Because most people today are bilingual, that is, they are language producers and comprehenders of at least two languages, the proposed model by Pickering & Garrod (P&G) cannot be generalized in its current version; it accounts only for a small minority of language users. Compared to earlier language-processing models, it is more complex – in particular, by integrating production and comprehension – with regard to predicting actions and the embodiment of language. However, it fails to account for the important aspects of increased processing demands and processing costs in bilingual language interaction. These additional demands and costs are observed even for bilinguals with native-like knowledge of the language (so-called highly proficient bilinguals) and cannot be attributed to a low level of proficiency. I consider these aspects most relevant for a truly integrated, up-to-date, and generalizable theory of language production and comprehension.
When comparing the performance of monolinguals and bilinguals on the same task, we can trace increased processing demands and processing costs with different measures, two of which I will describe. Such studies investigate the language production in only one language at a time, not during switching between languages and do not make claims about switch costs.
First, increased processing demands for bilinguals can be observed in functional magnetic resonance imaging (fMRI) studies in terms of modulation of the blood-oxygen-level-dependent (BOLD) signal. A recent study (Parker Jones et al. Reference Parker Jones, Green, Grogan, Pliatsikas, Filippopolitis, Ali, Lee, Ramsden, Gazarian, Prejawa, Seghier and Price2011) compared performance of monolinguals and highly proficient bilinguals on naming pictures and reading words aloud. They demonstrated that the same five left hemisphere areas sensitive to increasing demands on speech production in monolinguals showed higher activation in bilinguals. More specifically, during word retrieval and articulation, higher activation was found in dorsal precentral gyrus, pars triangularis, pars opercularis, superior temporal gyrus, and planum temporale. Word retrieval was more demanding for bilinguals than for monolinguals.
Second, processing costs for highly proficient bilinguals are also reported from simple word retrieval tasks. Reaction time latencies are longer for bilinguals than for monolinguals when they were naming pictures or producing a list of tokens from a common semantic category (e.g., Gollan et al. Reference Gollan, Montoya and Werner2002; Reference Gollan, Montoya, Fennema-Notestine and Morris2005).
Increased language-processing demands are attributed to the speaker's necessity of dealing with two languages. Parallel activation of both languages was observed at all stages of bilingual speech production (e.g., Guo & Peng Reference Guo and Peng2006). Consequently, task performance in one of two available languages involves competition of words of the unwanted language with those of the intended language. Therefore, lexical selection is a more demanding process for bilinguals than for monolinguals. As a result, retrieval of words is slower even in highly proficient bilinguals of all ages than in monolinguals (Sorace Reference Sorace2011).
Unbalanced bilinguals are better in one language compared with the other. Hence, they have a stronger and a weaker language. Usually their mother tongue is the stronger language whereas a second, later learned language is the weaker language, characterized by incomplete acquisition of lexicon and grammar. If the first language undergoes attrition (that is, language loss due to infrequent use of that language as a consequence of migration), this language becomes the weaker language and the second language develops to be the stronger language. In both cases, if language proficiency is rather low, it is even more demanding to produce that weak language. It requires more resources to inhibit the stronger, unwanted language (Meuter & Allport Reference Meuter and Allport1999). This means that using a weak language involves much language control to avoid unintentional switching to the stronger language. Several processing models for bilingualism (e.g., Green Reference Green1986) capitalize on the concept of language control and emphasize the problem of limited resource and processing capacities. Applying the proposed theory to a communicative setting with unbalanced bilinguals, I seriously question, in particular, the suggested obligatory action predictions and the effortless involvement in interactive language.
P&G acknowledge one condition under which predictions of one's own and other's actions can be difficult, namely, if they are “unrelated” (an example for a related action would be ballroom dancing; see target article, sect. 2.3, Joint Action). In my view, in the domain of language, the ease of prediction depends not only on relatedness, but also, and much more, on familiarity. Not only are predictions easier and more likely to be correct when communicative participants are familiar with each other and their communicative habits, but also the interlocutors must be familiar with the language they are using for the interaction, and with the culture of that language. It is certainly rather easy to predict actions of communication in a long-standing relationship (e.g., the best friend), but for an exchange student – age 16, foreign to a country, new to a host family, and with little knowledge either of the language spoken or of the family's speech habits – the success rate of other's action prediction is most probably low. Similarly, in foreign-language reading, cultural familiarity has been found to be crucial for comprehension. Studies on non-native reading revealed that if readers lack the relevant cultural knowledge, reading activities could not fully compensate for the discrepancy or help readers comprehend a text (Erten & Razı Reference Erten and Razı2009).
With all this in mind, it becomes clear that the language producer in the proposed model is ideal, even as a monolingual. There are moments when we suffer a snub from another person's action: we are so surprised that words fail us. Some seem to have the gift of gab, and are more quick-witted than others. For such situational or inter-individual variability known from everyday life, the current model does not account. It holds the view that interlocutors are perfectly coordinated and that there is no time gap between interlocutors' turns. Nonetheless, P&G have been aware of differences between native and non-native speakers or of the difference between speaking to an adult or a child (see sect. 4, General Discussion) to some degree, but they did not integrate these differences into their model. The authors might want to elaborate on situations and conditions when the construction of forward models, covert imitations, joint actions, and continuous prediction generation are not ideal, not even for the monolingual.
