Language Processing of Bilinguals

Effects of speed of word processing on semantic access: The case of bilingualism.

Clara D. Martin, Albert Costa, Benjamin Dering, Noriko Hoshino, Yan Jing Wu
Presented by: Josh Elliott

Introduction:

The study investigated the phenomena in which bilinguals have consistently displayed slower language processing than monolinguals in both their second and first language. Past studies display these deficits in a large array of tasks, including word list recognition (Golan et al, 2011), non-word detection (Soares & Grosjean, 1984), and lexical decision making. Given this information, the authors set out to tests whether or not discrepancies between the two groups processing speed affects the semantic representation of words, even when the task does not require semantic access. To test this, the authors use a paradigm developed by the lead author in a previous study, where English-Welsh bilinguals were presented with a stream English and Welsh words, and were instructed to indicate where or not the presented word in one language was more or less than 5 letters, and to disregard words in the other non-target language. Afterwards, data using event related potentials (ERP) indicated that semantic processing was involved in both the letter counting task and word disregarding task. (Martin 2009) A similar paradigm was used for the current experiment where a stream of English and pseudo words were read to monolinguals and bilinguals, testing the hypothesis that semantic activation of unattended English words is due to slow processing speed of bilinguals.

Methods:

The participants consisted of 15 native English monolingual speakers (13 female: mean age = 20.6 years ± 5.2) and 15 native English-Welsh speakers (9 females; 22.9 years ± 4.6) all of whom spoke both languages on a daily basis. The stimuli that were presented were 156 common English nouns and 156 pseudo words that were presented in a continuous stream and were all repeated 4 times throughout the experiment, resulting in a total of 1248 English words and pseudo words being presented. In addition, each word was presented as a filler word, or as critical stimuli, meaning that it was an English word paired with a semantically related or unrelated English word, or it was a pseudo word paired with a pseudo word. All words were presented for 200ms with 800ms in between each word, while participants were instructed one of hit a button for pseudo words longer than five letters, another button for pseudo words give letters or less, and to withhold a response if the word was an English word. To measure the process, button responses were used to measure reaction time accuracy, while ERP's were used to measure the 3 target conditions: English pairs with semantic links, English words with no semantic link, and pairs of pseudo words.

Results:
The author's hypothesis was confirmed, as the experiment showed that bilinguals had slower word processing than monolinguals, as demonstrated in the in button reaction time, as well as the ERP results on N2 (fig 1b. and fig 2b.), which indicating different processing speeds between groups. Because of the slower language processing exhibited by bilinguals they showed signs of the semantic priming effect for semantically related words, as shown by the N400 when comparing semantically related words with unrelated words between groups. These results suggest that bilinguals automatically process the semantics of a word, implying that being a fluent in another language changes how the brains work as a whole. In fact, more research has been conducted regarding how the brain networks of bilinguals differ from that of monolinguals and that the brains of bilinguals have more white matter tracts that connect areas of the brain associated with language (García-Pentón, Fernández, Iturria-Medina, Gillon-Dowens, & Carreiras, 2014). Brain imaging studies like this support the idea that being bilingual changes the way the brain communicates, but as far as semantic access speed, there are still questions regarding whether its slow word processing in particular that causes automatic semantic processing, as the authors suggest. A way to possibly test this is inducing slow word processing in a group, or to find a group with naturally slow word processing (such as individuals who might have forgotten their native language), and see if they display automatic semantic activation as the authors would suggest. There are also other possible explanations, such as whether or not it's due to specially learning multiple languages at a young age, such as all the participants in the study, and whether it's a phenomenon specific to spoken language. However, it's difficult to disprove, as the study did not mention any other hypotheses that explain automatic semantic access.

References:

Graph 1

Graph 2