[Draft version] KANTÉ I., In Y. Rolland et al.(eds) 2015.

Heritage and Exchanges: Multilingual and intercultural approaches in

training context, pp 25-42. UK: Cambridge Scholars Publishing.

Contrastive Linguistics: A “double” interface in L2/L3 acquisition?*

Issa KANTÉ, University of Reunion

Abstract
This paper claims that modern Contrastive Linguistics can be considered as a “double” interface between theory and
application and between various linguistic approaches. It shows how empirical problems, hindering the original
contrastive enterprise, led to the emergence of Error and Interlanguage analyses, which primarily focus on actual
problems occurring in language acquisition and use the findings to solve them. One of the key questions is whether
one can deal with learners’ errors in ignoring their L1. The paper carries out a historical outline demonstrating the
interconnections between Contrastive Linguistics and Error/ Interlanguage Analyses and their individual limitations.
Therefore, it argues that, despite their apparent opposition, they are rather complementary fields, with converging
points and common goals to second or third language acquisition. The study also examines how corpus-based
approach emerges as a converging empirical ground to these fields.
Keywords: First/second language, learner, corpora, Contrastive Linguistics, Error and Interlanguage Analyses.
Introduction
This paper aims to demonstrate that, in its current form, Contrastive Linguistics (CL) can be perceived as a
“double” interface between theory and application and between various linguistic fields (Error Analysis, Interlanguage
Analysis, Language Typology and Translation Studies). Sections one and two explore how empirical problems
hindering the original contrastive enterprise led to the emergence of Error Analysis, which primarily focuses on what
problems actually occur in second or third language (L2/L3) acquisition and then uses findings to solve them. The
recurrent question has been whether one can deal with learners’ errors in ignoring their mother tongue (L1). Not long
ago, some pedagogical approaches would ban any recourse to L1 in L2 or L3 teaching. Over the last decades, many
studies have re-established “interference/transfer” as a major factor in language acquisition.
Based on the interconnections between Contrastive Analysis, Error Analysis and Interlanguage Analysis and their
distinctive limits, we postulate in section three that, despite their apparent opposition, they are rather complementary
approaches of CL, with converging points and common goals regarding L2/L3 teaching and learning. For this
perspective, section four examines how CL can be regarded as a “double” interface between theoretical and empirical
linguistic studies. Furthermore, we underline how the corpus-based approach provides a more solid empirical base to
contrastive studies than an intuition-based one and thus has enabled the consolidation of CL approaches.
1. The historical background of contrastive studies
Contrastive Analysis (CA) can be defined as a sub-discipline of linguistics, which fosters a systematic comparison
of languages in order to emphasize and describe their similarities and differences. According to Gast (2011), such a
definition can be considered in a narrow sense, where CA is viewed as a branch of Comparative Linguistics, which
compares pairs of “socio-culturally1 linked” languages. This definition restricts the scope of CA to the comparison of
two closed languages (Spanish and Basque, for instance). It can also be viewed in a broad sense, where CA compares
two or more languages which need not be socio-culturally or genetically related. In his book Contrasting Languages.
The Scope of Contrastive Linguistics, Krzeszowski (1990: 10) defines CA in this broad sense. According to him, the
origin of contrastive studies can be traced back to the 11th century A.D. with Aelfric’s Grammar of Latin for native
English learners, with the implicit assumption that knowing the grammar of one language may facilitate the learning
of another one. However, in its current form, most scholars agree that modern CA began in the mid-1940s. As Fisiak
(1981a: 215) puts it, the Second World War aroused a great interest in foreign language learning, whereby CA
emerged as an important part of foreign language teaching methodology. The two prominent scholars were then
Charles Fries and Robert Lado, who sketched the CA program as follows:
The most efficient materials are those that are based upon a scientific description of the language to be learned,
carefully compared with a parallel description of the native language of the learner, (Fries 1945: 9).
The plan of the book rests on the assumption that we can predict and describe the patterns which will cause
difficulty in learning, and those that will not cause difficulty, by comparing systematically the language and
culture to be learned with the native language and culture of the student, (Lado 1957: p.vii, Preface).

*
Special thanks to Phil Wade and Eileen Williams-Wanquet for their advice and perceptive comments on earlier versions
of this paper. Any remaining errors are the author’s responsibility.
1. “Two languages can be said to be socio-culturally linked when (i) they are used by a considerable number of bi- or
multilingual speakers, and/or (ii) a substantial amount of ‘linguistic output’ (text, discourse) is translated from one language
into the other. According to this definition, contrastive linguistics deals with pairs of languages such as Spanish and Basque,
but not with Latin and (the Australian language) Dyirbal, as there is no socio-cultural link between these languages.” (Gast
2011)
1
 [Draft version] KANTÉ I., In Y. Rolland et al.(eds) 2015. Heritage and Exchanges: Multilingual and intercultural approaches in
training context, pp 25-42. UK: Cambridge Scholars Publishing.

