Cerebral Lateralization of Second Language
Faculty Mentor: Gregory A Thompson, Anthropology
It has been generally accepted that language shows left-hemisphere dominance in right-handed monolinguals. In contrast, current models of L2 lateralization reveal no such consensus. Some studies (Galloway and Scarcella 1982; Gordon and Zatorre 1981; Paradis 1992) have either found no evidence supporting increased right hemisphere involvement in L2 processing or indicated that bilinguals’ L2 is not any more bilaterally organized than their L1. Other studies (Ke 1992; Hoosain and Shiu 1989; Jia et. al. 2013) have shown the opposite, finding that the L2 is significantly less left-lateralized. Still other studies have suggested that there is simply too much inter-individual variation in L2 organization to make generalizations useful (Perani et al 1996). It may be the case that factors like proficiency level or age of acquisition, have the greatest effect on lateralization.
Overall, the research on this topic has mainly focused on adults with high attained proficiencies in their L2. Most studies have investigated L2s that are closely related to the L1. The literature could benefit from studies including subjects of different age groups, different proficiency levels, and more diverse L1/L2 diads. This study seeks to contribute evidence to a model of cortical representation of second language by administering a Dichotic Listening (DL) task to a group of Korean fifth-grade elementary school students learning English in school as a foreign language.
Subjects included twenty right-handed Korean children in the fifth grade enrolled in an English class at a public elementary school in Seoul, South Korea (English classes are part of the standard curriculum of Korean schools). The stimuli were presented in the form of an online Qualtrics survey distributed via a URL link. The survey first gathered demographic information, then presented a randomized series of eight listening tasks. The tasks involved listening to audio recordings of word lists in either English or Korean. There were four listening tasks for each language: two baseline tasks, which presented only one word list in one ear; and two dichotic tasks, which presented two different word lists simultaneously, one in each ear. Instructions for each recording asked the students to attend to a particular ear. On the following survey page, they selected the words they heard from the attended ear, choosing from among 10 distractor words (20 words total per question). Data analysis was completed on Microsoft Excel. Right ear advantage (REA) scores were calculated by subtracting the number of words correctly identified in the left ear from that of the right ear. For each language, baseline REA was compared to dichotic REA using a matched-pairs t-test to determine whether the distracting word list produced an effect. Then, dichotic REAs between languages were compared, also using a matched-pairs t-test.
No difference was found between dichotic REAs and baseline REAs. A matched-pairs t-test of English REA and Korean REA showed no significant difference between the languages (p = 0.23). English REA and English proficiency (measured as English baseline REA) showed a slight but insignificant positive correlation (r2 = .30485). But more surprising than these results is that the mean REA scores showed no evidence of lateralization to either side. The mean English REA was 0.45 (p = .35) and the mean Korean REA was -0.45 (p = .33), neither of which was significantly different from zero.
This study failed to find any evidence of language lateralization in either L1 or L2. Given that the literature clearly indicates a pattern of left L1 lateralization across cultures and languages, these results are surely due to limitations in the study design, and not to some bilateral language phenomenon in this population of Korean children.
Limitations include the relatively relaxed administration of the survey that resulted from assessing many students at one time. It is possible that some participants’ responses were skewed by consulting with their neighbors, listening to the recordings more than once, listening to the recordings while giving responses, or improper use of the headphones. Furthermore, past DL studies have used word lists of 40 to 60 words. This number was not practical for the current study, but as a consequence of the shorter lists, some lateralization could have been masked in the data.
Even if the present study had found significant differences in the data, a number of methodological concerns make interpreting DL data tricky. The DL method has been used in a variety of study designs to measure a handful of different (but related) cognitive functions. While differing REA scores have been shown to reflect lateralization to a significant degree (as confirmed by DL/fMRI combined studies), they are also affected by complex interactions among a many interwoven cognitive functions (Norrelgen 2012). Still other concerns must be borne in mind when interpreting DL data. Obler et al listed some parameters of interest in their 1982 review: “subject selection, language and stimulus selection, testing procedures, data analysis, and interpretation of results.” It may be the case that inter-individual variation along these parameters has been so high that averaging the data would not produce any interesting new information. In summary, researchers should practice extra rigor when administering DL tasks and interpreting the results.
The present study hypothesized that the L1 and L2 of Korean children learning English would show differing lateralization patterns. The DL task, administered to fifth-grade Korean classrooms in the form of a Qualtrics survey, was unable to find any evidence of this. Further research into the neural substrates of second language should focus less on hemispheric dominance and more on the factors that give rise to inter-individual variation.