Cross-Linguistic Influence – A Friend Or A Foe

Terminological Clarification

The concept cross-linguistic influence (CLI) was initially introduced by Kellerman and Sharwood Smith (1986) in order to address various types of linguistic phenomenon in the context ofL2 acquisition. This umbrella term was used to describe the relationships among languages and the way the constituents of one language are integrated into another language, leading to such linguistic phenomena as borrowing, transfer, avoidance, interference, etc. (Kellerman& Sharwood Smith, 1986).

In the later studies, the phenomenon of CLI was also explored in the context of L3 acquisition (Hammarberg, 2001; Hammarberg, 2010; Slabakova, 2016; Westergaard et al., 2017) and different patterns of CLI were found. For instance, it was established that in the area of phonology L1 is the only source language for CLI occurrence, thus, most learners even the most advanced ones, keep their L1-based accent in their oral production, whereas in the area of grammar, L2 influence is more frequent, especially if L3 learners are extensively exposed to L2 and have reached a proficiency level in it(Ringbom, 2001).

Moreover, the concept CLI, in its broader sense, isused to address numerous linguistic phenomena investigated in the context of L3 acquisition, e.g., lexical inventions (Dewaele, 1998), the phenomenon concerned with words “which are morpho-phonologically adapted to the TL [target language] but which are never used by native speakers”(1998, p. 471).

In literature, the term CLI is quite often replaced by the term transfer, while discussing relationships between previously acquired languages and L3 (Neuser, 2017). Interestingly, one of the earlier definitions of transfer , interreference , carried a negative connotation while pointing out the shortcomings of the linguistic transfer rather than its facilitative effect on L2 acquisition (Weinreich, 1953). The phenomenon used to be treated as “instances of language deviation from the norms of either language which occur in the speech of bilinguals as a result of their familiarity with more than one language” (Weinreich, 1953, p. 1). Hence, transfer from a previously acquired language to a target language used to be seen as an example of negligence (Neuser, 2017). Yet, Weinreich (1953) and Lado (1957) spoke of transfer as linguistic, psycholinguistic, and sociolinguistic phenomenon inseparable from language learning and use (Neuser, 2017), and today, the term interference , in its general sense, has been replaced by such concepts as linguistic transfer (LT) or cross-linguistic influence (CLI) (Neuser, 2017).

Nevertheless, certain unclarities regarding the definition of transfer are still existent, but despite some evident dissimilarities, they are not conflicting (Neuser, 2017). Firstly, it was claimed that t ransfer is a process (Weinreich, 1953) since a learner undergoes a process of identifying equivalent elements between formerly acquired languages and a target language (Neuser, 2017). However, this claim contradicts the one made by Schachter (1983) who advocated for transfer as being “a constraint on the learner’s hypothesis testing process” (p. 32), inasmuch as learners tend to design and test their own hypotheses regarding the new language based on their knowledge of previously acquired languages.

In the context of L2 and L3 acquisition, transfer can be also seen as a strategy applied by learners in order to surmount obstacles they encounter (Neuser, 2017) at the early phase of a target language acquisition (Hammarberg, 2001), e.g., “cope with the still too unfamiliar phonetic form of L3” (Hammarberg, 2001, p. 35). However, it was argued that not all learners apply transfer strategies while acquiring an additional language (Neuser, 2017).

Oldin (1989) also maintained that there is a certain lack of accuracy in defining the concept, pointing out that linguistic transfer is a result of either likeness or dissimilarity between formerly acquired languages and a target language, therefore providing an additional support to Jordens and Kellerman’s theory of psychotypology (1981).

Neuser (2017) suggests treating transfer as “a snapshot of processes occurring at that moment in time, given the particular circumstances surrounding the learner and the given task” (2017, p. 10), since every learner might employ various types of transfer dictated by circumstances.

As it becomes clear from the present chapter, the phenomenon of transfer has drawn much of linguists’ attention and produced numerous definitions. For the convenience of the present article, the terms linguistic transfer (LT), transfer and cross-linguistic influence (CLI) will be used interchangeably in accordance with varied sources this paper was based upon.