Both these quotes suggest that, in its original form, CA had two main concerns: (i) compare learners’ mother
tongue (L1) and the target language (L2) and (ii) predict areas of difficulty. Robert Lado, who was influenced by Fries
but also by his personal background,2 strongly advocates a systematic comparison of L1 and L2 and their cultures. The
underling idea is that if L1 and L2 are similar, there will be few or no learning difficulties – it can even be beneficial
for students who may produce correct patterns in L2, thus generating a “positive transfer.”3 But if L1 and L2 are
different, then learning difficulties are to be expected, learners will produce erroneous L2 patterns, resulting in a
“negative transfer.” Once the areas of difficulty are diagnosed, teaching materials should be tailored based on learners’
needs according to their L1 and culture and the L2 they are learning. In this perspective, CA was clearly designed to
be an applied enterprise, and the so-called “strong” or “predictive” CA hypothesis was born. Nevertheless, one should
bear in mind, as Fisiak (1981a: 215) reminds us, that before the mid-1940s, the early published contrastive studies
where predominantly theoretical – the applied aspect was rather peripheral.
2. The evolution and different phases of CL
2.1 Early Contrastive Analysis hypothesis
Over decades, CA has gone through different periods of “success-decline-success,” to use Granger’s (2003: 17)
term. From the 1940s up to the 1970s, its main agenda and goals were to:
- describe and compare learner’s L1 and L2,
- provide insight into similarities and differences,
- explain and predict problems in L2 learning,
- and develop course materials for language teaching.
Following Lado (1957), some scholars such as Banathy, Trager and Waddle (1966) advocate a strong version of
CA in stating that:
the change that has to take place in the language behavior of a foreign language student can be equated with the
differences between the structure of the student’s native language and culture and that of the target language
and culture. The task of the linguist, the cultural anthropologist, and the sociologist is to identify these
differences. The task of the writer of a foreign language teaching program is to develop materials which will be
based on a statement of these differences; the task of the foreign language teacher is to be aware of these
differences and to be prepared to teach them; the task of the student is to learn them. (p. 37)
As noted above, the purpose of CA in comparing the L1 and L2 is to predict the potential difficulties that learners
could face in L2 acquisition, and devise teaching materials to tackle these difficulties. Proponents of this approach
claim that interference from learners’ L1 is the prime cause, or even the unique cause of difficulty in L2 learning, as
clearly stated in the quote above. In reaction to this standpoint, Wardhaugh (1970: 124) rightly argues that such an
extreme version “is quite unrealistic and impracticable,” while Johansson (2008) observes that there are two main
reasons to the rejection of the early CA hypothesis. On the one hand, some contrastive linguists had exaggerated some
of their claims (strong version), and on the other hand some teachers had expected too much from this new field. As a
result, this challenge led to the development of Error Analysis (EA), which marks the first shift in CL evolution, from
a strong CA Hypothesis to a weak EA Hypothesis and later to Interlanguage Analysis (IA). These successive shifts
had then rhymed with a latent decline of CA facing virulent criticisms. CA was/is mainly criticized for
overemphasizing the role of transfer as a source of errors. As a matter of fact, many difficulties predicted by CA do
not show up in actual learner performance and many errors occurring in learner performance are not predicted by CA.
The underlying idea is that L1 interference is only one of the sources of errors, many problems are not language
specific and can be related to psychological, pedagogical, as well as other extra-linguistic factors. For example,
learners use analogical replacement or overgeneralization of L2 rules that they have only partially learnt, such as the
erroneous use of the -ed form in *he comed, instead of the irregular form he came.
Although transfer is not the single cause of language learning difficulties, it can however hardly be denied as a
major one. On the basis of increasing quantities of L2 data, recent studies4 have determined the phenomenon of
transfer as a major interlanguage factor that need be considered in L2/L3 acquisition process. As Krzeszowski (1990:
64) points out, learner’s L1 is crucial to understand an error like *these his children (ill-formed in English) versus te
jego dzieci (well-formed in Polish). Similarly, the influence of learner’s L1 is noticeable in the following sentence of a
French student learning English as L2.
... some people (er) conduce conduces us to a an hospital and (em) (er) the moral (er) in (er) that history is
that... [LONGDALE_LEARNER-CORPUS].

