Content uploaded by Jasmin Sadat
Author content
All content in this area was uploaded by Jasmin Sadat on Sep 21, 2015
Content may be subject to copyright.
The Routledge Encyclopedia of SLA
166: Bilingual language production - Lexical access and selection
Albert Costa, Jasmin Sadat, Clara D. Martin
Bilingual speakers have the remarkable ability to speak alternatively two different
languages with rare involuntary intrusions from the disregarded language to the
language in use (see Poulisse & Bongaerts, 1994). During language production,
bilinguals do not only have to select the appropriate words to convey information as
monolinguals do, but they also have to guarantee that those words belong to the
intended language. The extra cognitive process that bilinguals need to put at play during
language production is referred to as language control. This language control ability is
even more remarkable considering the speed and reliability with which speakers
produce language (about 4 words per second, with less than 1 error per 1000 words
produced). In the following, we will describe why and how bilingual speakers need
language control for lexical access and selection during language production. Then, we
will turn to the consequences of mastering two languages, in the linguistic but also in
the non-linguistic domain. Finally, we will briefly describe some current limitations of
our knowledge on the consequences of being a bilingual in more complex speech
production contexts.
How lexical access and selection are controlled in the bilingual brain?
Each time a speaker produces a word, at least three different levels of representations
are involved (e.g., Caramazza, 1997; Dell, 1986; Levelt, 1989): Production of a word
implies the retrieval of semantic, lexical and phonological information associated to the
word. One key assumption concerning the dynamics in the language production system
is spreading activation (Collins & Loftus, 1975): When a speaker intends to produce a
word (e.g., “apple”), the semantic information related to the intended concept is
activated (e.g., concept of apple), but semantic related information is also activated
(e.g., concept of pear, orange etc.). According to most speech production models, the
semantic related information which is activated spreads proportional activation to the
corresponding lexical and phonological representations.
In the current section, we will focus on two core components of language production:
lexical access and selection. Lexical access refers to the retrieval of the appropriate
lexical and phonological representations of a given concept. As several lexical
representations are activated during lexical access, selection mechanisms are needed to
select the appropriate one. This process is called lexical selection: the lexical
representation with the highest activation level will be selected among all activated
representations. The selection process is more difficult when lexical competitors are
highly activated, which results in longer naming latencies and/or spontaneous speech
errors.
This system of lexical access and selection becomes especially intriguing in the
bilingual brain as bilingual speakers have two different lexical representations
(translation words) corresponding to each concept. The main question is then to know
how lexical access and selection operate when two lexical and phonological
representations are associated to the concept a bilingual speaker wants to express. The
most straightforward hypothesis would be to assume that the disregarded language is
completely shut down while speaking in the “active” language. This would put bilingual
speakers in the same conditions as monolinguals each time they intend to speak in one
of their two languages. This hypothesis is unlikely as several experiments showed that
lexical and phonological representations of the disregarded language are activated while
speaking in the active language (for a review see Costa, 2005). Evidence for this
assumption comes for example from studies using cognate words. Cognates are
translation words with high phonological overlap across the two languages of a
bilingual (e.g., tomato – tomate [in Spanish]). Several studies have shown that cognate
words are produced faster in comparison to non-cognate words (i.e. apple – manzana [in
Spanish]; e.g., Costa, Caramazza, & Sebastián-Gallés, 2000). These findings suggest
that during bilingual word production, lexical and phonological representations of the
intended and unintended languages become activated. Lexical items of the two
languages spread activation to their phonological representations, thereby boosting the
processing of phonologically overlapping translations. Assuming this permanent co-
activation of both languages, the language production system of a bilingual speaker has
to benefit from some sort of control to select the lexical representation of the intended
language.
Two different accounts can explain how bilingual speakers avoid interference
from the disregarded language while speaking in the other language: According to the
Inhibitory Control Model (Green, 1998), lexical selection is language non-specific,
meaning that activated representations in the disregarded language would compete with
representations from the target language. In order to achieve successful language
production, an overall suppression of representations from the disregarded language
would be applied, ensuring selection of the lexical representation in the target language.
The alternative account assumes a language-specific selection mechanism (e.g., Costa &
Caramazza, 1999). According to this account, lexical selection would be a priori
restricted to lexical representations from the target language. Lexical representations in
the disregarded language would be ignored and not enter into competition.
Language proficiency seems to be the key factor influencing bilingual language
control mechanisms. Language control mechanisms put at play during speech
production are functionally different between high- and low-proficient bilinguals (Costa
& Santesteban, 2004): Low proficient bilinguals seem to largely rely on the active
suppression of their native language when speaking in their second language (language
non-specific selection mechanism). On the contrary, high proficient bilinguals would
automatically select their language output without experiencing much interference from
the unintended language (language specific selection mechanism).
