Constrain my complement to a specific type


#1

If I have a complementizer type and a clausal verb type, can I constrain the latter to only combine with the former (rather than any complementizer)? If yes, how? I forgot…

Below, can I constrain cverb-comps1-lex-item to only allow comps1-comp-lex-item as its complement?

clausalcomp-verb-lex := verb-lex & clausal-second-arg-trans-lex-item &
[ SYNSEM.LOCAL.CAT.VAL.COMPS < #comps >,
ARG-ST < [ LOCAL.CAT.HEAD noun ],
#comps &
[ LOCAL.CAT [ VAL [ SPR < >,
COMPS < > ],
HEAD comp ] ] > ].

cverb-comps1-verb-lex := clausalcomp-verb-lex.

comp-lex-item := raise-sem-lex-item & basic-one-arg &
[ SYNSEM.LOCAL.CAT [ HEAD comp &
[ MOD < > ],
VAL [ SPR < >,
SUBJ < >,
COMPS < #comps > ] ],
ARG-ST < #comps &
[ LOCAL.CAT [ HEAD verb,
MC -,
VAL [ SUBJ < >,
COMPS < > ] ] ] > ].

comps1-comp-lex-item := comp-lex-item.

This results in an error:

cverb-comps1-verb-lex := clausalcomp-verb-lex &
[ SYNSEM.LOCAL.CAT.VAL.COMPS < [ SYNSEM.LOCAL.CAT.HEAD comps1-comp-lex-item ] > ].

Error: “No possible type for features (SYNSEM LOCAL NON-LOCAL) at path (SYNSEM LOCAL CAT VAL COMPS FIRST)”.


#2

It’s never a good idea for anything other than the DTR position of a lexical rule to select for a lex type of any sort. And here that’s clearly not what you want because the verb is combining with a CP, not the C directly.

If you want to be able to select for some particular C, then there has to be a feature that projects from that C to the CP that verb can be sensitive to.

What is the motivating example here?


#3

Multiple complementation strategies. But I think I am getting ahead of myself; right now I am not 100% sure why I thought I needed something like that. I will come back once I test multiple strategies.


#4

The only examples I know of are cases where complementizers differ in some other property that we already have a feature for (whether v. that is SF ques v. prop; that vs. for-to is FORM …).


#5

Back to this question:

In Russian (and in English, for that matter, I think), there are verbs that can take a clausal complement with the complementizer “that” and then there are verbs that can take a clausal complement headed by the complementizer “how”, but not necessarily both:

(1) I thought that Ivan ran
(2) I saw how Ivan ran

but
(3) *I thought how Ivan ran

Other than the spelling of the complementizer, there is no difference in these clausal complements.

How do I achieve this? What can I say about the clausal verbs (see vs. think), do I need to use FORM here, e.g. [ FORM that, FORM how ]?


#6

In English, there’s a difference in meaning. “see” can take wh-complements, but the wh-word should definitely appear in the semantics:

I saw how Ivan ran
I saw where Ivan ran
I saw who ran
I saw what Ivan ran to

This suggests that there’s a difference in meaning for Russian: https://books.google.co.uk/books?id=nh0UDAAAQBAJ&pg=PA372 – would you agree with their judgement?

Supposing there are lexically-conditioned complementizers, FORM would seem sensible.


#7

Well… I’d say that with the verb see (we are talking about Russian here), there is no difference between “I see that…” and “I see how”, unless you stress the “how”. But still, “how” is impossible with “think”, and then with something like “know”, how must mean “how” and does not seem synonymous with “that”.

(Don’t seem to be able to see the page in that google book).


#8

It occurs to me now that what I am modeling are semantically empty complementizers, so, I probably should not be throwing “how” into the mix at all… Or only test it with verbs with which it appears exactly the same as “that”.


#9

I’m pretty sure that see and know in English can embed questions as well as propositions and that how in English is never a complementizer.


#10

Strange that you can’t see the page. Copying it here (Dixon, 2010):

Russian distinguishes between reporting a fact and describing an activity but, unlike English, it uses the same complement clause construction, simply employing different complementizers—čto for fact and kak for activity, as in (Barentsen 1996: 23–4):

(9) Len videl [čto Mardži igraet v kroket]
Len saw that Mary played croquet
(10) Len videl [kak Mardži igraet v kroket]
Len saw Mary play croquet

What do you think of those examples?


#11

A pair that might make the semantic contrast clearer:

I saw that the book was new
?I saw the book be new

Does this sentence sound okay with čto but weird with kak?

This example is based on Benjamin Whorf’s discussion of complementizer choice in Hopi, in “Language, Thought and Reality”.


#12

(10) is sort of ambiguous for me, meaning both the same as (9) and meaning also “Len saw the way Mardzhi played croquet”.

As for:

I saw that the book was new
?I saw the book be new

– I think I can say:

(1) Ya videl, chto kniga novaya
I saw that book new
`I saw that the book was new’

(2) Ya videl kak kniga byla novaya
I saw how book was new
`I saw that book being new" – or something like this…

(2) is not very frequent, but I think it is possible with the past tense copula. I can’t imagine “kak” here without the copula though.