2. “Lado himself was an English and Spanish bilingual, who was born in America of Spanish parents, grew up in Spain and
then went to college in the USA.” (Cf. Lennon 2008: 51)
3. Notice that some recent studies in L1/L2 acquisition, such as Narcy-Combes (2014) and Souliou (2014), rather view the
phenomenon of “transfer” as a cognitive “code activation” in the use of L2 or L3, instead of an actual “transfer” or
“interference” between these languages and the speaker L1.
4. Cf. James (1998) Gilquin (2008) and Granger (2013), among others.
2
 [Draft version] KANTÉ I., In Y. Rolland et al.(eds) 2015. Heritage and Exchanges: Multilingual and intercultural approaches in
training context, pp 25-42. UK: Cambridge Scholars Publishing.

The errors in this sentence are related to different grammatical levels (i.e. lexical, syntactic and phonological levels),
yet they have a common origin, which is the learner’s L1. As for the lexicon, instead of using the noun story in
English, the speaker uses history, which is unexpected in this context. This word has probably been employed due to
the influence of the French word histoire, which has two equivalents in English: history and story. Contrary to French,
in English the word story, which can have a modalization effect and an intra-linguistic reference5, is opposed to
history essentially referring to past events or a discipline. At the lexical, but also syntactic, level the learner employs
the verb conduce6 (with a Subject - Verb agreement error) where he/she should rather use take or drive. Again, one
can see the influence of the French verb conduire (quelqu’un quelque part = to take/drive someone somewhere).
Finally, there is another error deriving from a phonological interference of the learner’s L1 in which the initial h of
hopital is silent in French, contrary to English. As a result, the learner’s hesitation and his/her use of an can be
attributed to the L1 where there is a linking between the article un and the noun hospital (un hopital).
Beyond the fact that interference seems to be a key factor in L2 acquisition and in response to criticism regarding
error prediction and their non-occurrence, many studies7 substantiate that the non-occurrence of an error does not
invalidate the prediction. This phenomenon may even confirm the predictive power of CA in the sense that it can
provide evidence that learner avoid problematic structures and areas of difficulties. Fisiak (1981b: 7) argues that “the
value and importance of Contrastive Analysis lies in its ability to indicate potential areas of interference and errors.”
Nonetheless, if many errors occur in learners’ performance, which are not predicted by CA, we should assume that all
learning problems are not language specific. Therefore CA has to be viewed as one component among many other L2
methodologies. Facing this reality, the original CA hypothesis went through a decline in the 1970s and saw the
emergence of other approaches such as Error analysis (EA) and Interlanguage Analysis (IA) to address language
learning difficulties.
2.2 The era of EA hypothesis or the “first revival” of CA
As the strong CA hypothesis began to wane, a weaker version known as EA flourished in the 1970s. Pit Corder
(1967 – in “The significance of learners’ errors”) was one of the pioneers of this new approach, which can be defined
as a type of linguistic analysis that focuses on the errors learners make, in comparing the errors made in the target
language and the constructions of that target language, (cf. Khansir 2012: 1029). The original claim of EA was that,
instead of predicting learners’ difficulties, one should rather observe what problems actually occur through a
systematic study of learners’ errors. Once learning difficulties are revealed and analyzed based on learners’
productions, the results can be used to improve language teaching. Its main goals can be summed up as follows:
- identification of the areas of learning difficulty through error analysis,
- devising remedial measures, for instance, in preparing sequences of target language items in class rooms and text
books with the difficult items first and then the easier ones,
- establishing the relative degree of emphasis, explanation and practice required for the various items in L2/L3. (cf.
Khansir 2012)
Its methodology, as elaborated by Corder (1974), consists of five major steps: (i) the selection of a corpus of
language (L2 or L3 production of students), (ii) the identification of errors in the corpus, (iii) the classification of
identified errors, (iv) the explanation of the causes of the errors, and (v) the evaluation of the errors. Based on this new
orientation, instead of a pure predictive role, CA is assigned an explanatory role for language acquisition. In addition
to interlanguage errors, which had been the main or even unique focus of CA, EA has the advantage to deal with many
other types of errors. It brings to light intra-language and extra-linguistic errors related to overgeneralization or
teaching and learning strategies. The creed has since been this: EA focuses on attested errors from authentic data,
while CA deals with competence and theoretical issues8. Yet, as the latter, the former does have limitations.
According to Johansson (2008: 114), EA hypothesis can be criticized on both its methodological procedures and
theoretical problems which are mainly related to difficulties of identifying, quantifying, and explaining errors. Firstly,
the ways learner corpora are built are still diverging from one study to another (nature of data collection and the
selection of informants). Therefore, trying to draw statistically significant findings from such samples may be a
questionable practice. Secondly, Dulay et al. (1982: 141-143) points out that some EA studies tend to confuse
explanatory and descriptive aspects, viz. they confuse “the process and the product.” For instance, instead of
explaining why an error has occurred, some EA studies essentially classify errors into different categories and give
their frequency.
Furthermore, according to Schachter and Murcia (1977), the analysis of errors in isolation focuses on errors and
hence excludes the rest of the data from consideration. As EA concentrates too much on errors, it investigates less the
whole performance of the learner. Therefore, in this perspective one can claim that EA too is not in a position to
identify all the learning difficulties. Strong CA proponents might argue here that learners usually avoid the most