CLI in L3 Acquisition

As can be seen from the literature, CLI is a phenomenon caused by interaction between different languages in the mind of bilingual and multilingual speakers (Cenoz, 2001; Tremblay, 2006; Slabakova, 2016; Hammarberg, 2001; Ringbom, 2001; Neuser, 2017). From previous studies, it became evident that CLI might have either positive or negative effect on L3 acquisition (Ringbom, 2001), yet Flynn and others (2004) spoke of a neutral influence as well. Foremost, former studies indicate that transfer from L1 and L2 can occur in both comprehension and production processes (Cenoz, 2001; Ringbom, 2001)and in various areas of language; however, it appears in different ways and at sundry rates (Neuser, 2017; Ringbom, 2001). For instance, in L3 comprehension process learners of a target language make an attempt to associate a new lexical form with the one they are already familiar with either in L1 or L2, based on similarity or perceived similarity between the acquired languages and a target language (Ringbom, 2001; Tremblay, 2006), especially when L2 and L3 share certain features, e.g., common cognates (Ringbom, 2001). Thus, once learners presume that a proper equivalent is found, they tend to transfer both a form and a meaning either from their L1 or L2 into the target language (Ringbom, 2001). Surprisingly, very little research was conducted in the field of L3-comprehension, but from the available research data, it becomes rather clear that any type of linguistic similarity might facilitate the process of comprehension at its initial stages (Ringbom, 2001).

The same is applicable to the production process, i.e. L3 learners tend to resort to previously acquired languages while acquiring L3, yet learners’ proficiency in the formerly acquired languages as well as the extent of exposure to those languages, will affect the transfer (Tremblay, 2006). For example, in a study conducted among learners of German as L3, with English as L1 and French as L2, it was found that the level of L2 proficiency affects the frequency of L2 interference with L3 production (Tremblay, 2006). Moreover, L2 proficient learners apply their knowledge of L2 in order to cope with the challenges of L3 acquisition (Tremblay, 2006). However, unless L3 learners are proficient in L2, it will have a very insignificant influence on L3 production if any. This claim contradicts the one made by Ringbom (2001) who advocates for the primacy of L1 in any type of lexical transfer despite the proximity between L2 and L3. She also claims that learners’ L2 proficiency and the extent of exposure to L2 are of little significance in the case of a lexical transfer, e.g., phrasal verbs or semantic extensions, as leaners tend to mistakenly believe that L3 lexical items have a meaning “corresponding to what is most commonly the core meaning of the equivalent L1 word” (Ringbom, 2001, p. 62).Nevertheless, the psychotypological distance between the target language and any other languages a learner has previously mastered, will determine the extent of phonological and grammatical transfer (Ringbom, 2001).

On the other hand, Hammarberg (2001) maintains that L3 learners usually transfer lexical items from the language they have most recently acquired and probably use on a regular basis, e.g., the participant in Hammarberg’s study tends to use German, the most recently acquired language, as a source language. This claim corresponds with the one made by Williams&Hammarberg who also assert that at the early stages of L3 acquisition, a competent user of L2 will rely on his/her L2 knowledge, especially in the use of grammar (Ringbom, 2001). Yet, in the case of Hammarberg’s subject, similarity in phonetic structure, vocabulary and syntax (Cenoz, 2000) between L2 and L3 probably led to the cross-linguistic transfer when the research participant relied on her knowledge of German (L2) while acquiring Swedish (L3) (Hammarberg, 2001). Hammarberg’s study also adds to the knowledge that transfer most often occurs during the early stages of L3 acquisition when “new input is connected to already established representations in the L1 and L2”(Neuser, 2017, p.14), since the participant of the study leans on her knowledge of L2 at the primary steps of L3 learning, but the situation changes once she reaches a satisfactory level of proficiency in L3 (Hammarberg, 2001).

