Re: valency question

At 03:27 PM Thursday 5/10/01, you wrote:
>How do natlangs and conlangs handle making distinctions between sentences like
>these?  English uses word order (with intonation differences too, I think),
>but I wonder about other ways (are there any?).
>
>      John called the beaver ugly.
>      John called the ugly beaver.
>
>      Beth named José Gaia.
>      Beth named Gaia José.
>
>It doesn't seem to be a regular ditransitive... The problem is the valency
>change, from one object ("the ugly beaver") to two equivalent ones ("the
>beaver", "ugly").
The Latin lingo for this is "double accusative" (although in Latin the
verbs that take a double accusative are different than in English). "John
called the ugly beaver" is not a double accusative, but just a single
object modified by an adjective. In your second pair, both are double
accusatives.

Besides word order, the sense can be conveyed by prepositions or by a
difference in case between the two objects (as Latin uses the dative for
verbs of giving--a conlang could have a case for the second object of verbs
of naming).

The main thing is that you recognize you have a verb with two different
types of object, and you need a way to distinguish which is which. It's
really just a coincidence in your beaver example that one of the objects is
an adjective in a substantive role, rather than a noun, leading to a
different possibility for ambiguity.

>Also a different kind of sentence:
>
>      Mary thought about a cat eating.
>      Mary thought about eating a cat.
>
>This seems to be similar but with whether the verb "thought about eating"
>calls the noun "a cat", or whether it's the verb "thought about" calling the
>phrase "a cat eating".  [Are there two kinds of -ing verb here?]
My analysis: In the first sentence, "eating" is a verbal adjective
(gerundive in Latin parlance), modifying "cat" which is the object of
"thought about". In the second, "eating" is a verbal noun (gerund), which
itself is the object of "thought about"; "cat" is in turn the object of the
gerund. Verbal nouns and verbal adjectives generally have different forms;
in English they happen to be the same.


>Possibly:
>
>      Belmo considered all fireflies demons.
>      Belmo considered demons all fireflies.
>
>...although I think this is just copula dropping ("considered all fireflies
>[to be] demons") it might still be a problem in a language that does it like
>that?
I think these are double accusatives again; I take "consider" to be
semantically similar to your earlier examples "name" and "call". I would
say that the forms with "to be" are alternative formulations, using an
infinitive phrase rather than a noun for the second object.

Strangely, I can't seem to come up with a Latin equivalent of English "call
x y", meaning there is someone or something x, which one refers to or
addresses as y.

In considering what to do with this sort of thing in a conlang, my
inclination would be to group verbs that you would like to take two objects
into semantic classes, and decide how your conlang would treat each class,
either by a word-order convention, a case or prepositional convention, or
by recasting the idea (as by using clauses instead of objects).

Cheers, Tom


~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Tom Tadfor Little            tom@telp.com
Santa Fe, New Mexico (USA)
Telperion Productions        www.telp.com
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Reply