Meine Hobbys sind Müzik hören und Swimmen:
Lexical Transfer in L1 Turkish-L2 English Learners of L3 German
The study of cross-linguistic influence (CLI), or transfer, has
been central in the field of second language acquisition. Studies dealing with
CLI have, until recently, predominantly focused on the effects of a first
language (L1) on the acquisition of a second language (L2) and vice versa
(e.g., Kecskes & Papp 2000). Particularly over the past decade, however,
the scope of CLI studies has broadened to include multilingual acquisition
contexts, encompassing research studies that go beyond mere L1-L2 interaction
and take into account the complex transfer relations of L1-L2-L3-Lx
combinations. In spite of the fact that the study of CLI in multilingual
acquisition contexts is still relatively unexplored and possibly poses more
questions than it currently answers, it has been firmly established that not
only a learner's L1, but also his additional language(s) can act as a source of
influence in the acquisition of further languages (e.g., Cenoz 2001; König,
Cedden & Onaran 2005; Sağın Şimşek 2006; Tremblay 2006).
Research into multilingual CLI has dwelt upon a wide array of
areas, ranging from phonological influence (e.g., Bannert 2005; Beach, Burnham
& Kitamura 2001) to the transfer of discourse patterns (e.g., Kellerman
2001), metalinguistic awareness (e.g., Jessner 1999; O'Laoire, Burke &
Haslam 2000), pragmatic competence (e.g., Jorda 2005), syntax (e.g., König et
al. 2005), and lexis (e.g., Ringbom 2001), just to mention a few. A common
theme that has emerged from these studies is that previously learned languages
are activated in the process of acquisition of an additional language in
qualitatively and quantitatively different ways. Among a number of factors that
have been identified as significant in the choice of the background language
from which a given structure is transferred into the target language, such as
linguistic typology of languages, learners' age, learners' proficiency levels
in the languages involved, the status of the languages and recency (Sağın Şimşek 2006), psychotypology and
the tendency to transfer from a foreign
language into another foreign language, have emerged as particularly salient.
2. (Psycho)typology and the L2->L3 transfer preference
Psychotypology basically refers to the perceived distance between
languages by the language learner. Kellerman (1977, 1983) argues that
psychotypology is central to CLI in that a learner is more likely to transfer a
structure from one language to another if the two languages are perceived as
similar as regards the target structure. If, on the other hand, the two
languages are perceived as dissimilar, the learner will tend to avoid the
transfer of a particular structure from the source language to the target
language, a point that Kellerman (1983: 117) summarizes with his frequently
quoted statement "not everything that looks transferable is
transferable". Ringbom (2001: 65) lends support to this view, stating
"languages perceived to be similar (roughly=related) to the target
language naturally provide many more reference points for the learner than do
wholly unrelated languages." Thus, under this view, in a multilingual
acquisition context the learner is expected to prefer to transfer a target
structure to Language Z from Language Y, rather than from Language X, if the
perceived similarity between Languages Y and Z by the language learner is
higher than that of Languages X and Z - a view that has received affirmation in
the relevant literature (e.g., Hammarberg 2001; Ringbom 1983, 1987, 2001).
Ringbom (1983, 1987), for example, analyzed more than 10,000
student essays written by L1 Finnish-L2 Swedish and L1 Swedish-L2 Finnish
learners of L3 English in Finland. He found that independent of whether Swedish
or Finnish was the L1 of the learners, the source of transfer errors into
English was more often Swedish than Finnish - not only in cases in which
Swedish was the L1 of the learners but also when Swedish was the learners' L2.
In other words, Ringbom provided evidence for the view that the perception of
two languages as similar, such as in the case of Swedish and English (in
contrast to Finnish and English), may lead to comparatively more transfer
between the languages perceived as similar (even if both constitute foreign
languages) and override the L1 as a source of CLI. Williams & Hammarberg
(1998: 323) provide two possible reasons for such a non-native
language dominance in multilingual CLI that possibly go hand-in-hand with the
Differences in the acquisition mechanisms in L1 and L2, and a
reactivation of the L2 mechanism in L3 acquisition.