Even though similarities of forms among certain lexical items may be of a great assistance to L3 learners at the initial stages of L3 learning, one has to be aware of the fact that these similarities might also cause errors like “in the case of partially overlapping meaning or false cognates” (Neuser, 2017, p.14) or transfer of meaning as it happens with calques and semantic extensions(Ringbom, 2001). Also, in the case of deceptive cognates, L3 learners might use single lexical items from their L1 or L2, the ones they believe to be analogous in their form to the words in L3 (Ringbom, 2001), which will evidently result in an error, thus having a non-fasilitative effect on L3 acquisition. Another example of non-facilitative effect can be found in the case of code-switching, when learners resort to their L1 or L2, depending on their psychotypological similarity to L3, and transfer a whole lexical item into L3, sometimes altering its form according to the rules of L3, thus “producing hybrids or blends that do not exist in L3” (Ringbom, 2001, p. 60).

As opposed to the earlier statement which was arguing for the primacy of the most recently acquired language, Cenoz (2001) asserts that L3 speakers generally borrow lexical items from typologically close languages, i.e. either from L1 or L2 (Cenoz, 2001). In her study of Basque and Spanish speakers as their L1 and L2, in this order or other way around, who were in the process of learning English as L3, Cenoz (2001) noticed that learners used to transfer terms from Spanish, which is typologically closer to English, rather than Basque, which is typologically distinct to Indo-Europeans languages. This phenomenon could be clearly spotted among older students (9^th grade) probably due to the fact they have reached “the higher metalinguistic awareness” (Cenoz, 2001, p. 16) than their younger peers, and thus were able to perceive the typological distance between Basque and English. Another explanation could be found in the L2 status, whereas learners with Basque as their L1 tended to use Spanish for transfer due to its status as L2 in addition to its typological proximity to English (Cenoz, 2001). As for the Spanish native speakers, the findings were quite contradictory, i.e., despite Basque being their L2, and thus was expected to act as a preferred language due to its L2 status, Spanish had a stronger influence due to its typological proximity to English. Therefore, it can be assumed that perceived typological closeness between previously acquired languages and a target language has a greater influence on L3 production than L2 status (Kellerman, 1995; Cenoz, 2001).

L3 Acquisition Models

In order to explain the impact that L1 or/and L2 might have on L3 acquisition, varied models were defined, whereas three of them can contribute to the future investigation this paper is part of. The first one is the Cumulative Enhancement Model (CEM) (Flynn, Foley, & Vinnitskaya, 2004) which maintains that both of previously acquired languages might have either positive or neutral effect on L3 acquisition as in the study of L1 Kazakh / L2 Russian / L3 English adult learners. However, from this model it is not quite clear which of the factors, typological or structural proximity, plays the most crucial role in a positive transfer, (Mykhaylyk et al., 2015).

The second model to be considered is Rothman’s (2011) Typological Proximity Model (TPM), which is based on Kellerman’s (1979, 1983 in Westergaard et al., 2017) idea of psychotypology. The TPM addresses an impact of typological proximity (Westergaard et al., 2017) on L3 syntactic transfer, and points out that transfer occurs as a wholesale and might have facilitative as well as non-facilitative effect (Rothman, 2010). The data for this model was drawn from the comparative study of “successful […] L1 Italian learners of English as […] L2 at the low to intermediate proficiency level of L3 Spanish, and successful L1 English learners of L2 Spanish at the same levels for L3 Brazilian Portuguese” (Rothman, 2010, p.1). Both groups of participants demonstrated a greater knowledge of a semantic structure under study achieved due to typological proximity among L1/L2 and L3.

The third model is Linguistic Proximity Model (LPM) (Mykhaylyk et al., 2015; Westergaard et al., 2017) which includes some of the CEM and TPM elements, but at the same time suggests that “structural similarity at an abstract level should be considered an important factor in third language acquisition” (Westergaard et al., 2017, p.667).It also maintains that CLI occurs “on a property-by-property basis”(Westergaard et al., 2017, p.678) and might have facilitative as well as non-facilitative impact. Moreover, according to LPM, neither order of L1/L2 acquisition nor typological grouping have any impact on CLI (Mykhaylyk et al., 2015). This hypothesis was tested in a comparative study in which Norwegian-Russian bilinguals were compared with Norwegian and Russian monolinguals with respect to two particular syntactic constructions: the order of adverb-verb and subject-auxiliary inversion (Mykhaylyk et al., 2015).