You are on page 1of 36

14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023].

See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Language Learning 55:3, September 2005, pp. 379–414

Interlanguage Transfer of Function Words

Gessica De Angelis
University of Toronto

This study investigated the use of nonnative function


words in the written production of learners of Italian as a
third or fourth language with English, Spanish, or
French as native or nonnative languages. Results show
the frequent use of the French subject pronoun il (he) in
learners’ texts. The rate of subject insertion and omission
was thus analyzed. English and Spanish first language
(L1) speakers with knowledge of French were found to
use significantly more subject insertion than speakers
without knowledge of French. This suggests that (a) prior
exposure to a nonnative language informs learners’ choices
of surface structures to a significant extent and (b) learners
with the same L1 but different prior nonnative languages
develop some significant differences in their target
language knowledge.

One of the principal goals of research on crosslinguistic


influence (CLI) is to account for the role of prior knowledge in
interlanguage (IL) production and development. Hypotheses

I would like to thank the University of London Central Research Fund for
having provided financial assistance with a travel grant and the University of
Toronto Connaught Fund for further assistance. I would also like to thank Jean-
Marc Dewaele and Scott Jarvis for their excellent comments on an earlier draft of
this article. A previous version of this article was presented at the Third
International Conference on Third Language Acquisition and Trilingualism,
Tralee, Ireland, 2003.
Correspondence concerning this article should be addressed to Gessica
De Angelis, Department of French, German and Italian, University of Toronto
at Mississauga, 3359 Mississauga Road North, North Building 227,
Mississauga, Ontario L5L 1C6, Canada. Internet: gdeangel@utm.utoronto.ca

379
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
380 Language Learning Vol. 55, No. 3

about CLI must therefore be sufficiently comprehensive to be able


to account for native as well as nonnative linguistic influence.
Decades of intense research activity on CLI show that much
attention has been devoted to the study of first language (L1)
influence, whereas nonnative linguistic influence has been inves-
tigated to a much lesser extent. The potential influence of non-
native languages in target language production is indeed
acknowledged in various definitions and discussions of language
transfer and CLI that we find in the literature (Gass & Selinker,
1983; Jarvis, 2000; Odlin, 1989; Selinker, 1992; Sharwood Smith,
1994), but our understanding of what can and cannot be trans-
ferred from one nonnative language to another in the process of
acquisition remains quite limited to date. Establishing the role of
prior nonnative linguistic knowledge in IL production and devel-
opment is of relevance to research on CLI not only because we have
limited understanding of the phenomenon, but also because it
raises the questions of how native and nonnative languages inter-
act with each other in shaping IL knowledge and how such inter-
actions may ultimately inform learners’ choices in production.
The broad aim of the present study is to provide empirical
evidence in support of the claim that prior knowledge of non-
native languages may lead to some meaningful differences in
learners’ target language knowledge. The focus on the use of
function words in third (L3) or fourth (L4) language written
production is motivated by prior research findings reported in
the literature on multilingualism (Ringbom, 1987; Stedje, 1977;
Vildomec, 1963; Williams & Hammarberg, 1998), which suggest
that multilinguals seem to favor the use of function words from
their nonnative languages rather than their native language in
production, provided that the source and the target language are
typologically close to each other. I examine the claim further by
assessing what types of function words (conjunctions, determiners,
prepositions, or pronouns) are used in the written production of
learners of Italian as an L3 or L4. First, I assess the overall
frequency of use of nontarget content and function words from the
informants’ native and nonnative languages (English, Spanish, and
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
De Angelis 381

French). Second, I examine whether some nontarget and nonnative


function words may be used more frequently than others, coming
to identify the French subject pronoun as a high-frequency non-
target function word used in the learners’ written texts. Third, I
assess whether the frequent use of nontarget subject pronouns is a
chance occurrence or is determined by underlying crosslinguistic
constraints. I examine the overall rate of subject insertion and
omission in the Italian target language, proceeding to argue that
prior nonnative linguistic knowledge seems to inform learners’
choices in regard to surface structures to a significant extent.
As the present study focuses on the influence of previously
acquired nonnative languages upon an L3 or an L4, I do not
provide a review of the literature on L1 influence, unless other-
wise specified. I examine, however, some of the hypotheses of L1
influence that are being extensively referred to in the literature
on multilingualism and discuss whether they can account for
multilinguals’ behavior as well. One of these hypotheses is
Kellerman’s (1977, 1978, 1983, 1984, 1986, 1987) notion of ‘‘psy-
chotypology,’’ which is reviewed in the next section.

Typology and Psychotypology

In the literature on multilingualism, language similarity


has been repeatedly reported as one of the key factors in the
degree of influence likely to occur between two or more non-
native languages (Ahukanna, Lund, & Gentile, 1981; Cenoz,
2001; Clyne, 1997; Clyne & Cassia, 1999; De Angelis & Selinker,
2001; Dewaele, 1998; Selinker & Baumgartner-Cohen, 1995;
Stedje, 1977; Vildomec, 1963; Williams & Hammarberg, 1998).
The idea that language similarity is a crucial factor in CLI is
not a new one and has in fact been discussed for decades in the
CLI literature (Andersen, 1983; Gass & Selinker, 2001; Jarvis,
2000; Odlin, 1989). The role of language similarity in CLI, how-
ever, presents some new challenges when one aims to account for
the production of multilingual speakers who are familiar with
more than two languages that are similar to each other.
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
382 Language Learning Vol. 55, No. 3

One highly influential hypothesis that stems from research


on L1 influence is Kellerman’s (1977, 1978, 1983, 1984, 1986,
1987) notion of ‘‘psychotypology’’ or perceived language distance,1
which refers to the relations that learners perceive to exist
between the native language and the target language, that is,
the perceived linguistic distance between L1 and second language
(L2) forms or structures. Whereas in the recent literature on L1
influence, the term psychotypology has been largely replaced by
terms such as language distance and typological proximity
(Jarvis, 2000; Odlin, 1989; Ringbom, 1987), discussions on psy-
chotypology, as well as the use of the term, frequently appear in
the literature on multilingualism and the acquisition of languages
beyond the L2 (Bouvy, 2000; Cenoz, 2001; Cenoz & Jessner, 2000;
De Angelis & Selinker, 2001; Ecke, 2001; Kellerman, 2001;
Ringbom, 2001). When a hypothesis developed to account for L1
influence is used to account for nonnative linguistic influence and
multilinguals’ production, however, some considerations arise.
This section examines the notion of psychotypology in its original
formulation more closely and outlines some considerations that
are central to the study on function words that follows.
The notion of psychotypology emerged from studies on
idiomatic expressions conducted with Dutch L1 learners of
English (Kellerman, 1977, 1978) and a study also on idiomatic
expressions conducted with Dutch L1 learners of German
(Jordens, 1977). Kellerman’s major theoretical concern was to
be able to establish the potential transferability of idiomatic
expressions, which he claimed was determined by two interact-
ing constraints: psychotypology and prototypicality. He proposed
that, when two languages are perceived as being close to each
other (psychotypology), L1 features are likely to influence L2
forms, but it is the second constraint (prototypicality) that is
argued to be the stronger determinant of CLI, with the claim
that ‘‘the less representative of the prototypical meaning a usage
of a given form is, the lower its transferability’’ (Kellerman,
1987, p. 65). Of concern to the present discussion is whether
the notion of psychotypology, as originally formulated, can be
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
De Angelis 383