5. For a detailed account of these notions, see Ballier (2007) and Kanté (2011).
6. Even if the verb conduce exists in English, it is not used with the meaning “to take/drive someone somewhere.” It rather
means “to lead/contribute to a result or an effect.”
7. Among which James (1971), Schachter (1974) and Fisiak (1981b).
8. Cf. Krzeszowski (1990: 190).
3
 [Draft version] KANTÉ I., In Y. Rolland et al.(eds) 2015. Heritage and Exchanges: Multilingual and intercultural approaches in
training context, pp 25-42. UK: Cambridge Scholars Publishing.

difficult points and those difficulties cannot be revealed by EA if the data in hand do not include certain structures. For
example, if learners do not produce any passive verb forms, EA cannot state whether they have mastered the passive
or not. Interestingly, Gilquin (2008: 5-6), following Granger (1997 and 1998) examining the use of the passive form in
learners discourse, points out that in comparing varieties of learner English, it turns out that the underuse of the
passive is not limited to French-speaking learners or Swedish- and Finnish-speaking learners as in Granger (1997).
The phenomenon occurs in the production of learners from various L1 backgrounds. Therefore, Gilquin (2008)
postulates that the underuse of the passive seems to be a universal feature of interlanguage, in the sense that learners
generally prefer unmarked forms than marked ones. The hypothesis is that the tendency may even be reinforced when
the corresponding structure in the learners’ mother tongue is much less frequent than in English.
Another interesting case is the use of relative clauses. In her study, Schachter (1974) demonstrates that learners
whose native language has relative clauses (Persian and Arabic on the one hand) made more errors in relative clauses
than learners whose L1 does not “formally” have relative clauses (Japanese and Chinese on the other hand). However,
the analysis also reveals that, in the data, the Japanese and Chinese learners use fewer relative clauses than Arabic-
speaking and Persian-speaking learners. So here again, one can clearly see that errors may not tell the whole story,
avoidance will always be looming in learner interlanguage, as Lennon (2008) puts it. This claim gives credit to the
idea that one does need to take into account the whole performance of learners in comparison to what native speakers
do in a similar situation. It also shows that CA should be used hand in hand with EA if we aim at fully explaining
learners’ errors. In other words, no single approach/hypothesis, strong or weak, can tackle all the issues encountered in
L2 or L3 learning.
In this regard, since L1-L2 transfer has been recognized as a significant factor in language acquisition, CA
regained credibility thanks to its predictive power and explanatory role. Eventually, the apparent limits of EA have
given birth to the development of a new approach known as Interlanguage Analysis (IA) Hypothesis. As Granger
(2003: 18) puts it, the globalization of society led to an increased awareness of the importance of interlingual and
intercultural communication. Here begins what we call the second phase of CA revival.
2.3 The era of IA hypothesis or the “second revival” of CA
In his famous paper “Interlanguage,” Selinker (1972) builds on the error analysis approach of Corder (1967) and
claims that the language of L2 learners is itself a linguistic system independent of either L1 or L2, although influenced
by both. The term “interlanguage” refers to a language intermediate between the native and the target language,
(Lennon 2008). The underlying assumption is that the learner’s language can be considered as a distinct language
variety or system with its own characteristics and rules: it is a dynamic system moving toward the target language. The
focus of IA is to study this system in focusing on the process of learning from/between L1 to/and L2, but takes into
account the learner’s L1. While the original CA hypothesis is primarily concerned with the comparison of L1 and L2,
EA and IA hypotheses, the latter more than the former, are more concerned with what language learners actually do
throughout the learning process.
One of the goals of IA is to find out the features of nativeness and non-nativeness in the learner language. It
focuses on the learning process, in comparing L1 to L2 or L3, which gives an insight into the learner’s interlanguage
developing in the course of learning. As Johansson (2008) points out, IA studies address questions such as “where
does the learner go wrong?” But also “what does the learner get right?” and, most important, “what is the nature of
language learning?” and “how can we best help learners on their way to the target language?” Nonetheless, despite its
apparent completeness, criticisms9 underline that IA mainly concentrates on morpho-syntactic development of
interlanguage; it has failed to deal with semantic development. More importantly, IA studies have also failed to clearly
define the concept of “interlanguage.” For instance, the terminological and conceptual instability is perceptible in
Lennon (2008) when he examines the terms used to describe the phenomenon of “interlanguage.” Historically, the
following concepts have been employed: “transitional competence,” “approximative system” and “idiosyncratic
dialect.” Nowadays the terms “interlanguage” and “learner language” are privileged. Moreover, in IA procedures,
Spolsky (1989) suspects an inconsistency between a process and a competence model, viz. some studies employ a
competence terminology while carrying out a processing model.
To summarize the evolution of CL, one can notice that from the original CA hypothesis where the focus is on
learning difficulties to be predicted (predictive approach), research moves to a diagnostic approach to figure out
learner errors through learner productions, and then to a new approach where the whole learning process is under
scrutiny. Put differently, CA compares L1 with L2, while EA and IA, with different analytic methods, focus on what
language learners do in the learning process. So far, the question can be asked whether these three analytic methods
are fundamentally opposed or rather complementary approaches of Contrastive Linguistics (CL).
3. Are CA, EA and IA complementary fields of CL?
On the basis of the above-mentioned interconnections between CA, EA and IA, coupled with their methodological
limitations, one can claim that these three fields are complementary, with converging points and common goals to
L2/L3 acquisition. They should be regarded as different approaches toward the same purpose. In its applied

