On Fairness
Being fair to both children doesn't always mean giving the same thing to both. For example, Alex and I wouldn't be equally happy with a nice My Little Pony or a stupid set of Magic cards -- if we both got one, only one of us would be happy.
I tried to bring this up in class a couple months ago, when we were discussing designing experiments to test linguistic judgments. Say you want to compare how acceptable people find (1) vs. (2):
(1) She is easy to fight.
(2) She is eager to fight.
But you want to precede these sentences with a context where "she" is introduced, like "Mary is coming over, and she is really mad." That sortof makes sense with both, but it makes more sense with (2) than with (1). In order to avoid introducing variables, you might be tempted to use the same context for both sentences, but there's a trade-off. Maybe to really be "fair", you should use different contexts for both sentences. Then you could ensure that the transition between the context and the target sentence is always equally coherent.
Speaking of coherence, is this making sense?? No one in class understood what I was talking about.
posted by Liz Coppock @ 11:22 AM
3 comments
3 Comments:
Hmmm. Are there people who find (1) more acceptable than (2)? Because I find (2) perfectly acceptable on its own, whereas (1) seems to need a context. Normally, I would say "She is easy to beat" rather than "She is easy to fight."
But I think I see what you mean. People are assuming that to test the acceptability of the phrase, you would have to use both in the same context.
However, what you seem to be saying is that each phrase might be perfectly acceptable IN ITS OWN BEST CONTEXT, but not very acceptable in the other's best contest.
Hence the parallel to fairness. Also see Dostoevsky, Crime and Punishment. Something about tea and jailed gentlemen.
Never, ever See: Dostoevsky.
But the point about giving each sentence its own context when determining acceptability is really interesting. It's true that if you use the same lead-in for two sentences, it could favor one or the other, but it's also true that if you use different lead-in's, you have a weaker point of comparison. two variables instead of one.
How about giving both sentences both contexts - or both sentences four contexts?
I thought Liz's point was that setting up the same context for each phrase that you're testing for acceptibility doesn't really test accurately.
What have you learned about the acceptability of a phrase if you set up a context in which one phrase fits, but the comparison phrase doesn't fit? Do you mark the second one "unacceptable", even if it would work fine elsewhere? And if you set up a context where the second phrase fits well, but the first phrase fails, is the first phrase now the "unacceptable" one?
Post a Comment
<< Home