Attentional processes in children
and adults bilingual in Greek and Albanian.
Ladas Aristea K.¹, Siegal Michael², Vivas Ana B.³
¹ South East European Research Center (SEERC), 24 Prox. Koromila st., 54622, Thessaloniki, Greece,
² Department of Psychology, the University of Sheffield, Mushroom Lane Annex, Sheffield, UK,
³ Department of Psychology, City Liberal Studies, the International Faculty of the University of Sheffield, 24 Prox. Koromila st., 54622, Thessaloniki, Greece,
As bilingualism is now a universal demographic characteristic, the need for studying the cognitive processes involved in this phenomenon is considered highly important. Studying the attention processes involved in bilingualism offers a pathway towards this goal, as attention is the main cognitive mediator of language. Although numerous studies have investigated executive attentional control in bilingualism, the findings are contradicting and have methodological limitations. One of these limitations is the absence of control for SES that has been repeatedly shown to influence the performance of both monolinguals and bilinguals on cognitive tasks. In addition, two functions of attention, alertness and visuospatial orienting, have been largely ignored in studies of bilingualism. To our knowledge, only two studies have investigated the 3 attention networks (i.e. executive attentional control, alertness & orienting) in bilingual individuals. Additionally, the possible effect of bilingualism in the Joint Attention (JA) effect has never been investigated, although it can be assumed to be influenced by bilingualism. Hence, the aim of the present study is to investigate the 3 attention networks and JA in bilingual children and adults, taking into account the SES factor.
Keywords: bilingualism, attention, SES, Albanian, Joint Attention, language switching task
1. Introduction
As psycholinguistic research has shown, in order to understand fully the function of language we need to take into account other cognitive processes that are inherently involved in language use, such as attention (Garnham, 1985). Rather than trying to study language in a vacuum, it is more appropriate to examine specific language phenomena in order to unravel the exact cognitive processes underlying the language function. One of these phenomena is bilingualism (i.e. being proficient in two languages; Martin, 2006). In the contemporary world, bilingualism is considered to be the universal norm rather than the exception (Segalowitz & Frenkiel-Fishman, 2005). Hence, it seems necessary to fully understand the cognitive components of bilingualism, not only because it will allow for a more clear understanding of cognition (Bialystok & Martin, 2004) and the nature of language use itself (Kolers & Paradis, 1980), but also because it is a universal demographic characteristic (Arlida et al., 2000; Segalowitz & Frenkiel-Fishman, 2005).
Numerous studies have attempted to investigate the effect of bilingualism on cognitive processes (e.g. Bialystok et al., 2005; Bialystok & De Pape, 2009; Bialystok & Martin, 2004; Colzato et al., 2008; Costa, Hernández & Sebastian-Galles, 2008; Morton & Harper, 2007), supporting a ‘bilingual advantage” on the executive functions that include the control of attention. However, often studies conducted so far had several methodological limitations (Morton & Harper, 2007; Colzato et al., 2008). One of the most important concerns the lack of control for Socio-Economic Status (SES) of the participants. That is, the bilingual group may have had a higher socio-economic status than the monolingual group and so effects may be due to SES rather than bilingualism per sé. Importantly, SES has been repeatedly shown to influence the performance of both monolinguals and bilinguals on cognitive tasks (e.g. Mezzacappa, 2004; Siegal, Surian, Matsuo, Geraci, Iozzi, Okumura et al., 2010).
Moreover, while executive attention (e.g. conflict resolution; goal-directed attention) has been relatively well studied, the two other functions of attention, alertness and visuospatial orienting, have been largely ignored in studies of bilingualism. To our knowledge, only two studies have investigated the three attention networks (i.e. executive attentional control, alertness & orienting; Posner & Boies, 1971; Posner & Petersen, 1990) in bilingual individuals (Costa, Hernández & Sebastian-Gallés, 2008; Hernández, Costa, Fuéntes, Vivas & Sebastian-Gallés, 2010). Neither of these studies involved children. Hence, the aim of the proposed study is to investigate the three attention networks of bilinguals, taking into account the SES factor. To this end, Albanian-Greek bilinguals have been chosen as the target population who, according to the Greek Statistical Authority (2010), are of a rather low SES.
