Has the Wason Selection Task been tested in languages other than English in which the equivalent of the “if … then” construction works differently (if there are any such languages)?
This seems like an obvious thing to do, since people have argued over whether people’s “bad” performance on the task just reflects the colloquial meaning of “if … then” in English (in unfamiliar contexts), but I can’t seem to find anything about it on Google.
I don’t know, but I am so totally on board with research that checks whether cognitive biases are actually due to natural language not being formal logic.
I have no intuitions for the particular case you’re talking about, but I got really annoyed a while back because the conjunction fallacy just seemed to be people using Grice’s maxims. “Linda is a bankteller.” “Linda is a bankteller and a feminist.” Well, if Linda were a feminist, the first sentence would have mentioned it, so people actually take it to mean “Linda is a bankteller and not a feminist.”
Anyway, there seems to be a paper that makes the same argument I’m making, but I haven’t read it. And I don’t really know anything about the Wason Selection Task. So, uh, apologies for the off-topic response; this is just something I like ranting about.
That’s a very good point about the conjunction fallacy. I feel kind of skeptical of a lot of cognitive biases research because it seems like the tests are attempts to trick people, and I have a hard time saying the results generalize to actual reasoning when linguistic ambiguity or other “tricking” mechanisms aren’t present. (I mean, I’m sure people investigate that kind of thing — I just don’t see the distinction made in a lot of popular writing about the subject.)
ETA: The Wason Selection Task is particularly frustrating because the usage of “if … then” they’re looking for is common in certain disciplines, which is a confounder. Reportedly MIT students do far better at it than the general population — does that meant that MIT students are “more rational,” or just that they’re more likely to study logic and computer programming? (This is an obvious thing to control for, and probably has been controlled for, but I’ve only seen the raw statistics stated.)
Yeah. It seems like many people who are good a formal kinds of logic then forget that natural language is based on Implicature. This leads to much nonsense.