9. Cf. Spolsky (1989) and Jie (2008).

4
 [Draft version] KANTÉ I., In Y. Rolland et al.(eds) 2015. Heritage and Exchanges: Multilingual and intercultural approaches in
training context, pp 25-42. UK: Cambridge Scholars Publishing.

orientation, CA starts with a comparison of two languages and predicts areas of difficulty for L2 learners, while EA
and IA start the opposite way with learners’ data first and then study the types and frequency of errors and their
sources (i.e. inter-lingual, intra-lingual or simply extra-linguistics errors). Contrary to what is often taken for granted10,
these strategies are not mutually exclusive; they are rather converging methodologies. Put differently, the three fields
should be considered as different phases of the same continuum of CL with CA hypothesis and IA hypothesis standing
at the two extremes, as Figure 1 illustrates.

Figure 1 : Phases of CL evolution.

In their procedures, CA contrasts the learner’s L1 and the target language (L2 or L3), EA compares the learner’s
interlanguage with the target language, while IA takes all three systems into account, that is to say L1, interlanguage
and target language (L2 or L3). Thus, the elements of the continuum are viewed as different fields of modern CL, yet
with diverging methodologies and focus, but complementary and mutually reinforcing. In these last decades, many
studies, basically in corpus linguistics, have supported this view. Scholars such as Granger (1996 and 2013), Jarvis
(2000), Gilquin (2008), among others, pinpoint the necessity for CL to combine the strong and weak version of CA
into what they call the “Integrated Contrastive Model,” henceforth ICM. The backbone of this approach is a process
based on the comparison of learner data with a reference native corpus (Native Language, NL vs. Interlanguage, IL)
on one side, with data produced by learners from different mother tongues and culture on the other (IL vs. IL). If the
analysis reveals L1-specific errors, L1-L2 contrastive analysis is then needed for explanations (cf. section 4.2).
Finally, when it comes to error explanation, there appears another converging point between the three fields.
Though EA and IA are not primarily concerned with theoretical problems (such as the issue of “equivalence” in CA),
they do need a theoretical background in order to characterize and explain errors. Thereby, CA can bring insight into
both learners’ L1 and L2. As Halliday (1964) suggests, once the errors have been collected, the analysis needs to be
done in two ways: descriptively or comparatively. Although Halliday endorses the descriptive method11, he also
indicates that if the teacher believes that the only cause of the error is due to interference, the error can then be
explained “comparatively.” On the basis of this vision, the interconnection between the practical and theoretical
aspects of CA, EA and IA stands as another converging point, which raises the question whether CL is an interface
between these fields.
4. Contrastive linguistics as a “double” interface
4.1 Between theory and application
As we have seen above, contrastive studies in the mid-40s through the late 60s were originally pedagogically
oriented. Starting in the 80s, instead of doing a pure applied research, some CL studies have taken up a more
theoretical orientation. Thus, Fisiak (1981b: 2-3) distinguishes theoretical contrastive studies from applied contrastive
studies when he states that the former are performed for their own sake, while applied ones are performed for some
pedagogical application. As a theoretical enterprise, CL systematically compares two or more languages in order to
give an exhaustive account of the differences and similarities between them. It also aims at providing and describing
an adequate model for language comparison (to give a framework of comparable areas, correspondence, equivalence.
As a consequence, theoretical CL deals with morpho-syntactic considerations such as formal/structural congruence /

10. Cf. Dyson (2010) for example.

.
11. According to him this method yields a simpler correction and can be used in language classes with students from
different backgrounds.
5
 [Draft version] KANTÉ I., In Y. Rolland et al.(eds) 2015. Heritage and Exchanges: Multilingual and intercultural approaches in
training context, pp 25-42. UK: Cambridge Scholars Publishing.