We know that being a bilingual involves having specific language control mechanisms
to avoid interferences between the two languages during speech production. In the next
section, we will give an insight on the consequences of this specific control mechanisms
implemented in the bilingual brain, at the linguistic and non-linguistic levels.
Consequences of mastering two languages
Although being able to speak two languages has obvious advantages, bilinguals incur
some linguistic disadvantages in their two languages, meaning in the non-dominant (L2)
but also the dominant language (L1). First of all, speaking in a L2 is overall slower and
less accurate than in a L1 as observed in picture naming experiments (for a review see
Hanulová, Davidson, & Indefrey, 2010). This L2 naming disadvantage has been
observed in many different languages and across different age groups (e.g., Gollan,
Montoya, Cera, & Sandoval, 2008). Note that this disadvantage is observed even in
bilinguals who are highly proficient in L2 and acquired their L2 early in life (Ivanova &
Costa, 2008).
Strikingly, the bilingual disadvantage in picture naming is not only observed
when bilinguals have to produce speech in their L2, but also when they speak in their
first and dominant language (Ivanova & Costa, 2008). This means that mastering two
languages leads to a disadvantage in word production, whatever the proficiency and the
language in use.
Apart from picture naming, the bilingual disadvantage has been observed in various
language production tasks: Bilinguals experience more Tip-of-the-Tongue states than
monolinguals (i.e., feeling of knowing an infrequent object’s name, but being unable to
retrieve it immediately; e.g., Gollan & Silverberg, 2001). When participants have to
generate as many exemplars of a given semantic category (e.g., animals), bilinguals
have slower first response times and retrieve less category members than monolinguals
(e.g., Rosselli & Ardila, 2002). Importantly, when having to classify picture names into
categories, monolingual and bilingual speakers perform equally (Gollan, Montoya,
Fennema-Notestine, & Morris, 2005). This suggests that the bilingual disadvantage does
not originate at the semantic level, but rather at a post-semantic stage, during
lexicalisation. Up to now, it is unclear at which post-semantic stage the naming
disadvantage occurs. Based on neurophysiological data, some authors propose that the
bilingual naming disadvantage originates at a very late post-lexical level during speech
production (Indefrey, 2006) while others suggest that it originates at the lexical level
(Strijkers, Costa, & Thierry, 2010).
The bilingual disadvantage in language production might be theoretically explained in
two ways: It might stem from the fact that bilinguals have to constantly control two
languages, and thus be explained in terms of executive control. Given that bilinguals
can produce speech in two different languages, their disadvantage can also be explained
on a linguistic basis.
According to the executive control account, the bilingual disadvantage is the
consequence of applying language control mechanisms during speech production
(Green, 1998). Since lexical representations of the two languages are co-activated
during lexical access, bilinguals are constantly exposed to potential competition
between translations. To avoid interference between languages, bilinguals have to apply
some sort of language control. This additional processing would slow down lexical
access and selection in bilinguals compared to monolinguals.
The alternative but not mutually exclusive explanation assumes that the bilingual
disadvantage is a frequency effect in disguise (linguistic account; Gollan et al., 2008).
That is, bilinguals use each of their two languages less frequently than monolinguals use
their unique language. Therefore, frequency-of-use of the bilingual’s lexical
representations should be lower than those of monolingual’s. Given that word frequency
negatively correlates with the speed of lexical retrieval (Oldfield & Wingfield, 1965),
bilinguals would show a disadvantage in lexical access relative to monolinguals.
It is important to note that although bilinguals suffer certain linguistic disadvantages, it
is known for a few years that using two languages has positive collateral effects for the
development of various cognitive abilities besides language. The general argument is
that recruiting executive control processes every time bilinguals produce language
(language control mechanisms) results in more efficient development of these processes
relative to monolinguals. As a consequence, bilinguals might benefit from better general
executive control, even in the non-linguistic domain. This prediction has been
confirmed by several studies comparing the performance of bilinguals and
monolinguals at different stages of development in non-linguistic tasks requiring
executive control. It has been demonstrated that bilingualism enhances executive
control abilities such as conflict resolution, monitoring or task switching (see
publications on collateral effects of bilingualism; e.g., Costa, Hernandez, & Sebastián-
Gallés, 2008).
