http://bps-research-digest.blogspot.com … guals.html
"... The researchers recorded the eye movements of 45 bilingual Belgian students as they read sentences in their native Dutch tongue. The key finding was that they read Dutch words faster when the equivalent word in their second language, English, was similar or the same as the Dutch word. Specifically, they spent less time fixating on words like "piloot" ("pilot" in English) than on control words like "eend" (that's "duck" in English).
Van Aassche and her colleagues said this shows that even when bilinguals read sentence after sentence in their native tongue, access to words in their second language remains open, rather than switched off, thus having an effect on the way the native language is processed... "
kazelee wrote:
"access to words in their second language remains open, rather than switched off, thus having an effect on the way the native language is processed"
...Is a larger conjecture. To me, it seems like they are reaching a bit farther than they should in saying this.
From the abstract: "Thus, this study shows that one of people's most automated skills, reading in one's native language, is changed by the knowledge of a second language."
Hmm. Even if they're only reading the L1 word faster because they have effectively had more exposure via the L2 cognate, doesn't the benefit of that exposure depend upon activating the same associations that were used in the 'doubled' exposure to the variations? It would still be a bit of a different conjecture, as you say though, because they'd be relying on single-word cues that could trigger specific access, thus perhaps rather than a general switch it's better to think of activating certain paths in a network. Or something. That metaphor is kind of what threw me off originally. Also, relying on eye movements to judge reading speed, and the fact, as a commenter on the above page mentioned, that they're using languages with the same scripts. I'm sure there's better articles out there that look at something besides cognates, getting to the semantic level or somesuch, like this kind of research: http://www3.interscience.wiley.com/cgi- … 4/ABSTRACT