A desire to "suppress" the L1 on the basis of its
being "non-foreign" and to use a foreign, non-L1 language in the
acquisition of another foreign language.
Similar findings were obtained by Cenoz (2001), who presents the
results of a study on the acquisition of L3 English by 90 Spanish/Basque
bilingual children at three different age levels in the Basque country. In the
total subject group, Basque is reported to be the L1 for 44%, Spanish for 23%
and both Basque and Spanish for 32% of the students. Using Mayer's (1969)
wordless picture story Frog, where are you? Cenoz (2001) collected oral
production data from the child participants, which were analyzed for CLI at the
lexical level from Basque/Spanish into L3 English. The results revealed that
students tended to transfer more from Spanish (an Indo-European language) than
from Basque (a non-Indo-European language) into L3 English (an Indo-European
language), the transfer rates across age groups being 62%-87% and 13%-38% from
Spanish and Basque, respectively. In other words, more transfer was found to
take place between typologically similar languages. Furthermore, similar to the
findings in Ringbom (1983, 1987) stated above, it was found that the L1 Basque
children displayed a stronger preference to transfer from Spanish into English
than the L1 Spanish children; thus, transfer from L2 to L3 was preferred over
transfer from L1 to L3. A further relevant point that emerged in the obtained
results is that the students in the highest age group tended to transfer from
Basque the least. Cenoz (2001: 16f) explains this finding by establishing a
connection between the higher metalinguistic skills that the older students are
expected to possess and their resulting perception of similarity/dissimilarity
(i.e., psychotypology) between Basque, Spanish and English:
Older students are able to perceive that Basque and English are
typologically more distant than Spanish and English, and they could use Spanish
rather than Basque as a base language when acquiring English. Younger learners'
lower metalinguistic ability does not allow them to perceive objective
linguistic distance, and they find both Spanish and Basque terms as
In a recent study, Sağın Şimşek (2006) also found a similar
L2-over-L1 dominance, possibly also coupled with a psychotypology effect, in
the acquisition of L3 English by 14 L1 Turkish/L2 German learners. As part of
the study, Sağın Şimşek collected written data from her participants that were
analyzed for CLI in word order from Turkish/German into L3 English. The results
of her analyses revealed that in nearly 25% (149 tokens) of the total written
L3 English production data, it was possible to find L2 German-induced
word order features as in examples (1) and (2) below. Influence from L1
Turkish, on the other hand, was almost absent (5 tokens) in the L3 written
data. Sentence (1) below from Sağın Şimşek (2006), for example, constitutes a
clear illustration of how the learner has employed German topicalization,
placing the verb in the main clause (dance) immediately after the topicalized
adverbial (after the break) as required in German but not in English or
Turkish. Similarly, example (2) also illustrates clear CLI from L2 German since
in the main clause the verb give is in clause-initial position, exactly as
required in German main clauses preceded by subordinate clauses.
(1) After the break
dance 8-9 people.
(Example 2 in Sağın Şimşek 2006: 77).
(2) When we come to the
airport, give we the suitcase there fort and go to the plane.
(Example 27 in Sağın Şimşek 2006: 83).
On the basis of these findings, Sağın Şimşek (2006: 131)
identifies "typological similarities between the target language and the languages
already known as the most dominant factor [in multilingual CLI]", and
taking into consideration the many surface similarities between German and
English, in comparison to Turkish and English, she adds that "the subjects
are aware that their knowledge of Turkish will not be of use when learning
English". In other words, L2 German was found to be preferred as the base
language from which to transfer structures into L3 English, thus overriding L1
Turkish, most probably due to the similarities the learners perceived between
the two former languages and due to the pressing need to stay in a
'foreign language mode' and not revert back to the 'L1-mode'.
3. The present study
The aim of the present study was to analyze CLI at the lexical
level in the written L3 German productions of Turkish-English-German
trilinguals. As mentioned above, a common point that arises from earlier
studies conducted on multilingual CLI is that psychotypology has been
identified as a highly influential factor in determining from which previously
learned language a language user will transfer a given structure into a target
language. As summarized, the findings obtained from previous studies indicate
that learners employ certain beliefs as to which previous language is more similar
to the target language and try to restrict potential transfer actions as much
as possible to the language identified as more similar.
