Post written by François Grosjean.
In an earlier post, we saw that bilingual language production is a dynamic process which can operate in different language activation states depending on a number of factors (see here). These can be linguistic but also psycholinguistic and sociolinguistic such as who you are talking to, whether you are using the "right language" to talk about the subject in question, how well you know the language you are speaking, how recently you have spoken the other language, the presence of speakers of the other language(s), and so on.
Is this also true of language perception? Researchers have spent considerable time examining the way bilinguals listen to, or read, their languages, and for many years they came to the conclusion that perceptual processing is nonselective, that is, that all the bilingual's languages are involved in the processes that take place during the acts of listening to, or reading, just one language. But is this always the case?
Although many different experimental tasks have been used to study this question, I will concentrate on a relatively recent one, at least in the study of bilingual speech processing. It is the eye-tracking technique which allows the experimenter to see where the participant is looking while listening to speech. In a first study, researchers Michael Spivey and Viorica Marian asked their Russian-English participants to look at a board situated in front of them which contained a number of objects. For example, there was a target object, a stamp, which had to be moved; there were also, along with the stamp, a competitor object (a marker), or a control object (a ruler), as well as filler objects.
In the Russian language session, the participants heard sentences such as “Poloji marku nije krestika” (Put the stamp below the cross) and researchers examined whether they looked at interlingual competitor objects (in this case, the marker) whose English name ("marker") shared the same onset as the target object ("marku"). This was compared to what happened in a control condition where the name of the control object (e.g. "ruler") had no phonetic similarity with "marku".
The researchers found significantly more eye movements to the interlingual competitor objects than to the control objects which seemed to show that the word onset of the target object (e.g. "marku") not only activated Russian words but also activated English words. Based on this, they concluded that processing is nonselective, that is that both Russian and English were involved in the processing.
This first study led to many other studies by other researchers with different pairs of languages. Viorica Marian and Michael Spivey came back to this question four years later because they realized that the contextual factors in their first study had probably pushed their participants towards a bilingual mode of processing, thereby activating both languages, and hence encouraging nonselective processing. Among the factors they mentioned was the fact that the participants knew they were taking part in an experiment on bilingualism, they were tested by bilingual experimenters who were fluent in both languages, and the two languages (Russian and English) were tested in adjacent experimental sessions.
Thus, this time they attempted to put their participants in as close to a monolingual mode as possible, the kind of situation many bilinguals find themselves in on a daily basis (e.g. at work where no-one else speaks their other language(s)). They used different experimenters for the Russian and English sessions, the experimenters spoke only one language, and the participants only took part in the one or the other session.
The result they found this time (in the Russian session again) was that the competitor objects were not looked at more than the control objects. Hence, in this case, the other language had been totally put aside (deactivated) and processing only took place in Russian.
So, how does listening take place in bilinguals? What seems clear is that the incoming speech wave is processed by the language(s) that contain(s) elements of that input. This can indeed lead to nonselective processing when words from different languages have similar word beginnings, or when homophones and cognates are involved. If the input, however, contains elements of just one language, then only one language will process it ordinarily.
Other factors that play a role are "top-down" factors such as who is speaking and the accompanying context, both linguistic and situational. Sometimes these factors contradict the "bottom-up" information (the speech signal) as when a listener is shocked to hear the speaker use a language that is not expected. (Imagine coming out with a sentence in Spanish to your bilingual Spanish-English friend who has never ever heard you speak one word of Spanish). When this happens, the listener may even have to ask the speaker to repeat what was said.
Another important factor is the listener's fluency in the language being spoken. If it is the stronger language that is being processed, then the weaker language may not intervene as much or at all. However, if it is the other way around, that is, the weaker language is being processed, then there if a fair chance that the stronger language may be active, and may be influencing processing.
In sum, the bilingual's language perception system, much like its production counterpart, is dynamic and can operate in different activation states depending on a number of linguistic, psycholinguistic and sociolinguistic factors.
Photo of a woman and a man from Shutterstock.
Spivey, Michael & Marian, Viorica. 1999. Cross talk between native and second languages: Partial activation of an irrelevant lexicon. Psychological Science, 10, 281–284.
Grosjean, Francois (2013). Speech perception and comprehension. Chapter 2 in Grosjean, François & Li, Ping. The Psycholinguistics of Bilingualism (pp. 29-49). Malden, MA & Oxford: Wiley-Blackwell.
"Life as a bilingual" posts by content area.
François Grosjean's website.