Advertisement

Journal of Cultural Cognitive Science

, Volume 2, Issue 1–2, pp 45–57 | Cite as

Culture cues facilitate object naming in both native and second language: evidence from Bodo–Assamese bilinguals

  • Bidisha Som
  • Rekha Kalita
  • Ramesh kumar Mishra
Research Paper
  • 50 Downloads

Abstract

We examined if iconic cultural images that are related to a bicultural bilingual’s native and second language modulate lexical access during object naming. Bodo–Assamese bilinguals named line drawings preceded by a picture belonging to one of the cultures (Bodo or Assamese). Speakers named in blocks in which the cultural image was congruent, incongruent or neutral with the language used for naming. The cultural images were, however, irrelevant to the main task. Results showed that cultural images facilitated naming in both languages compared to neutral images. There was no difference in the extent of facilitation between the languages. Our findings suggest that iconic cultural images boost lexical access in both languages of bilinguals and facilitate naming.

Keywords

Bilingualism Naming Cultural image Facilitation Bicultural bilinguals 

Introduction

Most bilinguals are also bi-cultural (Grosjean 2014). That is, they represent the cultural symbolism related to both the languages they speak. Objects that represent the cultural heritage of one language may be very distinct from another language. Cultural symbols may include images of objects, iconic scenes that represent a culture, and musical notes, etc. For example, the image of the Taj Mahal is a unique cultural symbol to Indians, and a Chinese individual’s semantic memory system similarly must represent the great wall of China uniquely. A bilingual speaker who is bi-cultural will thus store distinct cultural symbols connected to each language. These cultural images can influence language processing by amplifying specific contexts. Cultural differences in bilinguals can influence core cognitive processing mechanisms, such as the expression of emotion (Marian and Kaushanskaya 2004). In the domain of psycholinguistic processes, in bi-cultural and bilingual speakers, cultural symbols related to each language can constrain/enhance the activation of languages. For example, the image of the Great Wall of China should boost the activation of Chinese lexicon, and the image of the statue of liberty should boost English lexical access in an American speaker. However, an important question is will the image of the Great Wall of China inhibit language production in English in a Chinese-English bilingual? Experimental investigations with object naming in bi-cultural-bilinguals have shown that faces or iconic cultural images directly constrain activation levels of languages.

Before we get to the issue of cultural icons influencing bilingual lexical activation, it’s important to discuss few other important strands in bilingualisms that have direct bearing on the findings and conceptualization of the study. Bilingual language activation is intrinsically linked to cognitive control. That is because when bilinguals activate two lexical items in parallel to one concept, they need to deactivate one and focus on the other which is task or discourse relevant. In the Indian context, many authors have studied the types of cognitive control in bilinguals. In the absence of true monolinguals, they have compared bilinguals differing in their second language fluency. For instance, Singh and Mishra (2012) showed that Hindi-English bilinguals with superior proficiency in English exercised greater inhibitory control in an oculomotor stroop task compared to the low proficient bilinguals. Similarly, Khare et al. (2013) in their study on a similar population found that high proficient bilinguals have lesser attentional blink compared to low proficient bilinguals. Costa and Sebastián-Gallés (2014) had already shown that bilinguals bring in greater cognitive control when executive control demands are high in an ANT task. More recently, Saint-Aubin et al. (2018) replicated the findings of Mishra et al. (2012) on a Canadian bilingual population and found no difference in attentional disengagement but overall superior executive control. Mishra et al. (2012) had found that Hindi-English high proficient bilinguals disengage attention faster from a stimuli using the Posner’s paradigm. In many studies Bialystok and Martin (2004), Bialystok et al. (2008) has shown that bilinguals show superior inhibitory control compared to monolinguals. However, considering the spread of bi-multilingualism in India, the appropriate comparison could be between different types of bilinguals who differ on their fluency and language use. Since fluency and language use including the bilingual context influences cognitive control mechanisms (Abutalebi and Green 2007). Although it is important to state that many have not found superior inhibitory control in bilinguals compared to monolinguals and there are replication failures (Saint-Aubin et al. (2018). Even then, the psycholinguistic mechanism of lexical activation is linked to control (Green 1998) without which bilinguals won’t be able to activate and deactivate their languages appropriate to context. In our study, we will show that cultural icons bring in or induce a certain language context by pushing the activation of one lexicon and then in consequence, bilinguals have to deactivate the other lexical item. In sum, the studies on language control in bilinguals with different cognitive psychological tasks are theoretically linked to studies on contextual activation by priming.

The other strand of studies that are very important to our theoretical concern are those which have shown language non-selecticve activation of lexicon. Singh and Mishra (2012) using visual world eye tracking showed that Hindi-English bilinguals activate phonological cohorts of translation equivalents when they hear words in one language. These activations are seen in both language directions. Originally, Marian and Kaushanskaya (2004) have shown that Russian-English bilinguals activate phonologically related word when input is in any one language. This bottom up phonological activation of related words has been taken as a signature of bilingual non-selective activation. Sunderman and Priya (2012) found that even highly proficient Hindi-English bilinguals activate translation equivalents spontaneously when they do a translation recognition task. Therefore, this unintentional activation of either translation equivalents or phonologically related words in the bilingual lexicon in any given context requires a control mechanism. The key issue is if such control mechanisms are language related or domain general (Green and Abutalebi 2013). With regard to the design of this study, cultural icons are certainly to create a competition between lexical items. This competition between lexicon is to be sorted out dynamically with the exercise of appropriate cognitive control mechanisms. Although we do not do any cognitive control tasks in this study, the studies reviewed here are relevant in understanding the overall mechanisms involved in the interplay of context and control.