Bilingualism effects beyond single word production
Up to now, most of the studies comparing linguistic performance of mono- and
bilingual speakers have explored simplified experimental contexts, in which single word
retrieval is required (e.g., picture naming and verbal fluency tasks). We do not know
whether the bilingual disadvantage is also present in contexts where multiword
utterances or sentences have to be produced. For example, in order to produce a simple
noun phrase in English (i.e., “the red car”), the speaker not only needs to retrieve the
two lexical items corresponding to the object (car) and the property (red), but also the
grammatical and syntactic rules of the language in use (correct determiner, order of the
words etc.). At present we still do not know whether and how these additional processes
(e.g., grammatical and syntactic encoding) modulate the bilingual disadvantage in
speech production. If it turns out that the bilingual disadvantage is minimized in more
complex utterances, the consequences of this disadvantage for the regular use of
language might be negligible. By contrast, if such disadvantages are maximized in
multiword utterances, then consequences of bilingualism for speech performance could
be even more important than previously thought. Future studies will have to describe in
full detail the mechanisms that are sensitive to bilingualism in complex linguistic
contexts.
CROSS REFERENCES
35: bilingualism;
36: bilingual processing and representation;
104: Executive control and frontal cortex;
117: frequency effects;
142: Inhibitory control;
165: Levelt's model of speech production and comprehension.
REFERENCES
Caramazza, A. (1997). How many levels of processing are there in lexical access?
Cognitive Neuropsychology, 14(1), 177-208.
Collins, A. M., & Loftus, E. F. (1975). A spreading-activation theory of semantic
processing. Psychological Review, 82(6), 407-428.
Costa, A. (2005). Lexical Access in Bilingual Production. Kroll, J.F. & De Groot,
A.M.B. (Eds.), Handbook of Bilingualism: Psycholinguistic Approaches. Oxford
University Press.
Costa, A., & Caramazza, A. (1999). Is lexical selection in bilingual speech production
language-specific? Further evidence from Spanish–English and English–Spanish
bilinguals. Bilingualism: Language and Cognition, 2, 231–244.
Costa, A., Caramazza, A., & Sebastián-Gallés, N. (2000). The cognate facilitation
effect: Implications for models of lexical access. Journal of Experimental
Psychology.Learning, Memory, and Cognition, 26(5), 1283-1296.
Costa, A., Hernandez, M., & Sebastián-Gallés, N. (2008). Bilingualism aids conflict
resolution: Evidence from the ANT task. Cognition, 106, 59–86.
Costa, A., & Santesteban, M. (2004). Lexical access in bilingual speech production:
Evidence from language switching in highly proficient bilinguals and L2
learners. Journal of Memory and Language, 50(4), 491-511.
Dell, G. S. (1986). A spreading-activation theory of retrieval in sentence production.
Psychological Review, 93(3), 283-321.
Gollan, T. H., Montoya, R. I., Cera, C., & Sandoval, T. C. (2008). More use almost
always a means a smaller frequency effect: Aging, bilingualism, and the weaker
links hypothesis. Journal of Memory and Language, 58(3), 787-814.
Gollan, T. H., Montoya, R. I., Fennema-Notestine, C., & Morris, S. K. (2005).
Bilingualism affects picture naming but not picture classification. Memory &
Cognition, 33(7), 1220-1234.
Gollan, T. H., & Silverberg, N. B. (2001). Tip-of-the-tongue states in Hebrew–English
bilinguals. Bilingualism: Language and Cognition, 4(01), 63-83.
Green, D. W. (1998). Mental control of the bilingual lexico-semantic system.
Bilingualism: Language and Cognition, 1(2), 67-81.
Hanulová, J., Davidson, D. J., & Indefrey, P. (2010). Where does the delay in L2 picture
naming come from? psycholinguistic and neurocognitive evidence on second
language word production. Language and Cognitive Processes,
Indefrey, P. (2006). A meta-analysis of hemodynamic studies on first and second
language processing: Which suggested differences can we trust and what do they
mean? Language Learning, 56: 279–304.
Ivanova, I., & Costa, A. (2008). Does bilingualism hamper lexical access in speech
production? Acta Psychologica, 127(2), 277-288.
Levelt, W. J. M. (1989). Speaking : From intention to articulation. Cambridge, Mass.:
MIT Press.
Oldfield, R. C., & Wingfield, A. (1965). Response latencies in naming objects. The
Quarterly Journal of Experimental Psychology, 17(4), 273-281.
Poulisse, N., & Bongaerts, T. (1994). First language use in second language production.
Applied Linguistics, 15, 36-57.
Rosselli, M., & Ardila, A. (2002). A cross-linguistic comparison of verbal fluency tests.
International Journal of Neuroscience, 112(6), 759-776.
Strijkers, K., Costa, A., & Thierry, G. (2010). Tracking lexical access in speech
production: Electrophysiological correlates of word frequency and cognate
effects. Cerebral Cortex, 20 (4), 912–928.
WORD COUNT (excluding REFERENCES): 1875