Thus, considering the fact that English and German, both being
Indo-European languages, are structurally more similar than Turkish (an Altaic
language) and German and, therefore, more likely to be perceived as such by
language learners, it was expected that the subjects in the present study would
predominantly transfer from L2 English, rather than from L1 Turkish, into L3
German. This expectation was also directly related to the observed tendency of
multilinguals analyzed in previous studies to transfer from a previously
learned L2 rather than from their L1 into another additional language, which
was found to act as an additional determining factor besides the psychotypology
effect. In other words, from the perspective of both psychotypological impact
and the previously observed tendency of learners to suppress non-foreign impact
and prefer non-L1 mechanisms in the acquisition of additional foreign languages
(cf. Williams & Hammarberg 1998), the likelihood that the L3 German written
production data of L1 Turkish/L2 English participants would predominantly
contain manifestations of lexical transfer from L2 English appeared much higher
- particularly considering the widely acknowledged view that "in no other
area ... is the importance of psychotypological factors, perceived
similarities, more in the foreground than lexis" (Ringbom 2001: 60).
The subjects of the present study were 174 L1 Turkish/L2 English
learners of L3 German, who were undergraduate students from various departments
at Middle East Technical University (METU) in Ankara/Turkey. 50 of the subjects
were males (29%) and 124 were females (71%). METU is one of the few
English-medium universities in Turkey; i.e., with the exception of additional
foreign language courses offered at the university (like German, French,
Italian), all lectures, examinations, etc. are conducted exclusively in
English. Therefore, all newly admitted students are required to take an
English proficiency exam administered by the university. On the basis of the
scores obtained from this examination, students are either allowed to start
studying at their intended departments or are placed into the Department of
Basic English where they receive intensive English tutoring for one year or,
if necessary, a maximum of two years. Thus, the L2 English level of all
participants in the present study was high enough to receive instruction, write
essays, give presentations and attend exams in English, which requires an
upper-intermediate to advanced English proficiency level in all four language
In addition, the subjects in the present study were also attending
elective German language classes offered by the Department of Foreign Language
Education for three hours per week. Basically being German proficiency classes,
these German courses also focus extensively on metalinguistic abilities,
explicit German grammar, and German socio-culture. The German proficiency level
of students enrolled in these courses may be roughly classified as
pre-intermediate to intermediate.
4.2 Data collection
282 authentic exam papers (midterm and final exams) written by the
174 subjects in the German courses they were either taking at the time of the
investigation or had taken in a previous semester were analyzed. The exams in
the elective German courses they were enrolled in included a number of
different question types, ranging from more controlled fill-in-the-blanks and
rewrite questions as in (3) and (4), respectively, to less controlled, freer
question types as in (5) and (6).
(3) Sample fill-in-the-blanks question:
Bitte ergänzen Sie sinngemäss:
1. Meine Eltern sprachen ganz leise, ........ das Kind nicht
2. Er hat nicht viel gelernt, ........ hat er die Prüfung
3. Mein Vater schickt mir jeden Monat Geld, ........ ich hier
(4) Sample "rewrite" question:
Sie haben hier positive Aussagen. Bilden Sie bitte negative
1. Ich warte auf Sarah.
2. Das ist ein Buch.
3. Ich mache meine Hausaufgaben. (plural).
(5) Less controlled question:
Ergänzen Sie die Sätze mit einer passenden Konjunktion.
1. Hans ist ganz unglücklich,
2. Der Film war so langweilig,
3. Wir können nur dann kommen,
(6) Sample "free" writing questions:
1. Schreiben Sie einen Aufsatz: Meine Familie (50 Wörter)
2. Schreiben Sie 5 Sätze zu folgendem Thema: "Was wäre
passiert, wenn Ihr Wecker heute morgen nicht geklingelt hätte und Sie verschlafen
The fact that the data was comprised of authentic exam papers
embodied some significant dimensions that need to be focused on at this point.