Visual images and faces linked to different cultures of the bilinguals either facilitate or inhibit language production (Li et al. 2013; see also Kroll and McClain 2013). However, it’s not clear from existing studies if cultural images should just facilitate name retrieval in the language they are tagged to or if they would also inhibit when there is incongruence (Hartsuiker 2015). Available evidence so far on this issue coming from immigrant Chinese-English bi-cultural-bilinguals remains mixed (Zhang et al. 2013; Li et al. 2013; see also Roychoudhuri et al. 2016). While some have shown faces representing native language inhibiting language production in L2 (Zhang et al. 2013), others have not found this (Li et al. 2013). It’s not clear if this discrepancy between facilitation and inhibition with regard to cultural images is related to speaker’s proficiency in those languages or the length of stay of the bilinguals in the foreign country or their everyday linguistic context.

Li et al. (2013) examined Chinese-English bilinguals on a naming task in English and Chinese in separate blocks in the presence of a Chinese or a Caucasian face. It was observed that Chinese faces facilitated naming in Chinese, and the same effect was observed for English naming with Caucasian faces. However, English naming was not disrupted when speakers saw Chinese faces. Thus, while facilitation was observed with congruent cultural images, inhibition with incongruent cultural image was not observed. Other authors have suggested that the disruption in L2 naming in the presence of L1 cultural image could be linked to L2 proficiency (Yang and Yang 2013). From these studies, it is not clear if the disruption seen in L2 production is linked to L1 cultural images or due to other variables such as language dominance, proficiency, and length of stay in the L2 dominant culture. Albeit, no previous study has examined these variables carefully so far. Cultural images related to languages influence language production selectively (Costa and Sebastián-Gallés 2014; Friesen and Jared 2012; Hoshino and Kroll 2008; Jared and Kroll 2001; Singh and Mishra 2013).

Zhang et al. (2013) examined language production in Chinese-English bilinguals in the presence of cultural images. These bilinguals had stayed in the USA for a long time and had assimilated into the culture which is largely L2 dominant. In such a language context, speakers must be prepared to speak the second language more than the first language. In this study, Chinese-English speakers were asked to name objects to a Chinese face (study 1) or in the presence of a Chinese cultural icon (study 2) in a simulated dialogue situation. In study 1, participants listened to audio clips of a conversation partner who spoke on campus-life topics in English. In study 2, participants described the Chinese or American cultural icons for 1 min each in English. Participants then completed a story-telling task making up a story about a culturally neutral image in the presence of the cultural icons. In both the studies, fluency in English was reduced in the presence of Chinese cultural faces or icons. These effects were further examined (study 3) in a task where participants judged if a phrase was related to an object. Naming performance in the second language (English) was reduced which may suggest that the native image disrupted language production in the second language. These authors did not examine L1 naming in the presence of L2 culture cues. These results suggested that native culture cues can inhibit second language production. This may arise since native culture cues can boost the activation of the native language which in turn can create larger competition for second language retrieval. However, from this study, it was not clear if these effects can be obtained in other bilinguals who have not migrated to another country with a different culture but are bicultural and residents in their country. To evaluate this, Roychoudhuri et al. (2016) asked Bengali—English bilinguals to name line drawings of everyday objects under both blocked and mixed naming conditions. At the time of the study, these bilinguals were staying away from their native culture (Bengali) and studying in a large cosmopolitan research university where the lingua-franca of communication is English. In the experiment, iconic images belonging to Bengali (L1) culture were displayed in the background throughout half of the trials. On the other half, neutral images were displayed in the background for comparison. It was observed that participants were slower in naming in the L2 in the presence of an L1 cultural image as opposed to a neutral image. This is probably because such an image leads to higher activation of the native language lexicon which in turn led to competition with L2 words and their retrieval. However, when the cultural image was congruent with the language, no facilitation was observed in both the studies. Therefore, while the competition account may explain slowness in L2 word production, it’s not clear why there was no facilitation. Both these studies suggest that native cultural images interfered with the production of L2 words in bilingual speakers who had been residing away from their native culture and were more immersed in an L2 culture.

Hartsuiker (2015) raises the point that it is not easy to predict if culture-language interactions will always lead to facilitation or inhibition or both. Therefore, it is important to examine these in a group of bilinguals with good L2 fluency while considering naming in both the languages in the presence of cultural images belonging to each language. It has to be noted that Zhang et al. (2013) did not examine the effect of L2 (English) cultural image on L1 (Chinese) naming whereas Li et al. (2013) had exposed their speakers to both L1 and L2 cultural images, and they observed only facilitation for Chinese cultural images. Further, it is important to examine bilinguals who are part of a more mixed bilingual situation, where they speak both the languages more or less equally. It is possible that native language images will inhibit second language production more in a context where the second language is the dominant language. This is because, in such a context, the native cultural image appears more salient and may capture attention more. The culture-language production effects cannot be understood without references to such factors of individual differences or to the overall language situation that gives rise to such effects in speakers.

The adaptive control hypothesis (Green and Abutalebi 2013) explicitly refers to the influence that contextual cues may exert on language selection in bilinguals in discourse situations. As per the hypothesis, bilingual speakers select the context-appropriate language considering the cues available in the environment. Thus, when a bilingual speaker chooses a particular language when he is aware of a certain interlocutor to whom he may address, he is using this cue to control language in a top-down manner. However, many times visual cues in the environment that are in some way tagged to specific languages of the bilingual may also influence language selection in a bottom-up manner. These cues can bias the activation level of a certain language by either facilitating that language or inhibiting the other language. In this context, cultural cues can facilitate or inhibit languages in a bottom-up manner. The effects can be observed in language production tasks when such cues are present. However, from the studies that have examined bilingual language production in the presence of language specific culture cues, it is not clear when facilitation or inhibition of each language is observed. Although, it should be noted that in such studies apart from the culture cue’s relationship with particular languages, other variables such as language proficiency, language context, and immigrant status do play a role.