regarded as sufficiently powerful to explain multilinguals’ beha-


vior as well. Hereafter I will therefore restrict the focus of the
discussion to the notion of psychotypology.
Multilinguals may be familiar with several languages that are
typologically close to the target language, a case in which instances
of CLI cannot be easily accounted for with the psychotypology expla-
nation alone. If psychotypology is taken to refer to multilinguals’
perception of typological proximity, it is required first of all that we
determine what forms or structures learners are expressing judg-
ments about, as it is what learners identify as similar or dissimilar
across linguistic systems that constrains their judgments, that is,
what is interlingually identified as similar across linguistic systems
and not what can be formally defined as similar across linguistic
systems. Weinreich (1953) first pointed out that bilinguals establish
equivalencies between languages where equivalence does not exist
and described the phenomenon as ‘‘a type of interference . . . which
does not involve an outright transfer of elements at all’’ (p. 7).
In the literature on multilingualism, we find mixed reports
on learners’ reliance on their prior knowledge of nonnative lan-
guages. Most studies show that multilinguals tend to rely on
linguistic information from nonnative languages that are ty-
pologically close to the target language, as psychotypology
would predict (Clyne, 1997; Clyne & Cassia, 1999; De Angelis
& Selinker, 2001; Dewaele, 1998; Ecke, 2001; Herwig, 2001;
Möhle, 1989; Ringbom, 1986, 1987, 2001, 2002; Vildomec, 1963;
Williams & Hammarberg, 1998). There are also some reports
that multilinguals rely on nonnative languages typologically
more distant from the target language, despite having knowl-
edge of nonnative languages typologically close to the target
language (Rivers, 1979; Schmidt & Frota, 1986).
Typology, in its broadest of definitions, is concerned with
the ‘‘classification of structural types across languages’’ (Croft,
1990, p. 1), including the study of crosslinguistic typological
patterns. Typology and psychotypology are associated whenever
learners express judgments of typological proximity. My concern
is to establish the possible relationship between the actual
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
384 Language Learning Vol. 55, No. 3

typological patterns across languages and learners’ perception of


typological proximity. I discuss performance data of multilingual
learners who are familiar with Romance and Germanic lan-
guages, and I specify which structures I regard as being typolog-
ically similar or dissimilar to each other.
As anticipated in the introductory section, the present
study initially examines whether some nonnative function
words may be used more frequently than others in L3 or L4
written production. We shall see that this initial assessment
shows a difference in the frequency of use of subject pronouns
from the French nonnative language. Given the difference in
patterns of subject use in all the languages that participants
are familiar with (English, Spanish, French, and Italian), the
study then proceeds by focusing on the subject category2 across
the four languages, and the typological similarities and differ-
ences that exist between these languages. French and English
have obligatory subjects, and so they are regarded as having typolog-
ically similar structures. Spanish and Italian do not have obliga-
tory subjects and are also regarded as having typologically similar
structures. It follows that French and Italian, although genetically
related and typologically close, display typologically dissimilar
structures, whereas English and French, which are typologically
more distant, display typologically similarstructures.

The Use of Nontarget Function Words in Target Language


Production

One area of inquiry in which results obtained from research


in L1 and L2 speech production and research in multilingualism
seem to point in the same direction is the use of function words.
Garrett (1975) first pointed out that the speech errors of mono-
lingual speakers hardly ever involve function words (see also the
discussion in Dell, 1995) and several explanations have been
provided to account for this phenomenon. Bock (1995) comments
that ‘‘competing explanations for this apparent invulnerability
to error fuel a controversy about whether these elements are
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
De Angelis 385

retrieved in much the same way as lexemes (e.g., Stemberg,


1984) or are specified in a manner that, in effect, bypasses the
need for lexical retrieval (Garrett, 1982)’’ (p. 187).
Some studies on bilingual speech production regard the use
of L1 content or function words in L2 speech as a form of borrow-
ing that speakers employ in order to compensate for their lack of
knowledge in the target language. The use of L1 forms in speech
is regarded as a compensatory strategy (Poulisse, 1997; Poulisse
& Bongaerts, 1994), the general argument being that the use of
L1 forms occurs because the L2 system is not as highly devel-
oped and automatized as the native language system.
Poulisse and Bongaerts (1994) examined the use of L1 con-
tent words (nouns, verbs, numerals, adjectives, and most adverbs)
and L1 function words (prepositions, determiners, conjunctions,
and pronouns) in the oral production of 45 Dutch learners of
English as an L2. Fifteen of their participants were 2nd-year
university students (ages 19 to 22), 15 were Grade 11 students
(ages 15 and 16), and 15 were Grade 9 students (ages 13 and 14).
The authors examined 771 unintentional language switches,
which were divided into nonadapted language switches and
morphologically or phonologically adapted language switches.
The distinction between intentional and unintentional switches
was made based on ‘‘hesitation phenomena and intonation’’
(Poulisse & Bongaerts, 1994, p. 43). Those items that were not
preceded by hesitation and did not carry a strong intonation were
regarded as unintentional language switches. The authors found
that Dutch learners of English in Grade 11 and in Grade 9 used
more L1 function words than L1 content words in their L2 speech.
Poulisse and Bongaerts (1994) argue that L1 function
words are likely to be used in L2 production because of their
frequency of occurrence in the L1, which facilitates access to
them. They also provide further explanations by drawing on
Giesbers’s (1989) proposal that content words are more likely
to be selected correctly than function words because they carry
more meaning than function words do. L2 learners need to focus
their attention on the most meaningful parts of speech, and it is
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
386 Language Learning Vol. 55, No. 3

therefore more likely that their attention will be directed toward


the selection of content rather than function words. Differential
attention results in reduced monitoring of the output and an
increase in the use of L1 function words in speech.
Research findings reported in the literature on the acquisi-
tion of languages beyond the L2 indicate that in oral and written
production, multilinguals tend to use function words from their
nonnative languages rather than their native language, pro-
vided that the source and the target language are typologically
close to each other. Vildomec (1963) initially pointed out that
multilinguals often use function words from nonnative lan-
guages in production, even when function words in the source
and the target language are dissimilar in sound. Stedje (1977, as
reported in Williams & Hammarberg, 1998) examined the oral
production of 55 Finnish L1 learners of German as an L3 with
Swedish as an L2 who had lived in Sweden for less than 4 years.
He found that learners mostly used nontarget function words
from the Swedish L2 rather than the Finnish L1. Williams and
Hammarberg (1998) also report that Stedje3 found some differ-
ences in behavior with learners of German who have knowledge
of two languages typologically close to the target language.
Finnish L1 speakers with knowledge of Swedish as an L2 tend
to use more Swedish function words than Swedish L1 speakers
with knowledge of English as an L2. The latter group uses more
function words from the English L2 rather than from the
Swedish L1, and this seems to suggest that in the presence of
languages typologically close to each other, learners tend to
favor the L2 over the L1 as a source of function words.
Ringbom (1987) analyzed about 11,000 written essays of
Finnish L1 learners of English with Swedish as an L2. He found
complete language shifts from the Finnish L1 to be relatively rare
for both content and function words. He recorded only five instances
of function words from Finnish, and three of these words were the
Finnish conjunction ja (and), which was primarily used between
proper names. Ringbom (1987) fully acknowledges the rarity of
shifts from the Finnish L1 for function words, which he ascribes to
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
De Angelis 387

