LINGUIST List 4.161

Mon 08 Mar 1993

Disc: Pro-Drop

Editor for this issue: <>

Directory

1. , Pro-drop and dative subjects
2. , Subject: 4.160 Pro-drop
3. , German Datives
4. , RE: 4.160 Pro-drop

Heidi <hharleyAthena.MIT.EDU> speculates on a correlation between
dative subjects and pro-drop. Unfortunately, here's a counterexample:
Lezgian, a Nakho-Daghestanian language spoken in the eastern Caucasus,
has dative subjects and is not pro-drop (cf. the detailed description
in my grammar of the language: Haspelmath, Martin. 1993. A grammar of
Lezgian. Berlin: Mouton de Gruyter.).
 The dative subject facts are very robust:The experiencer of verbs
like 'see', 'find', 'fear' is in the dative case but has all behavioral
subject properties. Personal pronouns are normally used and are obligatory
in careful (especially written) style, but they may be omitted in colloquial
style if the referent is clear (note that there is no agreement in Lezgian).
 This one counterexample does not preclude the possibility that there is
a statistical correlation between dative subjects and pro-drop, and in view
of the following considerations I would not be surprised if it turned out
that there is one: Dative subjects are a feature of the larger type of
"role-dominated" languages, as opposed to "reference-dominated" languages
like English (this distinction was developed by Robert Van Valin and
William Foley and was elaborated by Aleksandr Kibrik). Role-dominated
languages tend to code semantic roles, whereas reference-dominated
languages tend to code pragmatic roles such as topic. Role-dominated
languages prefer morphological expression, whereas reference-dominated
languages prefer word order as a means of expression. Thus, I expect that
role-dominated languages in general have richer morphology, including
richer verbal morphology, and subject-verb agreement is correlated with
pro-drop -- so dative subjects should be correlated with pro-drop.
 Lezgian and other Daghestanian languages do show rich morphology
(Lezgian has 16 cases), including rich verb morphology, but they happen not
to have person agreement (other Daghestanian languages have extensive
gender-number agreement, which Lezgian lost quite recently).
 So even though I had to mention this counterexample, I think there is
something to the correlation.

Martin Haspelmath (Free University of Berlin)

Joe Stemberger makes the point that according to Lightfoot all
that matters to a learner trying to decide if a language is
pro-drop or not is whether it doesn't or does have expletive
subjects. Now, does this mean that "pro-drop" is simply defined
to mean "not having expletive subjects", or is there an independent
definition and then we have a factual claim that languages which
have the one property also have the other?

Aside from just wanting to understand what the discussion is
about, I ask this because there is a Polish dialect which
has expletive subjects but which, as far as I know, does not
differ markedly from other Polish dialects on the point of
presence vs. absence of explicit subjects (which is what
I thought pro-drop was supposed to be all about). Now,
if I can have a clearer understanding of what the claims
are, I would be prepared to check the crucial facts and
report back (but first we have to know what IF ANY facts
are crucial, for if the whole thing is circular then
no facts will be crucial).

Mike Maxwell writes:

<< Finally, (contra Jon Aske) I don't see any conflict in principle
<< between the notion that English is [-pro-drop] in some genuine
<< sense, and the notion that English is incipiently pro-drop. In
<< theory, one generation of speakers might allow these pro-drop-like
<< constructions for some reason that has nothing to do with a
<< "pro-drop" parameter, while the next generation restructured their
<< internalized grammar to make English [+pro-drop] (with
<< concomitant, but perhaps minor, changes to the set of sentences
<< admitted). (The current grammar might include some sort of
<< inter-sentential coordination construction, for instance.)

This makes me wonder whether anybody knows of any real examples
of language change taking place in this fashion, which, apparently,
has been hypothesized many times (as I recall, Morris Halle used to believe
that this is how phonological change takes place, for example, though
I do not know whether he still does).

As for me, although I know of cases of reanalysis, they are never
very drastic and, more importantly perhaps, they do not neatly
align with generations.

For those of you who think that German datives cannot control reflexivization,

 Es faellt ihm schwer, sich zu konzentrieren

is perfectly fine.

Message 4: RE: 4.160 Pro-drop

It seems to me that Dick Hudson may be taking the right sort of line on
PRO-drop in English. This accords also with the other suggestion that it might
be a performance matter. This seems to me to be of importance.

If the distinction between competence and performance is upheld, then those
that support the absoluteness of +/-[PRO-drop] can argue that the English data
is a matter of performance, and therefore English is a -[PRO-drop] language.
The fact that it is only in matrix clauses that it happens would seem to me to
support the performance approach.

It is also an invalid argument to say that PRO-drop is operating in English
because sentences like

 Can't remember any of the previous examples.

are reasonably common, in that, as many have pointed out, sentences where the
AUX is also dropped are common too

 Forgotten what I was going to say.

To say that the first is PRO-drop is to say that another rule entirely is
operating in the second case. And that seems to me to be counter intuitive to
say the least.

Two other observations.

I am not an expert on German, Icelandic or any of the other language where the
question of "dative subjects" has been raised. Perhaps Halliday is right in
dividing up "grammatical subject", "logical subject" and "thematic subject" and
the discussion is actually ranging over different definitions of subject at
this point. Thus we need to make sure that not only do we all agree, as
Manaster-Ramer suggests, on a definition of PRO-drop, but also on one for
"subject".

It seemed to me that the suggestion is implicit that PRO-dropping inherently
revolves around the possibility of identifying subject person (and gender?)
through inflexion or some other device elsewhere in the string, typically on
the verb. If my memory serves me correctly, Chinese, with no relevant
inflexional characteristics at all, only overtly expresses the subject,
of necessity, where the subject changes. Am I right?

Mark Hilton
School of Languages
University of Westminster.