>>> [log in to unmask] 03/05/02 09:20pm >>>
#And Rosta wrote:
#>
#> (i) When a category exhibits both X-type and Y-type properties
#> (e.g. Subject-type and Object-type), the best way to analyse it
#> is in terms of the category being both an X and a Y.
#
#Hmmm... The duckbill platypus has both bird properties and mammal
#properties, yet it's chategorized as a mammal. This is not to say
#that you're not right, at least insofar as general syntactic
#methodology is concerned.

That's kind of a hot debate in biological taxonomy, isn't it. The majority
see classification as determined solely by evolution, so a taxonomy
is a family tree, while a majority see a taxonomy as a more general-
purpose classification that takes into account physical structure
and so on.

At any rate, the purpose of categories in syntax is simply to
model the data; if I classify a word as a common noun, all I'm
saying is it behaves like a common noun. But if in biology
I classify something as a mammal, I'm not just saying it has
fur, bears live young, lactates, etc.; I'm also saying that it is
an evolutionary descendent of the protomammal.

[...]
#> (iii) I don't see any place in the grammar for Er and Ee. Outside
#> the gramamr, word senses define their own participant role, so the
#> encylopedia entry for Kiss will include definitions of Kisser and Kissee.
#> Inside the grammar, some syntactic categories (such as dependencies)
#> have semantic content; dependencies with semantic content are
#> 'theta-roles'. There is no need for additional layers of structure to
#> mediate between syntax's theta roles and the interpretative interface
#> with pragmatics.
#
#Does this allow for syntactic bootstrapping? The experimental evidence
#is overwhelming that syntactic bootstrapping exists (in fact, adults
#appear to be better at it than children, which is not surprising).

I think it allows for a much better account of syntactic bootstrapping than
pretty much any alternative model does. Syntactic categories that have
intrinsic meaning are going to be easier to learn than ones that don't.
And a language-specific mediating level of macroroley things like
Er and Ee don't have to be learnt at all.

[...]
#> Do regular passives' subjects also create barriers to extraction,
#> and disallow secondary predication?
#
#I don't think so:
#
#- De que  color fueron  comprados los    muebles?
#  Of what color were-PL bought    the-PL furniture-PL
#  'What was the color of the furniture that was bought?'
#
#- Las    mujeres fueron  vistas borrachas.
#  The-PL women   were-PL seen   drunk.
#  'The women were seen drunk.'
#
#I think that these are both okay (though perhaps a bit stilted).
#
#> If No, then they would pattern with S=X, psych and unaccusatives.

So passives add to the overall prettiness of the picture.

[...]
#> Another way of looking at it might be:
#> S = 'subjecty' [no semantic content]
#> A = 'agenty' [has semantic content]
#> P = 'patienty' [has semantic content]
#
#In English, at least, the experimental evidence suggests that subject
#does have attentional content. There's still a lot of work left to do
#to make the claim stick, but it seems that subject = locus of attention.
#(I can explain the experiments, if anybody's interested.)

I wouldn't want to dispute this. I should have used a different term than
'semantic content', to make clearer a distinction between 'discoursal
meaning' and the sorts of meaning relevant to semantic roles and
linking.

#A nice test would be a Spanish psych verb like GUSTAR (with the
#unmarked DAT-V-SUBJ order). If the dative participant turned out to
#be attentionally more salient, then we might write the whole whole
#thing down to leftwardness, i.e., topics tend to be expressed
#first.

I wouldn't be surprised, though, if the dative of psych verbs has
some subjecty properties that the actual subject lacks. For example,
it could be that psych datives and ordinary subjects both occupy
topic position. 

--And