Because most people today are bilingual, that is, they are language producers and comprehenders of at least two languages, the proposed model by Pickering & Garrod (P&G) cannot be generalized in its current version; it accounts only for a small minority of language users. Compared to earlier language-processing models, it is more complex – in particular, by integrating production and comprehension – with regard to predicting actions and the embodiment of language. However, it fails to account for the important aspects of increased processing demands and processing costs in bilingual language interaction. These additional demands and costs are observed even for bilinguals with native-like knowledge of the language (so-called highly proficient bilinguals) and cannot be attributed to a low level of proficiency. I consider these aspects most relevant for a truly integrated, up-to-date, and generalizable theory of language production and comprehension.
When comparing the performance of monolinguals and bilinguals on the same task, we can trace increased processing demands and processing costs with different measures, two of which I will describe. Such studies investigate the language production in only one language at a time, not during switching between languages and do not make claims about switch costs.
First, increased processing demands for bilinguals can be observed in functional magnetic resonance imaging (fMRI) studies in terms of modulation of the blood-oxygen-level-dependent (BOLD) signal. A recent study (Parker Jones et al. Reference Parker Jones, Green, Grogan, Pliatsikas, Filippopolitis, Ali, Lee, Ramsden, Gazarian, Prejawa, Seghier and Price2011) compared performance of monolinguals and highly proficient bilinguals on naming pictures and reading words aloud. They demonstrated that the same five left hemisphere areas sensitive to increasing demands on speech production in monolinguals showed higher activation in bilinguals. More specifically, during word retrieval and articulation, higher activation was found in dorsal precentral gyrus, pars triangularis, pars opercularis, superior temporal gyrus, and planum temporale. Word retrieval was more demanding for bilinguals than for monolinguals.
Second, processing costs for highly proficient bilinguals are also reported from simple word retrieval tasks. Reaction time latencies are longer for bilinguals than for monolinguals when they were naming pictures or producing a list of tokens from a common semantic category (e.g., Gollan et al. Reference Gollan, Montoya and Werner2002; Reference Gollan, Montoya, Fennema-Notestine and Morris2005).
Increased language-processing demands are attributed to the speaker's necessity of dealing with two languages. Parallel activation of both languages was observed at all stages of bilingual speech production (e.g., Guo & Peng Reference Guo and Peng2006). Consequently, task performance in one of two available languages involves competition of words of the unwanted language with those of the intended language. Therefore, lexical selection is a more demanding process for bilinguals than for monolinguals. As a result, retrieval of words is slower even in highly proficient bilinguals of all ages than in monolinguals (Sorace Reference Sorace2011).
Unbalanced bilinguals are better in one language compared with the other. Hence, they have a stronger and a weaker language. Usually their mother tongue is the stronger language whereas a second, later learned language is the weaker language, characterized by incomplete acquisition of lexicon and grammar. If the first language undergoes attrition (that is, language loss due to infrequent use of that language as a consequence of migration), this language becomes the weaker language and the second language develops to be the stronger language. In both cases, if language proficiency is rather low, it is even more demanding to produce that weak language. It requires more resources to inhibit the stronger, unwanted language (Meuter & Allport Reference Meuter and Allport1999). This means that using a weak language involves much language control to avoid unintentional switching to the stronger language. Several processing models for bilingualism (e.g., Green Reference Green1986) capitalize on the concept of language control and emphasize the problem of limited resource and processing capacities. Applying the proposed theory to a communicative setting with unbalanced bilinguals, I seriously question, in particular, the suggested obligatory action predictions and the effortless involvement in interactive language.
P&G acknowledge one condition under which predictions of one's own and other's actions can be difficult, namely, if they are “unrelated” (an example for a related action would be ballroom dancing; see target article, sect. 2.3, Joint Action). In my view, in the domain of language, the ease of prediction depends not only on relatedness, but also, and much more, on familiarity. Not only are predictions easier and more likely to be correct when communicative participants are familiar with each other and their communicative habits, but also the interlocutors must be familiar with the language they are using for the interaction, and with the culture of that language. It is certainly rather easy to predict actions of communication in a long-standing relationship (e.g., the best friend), but for an exchange student – age 16, foreign to a country, new to a host family, and with little knowledge either of the language spoken or of the family's speech habits – the success rate of other's action prediction is most probably low. Similarly, in foreign-language reading, cultural familiarity has been found to be crucial for comprehension. Studies on non-native reading revealed that if readers lack the relevant cultural knowledge, reading activities could not fully compensate for the discrepancy or help readers comprehend a text (Erten & Razı Reference Erten and Razı2009).
With all this in mind, it becomes clear that the language producer in the proposed model is ideal, even as a monolingual. There are moments when we suffer a snub from another person's action: we are so surprised that words fail us. Some seem to have the gift of gab, and are more quick-witted than others. For such situational or inter-individual variability known from everyday life, the current model does not account. It holds the view that interlocutors are perfectly coordinated and that there is no time gap between interlocutors' turns. Nonetheless, P&G have been aware of differences between native and non-native speakers or of the difference between speaking to an adult or a child (see sect. 4, General Discussion) to some degree, but they did not integrate these differences into their model. The authors might want to elaborate on situations and conditions when the construction of forward models, covert imitations, joint actions, and continuous prediction generation are not ideal, not even for the monolingual.