In both Zhang et al. (2013; study 1) and Li et al. (2013), the authors had used Chinese or Caucasian faces to whom the speakers had to respond. This might have led the speakers to align with the faces, perceiving them as their interlocutors more actively (Pickering and Garrod 2004) when they saw the faces in front of them. In a recent study, (Woumans et al. 2015) also found that speakers showed facilitation in a naming task when they saw the faces of people with whom they had a simulated Skype conversation before. This suggests that faces triggered the corresponding language. This evidence also suggests that the bilingual lexical retrieval system is sensitive to faces of interlocutors including their cultural identities. It can also be said that faces representing a typical culture create a certain mode which then influences bilingual language processing. More recently such proposals have come from bilingual language control models (Green and Abutalebi 2013). Culturally salient items do not provoke such mechanisms when one looks at them, unlike faces and hence their impact on the speaker’s language production can be considered to be more subtle. In this study, we have used scenes from the cultural life that are an integral part of Bodo and Assamese language speaking communities.

In the present study, we asked Bodo–Assamese adult bilinguals residing in Assam to name pictures in both the languages. Bodo language belongs to the Tibeto-Burmese language family and is spoken in the Bodoland region.1 The center of the Bodo area are districts covered under Bodoland Territorial Council. These districts are within the state of Assam but enjoy autonomous administration (Fig. 1). This creates a scenario where the Bodos live within an Assamese dominated state while maintaining strong Bodo identity. Bodo children learn to speak Assamese as a second language. Assamese, an Indo-Aryan language, is the state language of Assam. Therefore, this case is unique in its socio-linguistic features. The Bodo–Assamese bilinguals thus live in Assam and speak Assamese as the lingua-franca yet maintain Bodo as their mother tongue and use the same for all within community interactions. These bilinguals are thus not similar to earlier studies with Chinese- English bilinguals who have been migrants to another country and have stayed long in the L2 dominant context away from their native culture for long. The Bodo–Assamese bilinguals are not immigrants to another country or culture, yet speak their L2 (Assamese) along with their L1 (Bodo) in everyday situations, albeit to different degrees of frequency. The Bodo–Assamese bilinguals thus maintain distinct linguistic and cultural values associated with both the languages. We wondered if one would expect the native culture cue to inhibit L2 naming in such a population. Further, no previous psycholinguistic studies have been done on these bilinguals. This led us to the hypothesis that such bilinguals should show priming by both cultural images in an object naming task.
Fig. 1

Bodo autonomous region and Assam

We presented Bodo and Assamese cultural pictures and neutral pictures as primes preceding the line drawings. The speakers were told that the cultural pictures were not relevant for the task. Speakers named the line drawings in blocks, so there was no language switching. This way we were able to make the presence of the cultural pictures not conspicuous to the naming act. We predicted that while naming in L1 against an L1 cultural image, naming should be facilitated compared to neutral images. Similarly, we expected that while naming in L2 against an L2 cultural image, naming should be boosted. We expected that when cultural images and the language in which naming has to be done do not match, some interference should be observed. However, considering the lack of consensus in past research on the direction of interference (if L1 cultural images inhibit naming in L2 or L2 images should inhibit naming in L1) we predicted interference for both situations albeit of a different magnitude. Hartsuiker (2015) raises this very point suggesting that although facilitation in naming by a culture prime is well predicted it is difficult to predict inhibition and the causes thereof. If the inhibition observed in L2 naming during the presence of L1 cultural images/faces in earlier studies was because of low language proficiency of the speakers, then we should not observe any inhibition in these bilinguals as they have excellent proficiency in L2, their dominant language. Further, as observed in Zhang et al. (2013) and Roychoudhuri et al. (2016), if long stay in a non-native cultural context leads to inhibition of L2 production, we should not observe such effects here as Bodo–Assamese bilinguals are well integrated into the L2 context. It is reasonable to expect that after speakers have long assimilated into the L2 culture and have developed good proficiency in it, the impact of L1 images should be negligible. The study was thus designed to examine the influence of cultural images on naming in both L1 and L2 in bilinguals who were L2 dominant and who had stayed long in a mixed context. Please note that Zhang et al. (2013) did not examine naming in both the languages and Roychoudhuri et al. (2016) did not use L2 cultural images. We were able to study language naming in both L1 and L2 in Bodo–Assamese bilinguals as they are distinct languages tied to distinct cultures.

Assamese and Bodo languages share many cognates. Although they belong to different language families, because of historical contacts, many words share phonological forms. Since cognates and non-cognates are processed differently during recognition, we analyzed them separately. We assumed that speakers might name cognates faster than non-cognates. However, we did not have any specific hypothesis concerning how a culture stimulus may interact with cognate status as such. However, cognates and non-cognates were presented randomly during the experiment.