the typological distance between Finnish and English. He also notes


that the infrequent use of function words from Finnish illustrates
‘‘the tendency noted by Vildomec (1963, p. 170) to shift function
words to a foreign language from another foreign language rather
than from the L1, even when no formal similarities between these
words exists’’ (p. 122). The majority of the complete language shifts
he recorded were in fact from the Swedish L2, a language typologi-
cally close to the English target language. Function words made up
18% of the total number of errors traceable to Swedish. Among the
most frequently used words, Ringbom (1987) notes 13 instances of
the word fast (although), 10 instances of men (but), and 8 instances
of och (and). Ringbom (1987) generally explains the use of L2 func-
tion words in L3 written production by drawing a distinction
between knowledge and control and arguing that ‘‘with function
words . . . the attention given to control procedures tends to slacken,
since the learner gives only peripheral attention to them, normally
focusing on those other words in his utterance which are commu-
nicatively the weightiest’’ (Ringbom, 1987, p. 128).
Additional evidence of the use of nonnative function words
in L3 production is provided by Williams and Hammarberg
(1998), who examined the Swedish production of an English L1
speaker with German as an L2 and identified several instances
of nonadapted language switches in the Swedish target lan-
guage. Williams and Hammarberg (1998) categorize their data
by drawing a distinction between words used with an intended
pragmatic purpose and words used without a clear pragmatic
purpose. Of most relevance to the present discussion are the
switches without an identified pragmatic purpose (WIPP) that
were identified. The following example illustrates a WIPP switch
in which the speaker uses the German preposition mit (the
percentage sign indicates a pause filler): ‘‘en tjuv % <mit>/med
en nickel; GLOSS: a thief % <with>/with a key’’ (Williams &
Hammarberg, 1998, p. 308). Most of the WIPP switches in the
participant’s Swedish production were function words, with 92% of
the WIPP switches from the German L2, 4% from the English L1,
and 4% from other L2s. In the presence of three typologically close
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
388 Language Learning Vol. 55, No. 3

languages, the learner thus favored the German L2 over the English
L1 for function words. In Hammarberg (2001) we also find the
argument that the frequent use of function words in WIPP switches
is an indication that the L2 is active during the production process.
The focus up to this point in the article has been on the use
of L2 function words in L3 production, but it must be noted that
L3 production can also be influenced by L1 prepositions in some
meaningful ways. In a study on morphological transfer, for
instance, Jarvis and Odlin (2000) show that Finnish L1 speakers
with English and Swedish as L2s and Swedish L1 speakers with
English and Finnish as L2s are influenced by L1 patterns in
their use of prepositions. Their study shows semantic transfer
from the L1 and, once again, the important role of typological
distance in influencing whether CLIs occur in L3 production.
The studies reported so far indicate that the use of L2 function
words in L3 oral or written production is strongly influenced by
typological distance. Learners tend to rely on the language that is
typologically closest to the target language. When both the L1 and the
L2 are typologically close to the target language, however, learners
seem to favor the L2 over the L1 as a source of function words.
The present study explores further the use of nontarget
function words in written production by drawing a distinction
between the types of function words that learners use and the
similarities and differences of typological patterns in all the
languages that participants are familiar with. From a typolog-
ical perspective, two main questions are thus addressed:
1. Do multilinguals tend to use function words from their
native or nonnative languages?
2. Does the knowledge of one or more nonnative languages
inform learners’ choices of surface forms and structures?
The first question is addressed by way of frequency counts.
The analysis of learners’ errors, however, can provide only par-
tial information regarding the extent to which CLI may manifest
itself, most notably because CLI is sometimes covert, as in the
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
De Angelis 389

case of avoidance (Schachter, 1974). The second question is thus


addressed by focusing on instances of subject insertion and omis-
sion in the performance data. Subject insertion here refers to
whether overt subjects (nouns and pronouns) are present in
learners’ written summaries. Details concerning the partici-
pants and research design are provided below.

Method

Participants

The participants were adult learners of Italian as an L3 or


L4 at the University of Pittsburgh (Pennsylvania) and the
University of Puerto Rico at Mayaguez. These universities were
chosen for a number of reasons. At the University of Pittsburgh a
large of number of the native-English-speaking students study
Italian as an L3 or additional language and also have some
knowledge of Spanish or French as an L2. Italian is also exten-
sively studied at the University of Puerto Rico, where the vast
majority of the student population speaks Spanish as a native
language and also has prior knowledge of English or English and
French as nonnative languages. Puerto Ricans begin to study
English at a very young age (the most frequently reported age is 5 or
6). They learn English at school and are not immersed in an English-
speaking environment. The study of French is less widespread.
Participants in the study were all enrolled in a 1st-year Italian
language course at their university. A total of 54 English L1 speak-
ers took part, of whom 37 had prior knowledge of Spanish and 17
had prior knowledge of French as a nonnative language. There were
also 54 Spanish L1 participants, of whom 45 had prior knowledge of
English and 9 had prior knowledge of English and French.
All participants completed a sociobiographical question-
naire in their native language (see English version in Appendix
A) in which they were asked to indicate the length of study in
each of their foreign languages. The English L1 participants’
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
390 Language Learning Vol. 55, No. 3

proficiency level in Spanish and French was initially measured


by way of the questionnaire. The participants’ answers were
then compared with their performance on a 30-item translation
task (see Appendix B), which showed their knowledge of Spanish
to be quite low. Those who gave no more than 10 correct answers
in the Spanish translation task were accepted as having a simi-
lar level of proficiency and were included in the final participant
pool. As the items to be translated consisted of basic vocabulary,
these participants’ performance on the translation task was
taken as an indication that their overall Spanish proficiency
level could be safely considered low.
Those participants who were not familiar with Spanish
were asked to carry out the translation task in French, and
French proficiency level was assessed in a similar fashion to
that for assessing Spanish proficiency. Learners’ poor perfor-
mance on the French translation task indicated that their profi-
ciency level could also be safely considered as low. Those who
gave no more than 10 correct answers in the French translation
task were accepted as having a similar proficiency level and
were included in the final participant pool.
In Puerto Rico, the participants’ knowledge of English was
measured using the years of study indicated on the sociobiograph-
ical questionnaire. As all learners reported having studied
English for between 12 and 16 years, their English proficiency
level was regarded as high. Those with previous knowledge of
French reported having studied French between 1 and 1.5 years.
Their proficiency level was thus regarded as low.
The data at both universities were collected at the end of the
first term, a week apart from each other. Learners were asked to
complete all tasks during their regular 50 min of class time.

Materials

All participants (English L1 and Spanish L1 speakers) were


asked to read the same text on the life of a university student
and write a summary of it in the target language (Italian). The
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
De Angelis 391

text was taken from the Spanish textbook ¿Sabias que?


(VanPatten, Lee, Ballman, & Dvorak, 1992) and was presented
to each participant in his or her native language. (The Spanish
version is presented in Appendix C and the English version,
translated from Spanish, in Appendix D.)
All participants were instructed not to go back to a task
once completed and not to use dictionaries. The researcher close-
ly monitored all participants to ensure that instructions were
being followed.

Results

The results of the study are presented in two subsections:


(a) use of nontarget function words from the L1 and the non-
native languages and (b) subject insertion and omission.

Use of Nontarget Function Words From the First and Nonnative


Languages

Tables 1 and 2 summarize the use of nontarget content and


function words in the written summary of English L1 speakers
with prior knowledge of Spanish as an L2 or IL (n ¼ 37) and of
Spanish L1 speakers with prior knowledge of English as a non-
native language (n ¼ 45). English L1 speakers (Table 1) use

Table 1

Number of occurrences and relative frequency of English L1 and


Spanish IL content and function words used in the written sum-
mary task
Content Relative frequency Function Relative frequency

English L1 79 (53.74%) 8 (18.60%)


Spanish IL 68 (46.26%) 35 (81.40%)
Total 147 (100%) 43 (100%)

Note. n ¼ 37.
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
392 Language Learning Vol. 55, No. 3

Table 2

Number of occurrences and relative frequency of Spanish L1 and


English IL content and function words used in the written summary
task
Content Relative frequency Function Relative frequency

Spanish L1 203 (98.54%) 66 (97.06%)


English IL 3 (1.46%) 2 (2.94%)
Total 206 (100%) 68 (100%)

Note. n ¼ 45.