Taking into consideration that the exam papers were going to be scored and
actually constituted part of their final grades, it was expected that the
participants would, where possible, refrain from transferring any lexical items
from either L1 Turkish or L2 English into their L3 German written productions
because of the straightforward reason that in either case the result would be a
loss in valuable points. Thus, it was expected that the number of transferred
lexical items would be much less when compared to earlier studies of CLI at the
lexical level. However, this was on the other hand seen as an advantage since
this restraining nature of the data collection method was likely to ensure that
participants would only resort to transfer at points where they really needed
to transfer and would work hard to retrieve German lexical items. In other words, it was
going to be possible to infer with some confidence that students exclusively
transferred lexical items from their L1 (Turkish) or L2 (English) when they
felt obliged to use a lexical item that was actually not part of their lexical
repertoire in L3 German; i.e., in case of a knowledge gap. It is a
well-established fact that language learners do not always transfer
intentionally from their previously learned language due to lack of knowledge, but may
also employ non-intentional transfer resulting from various strategies (e.g.,
Poulisse & Bongaerts 1994; Williams & Hammarberg 1998). In this study,
however, this was very unlikely to be the case because of the obvious concern
of the participants to achieve high scores.
4.3 Data analysis
The collected L3 German written production data were analyzed for
three types of lexical CLI:
(I) Full lexical switches, in which a complete lexical item from a
non-target language is used in the production of the target language (e.g., an
acceptable L1 Turkish word in an L3 German production). De Angelis & Selinker
(2001) and Ringbom (1987, 2001) in their analyses of CLI in spoken L3
production focused on this category as well, referring to it as lexical
interlanguage transfer and language switches, respectively.
Ringbom (2001: 64, Table 4.2) defines the underlying cause of this
type of CLI as "insufficient awareness of [an] intended linguistic
form" and Serindağ (2005: 12), in his study of Turkish-English-German CLI,
provides the following examples for this category (among others), in which
lexical items from L2 English (in italics) are used in the production of L3
(7) Wie large ist die Wohung?
(8) Es ist half zehn.
(9) Ich trinke morgens tea.
(II) Morphologically hybrid forms, which are lexical items in
which a free or bound morpheme from any of the three languages involved is
combined with a free or bound morpheme from another one of the three languages.
De Angelis & Selinker (2001: 53) present a number of interesting examples
of morphologically hybrid forms (which they call morphological interlanguage
transfer) in their analyses of CLI in spoken L3 Italian production, one of
which is presented under (10) below:
Example (10) comes from subject 2 in De Angelis & Selinker,
who is reportedly a native speaker of British English with Spanish and Italian
as additional languages. The example illustrates how the Spanish plural bound
morpheme -as has been attached to the Italian stem bomba (sing. for English bomb) to form the hybrid
form bombas, whereas the actual target form would be the Italian plural noun
(III) Orthographically hybrid forms, which are lexical items in
the target language that have been altered orthographically due to CLI and
manifest orthographic properties from the target language (German) and,
additionally, from (at least) one of the non-target languages
(Turkish/English). Since the collected L3 German data was in written form, this
category was also of importance because it was expected that the orthographic
properties of Turkish and/or English would certainly have an impact on the
accuracy of the participants' spelling of words in L3 German.
Serindağ (2005) also analyzed his L3 German data for this CLI
category. Below is one of the examples that he identified as orthographic CLI from
L2 English. As the example demonstrates, the orthographic cluster sh, which
frequently represents the phoneme [š] in English, has been substituted for the
orthographic sch cluster in the target form Flasche (English: bottle).