While we are investigating these bilinguals as having the knowledge of Boro and Assamese, that is probably not entirely a correct characterization of these people. Since India as a country is sociolinguistically multilingual (Mohanty 2006), people speak multiple languages in any given context and children also learn many languages. In the context of this study, they also may know some Bengali or another language of North-eastern India. However, they may not be proficient in those third and fourth languages. Since proficiency has been linked to both language activation and control (Mishra 2018), we have considered bilinguals with the two languages here they are proficient and use most often. It’s the quantum of bilingualism with any two languages in any given context that influences both the psycholinguistic and control mechanisms. These bilinguals therefore use Boro and Assamese on a regular basis and switch between them. More recently with regard to context and control processes, Green and Abutalebi (2013) suggest that in dual language contexts, as opposed to other contexts bilinguals bring in inhibitory control mechanisms. Bilinguals who live in dual language context, such as our bilingual participants here, constantly switch and shift between the two languages. Therefore, lexical items from both languages receive constant activations. However, depending on the context and the interlocutors, the speakers choose one language but they shift into the other as the context changes. This unique type of bilingualism of India is very specific and control studies should consider them.

Methods

Participants

Forty Bodo—Assamese bilinguals in the age range of 21–30 (mean age = 25.37 years, SD = 2.80) participated in the experiment. They were residents of the city of Guwahati which is a dominant Assamese area. However, it is a cosmopolitan city with multilingual speakers. All the participants were naive to the purpose of the experiment. Participants gave informed consent for their participation in the study. All the participants had Bodo as their native language and they had learned Assamese (L2) socially and/or through education. The mean age of acquisition of their L2 was 5.50 (0.82). The participants’ proficiency in L1 and L2 was judged through a language background questionnaire that had questions on their native language, other languages known, usage of the L1 and L2 in home, social, work and other various domains, percentage of exposure to each of the languages as well as their preference of each of the languages for various social and entertainment purposes. Participants also indicated proficiency in both the languages in terms of understanding and speaking on a scale of 1–5 (1 = poor and 5 = excellent) (Table 1). Only participants with a score of 4 and above were chosen for the study.
Table 1

Language proficiency and demographic details of the participants: 22 female and 18 male participants

 

Mean (SD)

Range

Age (in years)

25.37 (2.80)

21–30

Age of acquisition of L2 (in years)

5.5 (0.82)

5–7

Self-rating questionnaire scores

L1 mean (SD)

L2 mean (SD)

 

5 (0)

4.67 (0.47)

Stimuli

Six iconic (three Bodo and three Assamese) pictures representing each culture were selected from freely available web resources. The pictures measured 1789 × 869 pixels each (96 dpi). The pictures represented one dance form, one food item and one traditional item from each culture (Appendix 1). These pictures were selected from an initial list of ten pictures of each culture. Ten Bodo and 10 Assamese students rated these pictures for relatedness of these pictures to the respective cultures. These students did not participate in the main study. The selected Bodo and Assamese cultural pictures had a rating of 9.8 and 9.4 respectively in terms of how well they represented their respective cultures. Three pictures of comparable sizes not related to either of the cultures were used as neutral images. Participants also rated neutral pictures on their relatedness to either Assamese and Bodo cultures (Appendix 1). This rating (1.43) was significantly lower (p < 0.001) compared to the ratings received for Assamese and Bodo cultural images.

Ninety-six line drawings of common objects (45 natural + 51 manmade) measuring 300 × 300 pixels were selected as stimuli in the naming task (Appendix 2). Ten other Bodo–Assamese bilingual students from Guwahati University rated the drawings and their names and gave ratings on their agreement between the words and the respective line drawings on a scale of 1–10, with 10 being highest agreement with the name. The agreement between the drawing and the names in both Bodo and Assamese was high (9.2 and 9.8, respectively) and the difference between them was also significant (p = 0.02). Out of the 96 items, Forty-two of them had cognate names in Bodo and Assamese. The remaining items were non-cognates and had unique names in both the languages.

Procedure

Participants were given sufficient instruction about the experiment. The experiment was set up with the DMDX software (Forster and Forster 2003). Participants were seated in front of a laptop computer (with LCD monitor) at a distance of 60 cm. The monitor had a 1366 × 768 pixel resolution, and the screen refresh rate was 60 Hz. Each trial started with a fixation cross displayed for 1000 ms (Fig. 2). This was followed by a cultural image that stayed on the screen for 500 ms followed by a blank screen for 500 ms. The target picture then appeared and remained on the screen for a maximum of 3500 ms or till the speakers named them.
Fig. 2

Sequence of events on a sample trial

The participants were familiarized with the line drawings of the objects to be named as well as the cultural images before the experiment and were told that the same pictures would appear on the screen. During the familiarization phase, the expected names of all the pictures in L1 and L2 were told to the participants. They were asked to use the same names during the object naming experiment. They were told to name the line drawings as fast and accurately as possible. The experiment was done in two sessions: one session with naming in Bodo and another with naming in Assamese. To avoid any impact of the language of instruction on the result, instructions were given in Bodo when the target language was Bodo and Assamese when the target language was Assamese.

Design

There were one hundred and ninety-two trials in total. For half of the trials participants were asked to name in Assamese and the other half in Bodo in different sessions. The order of these sessions was counterbalanced. Each session was divided into three blocks: one with Assamese cultural image, one with Bodo cultural image and one with a neutral image. Trials were equally divided between the three blocks for each language. The presentation of these trials was randomized across participants. Congruency was established depending on the relationship between the language image and the cultural image. A trial was designated as “congruent” when the language used for naming matched the cultural image (Assamese naming in the presence of Assamese image, for example). Similarly, when there was a mismatch, the trial was designated as “incongruent” (Assamese naming in the presence of Bodo image). The trials with neutral cultural images were termed as “neutral” trials.