content words from both the English L1 (53.74%) and the


Spanish IL (46.26%). For function words, however, the preferred
source of linguistic information is the Spanish IL (81.40%)
rather than the English L1 (18.60%). The analysis of 45 summa-
ries written by Spanish L1 speakers with English IL (Table 2)
shows that Spanish L1 speakers resist using the English non-
native language for both content (1.46%) and function words
(2.94%). The Spanish L1 is clearly the preferred source of lin-
guistic information for both content (98.54%) and function words
(97.06%).
Tables 3 and 4 summarize the use of content and function
words in the written summaries of English L1 speakers with
French IL (n ¼ 17) and of Spanish L1 speakers with French and

Table 3

Number of occurrences and relative frequency of English L1 and


French IL content and function words used in the written summary
task
Content Relative frequency Function Relative frequency

English L1 29 (72.5%) 1 (9.09%)


French IL 11 (27.5%) 10 (90.91%)
Total 40 (100%) 11 (100%)

Note. n ¼ 17.
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
De Angelis 393

Table 4

Number of occurrences and relative frequency of Spanish L1 and


English or French IL content and function words used in the written
summary task
Content Relative frequency Function Relative frequency

Spanish L1 17 (80.95%) 11 (36.67%)


English IL 0 (0%) 1 (3.33%)
French IL 4 (19.05%) 18 (60.00%)
Total 21 (100%) 30 (100%)

Note. n ¼ 9.

English ILs (n ¼ 9). Results in Table 3 show that, in the case of


content words, English L1 speakers rely extensively on the
English L1 (72.50%) and to a lesser extent on the French IL
(27.50%). For function words, however, the preferred source of
linguistic information is the French IL (90.91%) rather than the
English L1 (9.09%). Seven out of the 10 French function words
used were the same word: the French subject pronoun il (he).
Results in Table 4 show that Spanish L1 speakers use nontarget
content words from either the Spanish L1 (80.95%) or the French
IL (19.05%), but not at all from the English IL. For function
words, the use of the English IL is minimal (3.33%) when com-
pared with that of the Spanish L1 (36.67%) and the French IL
(60.00%). Similar to the results obtained with English L1 speak-
ers with prior knowledge of French (Table 3), 12 out of the 18
French function words that Spanish L1 speakers used were the
same word: the French subject pronoun il (he).
The Spanish subject pronoun el and the French subject
pronoun il are formally similar to each other (the Italian third-
person singular subject pronoun is lui). The summaries of a total
of 91 participants with prior knowledge of Spanish as a non-
native language (n ¼ 37) and of Spanish as a native language
(n ¼ 54) were further analyzed in order to establish the rate of
use of the Spanish subject pronoun el (he). Table 5 shows the
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
394 Language Learning Vol. 55, No. 3

Table 5

Use of the Spanish subject pronoun el and the French subject pro-
noun il in the summaries of Spanish L1 and English L1 speakers
n Use of el Use of il

Spanish L1 and English IL 45 0 0


English L1 and Spanish IL 37 0 0
English L1 and French IL 17 0 7
Spanish L1 and English þ French IL 9 0 12

results of the frequency count and provides a comparison with


the frequency of use of the French subject pronoun il. All
instances in Table 5 refer to the use of il and el with the function
of subject pronoun in the Italian target language. Results show
that no participants, regardless of language combination, used
the Spanish subject pronoun el in their written summaries.
Neither Spanish L1 speakers with English IL (n ¼ 45) nor
English L1 speakers with Spanish IL (n ¼ 37) used the French
subject pronoun il, and the summaries of a total of 26 partici-
pants (17 English L1 and 9 Spanish L1 speakers) with French as
a nonnative language contained a total of 19 instances of the
French subject pronoun il.

Subject Insertion and Omission

The rate of subject insertion and omission in the Italian


target language was analyzed in a total of 108 summaries of
English L1 and Spanish L1 speakers. All overt subjects, whether
nouns or pronouns, were counted as instances of subject inser-
tion. The observed frequency of subject insertion or omission is
summarized in Tables 6 and 7. A one-sample Kolmogorov-
Smirnov Test4 showed the samples to be normally distributed,
with a significance value of .282 for English L1 speakers and of
.117 for Spanish L1 speakers.
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
De Angelis 395

Table 6

Number of occurrences of subject insertion or omission in the sum-


mary task of English L1 speakers with Spanish or French as non-
native languages
ILs n Number of subject insertions Number of subject omissions

Spanish 37 105 175


French 17 99 21

A one-way analysis of variance (ANOVA) with speaker


group as independent variable (French or Spanish as nonnative
languages) and rate of subject insertion as dependent variable
showed a highly significant effect, F(1, 52) ¼ 16.62, p < .0002,
Z2 ¼ .242.5 The 17 English L1 speakers with French as an add-
itional language used more subject insertion (M ¼ 78.36%,
SD ¼ 28.32) than the 37 English L1 speakers with Spanish as
an additional language (M ¼ 36.55%, SD ¼ 37.58).
Similar differences appear in the group of Spanish L1 par-
ticipants. A one-way ANOVA with speaker group as independent
variable (English or English and French as nonnative
languages) and rate of subject insertion as dependent variable
showed a significant effect, F(1, 52) ¼ 6.58, p < .014, Z2 ¼ .112.
The 45 Spanish L1 speakers with English as an additional lan-
guage used less subject insertion (M ¼ 30.95%, SD ¼ 26.32)

Table 7

Number of occurrences of subject insertion or omission in the sum-


mary task of Spanish L1 speakers with English and English þ
French as nonnative languages
Number of subject Number of subject
ILs n insertions omissions

English 45 86 224
English þ French 9 35 26
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
396 Language Learning Vol. 55, No. 3

90
80
% subject insertion

70
60
50
40
30
20
10
0
English L1 + English L1 + Spanish L1 + Spanish L1 +
French Spanish English English +
French

Figure 1. Mean proportions of subject insertion in the summary task of


English and Spanish L1 speakers with different additional languages.

than the 9 Spanish L1 speakers with English and French as


additional languages (M ¼ 57.83%, SD ¼ 39.70). Figure 1
graphs the proportion of subject insertion among the study’s
participants.

Discussion

The answer to the study’s first question (Do multilinguals


tend to use function words from their native or nonnative
languages?) is that the multilingual participants of the present
study used function words from both their first and nonnative
languages, provided that the source and the target language
were typologically close to each other. Of particular interest
are the results obtained from Spanish L1 speakers with prior
knowledge of French as a nonnative language (Table 4), as they
provide an insight into multilinguals’ behavior when three lan-
guages typologically close to each other (Spanish, French, and
Italian) are present in their minds. These learners used nontar-
get function words from both the Spanish L1 (11 instances) and
the French IL (18 instances), which suggests that both lan-
guages were perceived as close to the target language. As
noted, 12 out of the 18 French function words consisted of the
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
De Angelis 397

same word: the French subject pronoun il (he). The frequent use
of the French subject pronoun also emerged from the analysis of
the summaries written by English L1 speakers with prior knowl-
edge of French, where 7 out of the 10 nontarget function words
found consisted of the French subject pronoun il.
The frequent use of one type of function word over others
raises the question of whether it is a chance occurrence or is the
surface manifestation of nonnative linguistic influence. Before
proceeding to provide an answer to this question, I first examine
whether some of the existing explanations6 on the use of non-
target function words in L2 or L3 production can account for the
evidence produced in the present study.
One explanation for the use of L1 function words in L2 oral
production is provided by Poulisse and Bongaerts (1994), who
examined the oral performance of Dutch L1 learners of English,
focusing on the use of unmodified nontarget content and function
words in L2 speech. The authors found L1 function words to be
used more frequently than L1 content words in L2 speech and
proposed that this may occur because L1 function words are more
frequent than L1 content words. Function words also are shorter,
require less activation, and are consequently selected faster than
content words. An explanation based on frequency of occurrence
may plausibly account for the behavior of speakers with knowl-
edge of two languages only, as L1 function words can generally be
argued to be used more frequently than L1 content words in
speech, but it cannot account for the results of the present
study. As noted, participants made extensive use of nontarget
function words from their nonnative languages, and it would not
be feasible to account for these data by claiming that nonnative
function words may be used more frequently than L1 function
words, unless the nonnative language were the speakers’ most
dominant language and were in regular use. This was not the
case for the multilingual participants in the present study.
Poulisse and Bongaerts (1994) also argue that L1 content
words are more likely to be selected correctly than L1 function
words because content words carry more semantic weight and
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
398 Language Learning Vol. 55, No. 3

