Tami Gollan wrote that in an e-mail discussion we were having about cognates and I love the line, so I borrowed it as my title– and I agree, cognates are wicked cool!
At the bilingual workshop, Hans Stadthagen-González gave a talk about his research on cognates. We had fun discussing cognates and what you do with them in testing. Some of the questions that came up included:
- In language pairs where there are a lot of cognates do we eliminate them from tests?
If we eliminate them, then we end up with items that are non-cognates. If bilingual’s attention is divided between two languages then the non-cognates would be those that occur LESS often because their frequency for bilinguals would go down. So, that test would be HARDER. But, then the cognates might be EASIER because they’d be MORE frequent. Or, maybe they wouldn’t because of course the occur in both, but each occurs less. Is this making any sense?
- And what about cognates that are high frequency in one language, but low-frequency in another?
Even if the word is common in one language doesn’t mean it’s common in another. In our article on cognates we talk about the item “velocity” from the TOLD. In Spanish, the translation “velocidad” occurs much more frequently than it does in English. So, maybe kids who have exposure to it in Spanish will learn it in English as well, even if their monolingual peers don’t know it.
- How do we “control” for cognates?
I think a possibility is to index these by frequency in the language. But, then for bilinguals, frequency will be different. Is it even possible to do? What would the math look like? Should we even bother?
I’m asking more questions than I’m answering I know. But, there is still a lot to learn in this area– at least for me.