Garcia, X. . 1 , Martin, C. 1 , Breton, A. 3 , García, G. 1 & Costa, A. . 1, 2
1 Departament Tecnologies de la Informació i la Comunicació, University Pompeu Fabra, Barcelona, Spain
2 Institució Catalana de Recerques i Estudis Avançats, Universitat Pompeu Fabra, Barcelona, Spain
3 Institut des Sciences Cognitives, CNRS-Université de Lyon
The main goal of the present study was to investigate difficulties in sentence reading in a second (L2) compared to a first language.
Sentences that we hear or read can be meaningful or not (e.g., Mozart composed classical/orange music) and can also be true or false (e.g., Mozart composed classical/jazz music). Hagoort and colleagues (2004) tested the temporality of meaning and truth integration during sentence comprehension. They recorded Event-related Potential data from participants who had to read silently correct sentences (e.g., Mozart composed classical music), sentences with semantic violations (e.g., Mozart composed orange music) and sentences with world knowledge violations (e.g., Mozart composed jazz music). ERP data revealed that integration of both semantic and world knowledge during sentence reading happens in parallel (both violation effects observed 400 ms after the critical word presentation; N400 effect).
In the present study, we investigated semantic and world knowledge violations during reading in a second language (L2). We hypothesized that difficulties in reading in L2 might be due to difficulties in world knowledge integration: It might be that detecting meaningless sentences is as easy in L2 as in L1 while detecting false sentences is harder in L2 than in L1.
While recording electroencephalographic data, we presented Spanish-English late bilinguals with correct sentences, sentences with semantic violations and sentences with world knowledge violations. Participants had to read sentences silently. Semantic violations induced a N400 effect, as previously observed in monolinguals (Hagoort et al., 2004). While world knowledge also produced a N400 effect in monolinguals, the effect was delayed in bilinguals (P600 effect).
We concluded that L2 readers integrate semantic as monolinguals do, but integrate world knowledge later on in the processing stream. This delay in world knowledge integration might be one of the reasons why reading is more laborious in L2 than in L1.