learners focus their attention on the most meaningful parts of


speech. An argument based on semantic weight is also advanced
by Ringbom (1987), who found L2 function words to be used
more frequently than L1 function words in the written essays
of Finnish L1 learners of English with Swedish as an L2.
Ringbom (1987) argues that L2 function words are frequently
used in L3 production because learners focus their attention on
the most meaningful parts of speech. Since function words carry
less semantic weight than content words, learners generally
devote less attention to them in the production process. These
two proposals thus advance similar arguments based on the
different semantic weight of content and function words. In
Poulisse and Bongaerts (1994), the argument accounts for the
use of L1 function words in L2 speech, whereas in Ringbom
(1987), the argument accounts for the use of L2 function words
in L3 written production. What these proposals seem to share is
their focus on target language monitoring processes, which may
be affected by the semantic weight that content and function
words carry. Although semantic weight draws a useful distinc-
tion between the way content and function words may be treated
in the mind, the results of the present study show that learners
seem to be selective in their choices of nontarget function words,
favoring the use of nonnative pronouns over other types of func-
tion words (i.e., determiners, conjunctions, and prepositions).
Williams and Hammarberg (1998) proposed a model of L3
production that accounts for the use of nontarget language
words in performance. The model proposes that one of the speak-
ers’ nontarget languages becomes the preferred source of non-
target linguistic information by being assigned the role of default
supplier in the production process. Four factors determine which
of the speaker’s nontarget languages is assigned the role of
default supplier: proficiency, recency, typology, and L2 status.
The results of the current study provide support for Williams
and Hammarberg’s (1998) proposal, as learners seem to give
preference to one nontarget language over others that are avail-
able to them. In the presence of two languages typologically close
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
De Angelis 399

to the intended target, however, the preference seems to be less


clear, particularly with regard to function words.
Hammarberg (2001) argues that the frequent use of non-
target and nonnative function words in production is an indica-
tion that the L2 is activated in parallel to the target language
during the production process. The activation metaphor is fre-
quently discussed in the literature on multilinguals’ production
(De Angelis & Selinker, 2001; Dewaele, 1998, 2001; Dijkstra &
van Hell, 2002; Grosjean, 2001; Herwig, 2001; Singleton, 2002),
and some of the evidence these studies present suggests that
more than one language can be activated in parallel to the target
language in the production process. Although parallel language
activation is a plausible explanation for the general use of non-
target linguistic information in production, the evidence of this
study shows that some nontarget function words are used more
frequently than others, and this suggests that learners are selec-
tive in their choices. This is a difference that must be accounted
for by going beyond the language activation metaphor.
The explanations reviewed so far on the use of nontarget
function words in L2 or L3 production focus on frequency of
occurrence, semantic weight, attentional capacity, reduced mon-
itoring processes, and parallel language activation. None of
these proposals makes a distinction between the different types
of function words that learners use, a distinction that instead
emerged quite clearly from the results of the present study. This
takes us to the answer to the study’s second question: Does the
knowledge of one or more nonnative languages inform learners’
choices of surface forms and structures?
Overall results show that prior knowledge of French as a
nonnative language informs learners’ choices of surface struc-
tures in the Italian target language to a considerable extent. The
high rate of subject insertion in the summaries of English L1
speakers with previous knowledge of French is perhaps not
surprising, since both English and French are languages that
have obligatory subjects. These learners had never been pre-
viously exposed to a language that does not have obligatory
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
400 Language Learning Vol. 55, No. 3

subjects and may have consequently perceived the pattern of


subject insertion as transferable to the Italian target language.
The rate of subject insertion in the summaries of English L1
speakers with previous knowledge of Spanish is instead consid-
erably lower. The higher rate of subject omission suggests that
learners perceived the similarity that exists between the pattern
of subject omission in the Spanish nonnative language and the
Italian target language and regarded it as transferable to the
Italian target language.
A similar picture seems to emerge from the summaries of
Spanish L1 speakers. Learners with previous knowledge of
English as a nonnative language perceived the similarity that
exists between the native and the target language and regarded
the pattern of subject omission as transferable to the Italian
target language. The rate of subject insertion in the summaries
of Spanish L1 speakers with previous knowledge of both English
and French is, however, considerably higher, and this suggests
that learners were influenced by their knowledge of French.
Learners’ overall perceptions of the similarities that exist
between French and Italian may have led them to regard the
pattern of subject insertion as transferable to the Italian target
language, partially overriding the similarity they may have per-
ceived to exist between the Spanish native language and the
Italian target language.
The value of these results lies in the fact that participants
in both groups had the same ‘‘constants,’’ that is, the L1 and the
target language, but differed in their previous knowledge of
nonnative languages. The differential rate of subject insertion
in both groups suggests that the difference cannot be the result
of L1 influence, but that it is the result of nonnative linguistic
influence. The difference that learners perceived to exist
between the typological patterns of their native and nonnative
languages triggered or constrained subject insertion or omission
in the Italian target language. In some cases, it seems to have
facilitated the acquisition process; in others, it seems to have
delayed it.
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
De Angelis 401

The similarity or dissimilarity of typological structures and


how these may be perceived by multilingual speakers cannot be
entirely separated from learners’ overall understanding of the
similarities and differences that may exist between their lan-
guages. It cannot be assumed a priori, however, that whenever
two languages typologically close to each other are present in the
speaker’s mind and one of them is the speaker’s native language,
it is the native language that will have the most dominant role.
The results of the present study in fact show that in the presence
of three typologically similar languages (Spanish L1, French L2,
and Italian as the target language), learners are influenced by
their knowledge of French to a considerable extent.
It was previously pointed out that Kellerman’s notion of psy-
chotypology is frequently discussed in the literature on multilin-
gualism, where it is used to account for instances of CLI in
multilinguals’ performance. Although it can be said that the
results of the present study show psychotypology to be at work,
these results also indicate that when two languages ‘‘close’’ to the
target language (e.g., Spanish L1 and French L2) are present in
their minds, learners do not seem to perceive nontarget function
words belonging to these languages as equally transferable to the
target language. Most importantly, the results show that the
native language does not always have a privileged status. As pre-
viously noted, if the notion of psychotypology is taken to refer to
multilinguals’ perception of typological proximity, it is required
first of all that we determine what forms or structures learners
are expressing judgments about. The notion of psychotypology can
in fact lead to fruitful generalizations related to nonnative linguis-
tic influence when used to refer to specific forms or structures that
learners are expressing judgments about rather than to the overall
similarity that learners may perceive to exist between their lan-
guages. A notion that seems to be better suited to accounting for
the general similarities and differences that learners perceive to
exist between the target language and any other language present
in their mind is proposed in Ringbom (2002), which argues for the
existence of an overall level of transfer, which is determined by
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
402 Language Learning Vol. 55, No. 3