Data analysis

Naming latency was measured as the temporal gap between the appearance of the line drawing and the triggering of the voice key. Naming data was recorded and accuracy and response time of the recorded data was checked by using CheckVocal (Protopapas 2007). This programme presents each recorded response audiovisually (as a waveform, spectrogram and sound is played out) along with the corresponding printed correct response and registered response time. The vocal responses were audiotaped to analyze the errors later. Voice-key malfunctions and incorrect use of the word to name the picture were considered as errors (15.6%) and removed from further analysis. Naming latencies below 250 ms (0.1%) were discarded as outliers. No upper limit was fixed for naming latencies based on the analysis procedure of several previous object naming studies (Gollan et al. 2014; Roychoudhuri et al. 2016; Bhatia et al. 2017). After the outlier removal, the latency distribution was found to have a skewness of 1.07 and kurtosis of 0.84. Repeated measures ANOVA was performed with naming latency as the dependent variable. Cognate status (Cognates, Non-cognates), Congruency (Congruent, Incongruent, Neutral) and language (Bodo, Assamese) were treated as within-subject factors (See Table 2 for average naming latency in each condition). Similar analysis was for performed on error trials with the same factors. Statistical analysis was done both by subject (F1) and by item (F2).
Table 2

Average naming latencies and error percentages for congruent, incongruent and neutral conditions for both L1 (Bodo) and L2 (Assamese) naming

 

Cognates

Non cognates

Congruent

Incongruent

Neutral

Congruent

Incongruent

Neutral

RT (ms)

Assamese

1186 (193)

1244 (191)

1312 (248)

1297 (213)

1283 (230)

1331 (234)

Bodo

1280 (193)

1301 (270)

1388 (293)

1221 (215)

1271 (219)

1292 (288)

Errors (%)

Assamese

2.2 (1)

2.9 (2)

1.9 (2)

3.8 (3)

1.8 (1.4)

3.2 (2.8)

Bodo

3.1 (3)

2.1 (2)

2.9 (2)

1.9 (1.7)

2.9 (2.9)

2.4 (2.9)

SD given in brackets

The values for Cognate and noncognate items are given separately

Results

Naming latency

There was no significant effect of Cognate statuts, F1 (1, 39) = 0.03, p =0.86, η2= 0.01 and F2 (1, 180) = 0.17, p = 0.68, η2= 0.001. However, congruency of the cultural image with the language used for naming influenced naming latency, F1 (2, 78) = 29.88, p < 0.001, η2= 0.43, F2 (2, 180) = 1.57, p = 0.21, η2= 0.02 (Fig. 3, left panel). Participants were faster responding on congruent trials (1246 ms, SE = 29.89) compared to incongruent (by 29 ms) and neutral (by 85 ms) trials, p = 0.002 and p < 0.001 respectively (Table 2). Responses on incongruent trials (M = 1275 ms, SE = 31.55) were also faster (p < 0.001) compared to neutral trials (M = 1331 ms, SE = 37). There was no effect of language in the subject-wise analysis, F1 (1, 39) = 0.55, p =0.46, η2= 0.01 and in the item-wise analysis F2 (1, 180) = 0.08, p =0.78, η2<0.001. The interaction between language and cognate status was significant both by-subjects, F1 (1, 39) = 23.94, p < 0.001, η2= 0.38 and by-items, F2 (1, 180) = 3.66, p =0.05, η2= 0.02. Pairwise comparisons showed that participants were faster (p = 0.002) naming in Assamese (M = 1247 ms, SE = 30.75) compared to Bodo (M = 1323 ms, SE = 39.6) only for cognate items but not for noncognate items (p = 0.14). Further, the interaction between cognate status and congruency was also significant by-subjects, F1 (2, 78) = 4.26, p = 0.018, η2= 0.1 but not by-items, F2 (2, 180) = 0.45, p =0.64, η2= 0.005. Pairwise comparisons showed that RTs on congruent and incongruent trials were faster compared to neural trials for both cognate (p < 0.01) and non-cognate items (p < 0.02). The interaction between congruency and language was not significant, F1 (2, 78) = 0.19, p = 0.83, η2= 0.005 and F2 (2, 180) = 0.009, p =0.99, η2<0.001. Nonetheless, since it was important for our hypothesis, we performed pairwise comparisons on this interaction. The analysis revealed that responses on congruent and incongruent trials were faster compared to neutral trials, for Assamese naming (p < 0.001) as well as Bodo naming (p < 0.01). The difference between congruent and incongruent trials was significant (p = 0.02) only for Bodo naming. The three-way interaction between cognate status, language and congruency was nonsignificant both by-subjects (p = 0.11) and by-items (p = 0.93).
Fig. 3

Naming latencies for cognate (left panel) and noncognate (right panel) items as a function of the language in which the picture was named and the congruency with the culture image

