An Introduction to Applied Linguistics

e mi dicevo: ‘ce que je vois là’ (and I was telling to myself: ‘that what I see there’). 
An example of an unacceptable switch from their set is: mais c’était interrotto (but 
it was interrupted). The data show that the processes involved in unacceptable 
switches were different in terms of brain areas involved, and the authors suggest 
that acceptable switches are easier to process while the unacceptable ones are 
processed as violations. The outcomes do not seem to support the patterns of 
switching costs found by Meuter and Allport (1999) and Jackson et al. (2001).

---

135
Psycholinguistics
One problem with the data on switching costs presented so far is that they were 
gathered in an experimental setting. Grosjean and Miller (1994) have argued that 
in natural CS there are no switching costs. The costs found are the result of the 
experimental setting rather than from the switching mechanisms involved. In 
normal interaction, bilingual speakers who are used to switching are never forced 
to switch unexpectedly, the switches are integrated in the production process. 
So far there are no data on the switching costs in natural CS speech. Preliminary 
data on ‘heavy code-switchers’ (data from a study on long term Dutch migrants 
in Australia, de Bot and Clyne in prep., see also Broersma, Isurin, Bultena and de 
Bot 2009) suggest that for such experienced switchers, pauses between words in 
the same language are similar to those for pauses between words from different 
languages. Such analyses are complex though, since it is quite often totally unclear 
to what language elements in CS speech belong. As Clyne (1987) argues, extensive 
CS often involves ‘compromise forms’, words that show traces of more than one 
language.
CS in dialogue
CS has typically been studied from a monologue perspective, in line with the 
general trend in the study of language production. Kootstra, van Hell and Dijkstra 
(2009) propose a change in perspective by using an extension of the interactive 
alignment model by Pickering and Garrod (2004). In this model, dialogue is taken 
as the basic unit of analysis. Kootstra et al. argue that CS typically takes place 
in interaction and that the study of CS in dialogue is ecologically more valid 
than taking a monologue perspective. They also show that it is possible to gather 
experimental data on CS using the so-called ‘confederate paradigm’ in which 
interactional data are gathered in a setting in which one of the interactants is 
actually manipulating the conversation to prime certain types of language use. 
This technique has been used extensively in studies on syntactic priming within 
and between languages. Hartsuiker, Pickering and Veldkamp (2004) carried out 
a syntactic priming experiment in which participants who were intermediate to 
high level Spanish learners of English had to describe pictures after being provided 
with a verbal cue from a confederate. The cues were given in Spanish, while the 
descriptions of the pictures had to be given in English. The critical contrast was 
the use of active versus passive constructions by the confederate and the effect 
this had on the proportion of use of passives by the participants. The results 
showed that the confederate’s use of the passive in Spanish had an impact on the