Of interest would be to also explore how these attention networks mature in the bilingual mind, which would give a more complete view of the bilingual effect on attention. Thus, this study is suggested to include both children and adults. As a measure of the three main attentional networks, the ANT task has been chosen, which has been demonstrated to validly and reliably measure the three networks of attention (e.g. Rueda et al., 2005), and is available in two versions: one for children (ANT Child task; Rueda et al., 2004) and one for adults (ANT; Fan et al., 2002).
In addition, we will investigate the effect of type of spatial cue on the bilingual advantage. That is, in the two studies (Colzato et al. 2008; Hernández et al., 2010), that investigated visuospatial orienting, the cues used to direct attention were non-symbolic, exogenous cues (i.e. luminance of a peripheral box), although it has been claimed that the “bilingual advantage” seems to generalize to other tasks involving symbolic stimuli (e.g. Bialystok & De Pape, 2009). Unlike exogenous cues, symbolic cues (e.g. arrows, words, faces, gaze) require a level of interpretation to direct attention, and depend more likely on endogenous attention processes (Frischen, Bayliss & Tipper, 2007). Thus, it could be the case that lack of a bilingual advantage on attentional orienting could be attributed to the nature of the cues used in those tasks. We will employ the Joint Attention (JA) paradigm (e.g. Frischen & Tipper, 2004) which has shown that there is a benefit in performance when a target appears in the location indicated by the gaze (i.e. schematic faces gazing right/left). However the unique case of gaze cues, which are claimed to elicit both endogenous and exogenous orienting of attention (e.g. Frischen et al., 2007) has not been investigated yet in bilingual individuals and this is another reason why studying bilingualism via a JA task is very interesting and unique.
In addition to attentional orienting by a visual cue, another bottom-up attention phenomenon which can be tested by a JA paradigm is Inhibition Of Return (IOR; avoid to re-focus attention on an already processed target; Posner& Cohen, 1984). Interestingly, only two studies have tested the IOR effect in bilinguals (Colzato et al., 2008; Hernández et al., 2010). According to Colzato et al. (2008), bilingualism seems to modulate IOR, as bilinguals exhibited an IOR effect in much longer SOAs than expected (i.e. in 700ms, whereas should be present in about 300ms; Colzato et al., 2008). According to the efficient search account of IOR , whereby IOR serves to avoid possible needless re-processing of an already processed location (e.g. Fuentes, Vivas, & Humphreys,1999), such inflexible IOR ability is not beneficial, thus supporting a bilingual disadvantage in this type o inhibition. However, other studies investigating types of inhibition other than IOR (e.g. Bialystok, 2006; Bialystok & De Pape, 2009) claim that bilinguals possess a more efficient inhibitory ability compared to monolinguals, that results in a bilingual advantage in inhibition.
This debate on the nature of a bilingual advantage in inhibition can be elucidated further by employing the JA task in bilinguals, which elicits IOR. If this “bilingual advantage” generalizes to IOR, according to the aforementioned efficient search account of IOR bilinguals should manifest IOR earlier than monolinguals, thus executing a visual search of space more efficiently. By contrast, if bilinguals exhibit IOR in much longer SOAs, as was the case in the study of Colzato et al.(2008), they will display less flexible inhibitory ability, which regarding the efficient search account of IOR is not beneficial.
Another innovation in the present study, as compared to previous studies on bilingualism, concerns how level of bilingual skill is measured. Most studies of the bilingual effect on cognition (e.g. Costa, Hernández & Sebastian-Gallés, 2006; Garrat & Kelly, 2008; Portocarrero et al., 2007) have used a self-report questionnaire on language use and background to measure the participants’ level of bilingualism. However, according to the review of Mindt et al. (2008), the subjectivity of such a measure should not be underestimated. Thus, we will employ a more objective measure of bilingual skill, the numerical language-switching task (Meuter & Allport, 1999). This task appears to be valid as an index of language proficiency in bilingual individuals (Costa & Santebastian, 2004; Meuter & Allport, 1999). However, to our knowledge this task has not been employed in bilingual children, which calls for the need to adapt the language switching task appropriately to match the abilities of young children.
2. Method
Participants
Participants will be divided in three age-groups, one for younger children (ranging from 60 to 84 months of age), one for older children (from 85 to 120 months old) and one for adults (age range of 18 to 42 years old), with 50 participants in each age group (25 bilingual and 25 monolingual). Children will be recruited from randomly selected schools in the western area of Thessaloniki, after a license has been granted from the Greek Ministry of Education, Religion and Lifelong Learning and after their parents have provided an informed consent for their children’s participation. Bilingual adults will be recruited from the Albanian associations and monolingual adults will be recruited from the numerous Greek folk associations situated in Thessaloniki and we will also obtain written informed consent.