correspondence, semantic and phonological issues of equivalence or comparability. It seeks to isolate systems,
subsystems, constructions or rules to be compared with their cross-linguistic “equivalents” (cf. Krzeszowski 1990).
Another key theoretical issue is the notion of linguistic universal categories and how they are realized in the contrasted
languages (cf. Fisiak 1981b).
As an applied enterprise, CL is related to language teaching and the development of teaching/learning materials.
According to Fisiak (1981b), applied CL studies are often inspired by the findings of theoretical CL ones and select
whatever information is necessary for a specific purpose. The former devotes more attention to surface representations
than theoretical issues. Furthermore, as part of Applied Linguistics, applied CL is inspired by several other disciplines,
including theoretical, descriptive and comparative linguistics, psycholinguistics, sociolinguistics, and psychology of
learning and teaching (cf. Krzeszowski 1990).
Considering the scopes of the two approaches, we can notice that the interest in language comparison extends
beyond practical applications since it pinpoints what is general to languages and what is specific to a particular
language. Therefore, CL (theoretical or applied) contributes not only to the understanding and learning of languages in
general, but also to the study of individual languages it compares. In this regards, CL stands as an interface between
theory (general descriptive linguistics, language typology or comparison) and application (EA and IA for instance).
Based on CL methodology (Description – Juxtaposition – Comparison), Krzeszowski (1990: 10, 35-37) postulates that
any CL approach is necessarily dependent on theoretical linguistics, “since no exact and reliable exploration of facts
can be conducted without a theoretical background which provides concepts, hypothesis and theories.” For instance,
the earlier contrastive studies were conducted in the structural framework and many recent studies have embraced the
cognitive framework. Moreover, CL is also dependent on descriptive linguistics since no comparison of languages or
any elements is possible without their prior description. From this perspective, Krzeszowski (1990:10) defines CL as:
an area of linguistics in which a linguistic theory is applied to a comparative description of two or more
languages, which need not be genetically or typologically related. The success of these comparisons is strictly
dependent on the theory applied.
In this definition, Krzeszowski (1990:10) clearly makes of theory the cornerstone of CL. Nevertheless, beyond
theory and methodology, corpus-based approach is another converging factor between theoretical and applied CL,
but also toward Translation Studies (TS). Nowadays, corpora are more and more crucial to all fields of linguistics
and beyond.
4.2 Corpus-based approach as a common ground to CL studies and TS
The development of corpora has benefited many fields of linguistics, mainly CL and Translation studies12. One of
the weaknesses of early stages of linguistic studies is due to intuition-based judgments. With the emergence of corpus
linguistics, CL has gained momentum through corpus data, which enable a more objective and reliable empirical
investigation. Given that some CL approaches focus on matters of performance, which requires authentic and more
quantitative data, the use of multilingual corpora inevitably becomes an integral part of this discipline in few decades.
Mostly adopting a cross-linguistic approach (in the broad sense of the term), contrastive linguists employ the different
types of corpora according to the purpose of their study. In a more theoretical approach, CL can use comparable
corpora containing original texts in two or more languages. The focus will be on unilingual texts sharing similar
content and features in order to examine how a particular linguistic feature or cultural phenomenon is expressed in
different languages. A CL study can either be interested in theoretical issues such as translation “equivalence” or in
practical goals such as the application of its results in devising translation manuals, training programs, machine
translation, etc. For this perspective, CL needs translation (or parallel) corpora, which consist of original texts in one
language and their translations in another. When it comes to translation corpora, CL and TS cover a common ground
and do converge. Authors such as Baker (1993), Chesterman (1998), Ramón García (2002) and Granger (2003)
pinpoint the close and increasing relation between CL and TS. For instance, according to Chesterman (p.39) “a sharp
dividing line between Translation Theory and Contrastive Analysis is not well motivated.” He argues that they have a
common interest “in seeing how ‘the same thing’ can be said in other ways, although each field uses this information
for different ends.”
Focusing on learner language, CL (EA and IA in particular) needs another type of corpus which compiles learner
interlanguage data, i.e. learner corpus. This type of corpus, but also the corpus-based approach in general, has become
a common ground to all fields of CL (CA, EA, IA). The development of electronic corpora has a huge potential to
bring together these fields. This new perspective is strongly supported by the ICM (Integrated Contrastive Model) put
forward by Granger (1996) and Jarvis (2000). This model suggests the followings steps and procedures:
- Step 1 consists of comparing two languages and makes predictions about learners’ interlanguage.
- Step 2 examines learners’ data to look for traces of L1 (transfer) in the data.
- Step 3 tests the accuracy of the predictions and establishes the (potential) presence of transfer.
ICM brings together the three areas of CL, in the sense that it deals with all the three varieties of languages (i.e. L1, L2
and interlanguage). Based on Granger (1996) and Jarvis’s (2000) model which combines contrastive and interlanguage

