LINGUIST List 6.549

Wed 12 Apr 1995

Sum: Three argument verbs

Editor for this issue: <>


  1. , Three argument verbs

Message 1: Three argument verbs

Date: Sat, 08 Apr 1995 22:45:08 Three argument verbs
From: <>
Subject: Three argument verbs

Here are three of the many responses I got on the quesition I posed about verbs
like "consider" as three-argument verbs where the two internal arguments are in
a copula relationship. One of the respondents who agreed with me stated that
one could view these arguments as 1] theme and 2] goal. This makes the most
sense to me, bearing in mind that I don't know much about case theory and have
a hard time buying what little I do know. The first respondent below mentions
that it could work just fine as a rule, but that it wouldn't really explain
anything. I think the second respondent makes a good case for the usefulness of
the rule as applied to natural-language processing. I haven't completely
figured out the third, which is a look at this problem as dealt with in Korean,
but it sounds interesting and I intend to go back and look it over again.
Once again, I'm no syntactician (or syntactologist or whatever you choose to
call it) but it's an interesting thing to think about.
Thanks for all your responses! David Harris, Georgetown U.
>From IN%""
Subj RE: A rethinking of small phrases as phrases containing three-argument
Dear David Harris,

As you assume, the class of verbs taking small clause complements are a
relatively small class of verbs. I think it is the same class of verbs
that are usually called E(xceptional) C(ase) M(arking) verbs. Thus they
can also have an "accusative with infinitive" (or whatever is the "school
grammar" term for it). A relevant pair would be:

(1) I saw [Peter naked]
(2) I saw [her steal his clothes]

About your rule, it seems to me it might work just fine. The only problem
is that, as far as I can see, it doesn't explain anything. It just
stipulates a property of these verbs. I think a generative grammar should
do more than that. Thus, the standard "GB" explanation for the limited
distribution of constituents like those bracketed in (1) and (2) above is
that the matrix verb selects a complement of a category that doesn't block
case marking from the matrix verb, ie. IP, instead of CP. This, of course,
also reduces to a stipulation, but, I think, a more "innocent" one than


Oeystein Nilsen.
>From IN%"Marion.KeeA.NL.CS.CMU.EDU"
Subj RE: A rethinking of small phrases as phrases containing three-argument

Sure, that's the kind of thing we do in Natural Language Processing.
On my current project, we needed a few three-valued verbs like that.
We classed them as such in our English lexicon. When our parser
encounters a structure that fits the three-valued pattern, it triggers
rules we wrote for those verb classes, which handle the analysis of
those uses of those verbs. The trick is to make sure you define the
three-valued contexts precisely enough to avoid firing the rules when
they shouldn't apply, or not firing them when they should. That
generally requires semantic classification not only of the verbs in
question, but of the nouns and adjectives that can combine with them in
three-valued structures. We generally apply lexical semantics in
classifying the items in our lexicon.

If I understood what theories of syntax are really looking for (what is
the Holy Grail of that field), then I might be able to explain to you
why this analysis is not one that's likely to arise in syntactic theory.
Unfortunately, I've never understood why the field of syntax does what it
does (or why so much status accrues to doing it in certain highly
abstract ways.) So if someone else sends you a coherent explanation
of why the three-valued "hack" you describe is not good syntax, please share
it! (probably it has something to do with mere description not being
the same thing as explanation.)
--Marion Kee
>From IN%""
Subj Different Proof
Dear Mr. Harris:
I have read the opinion in the "LINGUIST List" on 'I consider him a friend'.
As a Korean Linguist, I can add many data on the three argument thesis
in which they are hierachically structured.
In Korean, we have minimal pairs such as; I consider him a friend, I consider
him [__ is a friend], and I consider __ [he is a friend] with different senses.
I have called this sort of verbs a verb of cognition/evaluation which logically
requiries three arguments at least such as for experiencer, theme, and oblique
theta(among others, goal). The sentence "I consider him a friend" should be
analysed as "I consider [him] [a friend]" in which the brackets are theme and
goal. The theta of goal in other examples above are realized in CP(sentence)
rather than NP(noun) but deserve the same theta of goal. It is normal to say
Noun Phrase is converted with ease to CP(full sentence) so that we hold that
the theta of goal can be Noun or Sentence but with a slight difference of
sense(intensional meaning).

In Korean, the null theme argument as in "I consider __ [he is a friend}" is
assumed to refer to the situation at hand both the speaker and hearer share
in common which we call "the situation null theme." This case is contrasted
with the overt theme case as in "I consider [him] [....]" in the sense that
the former implies nothing of the evaluation of the state of the theme
argument at all but the letter something of negative aginst the state of
goal argument. So, the two cases with "him" imply that others (sometimes
including the speaker as well) know the man refered is not genuine friend
(for instance, the enemy state or other very afar-relation state) at all and
the speaker tries to regard the different mental state of making a
friendhood or friendship for him.

So far, I explain in brief fashion that Korean language has full three
argument structure for evaluation verbs with the theta structure of experiencer
, theme, and goal. The theme argument realized in two modes with overt form
or covert form both of which are contrasted in their intensional meanings.
And the goal argument is realized in sentence or noun but I did not add
any explanation of the difference between them yet because some complicated
issues are involved here. I can say something on the ground of Korean data
that the noun type is higher than the sentence type in the sense that a sentence
has to add some affix to be a noun type.

Due to my ignorance of the world trends, I could not see any three argument
idea for evaluation verbs yet(probably English parameter shows only the over
t theme and this is why they tend to analyse it "a small clause") and they
never explain why sentences are changed to nouns. This should be treated
fairly through the global eyes. I will send Korean real data if you are
interested in.

Let me add my request. I am now looking for a host who sponsor me a visiting
professor in the States. If you have a valuable information, please let me
know it.

Best regards,
 Sincerely, Jee-hong Kim

Jee-hong Kim (Ph.D.)
Chairman, Assosiate Professor
Department of Korean Language Education
Gyeong-sang National University
Jin-ju, 660-701
Republic of Korea(South Korea)
Phone #: 0591-751-5586(office)
 0591-751-5578(linguistic main office)
Fax. #: 0591-751-6117(public use)
E-mail address:
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue