Thursday, 8 October 2009

This one of those times I hate being a linguist

Non-linguists, ignore this post.

I was on Facebook recently (when am I not?) when a friend posted a link to an event with the warning to go to said event or "I will shun you like an Amish child with an iPod". Seeking clarification, I asked whether that meant he intended to shun me "as he would shun an Amish child with an iPod" or "as an Amish child with an iPod would shun me". After initially suggesting "both", he settled on "as an Amish child with an iPod would be shunned by his (or her) community".

I had considered the last possibility, but discarded it as a) not funny, and b) not following from the initial claim.

But thinking about it, I couldn't decide why I thought it was an bad reading of the claim. I think it's because it involves the supposition of an entity (the Amish community) which isn't instantiated (that isn't the right word, but follow me) in the discourse. But why isn't it? Specifying the child as Amish surely presupposes an Amish community, doesn't it?

So maybe it is, and I am being misled by the syntax, or maybe it's the semantics, or some combination thereof. Experts feel free to chime in on that question too. "Shun" is a two-place predicate, involving a shunner and a shunnee, who (as the speaker and addressee) are clearly instantiated into the discourse in the possible world in question in "or I will shun you". So my readings of the (elliptical) clause "like an AcwaiP" involve filling in the predicate (to shun) and the other role, i.e. I read the clause as "like [X will/would shun] an AcwaiP" or "like an AcwaiP [will/would shun" Y]. And my preference, and perhaps it's only me, is to coreference the empty NP with one of the available antecedents (the speaker, or the addressee). There's a syntactic contraint here, in that the established entities (the speaker and the addressee) cannot be switched, i.e. there is no reading of "like an AcwaiP [would shun the speaker]" or "like [the addressee would shun] an AcwaiP".

So my questions are:

a) Am I the only one with these intuitions?
b) Is this a matter of discourse operations (or Gricean-type information construal issues)?
c) What *is* the syntax/semantics of this situation and its solution?

and for good measure

d) What is the syntax/semantics of [like an AcwaiP] w/r/t the rest of the clause? I suppose it's a manner, which makes the whole thing an adverbial, but heck if I know how to draw the tree.

1 comment:

D. Sky Onosson said...

It's things like this that are making me appreciate my surprising (to me) move away from syntax/discourse towards phonology/phonetics.

I remember once talking to Kevin about some assignment or other, and telling him I was interested in "why" such-and-such was happening. He told me, don't worry about the why, just the what (not an exact quote).