12. For more arguments on this issue, see Chesterman (1998) and Ramón García (2002: 397-400).
6
 [Draft version] KANTÉ I., In Y. Rolland et al.(eds) 2015. Heritage and Exchanges: Multilingual and intercultural approaches in
training context, pp 25-42. UK: Cambridge Scholars Publishing.

analyses, Gilquin (2008: 3) worked out a more consolidated model (Detection-Explanation-Evaluation – DEE).
Gilquin’s model has the merit of (i) highlighting the interconnections between CA, EA and IA, (ii) making use of the
three varieties of languages in the learner’s linguistic environment, and (iii) carrying out multiple and complementary
corpus-based procedures. The DEE model outlines the use of the following corpora and analytic steps:
- Comparable corpora to compare original L1 and original L2 texts, e.g. French original texts versus English original
texts.
- Translation corpora to compare source L1 and translated L2 texts or vice versa, e.g. French texts versus their
translation in English or vice versa.
- Comparable corpora and Learner corpora either to compare native L1 and interlanguage, e.g. native French texts
versus L2 learners’ English texts, or to compare native speakers’ texts in L2 with learner interlanguage, e.g.
native English texts versus L2 Learners’ English texts.
- Learner corpora to compare, on the one hand, interlanguage texts by learners from the same L1, e.g. French L1
students learning English versus French L1 students learning English, and on the other hand to compare
interlanguage texts by learners from different L1s, e.g. French L1 students learning English versus Spanish L1
students learning English and versus German L1 students learning English (English as L2 for of all them).
The challenge for future CL studies is to be able to apply such a complex empirical procedure. If they are successful,
corpora will have eventually brought together the so-called distinct CL fields, as it has already done for CL and TS in
using translation corpora (see Granger 2003 and Ramon Garcia 2002)
5. Conclusion
Since the mid-1940s, CL has seen its original hypothesis heavily criticized and doomed to disappear. Rather than
damaging CL, internal criticisms and recurrent shifts actually enabled it to adapt, to broaden its scope and to tone
down its strong hypothesis. The emergence of corpora has accelerated and strengthened CL adjustment and revival. At
the end of the day, CL stands as an interface between theoretical and applied inspirations and is taking new steps
toward other fields such as Translation Studies. In the meantime, the continuum keeps evolving, since in addition to
the three basic CL phases (CA, EA and IA), a new consolidating methodology (ICM) has been on the rise in this
decade. So far, CL’s continuum can be represented as follows:

Figure 2 : Current continuum of CL approaches.

References
Baker, M. (1993). “Corpus Linguistics and Translation Studies. Implications and Applications.” Francis Baker &
Elena Tognini-Bonelli (1993), 233-250.
Ballier, N. (2007). “La complétive du nom dans le discours des linguistes.” In D. Banks (ed.), La coordination et la
subordination dans le texte de spécialité. Paris: l’Harmattan, 55-76.
Banathy, B. H., Trager, B. C. and Waddle, C. D. (1966). “Use of Contrastive Data in Foreign Language Coarse
Development.” In Valdman (ed.) Trends in Language Teaching. New York: McGraw-Hill.
Chesterman, A. (1998). Contrastive Functional Analysis. Amsterdam & Philadelphia: John Benjamin’s.
Corder, S. P. (1974). “Error Analysis.” In Allen, J.L.P. and Corder, S.P. Techniques in Applied Linguistics. Oxford:
Oxford University Press.
Dulay, H. C. et al. (1982). Language Two. New York: Oxford University Press.
Dyson, B. (2010). “Learner language analytic methods and pedagogical implications.” Australian Review of Applied
Linguistics, vol. 33, No 3 (2010)

7
 [Draft version] KANTÉ I., In Y. Rolland et al.(eds) 2015. Heritage and Exchanges: Multilingual and intercultural approaches in
training context, pp 25-42. UK: Cambridge Scholars Publishing.

