From Polly’s L&P paper The (dis)organization of grammar: 25 Years (2000):
[…] new complications arise the minute one moves away from the ‘running in tandem’ of the syntax and semantics.
I’m totally smitten with direct compositionality and have been following it since being introduced to Polly’s work two years ago. And yet, it seems I still have some things to learn!
In an attempt to simplify my meaning of the relative clause and yet still allow stacking/other domain restriction to happen, I posited this rule in my second-to-last set of notes on RCs:
   If there is an expression of the form <[α]; RC; [[α]]>, then there is an expression of the form <[α]; RC/RRC; λQe,t[λDe,t[λPe,t[Q(α ∩ D)(P)]]]>.
The idea I had in mind was just to lift the semantic value of the original RC to the more complicated one that I had been pushing. And, sure, it works, but as Polly noted to me, it’s also pretty goofy. The syntactic category of RC is being lifted to take another RC…and yet the semantics seems to be doing something entirely different! In part, this was just sloppiness on my part due to haste (and admittedly, I felt that something was wrong with the idea of giving the shifted RC a syntactic category that took another RC while thinking that it could first combine with the quantifier…actually, I knew something was wrong, I just couldn’t pin down what); but, it was also something that I’d never thought much about before. When we say that the syntax and the semantics work in tandem, we mean precisely that! So when you posit a shift rule, the syntactic and semantic shifts should make sense with respect to each other! That is, the more elegant (and simpler and right-er and and …) rule would be:
   If there is an expression of the form <[α]; RC; [[α]]>, then there is an expression of the form <[α]; RC/RRC; λRe,t[α ∩ R]>.
Ach, life :-P
emma :: Mar.01.2007 ::
misc, semantics, linguistics ::
No Comments »