Material and Procedure
The proposed attention tasks, ANT (Child & adult version) and JA and the language-switching task will be in a computerized form, thus enabling easy administration and precise recording of reaction times in milliseconds. In the ANT Child task (see Figure 1), children will be asked to feed the hungry central fish that will be appearing on screen as soon as possible, by pushing either the left or the right mouse-button, according to the orientation of that fish. Additionally, they will be instructed that sometimes the fish will be appearing alone and sometimes it would be swimming with other fish; in such a case, they will be told to focus on the central fish and press the button corresponding to that fish’s orientation only, so as to feed it. The ANT for adults will be identical to the Child ANT, with the exception that the fish-stimuli will be replaced by arrows and participants will be asked to respond to the direction of the central arrow by pressing the corresponding mouse-button.
Fig.1. An example of a typical ANT Child trial as well as all the experimental conditions and the stimuli used in the ANT Child task (adopted from Rueda et al. 2004).
The JA task (see Figure 2) will be a detection task, executed by adults only. Participants will be instructed to fixate at the center of the screen throughout the procedure and will be encouraged to respond by pressing the space bar as soon as they see the target but without making anticipated responses. We will include catch trials to ensure that the participants will follow these instructions. On each trial, the central fixation point will appear for 1000ms. This will be followed by the same fixation flanked by 2 boxes during another 1000ms. After, the cross will be replaced by a face-stimulus with averted gaze towards the left or right for 150ms. Depending on the SOA (Stimulus Onset Asynchrony) condition, after a further interval of 350, 550, 1500 or2550ms (350, 550, 1500, 2550 SOA) the cue will be replaced by the same face stimulus, with direct gaze during 150 (the central cue). Finally the target will appear at the same location of the gaze or at the opposite location. In some of the trials (catch trials) the target will not appear and participants will be instructed to withhold their response and wait until the next trial begins.
Fig.2. Sequence of events in a typical trial. In this example-trial, the averted gaze is directed to the left and the target appears at the uncued location (right).
The measure of bilingual skill will include Arabic digits (1-9) as target stimuli which participants will be asked to read aloud in either language. The background will be either a Greek or an Albanian colored flag (depending on the condition), serving as the language cue by prompting participants to read the digit in the analogous language. For adults, digits will be presented in short sequences (“lists”), varying unpredictably in length from 5 to 14 items. Within each list, generation of digits will be random; no number will be presented twice in a row. Trials will be of two types: (1) trials where the language of response will be the same, either L1 or L2, as in the previous trial (non-switch trials), and (2) trials where the language of response, either L1 or L2, will be different than the language used in the preceding trial (switch trials). Number of switching trials within the lists will vary (from 0 to 4 switch-trials). Response in L1 will be required for half of the trials and in L2 for the other half.
Due to children’s young age and the task demands, considered possibly too high for their age (M. Siegal, personal communication, April 10 2010), instead of digit lists thus random presentation of digits, the target stimuli will be presented in sequence (from 1 to 10). Additionally, the language cue will change in each trial, thus all trials will require a switch from L1 to L2 and vice-versa.
Since a measure of intelligence has not yet been standardized for the Albanian population (e.g. Zimmerman, Connolly, Bozo, Bridson, Rohner & Grimci, 2006), the following intelligence tests will be used to match the bilingual and monolingual groups. The Raven’s Standard Progressive Matrices (SPM; Raven, 1958) and the Raven’s Coloured Progressive Matrices (CPM; Raven, Court & Raven, 1986) will be used to measure general intelligence in adults and children respectively, which is considered to be a relatively culture-free, reliable and valid measure of Spearman’s g (Raven, 2000; Wicherts et al., 2010). As an indicator of verbal intelligence, the Vocabulary subtest of the Weschler Adult Intelligence Scale- version III (WAIS-III; 1997) and the Weschler Intelligence Scale for Children-version III (WISC-III; 1991) will be administered. As norms for this test are not available yet for Albanians, the teachers of the bilingual children will be asked to provide ratings of the children’s Albanian vocabulary, where applicable, which will then be correlated with the Greek norms for this test. Additionally, a psychologist of Albanian nationality will score the children and the adults on the Albanian vocabulary test, from the participants’ recorded answers.