The data analysis revealed a total of 124 instances of CLI at the
lexical level falling under the three categories outlined above. As can be seen
in Table 1 below, out of these 124 items, only 8 instances (7%) were identified
as CLI from L1 Turkish, whereas the remaining 116 (93%) were identified as CLI
from L2 English. In other words, the participants clearly regarded English and
German as more similar, predominantly taking L2 English as the source language
from which to transfer a lexical structure into L3 German. It possibly needs to
be pointed out once more at this point that the analyzed data source consisted
of authentic exam papers, which meant that no matter which language the
students transferred from, they knew that the result would be a loss of
valuable points since their instructors expected them to produce exclusively
German lexical items. However, regardless of this fact, the students
nevertheless tended to avoid using Turkish lexical items in cases in which a
knowledge gap existed in their L3 German and instead provided lexical items which
they transferred from L2 English. Thus, this finding goes hand-in-hand with
results of earlier studies in which it was found that the perception of two
additional languages as being similar may easily lead to comparatively more transfer
between these similar languages and may override the L1 as a source of CLI, as
is the case in the present study. This finding also provides support for the
view of Williams & Hammarberg (1998: 323), who, as mentioned at the outset
of the present study, state that the underlying reason for this widely observed
tendency might also be the desire to "suppress" the L1 on the basis
of its being "non-foreign" and to use a foreign, non-L1 language in
the acquisition of another foreign language instead.
Table 1: CLI from L1 Turkish
and L2 English across types of lexical CLI.
Full lexical switches
Morphologically hybrid forms
Orthographically hybrid forms
As the analysis of the individual utterances below illustrates,
this quantitative overall dominance of L2 English rather than L1 Turkish as the
source language for CLI on L3 German also clearly manifested itself in the
5.1 Full lexical switches
Full lexical switches constituted 55 (44%) of the total 124
identified instances of CLI at the lexical level. While 48 (87%) out of this
subset of 55 full lexical switches were identified as CLI from L2 English, only
7 tokens (13%) were analyzed as CLI from L1 Turkish. With the exception of one
item, the striking majority (98%) of full lexical switches from L2 English
were, as exemplified in samples (12)-(20) below, lexical items that embodied
important formal (phonological and/or orthographic) and semantic similarities
to the German target items. The only exception to this pattern is sample (21)
below, in which the English lexical item ache was substituted for the German
lexical item Schmerzen despite the fact that no formal similarity whatsoever
existed between them.
(12) Sie ging ins Bath (Bad) und wusch ihre Hände.
(13) Wenn ich nicht studieren wüde, würde ich ein singer (Sänger)
(14) Wenn ich studiere, kann ich ein guter engineer (Ingenieur)
(15) Ich arbeite, dabei höre ich music (Musik).
(16) Ich brauche einen Paper (Papier), ein Bleistift, ein
Radiergummi und eine Briefmarke.
(17) Ich möchte eine liter milk (Milch), bitte.
(18) Er ist krank. Er hat eine wound (Wunde) und schmerzen.
(19) Um 8 Uhr geht sie ins Bed (Bett) .
(20) Ich war krank gestern, da ich das ice (Eis) ass.
(21) Ein Mensch der so krank ist muss viel Ache (Schmerzen) haben.
A possible explanation for this tendency to use non-target language lexical
items that bear a significant amount of similarity to the target lexical item
is provided by De Angelis & Selinker (2001), who extend Dell's (1995) view
on monolingual processing that phonologically and semantically related words in
L1 processing may be activated in the lexical retrieval. In the same vein, De
Angelis & Selinker claim that phonologically and semantically similar
lexical items may also be activated cross-linguistically as a compensation
activity for weak or absent target language knowledge, which they regard as a
possible explanation for the established high impact of cross-linguistic
similarities on CLI. Though rather speculative, this explanatory approach
appears to have a grain of truth since, as mentioned at the outset of the
present study, formal (typological) similarities have indeed been found to act
as an important factor in CLI in that more similarities between languages
result in more CLI between them.
This can also be taken as part of the reason for why full lexical
switches from L1 Turkish were so infrequent compared to L2 English. As
mentioned above, full lexical switches from Turkish were only 7 in number (the
noun Müzik for German Musik was used three times). As illustrated in samples
(22)-(25), two of these Turkish lexical items (kanser-cancer; müzik-Musik-music)
displayed a significant degree of similarity to their English and/or German
counterparts, which again points to the likelihood of the impact of
cross-linguistic similarity on CLI.
(22) Es ist ratsam dass ich nicht rauchen wil ich kann Kanser
(German: Krebs, English: cancer) werden.
