Reentrancy in feature structures

[Carl Pollard]

Hi again,

MIke Maxwell replied to John Beavers:

>
As a lurker on this list, let me ask a couple naive questions.  First,
it seems to me that the meaning of "identical" in Luis Casillas'
definition is a crucial issue here.  Dredging something out of my old
lisping days, I recall that Lisp made a distinction between (at least)
two notions of "identical", often labeled "eq" and "equal".  One (I
can't recall which) required identity of content (so e.g. two strings
were equal under this definition if they were "spelled" the same), while
the other required identity of reference (i.e. they were the same memory
location).  I presume that condition 4 requires the latter (stronger)
definition, correct?  In which case the equivalency really is just the
same as re-entrancy, correct?
>>

Right, the difference between EQ and EQUAL is similar to the
difference between token identity amd isomorphism of substructures.


>
Second (and assuming the answer to the first question is "yes"), I can
think of one obvious case where this sort of thing makes a difference,
namely with co-reference.  E.g. in some language where verbs are
marked for agreement with subject and object, you would get a
reflexive form only where the subject and object agreement features
were identical in the strong sense.  (Otherwise you couldn't
distinguish the translation equivalent of "He saw him" from "He saw
himself.")
>>

Exactly. See my previous message. The point is better made though using
inanimate NPs and a language where the agreement features are not
semantically motivated, since then it is harder to say that it's
all semantics and has nothing to do with syntax.

>
Are there other cases where the strong notion of identity is required,
besides (person) co-reference?
>>

I hope not, since in my current research I have foresaken the
graphical implementation of feature structures in favor of a treatment
in terms of higher order logic. In the models this means feature
structure types are indexed product types and the feature structures
themselves are members of those types.  Thus there is only plain old
equality (EQ), nothing corresponding to EQUAL. This makes some things
(such as formalizing set values and distinguishing neutralization from
ambiguity) much easier, but also means that coindexing cannot be
handled the same way as in HPSG.

Carl