bruner, J., Oliver, R. and Greenfield, P. Studies in Cognitive growth, Wiley.
CAZDEN, C. (1965) Environmental assistance to the child’s acquisition of grammar, unpublished PhD disserta-
tion, Harvard University.
CAZDEN, C. (1972) Child Language and Educution, Holt, Rinehart & Winston.
CAZDEN, C., CANCINO, H., ROSANSKY, E. and SCHUMANN, J. (1975) Second language acquisition
sequences in children, adolescents and adults, Final Report, United States Department of Health, Education
and Welfare.
CHOMSKY, C. (1969) TheAcquisition of Syntuxin Childrenfrom 5 to IO, MIT Press.
CHOMSKY, N. (1966) Linguistica theory, In Mead, R. (ed.), Lunguuge Teaching: broader contexts, Northeast
Conference on the Teaching of Foreign Languages.
COOK, V. (1969) The analogy between first- and second language learning, IRAL 7, 207-216.
COOK, V. (1971) The analogy between first and second language learning, In: Lugton, E. and Heinle, C. (eds)
Toward u Cognitive Approuch to Second language Acquisition, Center for Curriculum Development Inc.
COOK, V. (1973) The comparison of languge development in native children and foreign adults, IRAL 11, 13-29.
COOK, V. (1980) Cognitive processes in second language learning. In Nehls, D. (ed.). Studies on language uc-
quisition. Julius Groos Verlag.
THE LI = L2 HYPOTHESIS: A RECONSIDERATION 23

CORDER, S.P. (1881) Language distance and the magnitude of the language learning task. In Corder, S.P. Er-
ror Analysis and Interlanguage, Oxford University Press.
CRYSTAL, D. (1976) Child Language Learning and Linguistics. Arnold.
D’ANGLEJAN, A. (1978) Language learning in and out of classrooms. In Richards, J. (ed.), Understanding
Second and Foreign Language Learning, Newbury House.
DE VILLIERS, J. and DE VILLIERS, P. (1973) A cross-sectional study of the acquisition of grammatical
morphemes in child speech, Journal of Psycholinguistic Research 2,267-278.
DULAY, H. and BURT, M. (1973) Should we teach children syntax?. Language Learning 23,245-258.
DULAY, H. and BURT, M. (1974a) You can’t learn without goofing: an analysis of children’s second language
‘errors’. In Richards, J. (Ed.), Error Analysis, Longman.
DIJLAY, H. and BURT, M. (1974b) Natural sequences in child second language acquisition, trmguage Learning,
24,37-53.
EISENSTEIN, M., BAILEY, N. and MADDEN, C. (1982) It takes two: contrasting tasks and constrasting
structures, TESOL Quarterly 16,3.
ELLIS, R. (1982a) Discourse processes in classroom second language development, unpublished PhD disserta-
tion, University of London.
ELLIS, R. (1982b) The origins of interlanguage, Applied Linguistics 3, 3.
ERVIN-TRIPP, S. (1970) Structure and process in language acquisition, in Alatis, J. (ed.), Monograph Series
on Languages and Linguistics.
ERVIN-TRIPP, S. (1974) Is second language learning like the first?, TESOL Quarter/y 8.2.11 l-127.
FELIX, S. (1978) Some differences between first and second language acquisition. In Waterson, N. and Snow,
C. (eds.), The Development of Communication, Wiley.
FILLMORE, C. (1968) The case for case. In Bach, E. and Harms, R. (eds.), Universals of Linguistic Theory,
Holt, Rinehart&Winston.
GEORGE, H. (1972) Common Errors in Language Learning Insightsfrom English, Newbury House.
GREENFIELD, P. and DENT, C. (1980) A developmental study of the communication of meaning: the role of
uncertainty and information. In Nelson, K. (ed.), Children’s language, Vol 2, Gardner Press.
GREENFIELD, P. and SMITH, J. (1976) The Structure of Communication, Academic Press.
HALLIDAY, M.A.K. (1976) Learning How to Mean, Arnold.
HALLIDAY, M. A. K. (1978) Language as a Social Semiotic, Arnold.
HATCH, E. (1974) Second language learning-universals. Working Papers on Bilingualism 3, l-18.
HATCH, E. (1978a) Discourse analysis and second language acquisition. In Hatch, E. (ed.) Second Language
Acquisition, Newbury House.
HATCH, E. (1978b) Discourse analysis, speech acts and second language acquisition. In Ritchie, W. (ed.) Se-
cond Language Acquisition Research, Academic Press.
HATCH, E. (1983) Psycholinguistics, Newbury House.
JAMES, C. (1980) ContrastiveAnalysis, Longman.
JAMES, C. (1981) The transfer of communicative competence, in Fisiak, J.(ed.) Contrastive Linguistics and
the Language Teacher, Pergamon Press.
KLIMA, E. and BELLUGI, U. (1966) Syntactic regularities in the speech of children, in Lyons, J. and Wales, R.
(eds.), Psycholinguistic Papers, Edinburgh University Press.
KRASHEN, S. (1976) Formal and informal linguistic environments in language acquisition and language learn-
ing. TESOL Quarter/y 10, (2), 157-168.
KRASHEN, S. (I 98 1) Second Language Acquisition and Second Language Learning, Pergamon Press.
LARSEN-FREEMAN, D. (1978), Evidence of the need for a second language index of development. In Ritchie,
W. (ed.), Second Language Acquisition Research, Academic Press.
LONG, M. (1983) Native speaker/non native speaker conversation and the negotiation of comprehensible input.
Applied Linguistics 4, (2), 126-141.
MCLAUGHLIN, B. (1978) Second Language Acquisition in Childhood, Lawrence Erlbaum.
MACNAMARA, J. (1973) Nurseries, streets and classrooms: some comparisons and deductions. Modern Language
Journal 57. 250-254.
24 ROD ELLIS

PIENEMANN, M. (1980) The second language acquisition of immigrant children. In Felix, S. (ed.). Second
Language Development, Gunter Narr Verlag.
SCOLLON, R. (1976), Conversations with a one year old. University of Hawaii.
SINCLAIR DE ZWART, H. (1973) Language
acquisition and cognitive development. In Moore, T. (ed.),
Cognitive Development and the Acquisition of Language, AcademicPress.
SLOBIN, D. (1973), Cognitive prerequisites for the development of grammar. In Ferguson. C and Slobin, D.
(eds.), Studies of Child Language Development, Holt, Rinehart & Winston.
WAGNER-GOUGH, _I. (1975) Comparative studies in second language learning, CAL-ERIC/CLL Series on
Languages and Linguistics, 26.
WELLS, G. (1974) Learning to code experience through language. Journal ofChild Language, 1, (2). 243-269.
WELLS, G. (1980) Apprenticeship in meaning. In Nelson, K. (ed.), Children’s language, Vol. 1, Gardner Press.
WELLS, G. et al. (1981) Learning through Interaction, Cambridge University Press.
WODE, H. (1976) Development sequences in naturalistic L2 acquisition. Working Papers on Bilinguaksm 11, 1-3 1.
WODE, H. (1978) The Ll vs L2 acquisition of English interrogation. Working Papers on Bilingualism 15, 37-57.
WODE, H. (1980) Learning a Second Language I. An Integrated View of Language Acquisition, Gunter Narr
Verlag.

You might also like