(23) Meine Hobbys sind Müzik (German: Musik, English: music) hören
(24) Ich gehe jetzt shon zur Uni um meine Vize (German: Prüfung,
English: exam) zu bestehen.
(25) Jemand der so gut Englisch spricht muss başarılı (German:
erfolgreich, English: successful) sein.
5.2 Morphologically hybrid forms
Morphologically hybrid forms constituted only 10 (8%) of the total
124 identified instances of CLI at the lexical level. A striking fact that was
observed was that the few morphologically hybrid forms in the present study
were exclusively formed out of morphemes from L2 English and L3 German and did
not include a single morpheme transferred from L1 Turkish. In other words,
morphological transfer from L1 Turkish was entirely inexistent.
It was in fact expected from the outset that examples of CLI
falling in this category would be rather few on account of the fact that
especially the transfer of bound morphemes is described in the relevant
literature as rare (Odlin 1989), though not totally non-existent, as illustrated
in De Angelis & Selinker (2001) and Ringbom (2001). However, as is
illustrated in the samples below, the participants in the present study did not
transfer any bound morphemes from their L2 English into their L3 German at all.
Instead, they either merged a stem from L2 English with a bound morpheme from
L3 German as in examples (26)-(30) or constructed a nominal compound by using a
free morpheme from the two languages each, as in (31) and (32).
(26) Ich kann das nicht rightig (richtig) schreiben.
(27) Wir abandonieren (verlassen) die Schule um 17:00.
(28) Er will den Nobel Preis winnen (gewinnen).
(29) Meine Hobbys sind Müzik hören und Swimmen (schwimmen).
(30) Ich habe viel zu viel Arbeit, aber trotzdem gehe ich dancen
(31) Wenn ich wache auf im Morgen, ich trinke Apfeljuice
(Apfelsaft) und esse eine Banane.
(32) Man braucht einen Topf für eine Tomatosuppe (Tomatensuppe).
5.3 Orthographically hybrid forms
Orthographically hybrid forms, which represented the most frequent
category of CLI in the present study, constituted 48% (59 tokens) of the total
124 tokens identified. Among these, only one, sample (43), was directly
attributable to influence from L1 Turkish orthography. The remaining 58
reflected the direct influence of the learners' knowledge of L2 English
orthography. Samples (33)-(42) below clearly display how frequent orthographic
clusters in English were preferred. In some instances, it was very obvious that
the effect of English orthography was even preferred when an alternative form
in the Turkish orthography was available. The sh cluster exemplified in samples
(33), (35) and (42), which is widespread in high-frequency words of English
like shoe, bush, fish, etc., for example, was preferred to using the Turkish
letter “ş” (as in şemsiye, şişe, ışık, etc.) which represents the
same phonetic realization [š]. The strong effect of the English orthographic
system was also evident in the frequent use of the th cluster as in highly frequent
English words such as three, brother, and father, exemplified in (34), (40) and
(41), and the gh cluster as in English words like weight, light, fight,
exemplified in (36)-(39).
(33) Ist die Milch frish (frisch)?
(34) Was felth (fehlt) ihnen?
(35) Um 11:00, sie ging Bazaar und einkaufte das Geshenk
(Geschenk) für ihre Freundin.
(36) Um 8:30 flieght (fliegt) die Machiene nach New York ab.
(37) Nein, ich steighe (steige) lieber um.
(38) Sie ruft an und fraght (fragt) nach den Preisen.
(39) Er seight (sieht) aus wie jung.
(40) Sie thrinkt (trinkt) eine Cola.
(41) Mein Bruther (Bruder) arbeitet bei Akbank.