Errors

Cognate status had no significant effect, F (1, 39) = 0.92, p = 0.34, η2= 0.02. There was no significant effect of congruency either, F (2, 78) = 1.66, p = 0.2, η2= 0.04. Language did not have a significant effect, F (1, 39) = 0.06, p = 0.80, η2= 0.002. The interaction between language and cognate status was significant, F (1, 39) = 10.45, p = 0.003, η2= 0.21. Pairwise comparisons showed that errors while naming in Assamese were higher (p = 0.014) for noncognate items (M = 2.9%, SE = 0.3) compared to cognate items (M = 2.4%, SE = 0.2). The two-way interaction between language and congruency was marginally significant, F (1, 39) = 2.68, p = 0.07, η2= 0.06. The three-way interaction between language, congruency and cognate status was also significant, F (2, 78) = 25.75, p < 0.001, η2= 0.4. To examine these interactions further, separate language * congruency models were created for cognate and noncognate items. For cognate items, wwhen participants named in Bodo, the percentage of errors on incongruent trials (M = 2.1%, SE = 0.3) was lesser than the errors on congruent trials (M = 3.1%, SE = 0.5) and neutral trials (M = 2.9%, SE = 0.4), p = 0.02 and p = 0.024, respectively. In contrast, while naming in Assamese, participants made higher percentage errors on incongruent trials (M = 2.9%, SE = 0.4) compared to congruent (M = 2.2%, SE = 0.2) and neutral trials (M = 2%, SE = 0.3), p = 0.009 and p = 0.01, respectively. For noncognate items, the percentage of errors on congruent trials (M = 3.8%, SE = 0.5) was higher (p = 0.03) and percentage of errors on incongruent trials (M = 1.8%, SE = 0.2) was lower (p = 0.001) compared to neutral trials (M = 3.2%, SE = 0.4) during naming in Assamese. However, while naming in Bodo, participants made higher percentage of errors on incongruent trials (M = 3%, SE = 0.5) compared to congruent (M = 1.9%, SE = 0.3) and neutral trials (M = 2.4%, SE = 0.5), p = 0.003 and p = 0.03 respectively.

Discussion

This study examined the influence of cultural images on object naming in Assamese and Bodo bilingual speakers. Speakers named objects that were presented following the cultural images linked to either Assamese or Bodo. Cultural images facilitated responses in the both languages with respect to the neutral images. Speakers were faster in naming when the cultural image corresponded to the language they used to name. The facilitation that we observed can be explained by assuming that a particular cultural image accentuated the activation of names in that language. Given the fact that bilingual speakers activate names in both the languages in parallel, such cultural images can boost the activation in one language. Therefore, naming latencies are faster in that language. This pattern of results is also consistent with those obtained by Li et al. (2013) with Chinese-English bilinguals. In that study, Chinese faces had facilitated naming in Chinese, but the authors had not obtained any inhibition for L2. However, in the Zhang et al. (2013) study, Chinese-English bilinguals were slower in naming when they were exposed to Chinese faces/images. The differences between these two studies could lie in the use of different methods and also participant related differences. In an earlier study, Jared et al. (2013) observed that visual representation of objects (Chinese cabbage vs. English cabbage) influenced naming in Chinese and English. Speakers were faster when the visual representation matched with the name to be used. The authors proposed the dual coding hypothesis that links specific visual representations of objects to specific language nodes in the semantic memory. Our study did not find any inhibition when the image was incongruent with the language used for naming. Further, we observed that participants were faster on incongruent trials compared to neutral trials.

We also observed a speed-accuracy tradeoff with respect to the influence of cultural images on naming. Errors were lower on trials with neutral images compared to cultural images for Assamese naming (cognate items) and Bodo naming (noncognate items). Thus, the facilitation due to cultural images observed on latency was not reflected in the accuracy of naming. Although participants were instructed to be both fast and accurate while naming, it is possible they prioritized being fast thus paying a cost with accuracy.

We had both cognates and non-cognates names for production. However, cognate status did not influence naming in any way. It might be argued that for cognate items, cultural images could have activated either of the languages thus resulting in faster responses even on incongruent trials compared to neutral trials. But if this was the case, we should not have observed faster RT on incongruent trials even for noncognate items. Thus, an alternate explanation is that the so-called neutral images are not neutral as such. As these images have no identifiable objects in them and are unfamiliar to the participants when compared to the cultural images, viewers may take a longer time to process them than incongruent images. Hence, what can be considered as a “neutral” culture cue is an important question. We believe that this issue calls for further investigation.

Much of these effects have to do with the bilingual speakers themselves and their language context in terms of where they live. The previous studies with Chinese-English bilinguals were on immigrants who lived in the L2 dominant language environment (Li et al. 2013; Zhang et al. 2013). Furthermore, they were away from their native culture for an extended time. Our speakers came from a contact language area where they were exposed to both L1 (Bodo) and L2 (Assamese) constantly. Furthermore, they were very highly proficient in L2 (Assamese). It has been suggested that the inhibition observed in Zhang et al. (2013) could be because of the low L2 proficiency of speakers. Hartsuiker (2015) observed that the inhibition seen in the Zhang et al. (2013) study could be explained in terms of the effect of language immersion. The Chinese-English speakers lived in the L2 dominant context, and thus the L1 images were less salient in the environment.

In our previous study (Roychoudhuri et al. 2016) where inhibition of L2 in the presence of L1 cues was observed, the participants were studying in a predominately L2 context. In such a language context, the opportunity to speak L1 as well as the visual availability of L1 cues is less. This may make L1 culture cues (they could be faces or scenes or cultural objects) very salient. Such cues then strongly activate the native language leading to inhibition of L2. On the other hand, in this study, the Bodo–Assamese bilinguals lived in a mixed language context and also were not away from their native culture. They spoke their L1 quite often and also were regularly exposed to cultural cues associated with this language. Importantly, these speakers had not been immigrants in another culture for long and thus lacked the same degree of immersive-ness in a non-native culture as the Chinese-English, or the Bengali-English bilinguals reported earlier. Therefore, for these bilinguals the degree to which the native culture cue could influence language selection in a bottom-up manner was different.

