LINGUIST List 4.209

Sun 21 Mar 1993

Disc: Pro-drop

Editor for this issue: <>

Directory

1. Mike Maxwell 6369, Pro-drop
2. , initial material drop
3. H.Stephen Straight, Phonological PRO-DROP

Re the on-going pro-drop issue, Douglas Purl <dcpselway.umt.edu>
writes:

> It would help in this discussion if someone would lay out
> rigorous ground rules. For example, pro-drop of objects has
> been mentioned lately... Now I can define an object as the
> necessary condition of a transitive verb... Or, I could define
> objects as merely sufficient conditions for transitive verbs,
> and posit a complex set of rules for determining transitivity
> with suppressed objects... If an object (or subject, etc.) is
> the necessary condition for a certain feature (say a verb), then
> there can be no pro-drop, can there?

Good question. If only I had a good answer... I guess this is an
issue that keeps coming up in this discussion; A says "B's theory
of pro-drop doesn't work because of language X", and B says
"language X doesn't have pro-drop, it has something else."

Let me toss out a sufficient (but not necessary) condition to
decide whether a verb is transitive in a language which allows
null pronominalization in object position. Many of the languages
which I would judge as being object pro-drop, e.g. Shuar (a
Jivaroan language of Ecuador) have object agreement marking on the
verb (although usually third person is unmarked). These are
clearly affixes, not clitics (e.g. they go inside other things
which are pretty clearly affixes). (One might argue about whether
object clitics in Romance languages constitute "real" pronouns,
but I'll avoid that issue.) So if a verb can take object
agreement, it's transitive, otherwise not. (This assumes we can
distinguish affected objects or other pseudo-objects.)

Alternatively, consider a language for which no verb requires
explicit object pronouns. On the assumption that a language must
have at least some transitive verbs, such a language must be
object pro-drop, even though we might argue as to whether some
particular verb is transitive. (However, someone might argue that
such a language has a productive lexical rule of
detransitivization--in fact, that might be one account of object
pro-drop.)

Mike Maxwell maxwella1.jaars.sil.org



--Boundary (ID fDcHLNvUfI72uXnkdUWqOQ)--

The analysis of alleged English pro-drop as the omission of pragmatically
recoverbale initial material receives support from German, where elements
in the forefield (before the finite verb in complementizerless declarative
clauses) may be omitted when pragmatically recoverable, independently of whether
they are subjects, objects, or obliques:

Koennte interessant sein.
'(That) could be interesting.'

Hab ich gehoert. (< Das hab ich gehoert.)
'(That) I have heard.'

Bin ich gegen. (< Da bin ich gegen. = Dagegen bin ich.)
'(That) I'm against.'

Omission in non-initial position is never possible:

*Ich hab gehoert.
*Ich bin gegen.

My feeling is that initial objects and obliques are even easier to
omit than initial subjects, but one would have to check this systematically.
I think such cases have been treated in the theoretical literature
in the last ten years, but I don't know where.

Martin Haspelmath, Free University of Berlin

The combination of Richard Hudson's, Richard Ogden's, and Leo Connolly's
contributions on alleged instances of pro-drop in English all bring to a head
the problem lurking in any deletion-dependent account of language. Even if a
strictly phonological account of English pro-drop as apocope or truncation or
non-production depends critically on the claim that something is present for
the speaker (covertly) and presumably recovered by the listener (covertly)
during the process of speech emission and reception that occurs between them.
So, if the production of the initial pronoun (and auxiliary) is absent (whether
because it was not present to begin with or because it was deleted before it
got to Broca's area or because Broca's area can be inhibited--presumably by a
constantly premonitoring Wernicke's area--to result in non-production of
non-essential beginning words), then the evidence for the "missing" stuff might
come from two different sources. First, there's the evidence of language
production: careful analysis of speech data might reveal barely perceptible or
even imperceptible but physically measurable vestiges of the missing material.
Second, there's the evidence of language reception: listener's might treat the
truncated message identically in every respect to it non-truncated counterpart,
even to the extent of believing that the missing material had been present, but
at least to the extent of strongly inferring its "existence". In either case,
however, the phonological "explaining away" of alleged instances of pro-drop in
English depends upon a process of interpretation by listeners that must in many
respects parallel what someone who listens to a "true" pro-drop language must
do. So maybe English is pro-drop for listeners but not for speakers?

H. Stephen Straight, Anthropology and Linguistics, Binghamton University
 E-mail: <sstraighbingvaxa> or <sstraighbingvaxa.cc.binghamton.edu>
 Voice: 607-777-2824 FAX: 607-777-2477 (Anthro Dept), 607-777-2889 (LxC)