(42) Ich gehe jetzt shon (schon) zur Uni um meine Vize zu
(43) Ich mag Müsik (Musik) hören, Bücher lesen, fernsehen und mit meiner
6. Discussion and conclusion
Overall, the results of the present study have demonstrated that
the Turkish-English-German trilingual participants under investigation clearly
preferred to transfer lexical items from their L2 English rather than from
their L1 Turkish into L3 German. As was shown, independent of type of CLI, the
great majority of lexical items transferred into written L3 German was
transferred from L2 English (116 out of 124) rather than from L1 Turkish (8 out
of 124). This was in fact expected in the light of findings of previous studies
into multilingual CLI since, as mentioned at various points throughout this
article, studies conducted on CLI to date have pointed to the strong impact of
two independent factors (among others), which apparently were in strong
interaction in the present study: psychotypology, the perceived distance
between languages by a language learner, and the tendency of language learners
to suppress their L1 in foreign language environments due to the belief that
using non-L1 structures would constitute a better strategy in
'foreign-language' environments (Williams & Hammarberg 1998).
In the present study, both of these above-mentioned highly
plausible factors were embodied in the L2 English of the participants. Their L2
English was very likely to be perceived as more similar to the target language
L3 German due to the various formal features that these two languages share and
at the same time constituted the only possible source language to fall back
upon in instances where the L1 was going to be suppressed in order to use
non-L1 structures. Therefore, it might be possible to refer to an 'augmented'
L2 English effect in the present study due to the probable interaction of these
two factors, which reflected itself in the almost complete absence of CLI from
L1 Turkish in the L3 German data. As mentioned before, the students were well
aware of the fact that any non-German structure, no matter if English or
Turkish, would lead to the loss of valuable points on the exam. However,
regardless of this notion, the participants displayed a distinct preference for
transferring items from L2 English, which clearly speaks for the view that the
participants most probably felt it would be more reasonable to stay in the 'foreign-language
A relevant, striking observation that was not part of the original
focus of investigation, but emerged in the course of the analyses of the exam
papers of the participants relates to little notes that the participants were
found to have taken on the papers during their exams. It was observed that
these individual notes, basically translations or vague definitions of unknown
or lesser known German lexical items scribbled in the immediate environment of
unknown target words within question stems, were almost exclusively written in
English. In other words, the participants tried to refrain from using L1
Turkish and instead preferred using L2 English even in instances that they knew
would actually not constitute part of the evaluation process. This observation
can be taken as further evidence for the tendency of language
learners outlined above to stay in the foreign language mode in the learning of additional
In conclusion, the results of the present study have further confirmed
the findings of earlier studies on CLI in multilingual acquisition contexts
that have underscored the impact of (psycho)typology and the tendency of
language learners to stay in a 'foreign-language' mode in the processing of
additional languages. However, as mentioned at the outset of the present study,
the field of multilingual CLI is still so young and comparatively unexplored that
further studies may yield clues to different directions that need to be taken
to arrive at more valid and reliable results. Therefore, needless to say, it is
absolutely required to increase the number of studies on multilingual
acquisition and to vary the languages involved to be able to arrive at a
clearer picture of the processes involved.
Bannert, Robert (2005), Phonological Interferences in the Third
Language Learning of Swedish and German (FIST). Vortrag bei FONETIK 2005,
Beach, Elizabeth; Burnham, Denis & Kitamura, Christine (2001),
Bilingualism and the relationship between perception and production:
Greek/English bilinguals and Thai bilibial stops. International Journal of
Bilingualism 5: 2, 221-235.
Cenoz, Jasone (2001), The effect of linguistic distance, L2 status
and age on cross-linguistic influence in third language acquisition. In: Cenoz,
Jasone; Hufeisen, Britta & Jessner, Ulrike (Hrsg.) (2001), 8-20.
Cenoz, Jasone; Hufeisen, Britta & Jessner, Ulrike (Hrsg.)
(2001), Cross-linguistic Influence in Third Language Acquisition:
Psycholinguistic Perspectives. Clevedon: Multilingual Matters.
De Angelis, Gessica & Selinker, Larry (2001), Interlanguage
transfer and competing linguistic systems in the multilingual mind. In: Cenoz,
Jasone; Hufeisen, Britta & Jessner, Ulrike (Hrsg.) (2001), 42-58.
Dell, Garry S. (1995), Speaking and misspeaking. In: Gleitman,
Lila R. & Liberman, Mark (Hrsg.) (1995), Language. An Invitation to
Cognitive Science. Vol. 1. Cambridge, MA: The MIT Press, 183-208.
