AW: Trees, pheno, tectogrammar

[Carl Pollard]

Hi Tibor,

>
yes, agreed, but I thought the conceptual advantage of a Dowty-ian
Pheno/Tecto distinction would be that both levels can be worked on
independently. If syntax must mediate between these levels, and we have
syntax anyway, I don't think that wee need these additional levels.
>>

That is a little like saying we don't need the sine and cosine because
the lengths of the sides mediate between them! Think of pheno and
semantics as INTERPRETATIONS of the syntax, not as other "levels".
In that respect it is like GB/MP. But the interpretations of the syntax
are not sequences of movements, but rather structure-preserving
functors (in Lambek's sense). In terms of the typed lambda calculus,
they are translations that preserve all the TLC term constructors
and type constructors

>
So, to paraphrase your reaction: my observation is damaging for Dowty's
initial proposal, for Kathol's thesis, for Kathol/Kasper/Pollard on
extraposition, right?
>>

I'll have to go back and look.


Carl