Our study is also different from previous studies methodologically. While the Zhang et al. (2013, study 1) and Li et al. (2013) study had used faces; we used pictures of culturally salient items. Faces are more salient as visual objects and are known to hold attention. Further in these studies, speakers were asked to have a simulated dialogue with the cultural images, which must have brought a conversational context into the experiment. In our case, speakers were told that the scenes were not relevant for the main task. Thus, speakers did not directly engage with them or align their languages with them as they would do in a discourse context. Therefore we assume that the effect we observed was more subtle and indirect than previous studies. These results, therefore, show that even images that are present in the environment can subtly influence bilingual language selection and production when they may not be paying attention. These results thus extend and enrich previous observations.

In the introduction we discussed studies that have shown superior inhibitory control in bilinguals who use a certain language more fluently. Thus, fluent bilinguals use language more and also use inhibition more consistently. We also discussed studies that have shown language non-selective activation during bilingual language processing. These two strands of research converge theoretically in suggesting that brain networks engage in inhibition of context inappropriate language during processing and this then enhances overall bilingual cognitive control (Green and Abutalebi 2013). Our results show that cultural icons trigger language activations when they match with the language intended to be used for production. This priming has to be broadly understood at the interface of lexical activation and control. However, we used a language production task and this task calls for specific control mechanisms that have been well studied before (Costa and Sebastián-Gallés 2014). We found an effect of the cues that were facilitated by pushing the activation level of languages they were connected to. However, the absence of inhibition that would be expected based on previous findings can be explained in different ways. Methodologically the task demands here were different from the non-linguistic studies where control has been investigated with executive control tasks (Singh and Mishra 2012). At the moment theoretical developments are underway that are linking contextual effects, lexical activations and control (Mishra 2018). This study is preliminary in showing a lesser studied population the extent to which the bilingual cognitive system is flexible and sensitive to visual cues that are linked to both psycholinguistic and executive control mechanisms.

One of the possible limitations of this study was that cultural images were presented for a certain duration before the picture naming objects, unlike the study by Roychoudhuri et al. (2016) in which the cultural images stayed in the background throughout the trial thereby making them less salient. Thus, in this case, the cultural images could have still acted as cues which the participants might have explicitly associated with a particular language leading to the observed effects on naming latency. Further, naming was tested only in blocked but not in the mixed condition. Language switching is an essential component of bilingual language processing; thus it would be interesting to see in future studies how the cultural images influence the control mechanisms involved in language switching.

In sum, we have shown evidence that cultural images facilitate language production in bilinguals. However, we did not find evidence of inhibition when the cultural images and the naming language did not match. Our study shows that such effects can be seen in a non-immigrant population who live in a mixed language context.

Footnotes

  1. 1.

    The Bodos are part of one of the largest indigenous populations of North East India. They were granted Bodoland Territorial Council after years of agitation, political unrest and insurgency for greater autonomy. The BTC has autonomous administrative jurisdiction over the Bodo dominated districts of western Assam. Unlike many other indigenous populations in India, the Bodos show high language maintenance as evident in good number of publication in the language. Their long struggle has also seen the language getting included in the eighth schedule of the Indian constitution.

Notes

Acknowledgement

The current research work was carried out with funding from Department of Science and Technology Grant [Government of India] awarded to Bidisha Som [grant no: SR/CSI/86/2012].

Compliance with ethical standards

Conflict of interest

On behalf of all authors, the corresponding author states that there is no conflict of interest.