Hammarberg, Björn (2001), Roles of L1 and L2 in L3 production and
acquisition. In: Cenoz, Jasone; Hufeisen, Britta & Jessner, Ulrike (Hrsg.)
Jessner, Ulrike (1999), Metalinguistic awareness in multilinguals:
cognitive aspects of third language learning. Language Awareness 8: 3 & 4,
Jorda, Maria Pilar Safont (2005), Pragmatic production of third
language learners of English: A Focus on Request Acts Modifiers. International
Journal of Multilingualism 2: 2, 84-104.
Kecskes, Istvan & Papp, Tünde (2000), Foreign Language and
Mother Tongue. Mahwah, NJ: Lawrence Erlbaum Associates.
Kellerman, Eric (1977), Towards a characterization of the strategy
of transfer in second language learning. Interlanguage Studies Bulletin 2: 1,
Kellerman, Eric (1983), Now you see it, now you don't. In: Gass,
Susan M. & Selinker, Larry (Hrsg.) (1983), Language Transfer in Language
Learning. Rowley: Newbury House, 121-134.
Kellerman, Eric (2001), New uses for old languages:
cross-linguistic and cross-gestural influence in the narratives of non-native
speakers. In: Cenoz, Jasone; Hufeisen, Britta & Jessner, Ulrike (Hrsg.)
König, Wolf; Cedden, Gülay & Onaran, Sevil (2005), Language
production in Turkish-German-English trilinguals. International Journal of
Multilingualism 2: 2, 135-148.
Mayer, Mercer (1969), Frog, where are you? New York: Dial Books
for Young Readers.
O'Laoire, Muiris; Burke, Mary & Haslam, Mary (2000), From L2
to L3/L4: an investigation of learner's metalinguistic awareness and learner
strategies. Teangeolas Journal of the Linguistic Institute of Ireland 38/39,
Odlin, Terence (1989), Language Transfer. Cross-linguistic
Influence in Language Learning. Cambridge: Cambridge University Press.
Poulisse, Nanda & Bongaerts, Theo (1994), First language use
in second language production. Applied Linguistics 15: 1, 36-57.
Ringbom, Hakan (1983), Borrowing and lexical transfer. Applied
Linguistics 4: 3, 207-212.
Ringbom, Hakan (1987), The Role of First Language in Foreign
Language Learning. Clevedon: Multilingual Matters.
Ringbom, Hakan (2001), Lexical Transfer in L3 Production. In:
Cenoz, Jasone; Hufeisen, Britta & Jessner, Ulrike (Hrsg.) (2001), 59-68.
Sağın Şimşek, S. Çiğdem (2006), Third Language Acquisition.
Turkish-German Bilingual Students' Acquisition of English Word Order in a
German Educational Setting. Münster: Waxmann Verlag.
Serindağ, Ergün (2005), Zur Relevanz des bewussten Einsatzes des
Englischen im Unterricht "Deutsch als zweite Fremdsprache" bei
Muttersprachlern des Türkischen. Zeitschrift für Interkulturellen Fremdsprachenunterricht
10: 2, 16 S. [Online: zif.spz.tu-darmstadt.de/jg-10-2/beitrag/Serindag4.htm.
16. Mai 2007.]
Tremblay, Marie Claude (2006), Cross-linguistic influence in third
language acquisition: the role of L2 proficiency and L2 exposure. Cahiers
Linguistiques d'Ottawa/Ottawa Papers in Linguistics 34, 109-119.
Williams, Sarah & Hammarberg, Björn (1998), Language switches
in L3 production: implications for a polyglot speaking model. Applied
Linguistics 19: 3, 295-333.
2007 Zeitschrift für Interkulturellen Fremdsprachenunterricht
Kırkıcı, Bilal (2007), Meine Hobbys sind Müzik hören und Swimmen:
Lexical Transfer in L1 Turkish-L2 English Learners of L3 German.
Fremdsprachenunterricht [Online] 12: 3, 12 S.