‘‘how much cross-linguistic similarity the learner can generally


perceive, beginning from a common alphabet and phonemes in
common over the division into grammatical categories (case, gen-
der, word classes) to the number of cognates and other lexical
similarities’’ (p. 1).
In the present study the rate of subject insertion and omis-
sion was analyzed in order to assess whether the high frequency
of use of the French subject pronoun il (he) was the surface
manifestation of CLI or a chance occurrence. Overall results
seem to suggest that the frequent use of the French nontarget
function word il in the participants’ Italian production is the sur-
face manifestation of nonnative linguistic influence rather than a
chance occurrence. There are some limitations to this interpreta-
tion, however, as other factors may have also been at play.7
The frequent use of the French subject pronoun il (he) may
be partly due to an intralingual overgeneralization process, as
the word il is also a definite article in the Italian target lan-
guage. The presence of words with a different meaning in the
source and the target language has been previously shown to
have an effect on learners’ target language production. Ringbom
(1987), for instance, found that his participants (Finnish L1 and
Swedish L2) used 13 instances of the Swedish word fast
(although) and 10 instances of the Swedish word men (but) in
their English essays. He argued that the existence of the words
fast and men in English may have contributed to their frequent
use in the target language, even though these words do not
share the same meaning in English and Swedish.
Another issue to consider relates to the potential influence
of Spanish. The Spanish subject pronoun el (he) has no formal
counterpart in Italian, but in Spanish the word el functions as a
definite article (the) as well as a subject pronoun (he). There is
then the possibility that those participants who had prior knowl-
edge of Spanish as a native or nonnative language may have
equated the function of the Italian word il with the function of
the Spanish word el (definite article and subject pronoun). The
data available seem to provide mixed support for this proposal.
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
De Angelis 403

Some of the participants (English L1, French L2, and Italian L3)
used the French subject pronoun il but were not familiar with
Spanish at all, and those learners with Spanish L1 and English
L2, and with English L1 and Spanish L2, never used the word il
as a subject pronoun. On the other hand, a closer look at the
results presented in Table 5 shows that 17 English L1 speakers
with French IL used 7 instances of the French subject pronoun
il, and the smaller group of 9 Spanish L1 speakers with both
English and French as ILs used a total of 12 instances of the
French subject pronoun il. The data then seem to suggest that
the potential influence from Spanish cannot be ruled out com-
pletely, as learners with knowledge of Spanish seem to have
used the French subject pronoun il more frequently than those
who had no prior knowledge of Spanish. In other words, there is
the possibility that prior knowledge of Spanish, together with
prior knowledge of French, may have increased the likelihood
that the word il was used with the function of subject pronoun in
the Italian target language. Similar combined effects are also
reported in Ringbom (1987) and Dewaele (1998). The frequency
data of this study are clearly limited and can provide only some
partial answers in this regard. Further studies would be needed
to assess the extent to which interlingual processes, as well as
the combined knowledge of typologically close languages, may
influence L3 or L4 production and may ease the path of CLI as
discussed.

Conclusion

The present study examined the use of nonnative function


words in the written production of learners of Italian as an L3
or additional language with English, Spanish, or French as
native or nonnative languages. The study aimed to establish
(a) whether multilinguals tend to use nontarget function words
in written production from their L1 or their nonnative languages
and (b) whether prior knowledge of one or more nonnative lan-
guages informs learners’ choices of surface forms and structures.
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
404 Language Learning Vol. 55, No. 3

Results show that Spanish L1 and English L1 learners of


Italian as an L3 or L4 rely extensively on their nonnative lan-
guages for function words, provided that the source and the
target language are typologically close to each other. Results
also show that learners seem to be selective about the type of
nonnative function words they use in written production.
English L1 and Spanish L1 speakers with prior knowledge of
French made extensive use of the French subject pronoun il (he)
in their written summaries, a result that was taken as an initial
indication that learners’ choices may have been the result of
nonnative linguistic influence rather than a chance occurrence.
The overall rate of subject insertion and omission was examined
in the texts written by Spanish L1 and English L1 speakers. A
one-way ANOVA showed the rate of subject insertion to be sig-
nificantly higher in the summaries of English L1 and Spanish L1
speakers with knowledge of French as a nonnative language.
These results suggest that (a) some nontarget and nonnative
function words may be more transferable than others; (b) prior
knowledge of a nonnative language seems to inform learners’
choices of surface forms (function words); and (c) prior knowl-
edge of a nonnative language seems to inform learners’ choices of
surface structures (subject insertion and omission).
Overall findings suggest that multilingual learners with
the same L1, but a different L2, may develop some differences
in their target language knowledge and therefore that prior
knowledge of a nonnative language is a central variable for the
study of CLI and IL development. Whether hypotheses of L1
influence, and more broadly of CLI, aim to be predictive or to
hypothesize tendencies, they will need to take into account that
target language knowledge may also be affected by learners’
prior knowledge of one or more nonnative languages. Further
research is needed to clarify to what extent the interaction
between native and nonnative linguistic knowledge may influ-
ence target language production and development and what the
broader implications for research on CLI may be. At present it
can only be said that hypotheses of CLI cannot be based on
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
De Angelis 405

speakers of two languages only. Prior nonnative linguistic


knowledge must also be accounted for, as it seems to influence
target language production and development in some meaning-
ful ways.
Revised version accepted 18 January 2005

Notes
1
The term psychotypology is proposed in Kellerman (1983), but the idea of
perceived linguistic distance is extensively discussed in Kellerman’s earlier
work (Kellerman, 1977, 1978).
2
Most discussions on subjects and CLI focus on the prodrop parameter and
the clustering of properties such as subject-verb inversion or that-trace
effects (Cook, 1993; Gass & Selinker, 2001), or on the initial/final state
debate (White, 2000). To the best of my knowledge, most studies on
Universal Grammar (UG) and CLI address research questions related to
L1 influence rather than nonnative linguistic influence. The present study
does not focus on L1 influence and does not address UG-related questions.
As will become apparent, its findings are of relevance to ongoing debates on
CLI and UG, but a discussion on the topic is too complex to be dealt with in
this article and will form the body of a different article.
3
Personal communication between Williams and Hammarberg and Stedje.
4
Values of less than .05 on this test indicate that the observed distribution
does not correspond to the theoretical distribution.
5
According to Cohen (1992), values between 2% (.02) and 12.99% (.1299)
suggest a small size effect, and those between 13% (.13) and 25.99% (.2599)
indicate a medium size effect.
6
The proposals reviewed in this section refer to oral as well as written
production. Even though learners have more time to monitor their output
when writing, at present we do not have clear evidence that nontarget
function words are used less frequently in written production.
7
I thank the anonymous Language Learning reviewer for his or her insight-
ful observations.

References

Ahukanna, J. G. W., Lund, N. J., & Gentile, R. J. (1981). Inter- and intra-
lingual interference effects in learning a third language. Modern
Language Journal, 65, 281–287.
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
406 Language Learning Vol. 55, No. 3