References

  1. Abutalebi, J., & Green, D. (2007). Bilingual language production: The neurocognition of language representation and control. Journal of Neurolinguistics., 20(3), 242–275.CrossRefGoogle Scholar
  2. Bhatia, D., Prasad, S. G., Sake, K., Mishra, R. K., Schiller, N. O. (2017). Task irrelevant external cues can influence language selection in voluntary object naming: evidence from hindi-english bilinguals. PLOS One, 12(1), e0169284.CrossRefPubMedPubMedCentralGoogle Scholar
  3. Bialystok, E., & Martin, M. M. (2004). Attention and inhibition in bilingual children: Evidence from the dimensional change card sort task. Developmental Science., 7(3), 325–339.CrossRefPubMedGoogle Scholar
  4. Bialystok, E., et al. (2008). Cognitive control and lexical access in younger and older bilinguals. Journal of Experimental Psychology. Learning, Memory, and Cognition, 34(4), 859.CrossRefPubMedGoogle Scholar
  5. Costa, A., & Sebastián-Gallés, N. (2014). How does the bilingual experience sculpt the brain? Nature Reviews Neuroscience, 15(5), 336–345.CrossRefPubMedPubMedCentralGoogle Scholar
  6. Forster, K. I., Forster, J. C. (2003). DMDX: A Windows display program with millisecond accuracy. Behavior Research Methods, Instruments, & Computers, 35(1), 116–124.CrossRefGoogle Scholar
  7. Friesen, D. C., & Jared, D. (2012). Cross-language phonological activation of meaning: Evidence from category verification. Bilingualism: Language and Cognition, 15(01), 145–156.CrossRefGoogle Scholar
  8. Gollan, T., Cera, C., Ferreira, V. C., & Flett, S. (2014). Translation-priming effects on tip-of the-tongue states. Language and Cognitive Processes, 29(3), 278–288.PubMedGoogle Scholar
  9. Green, D. W. (1998). Mental control of the bilingual lexico-semantic system. Bilingualism: Language and cognition, 1(02), 67–81.CrossRefGoogle Scholar
  10. Green, D. W., & Abutalebi, J. (2013). Language control in bilinguals: The adaptive control hypothesis. Journal of Cognitive Psychology, 25(5), 515–530.CrossRefPubMedGoogle Scholar
  11. Grosjean, F. (2014). Bicultural bilinguals. International Journal of Bilingualism, 1367006914526297.Google Scholar
  12. Hartsuiker, R. J. (2015). Visual cues for language selection in bilinguals. Attention and vision in language processing (pp. 129–145). Berlin: Springer India.CrossRefGoogle Scholar
  13. Hoshino, N., & Kroll, J. F. (2008). Cognate effects in picture naming: Does cross-language activation survive a change of script? Cognition, 106(1), 501–511.CrossRefPubMedGoogle Scholar
  14. Jared, D., & Kroll, J. F. (2001). Do bilinguals activate phonological representations in one or both of their languages when naming words? Journal of Memory and Language, 44(1), 2–31.CrossRefGoogle Scholar
  15. Jared, D., Poh, R. P. Y., & Paivio, A. (2013). L1 and L2 picture naming in Mandarin-English bilinguals: A test of bilingual dual coding theory. Bilingualism: Language and Cognition, 16(02), 383–396.CrossRefGoogle Scholar
  16. Khare, V., et al. (2013). Bilingualism and the increased attentional blink effect: Evidence that the difference between bilinguals and monolinguals generalizes to different levels of second language proficiency. Psychological Research, 77, 728–737.CrossRefPubMedGoogle Scholar
  17. Kroll, J. F., & McClain, R. (2013). What bilinguals tell us about culture, cognition, and language. Proceedings of the National Academy of Sciences, 110(28), 11219–11220.CrossRefGoogle Scholar
  18. Li, Y., Yang, J., Scherf, K. S., & Li, P. (2013). Two faces, two languages: An fMRI study of bilingual picture naming. Brain and Language, 127(3), 452–462.CrossRefPubMedGoogle Scholar
  19. Marian, V., & Kaushanskaya, M. (2004). Self-construal and emotion in bicultural bilinguals. Journal of Memory and Language, 51(2), 190–201.CrossRefGoogle Scholar
  20. Mishra, R. K., Hilchey, M. D., Singh, N., Klein, R. M. (2012). On the time course of exogenous cueing effects in bilinguals: higher proficiency in a second language is associated with more rapid endogenous disengagement. Quarterly Journal of Experimental Psychology, 65(8), 1502–1510.CrossRefGoogle Scholar
  21. Mishra, R. K. (2018). Bilingualism and Cognitive Control. Berlin: Springer.CrossRefGoogle Scholar
  22. Mohanty, A. K. (2006). Multilingualism of the unequals and predicaments of education in India: Mother tongue or other tongue. Imagining multilingual schools. 262–283.Google Scholar
  23. Pickering, M. J., & Garrod, S. (2004). Toward a mechanistic psychology of dialogue. Behavioral and brain sciences, 27(02), 169–190.PubMedGoogle Scholar
  24. Protopapas, A. (2007). Check Vocal: A program to facilitate checking the accuracy and response time of vocal responses from DMDX. Behavior Research Methods, 39(4), 859–862.CrossRefPubMedGoogle Scholar
  25. Roychoudhuri, K. S., Prasad, S. G., & Mishra, R. K. (2016). Iconic native culture Cues inhibit second language production in a non-immigrant population: Evidence from Bengali-English Bilinguals. Frontiers in Psychology, 7, 1516.CrossRefPubMedPubMedCentralGoogle Scholar
  26. Saint-Aubin, Jean, Hilchey, Matthew, Mishra, Ramesh, Singh, Niharika, Savoie, Dominique, Guitard, Dominic, et al. (2018). Does the relation between the control of attention and second language proficiency generalize from India to Canada? Canadian Journal of Experimental Psychology, 72, 10.Google Scholar
  27. Singh, N., & Mishra, R. K. (2012). Does language proficiency modulate oculomotor control? Evidence from Hindi-English bilinguals. Bilingualism: Language and Cognition, 15, 771–781.CrossRefGoogle Scholar
  28. Singh, N., & Mishra, R. K. (2013). Second language proficiency modulates conflict-monitoring in an oculomotor Stroop task: evidence from Hindi-English bilinguals. Frontiers in Psychology., 4, 322.PubMedPubMedCentralGoogle Scholar
  29. Sunderman, G. L., & Priya, K. (2012). Translation recognition in highly proficient Hindi, English bilinguals: The influence of different scripts but connectable phonologies. Language and Cognitive Processes., 27(9), 1265–1285.CrossRefGoogle Scholar
  30. Woumans, E., Martin, C. D., Bulcke, C. V., Van Assche, E., Costa, A., Hartsuiker, R. J., et al. (2015). Can faces prime a language? Psychological Science, 26(9), 1343–1352.CrossRefPubMedGoogle Scholar
  31. Yang, S., & Yang, H. (2013). Does bilingual fluency moderate the disruption effect of cultural cues on second-language processing? Proceedings of the National Academy of Sciences, 110, E4403.CrossRefGoogle Scholar
  32. Zhang, S., Morris, M. W., Cheng, C. Y., & Yap, A. J. (2013). Heritage-culture images disrupt immigrants’ second-language processing through triggering first-language interference. Proceedings of the National Academy of Sciences, 110(28), 11272–11277.CrossRefGoogle Scholar

Copyright information

© Springer Nature Singapore Pte Ltd. 2018

Authors and Affiliations

  1. 1.Indian Institute of Technology GuwahatiGuwahatiIndia
  2. 2.North Eastern Regional Institute of ManagementGuwahatiIndia
  3. 3.Centre for Neural and Cognitive SciencesUniversity of HyderabadHyderabadIndia

Personalised recommendations