Fisiak, J. (Ed.). (1981a). Theoretical issues in Contrastive Linguistics. Current Issues in Linguistic Theory 12.
Amsterdam/Philadelphia: John Benjamin’s.
Fisiak, J. (Ed.). (1981b). Contrastive Analysis and the Language Teacher. Oxford: The Pergamon Institute of English.
Fries, C. C. (1945). Teaching and Learning English as a Second Language. Ann Arbor: University of Michigan Press.
Gast, V. (2011). “Contrastive analysis: Theories and methods.” In Kortmann, B. and J. Kabatek (ed.): Dictionaries of
Linguistics and Communication Science: Linguistic theory and methodology. Berlin: Mouton de Gruyter. Accessed
via http://www.personal.uni-jena.de/~mu65qev/papdf/contr_ling_meth.pdf
Gast, V. (2013). “Contrastive Analysis.” In Michael Byram and Adelheid Hu (eds.), The Routledge Encyclopedia of
Language Teaching and Learning, pp. 153-158. 2nd Edition. London: Routledge.
Gilquin, G. (2008). “Combining contrastive and interlanguage analysis to apprehend transfer: detection, explanation,
evaluation.” In G. Gilquin, S. Papp & M. B. Diez-Bedmar (eds.), Linking up Contrastive and Learner Corpus
Research, pp. 3-33. Amsterdam & Atlanta: Rodopi.
Granger, S. (1996). “From CA to CIA and back: An integrated approach to computerized bilingual and learner
corpora.” In: K. Aijmer, B. Altenberg and M. Johansson (eds) Languages in Contrast. Papers from a Symposium
on Text-based Cross-linguistic Studies. Lund 4-5 March 1994. Lund: Lund University Press. 37-51.
Granger, S. (1997). “Automated retrieval of passives from native and learner corpora: precision and recall.” Journal of
English Linguistics, 25(4): 365- 374.
Granger, S. (1998). “Interlanguages in contrast: translationese vs learnerese.” Paper presented at the conference
Languages in Contrast. Information structure in parallel texts, University of Oslo, 30-31 October 1998.
Granger, S. (2003). “The corpus approach: a common way forward for contrastive linguistics and translation studies.”
In S. Granger, J. Lerot, & S. Petch-Tyson (Eds.), Corpus-based approaches to contrastive linguistics and
translation studies (pp. 17-29). Amsterdam & Atlanta: Rodopi.
Granger, S. (2013). “The passives in learner English: Corpus insights and implications for pedagogical grammar.” In
S. Ishikawa (Ed.), Learner corpus studies in Asia and the world –Vol 1 (pp. 5-15).
James, C. (1971). “The exculpation of contrastive linguistics.” In Nickel, G. (ed.), Papers in Contrastive Linguistics,
53-68. Cambridge: Cambridge University Press.
James, C. (1998). Errors in Language Learning and Use: Exploring error analysis. London: Longman.
Jarvis, S. (2000). “Methodological rigor in the study of transfer: identifying L1 influence in the interlanguage
lexicon.” Language Learning, 50(2): 245- 309.
Jie, X. (2008). “Error Theories and Second Language Acquisition.” US-China Foreign Language. 6, 35-42.
Johansson, S. (2008). Contrastive analysis and learner language: A corpus-based approach. Oslo: University of Oslo.
Kanté, I. (2011). “La complétive nominale finie: entre syntaxe et sémantique – une étude contrastive anglais-français.”
Thèse de doctorat, sous la direction de Nicolas Ballier, Université Paris 13.
Khansir, A. A. (2012). “Error Analysis and Second Language Acquisition.” Theory and Practice in Language Studies,
vol. 2, No 5, 1027-1032.
Krzeszowski, Tomasz P. (1990). Contrasting Languages. The Scope of Contrastive Linguistics. Berlin: Mouton.
Lado, R. (1957). Linguistics across Cultures. Ann Arbor: University of Michigan Press.
Lennon, P. (2008). “Contrastive Analysis, Error Analysis, Interlanguage.” In S. Gramley and V. Gramley (eds)
Introduction to Applied Linguistics. Bielefeld: Aisthesis, pp. 51-60.
Narcy-Combes, J.-P. (2014). “De la complémentarité des domaines pour un didacticien des langues: quelle place pour
les multilittéracies.” FLuL - Fremdsprachen Lehren und Lernen 43:2, pp. 29-42.
Ramón García, N. (2002). “Contrastive Linguistics and Translation Studies Interconnected: The Corpus-Based
Approach.” Linguistica Antverpiensia, 1, pp. 393-406.
Schachter, J. (1974). “An error in error analysis.” Language Learning, vol. 24 no 2, pp. 205- 214.
Schachter, J. and Mrianne, C. M. (1977). “Some Reservations Concerning Error Analysis.” TESOL Quarterly. 2.4,
pp.441-451
Selinker, L. (1972). “Interlanguage.” International Review of Applied Linguistics in Language Teaching. 10(3): 209-
231
Souliou, L. (2014). “L’activation d’une autre langue que celle attendue : Pratiques et représentations des apprenants
d’une troisième langue.” Thèse de doctorat sous la direction de Jean-Paul Narcy-Combes, Sorbonne Nouvelle -
Paris 3.
Spolsky, B. (1989). Conditions for Second Language Learning. Oxford: Oxford University Press.
8