Andersen, R. (1983). Transfer to somewhere. In S. Gass & L. Selinker


(Eds.), Language transfer in language learning (pp. 177–201). Rowley,
MA: Newbury House.
Bock, K. (1995). Sentence production: From mind to mouth. In J. L. Miller
& P. D. Eimas (Eds.), Speech, language, and communication (pp.
181–216). San Diego, CA: Academic Press.
Bouvy, C. (2000). Towards the construction of a theory of crosslinguistic trans-
fer. In J. Cenoz & U. Jessner (Eds.), English in Europe: The acquisition of a
third language (pp. 143–156). Clevedon, UK: Multilingual Matters.
Cenoz, J. (2001). The effect of linguistic distance, L2 status and age on
cross-linguistic influence in third language acquisition. In J. Cenoz,
B. Hufeisen, & U. Jessner (Eds.), Cross-linguistic influence in third
language acquisition: Psycholinguistic perspectives (pp. 8–20).
Clevedon, UK: Multilingual Matters.
Cenoz, J., & Jessner, U. (Eds.). (2000). English in Europe: The acquisition
of a third language. Clevedon, UK: Multilingual Matters.
Clyne, M. (1997). Some of the things trilinguals do. International Journal
of Bilingualism, 1(2), 95–116.
Clyne, M., & Cassia, P. (1999). Trilingualism, immigration and relatedness
of languages. I.T.L. Review of Applied Linguistics, 123–124, 57–74.
Cohen, J. (1992). Statistical power analysis for the behavioral sciences. New
York: Wiley.
Cook, V. (1993). Linguistics and second language acquisition. Houndmills,
UK: Palgrave.
Croft, W. (1990). Syntactic categories and grammatical relations. Chicago:
University of Chicago Press.
De Angelis, G., & Selinker, L. (2001). Interlanguage transfer and competing
linguistic systems in the multilingual mind. In J. Cenoz, B. Hufeisen, &
U. Jessner (Eds.), Cross-linguistic influence in third language acquisition:
Psycholinguistic perspectives (pp. 42–58). Clevedon, UK: Multilingual
Matters.
Dell, S. G. (1995). Speaking and misspeaking. In L. R. Gleitman &
M. Liberman (Eds.), Language: An invitation to cognitive science (2nd
ed., Vol. 1, pp. 183–201). Cambridge, MA: MIT Press.
Dewaele, J.-M. (1998). Lexical inventions: French interlanguage as L2
versus L3. Applied Linguistics, 19(4), 471–490.
Dewaele, J.-M. (2001). Activation or inhibition? The interaction of L1, L2
and L3 on the language mode continuum. In J. Cenoz, B. Hufeisen, & U.
Jessner (Eds.), Cross-linguistic influence in third language acquisition:
Psycholinguistic perspectives (pp. 69–89). Clevedon, UK: Multilingual
Matters.
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
De Angelis 407

Dijkstra, T., & van Hell, J. V. (2002). Testing the language mode hypothesis
using trilinguals. In J. Ytsma & M. Hooghiemstra (Eds.), Proceedings of
the Second International Conference on Trilingualism. Leeuwarden,
Holland: Fryske Academie[CD-ROM].
Ecke, P. (2001). Lexical retrieval in a third language: Evidence from errors and
tip-of-the-tongue states. In J. Cenoz, B. Hufeisen, & U. Jessner (Eds.),
Cross-linguistic influence in third language acquisition: Psycholinguistic
perspectives (pp. 90–114). Clevedon, UK: Multilingual Matters.
Garrett, M. F. (1975). The analysis of sentence production. In G. H. Bower
(Ed.), The psychology of learning and motivation (pp. 133–177). San
Diego, CA: Academic Press.
Garrett, M. F. (1982). Production of speech: Observations from normal and
pathological language use. In A. Ellis (Ed.), Normality and pathology in
cognitive functions (pp. 19–76). London: Academic Press.
Gass, S., & Selinker, L. (Eds.). (1983). Language transfer in language
learning. Rowley, MA: Newbury House.
Gass, S., & Selinker, L. (2001). Second language acquisition: An introduc-
tory course. Mahwah, NJ: Erlbaum.
Giesbers, H. (1989). Code-switching tussen Dialect en Standaardtaal [Code-
switching between dialect and the standard language]. Amsterdam:
P. J. Meertens-Instituut.
Grosjean, F. (2001). The bilingual’s language modes. In J. L. Nicol (Ed.),
One mind, two languages: Bilingual language processing (pp. 1–22).
Oxford: Blackwell.
Hammarberg, B. (2001). Roles of L1 and L2 in L3 production and acquisi-
tion. In J. Cenoz, B. Hufeisen, & U. Jessner (Eds.), Cross-linguistic
influence in third language acquisition: Psycholinguistic perspectives
(pp. 21–41). Clevedon, UK: Multilingual Matters.
Herwig, A. (2001). Plurilingual lexical organisation. In J. Cenoz, B. Hufeisen, &
U. Jessner (Eds.), Cross-linguistic influence in third language acquisition:
Psycholinguistic perspectives (pp. 115–137). Clevedon, UK: Multilingual
Matters.
Jarvis, S. (2000). Methodological rigor in the study of transfer: Identifying L1
influence in the interlanguage lexicon. Language Learning, 50(2), 245–309.
Jarvis, S., & Odlin, T. (2000). Morphological type, spatial reference, and
language transfer. Studies in Second Language Acquisition, 22,
535–556.
Jordens, P. (1977). Rules, grammatical intuitions and strategies in foreign
language learning. Interlanguage Studies Bulletin, 2(2), 5–7.
Kellerman, E. (1977). Towards a characterization of the strategy of transfer
in second language learning. Interlanguage Studies Bulletin, 2, 58–145.
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
408 Language Learning Vol. 55, No. 3

Kellerman, E. (1978). Giving learners a break: Native language intuitions


as a source of prediction about transferability. Working Papers on
Bilingualism, 15, 59–92.
Kellerman, E. (1983). Now you see it, now you don’t. In S. Gass &
L. Selinker (Eds.), Language transfer in language learning (pp.
112–134). Rowley, MA: Newbury House.
Kellerman, E. (1984). The empirical evidence for the influence of the L1 in
interlanguage. In A. Davies, C. Criper, & A. P. R. Howatt (Eds.),
Interlanguage (pp. 98–122). Edinburgh, UK: Edinburgh University
Press.
Kellerman, E. (1986). An eye for an eye: Crosslinguistic constraints on
the development of the L2 lexicon. In E. Kellerman & M. Sharwood
Smith (Eds.), Crosslinguistic influence in second language acquisition
(pp. 35–48). Oxford, UK: Pergamon Press.
Kellerman, E. (1987). Aspects of transferability in second language acquisi-
tion. Unpublished manuscript.
Kellerman, E. (2001). New uses for old language: Cross-linguistic and cross-
gestural influence in the narratives of non-native speakers. In J. Cenoz,
B. Hufeisen, & U. Jessner (Eds.), Cross-linguistic influence in third
language acquisition: Psycholinguistic perspectives (pp. 170–191).
Clevedon, UK: Multilingual Matters.
Möhle, D. (1989). Multilingual interaction in foreign language production.
In W. H. Dechert & M. Raupach (Eds.), Interlingual processes (pp. 179–
184). Tubingen, Germany: Gunter Narr Verlag.
Odlin, T. (1989). Language transfer: Cross-linguistic influence in language
learning. Cambridge, UK: Cambridge University Press.
Poulisse, N. (1997). Language production in bilinguals. In A. M. B. de Groot
& J. Kroll (Eds.), Tutorials in bilingualism: Psycholinguistic perspectives
(pp. 201–224). Mahwah, NJ: Erlbaum.
Poulisse, N., & Bongaerts, T. (1994). First language use in second language
production. Applied Linguistics, 15(1), 36–57.
Ringbom, H. (1986). Crosslinguistic influence and the foreign language
learning process. In E. Kellerman & M. Sharwood Smith (Eds.),
Crosslinguistic influence in second language acquisition (pp. 150–162).
Oxford, UK: Pergamon Press.
Ringbom, H. (1987). The role of the first language in foreign language
learning. Clevedon, UK: Multilingual Matters.
Ringbom, H. (2001). Lexical transfer in L3-production. In J. Cenoz,
B. Hufeisen, & U. Jessner (Eds.), Cross-linguistic influence in third
language acquisition: Psycholinguistic perspectives (pp. 59–68).
Clevedon, UK: Multilingual Matters.
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
De Angelis 409

Ringbom, H. (2002). Levels of transfer from L1 and L2 in L3 acquisition.


