Prevailing approaches do not have a computational lexicon

Shalom Lappin lappin at DCS.KCL.AC.UK
Sun Sep 22 09:18:31 UTC 2002


Hi Ash,
   I wonder why you are surprised by this phenomenon. There are two points
   worth noting here. (i) It is very common for theoretical linguists
   working in the minimalist framework to follow Chomsky in referring to
   syntaxas "the computational component" of the grammar while assiduously avoiding
clear specification of the computational properties of the syntactic
operations that they posit as elements of this component. I have found
that when pressed on this and related issues many minimalists tend to
retreat to the claim that they are only concerned with a theory of
competence, and so they are not responsible for a computational account of
the grammar, which they assign to impelmentation or performance. The
question remains, then, in what sense they are describing syntax as a
computational component of the grammar. (ii) The widespread lack of
awareness (indifference?) to work going on in other theoretical frameworks
and to research in computational linguistics is legion, and it has been
noted many times. From my own experience in these matters I have found
that regardless of how often one raises these questions, it has little
impact on working practise in large areas of the field. It is probably
best simply to set them aside in order avoid pointless irritation and to
use the time to pursue productive research. Regards.                               Shalom

>
> Dear list members,
>
> I draw your attention to the first line of the abstract included below
> for a talk by Tanya Reinhart at MIT:
>
> "The theoretical background assumption is that UG includes a
> computational lexicon, in which operations can apply (contra to
> prevailing approaches)."
>
> I realize that this is only from an abstract for a talk, but I don't
> understand a) the implication that this is a new approach, or b) the
> assertion that it is contrary to prevailing approaches.
>
> Even if "prevailing approaches" were meant to be construed as
> "Minimalism" (which it often is), I think this would still be puzzling,
> since
> Minimalism also claims to feature a computational lexicon and wants
> variation to be located in the lexicon, although no substantive theory
> of lexical organization or variation has been offered, in my honest
> opinion.
>
> Ash
>
> P.S. This message is being BCCed to Tanya Reinhart (BCC so she doesn't
> get spammed by "reply-to-all").
>
> -------------------------------------------------------------------------
> Abstract
>
>
> Thematic arity operations: Parametric variations.
>
> Tanya Reinhart (work with Tali Siloni)
>
> The theoretical background assumption is that UG includes a
> computational lexicon, in which operations can apply (contra to
> prevailing approaches). The focus of the talk will be the operation of
> reflexivization.  We argue that the massive linguistic variations found
> with reflexivization can be reduced to two parameters:  a. Does the
> operation apply in the lexicon (English, Hebrew, Dutch) or in the
> syntax (Romance, German)?  b. Is the accusative case structural
> (French,
> Italian, Dutch, German) or only thematic (Spanish, Hebrew, English)? In
> structural accusative languages the Auxiliary in unaccusative
> derivations is be, in thematic accusative languages, it is have. The
> syntax-lexicon Parameter setting makes use of data like the following:
> In the syntax setting, reflexivization is possible into ECM subjects.
> (In the lexicon setting - it is not). In the lexicon setting, reflexive
> nominalization is available. (In the syntax setting  it is not).



More information about the LFG mailing list