In J. Ytsma & M. Hooghiemstra (Eds.), Proceedings of the Second
International Conference on Trilingualism. Leeuwarden, Holland:
Fryske Academie [CD-ROM].
Rivers, W. M. (1979). Learning a sixth language: An adult learner’s daily
diary. Canadian Modern Language Review, 36(1), 67–82.
Schachter, J. (1974). An error in error analysis. Language Learning, 24,
205–214.
Schmidt, R. W., & Frota, S. N. (1986). Developing basic conversation ability
in a second language: A case study of an adult learner of Portuguese. In
R. Day (Ed.), Talking to learn: Conversation in second language acquisi-
tion (pp. 237–326). Rowley, MA: Newbury House.
Selinker, L. (1992). Rediscovering interlanguage. London: Longman.
Selinker, L., & Baumgartner-Cohen, B. (1995). Multiple language acquisi-
tion: ‘‘Damn it, why can’t I keep these two languages apart?’’ Languages,
Culture and Curriculum, 8(2), 115–121.
Sharwood Smith, M. (1994). Second language learning: Theoretical founda-
tions. London: Longman.
Singleton, D. (2002). Crosslinguistic interactions in the multilingual lexi-
con. In J. Ytsma & M. Hooghiemstra (Eds.), Proceedings of the Second
International Conference on Trilingualism. Leeuwarden, Holland:
Fryske Academie [CD-ROM].
Stedje, A. (1977). Tredjespråksinterferens i fritt tal—en jämförande studie
[Third language interference in free speech—a comparative study]. In R.
Palmberg & H. Ringbom (Eds.), Papers from the Conference on
Contrastive Linguistics and Error Analysis, Stockholm and Åbo, 7–8
February 1977. Åbo, Sweden: Åbo Akademi.
Stemberg, J. P. (1984). Structural errors in normal and agrammatic speech.
Cognitive Neuropsychology, 1, 218–313.
VanPatten, B., Lee, J. F., Ballman, T. L., & Dvorak, T. (1992). ¿Sabias que?
[Did you know that?] New York: McGraw-Hill.
Vildomec, V. (1963). Multilingualism. Leyden, The Netherlands: A. W. Sythoff.
Weinreich, U. (1953). Languages in contact. New York: Linguistic Circle of
New York.
White, L. (2000). Second language acquisition: From initial to final state. In
J. Archibald (Ed.), Second language acquisition and linguistic theory
(pp. 130–155). Oxford, UK: Blackwell.
Williams, S., & Hammarberg, B. (1998). Language switches in L3 produc-
tion: Implications for a polyglot speaking model. Applied Linguistics,
19(3), 295–333.
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
410 Language Learning Vol. 55, No. 3

Appendix A

Questionnaire

Guidelines: Please complete the following section in English:


1. Name:

2. Native language:
3. How many semesters of Italian have you completed at the
University of Pittsburgh?
4. How long have you studied Italian?
5. What level of Italian language are you currently taking at
the University of Pittsburgh?
6. Where did you learn Italian? Please place a check mark
where applicable:
 I learned it at school in the United States.
 I learned it at home with my family.
 I studied in Italy for a total of _____ years and _____
months.
 I was _____ years old when I started to study or learn
Italian.
 Other ______________________________________________.

Foreign Languages

Guidelines: Please complete the following section in English:


1. Please place a check mark next to the languages you have studied
or learned (besides Italian), even if you have studied them only for a
short time:
 Spanish
 French
 German
 Other (please specify) __________
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
De Angelis 411

2. If you had to judge similarity between languages, would you say


that Spanish and Italian are more similar than English and Italian?
Please indicate your answer below:
 Yes
 No
3. If you have ever learned or studied a foreign language besides
Italian, even if it was just for a short time, please indicate the
following:
Language: Spanish
Months/Years of Study:__________
Context:
 I learned it at school in the United States.

 I learned it at home with my family.


 I studied in a Spanish speaking country for a total of _____
years and _____ months.

 I was _____ years old when I started to study Spanish.


 Other _________________________________________________.
Language: French
Months/Years of Study:__________
Context:

 I learned it at school in the United States.


 I learned it at home with my family.

 I studied in a French speaking country for a total of _____


years and _____ months.

 I was _____ years old when I started to study French.


 Other _________________________________________________.
Language:_______________
Months/Years of Study:__________
Context:
 I learned it at school in the United States.
 I learned it at home with my family.
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
412 Language Learning Vol. 55, No. 3

 I studied in _____ for a total of _____ years and _____ months.


 I was _____ years old when I started to study ____________.
 Other _________________________________________________.
Language:_______________
Months/Years of Study:__________
Context:

 I learned it at school in the United States.


 I learned it at home with my family.
 I studied in _____ for a total of _____ years and _____ months.
 I was _____ years old when I started to study ____________.
 Other _________________________________________________.

Appendix B

Translation Into Spanish

Example:
ENGLISH SPANISH
to
register ¼matricularse
1. (to) dry ¼
2. monkey ¼
3. pineapple ¼
4. (to) wet ¼
5. jewel ¼
6. (to) fight ¼
7. clerk ¼
8. apple ¼
9. (to) get up ¼
10. heating ¼
11. (to) go out ¼
12. (to) go down ¼
13. leg ¼
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
De Angelis 413

14. clothing ¼
15. landlord ¼
16. bed ¼
17. brother ¼
18. brush ¼
19. sandwich ¼
20. thin ¼
21. lazy ¼
22. fireman ¼
23. (to) grill ¼
24. hurry ¼
25. forward ¼
26. yellow ¼
27. last name ¼
28. parents ¼
29. throat ¼
30. (to) get married ¼

Appendix C

Una historia verdadera

Pobre Juanito. Anoche se quedó en casa sin tener mucho que


hacer. Preparó su cena (un sandwich y una ensalada), comió y
después estudió unas cuantas horas. A las 10.00 fue al café
Casablanca, pero no encontró a ninguno de sus amigos. Después
de quince minutos, volvió a casa. Miró las noticias en el canal 4 (cree
que los reporteros de ese canal son muy buenos) y luego se acostó.
Pero Juanito no se durmió inmediatamente. Se levantó, buscó entre
sus libros una novela de Stephen King y comenzó a leer. Gracias al
insomnio y a una novela muy interesante, se pasó toda noche
leyendo. Leyó hasta la última página del libro. Cuando miró el
despertador, exlamó: ‘‘¡Ay no! ¡Es hora de ir a la clase de quı́mica!’’
Se tomó trez tazas de café (para no dormirse en clase) y corrió
al edificio de Ciencas Naturales. Cuando llegó, notó algo raro:
‘‘¿Dónde están los otros estudiantes? ¿Por qué no hay nadie aquı́?’’
Entoncés recordó la fecha y pensó: ‘‘¡Hoy es dı́a de fiesta! ¡No hay
clases!’’
14679922, 2005, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.0023-8333.2005.00310.x by University Of Szeged, Wiley Online Library on [28/02/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
414 Language Learning Vol. 55, No. 3

Appendix D

A True Story

Poor Juanito! Last night he was at home without much to do.


He prepared his dinner (a sandwich and a salad), ate and then
studied for a few hours. At 10:00 he went to the Casablanca coffee
shop, but he did not come across any of his friends. After fifteen
minutes, he returned home. He watched the news on Channel 4 (he
thinks reporters on Channel 4 are very good) and then went to bed.
But Juanito did not fall asleep immediately. He got up, searched
among his books for a Stephen King novel and started to read.
Thanks to insomnia and to a very interesting novel, he spent all
night reading. He read until the last page of the book. When he
looked at the alarm clock, he exclaimed: ‘‘Oh, no! It is time to go to
my chemistry class!’’ He drank three cups of coffee (in order not to
fall asleep in class) and ran to the Natural Sciences building. When
he arrived, he noticed something strange: ‘‘Where are all the
students? Why is nobody here?’’ Then he remembered what day it
was and thought: ‘‘Today is a holiday! There are no classes!’’

You might also like