Re: Interaction: middle and imperfect

From: Mari Broman Olsen (
Date: Fri May 30 1997 - 09:52:39 EDT

Jonathan writes:
> OK, I decided to check out the present of ERCOMAI, which also appears only
> in the middle. Is this right - with imperfective aspect, ERCOMAI occurs only
> in the middle, but with perfective aspect it never occurs in the middle? Are
> there other verbs that act this way? What does it all mean?
> >I've never thought of interactions between voice and mood,
> >but I just ran into an interesting fact: in the imperfect,
> >ERCOMAI always occurs in the middle (11 times); in the aorist,
> >it never occurs in the middle (168 times, all active). Being
> >a naturally speculative type, I wonder if this could have to
> >do with the force of the middle and the aspect of the imperfect
> >both reinforcing a more vivid depiction of the actual event of coming?
> >
> >I hadn't expected this kind of interaction between tense and mood. Does
> this kind of pattern occur with other verbs, too? Has anybody written
> anything useful about it?

Interesting! That correlation would be expected (and not, say, middle
and perfective), under the theory of Transitivity presented by Hopper
and Thompson (in Language 1980), in which properties of HIGH
transitivity (including active voice, perfective aspect, definite
direct objects) covary together cross-linguistically, as do properties
of LOW transitivity (passsive voice, imperfective aspect and
indefinites, e.g.). For a paper applying theories of this type (and a
list of Hopper and Thompson's properties) see my web page (Olsen and
Resnik paper). As pointed out for the middle in Greek

Incidentally, in the literature on romance languages that I am
currently reading, the "middle" is treated not as between active and
passive, but as between transitive and intransitive. So there is
syntactic-semantic support for a relation potentiating Carl's proposed
development and change among other constructions in the space between
transitive and intransitive, i.e. between passive (agent argument may
be left unexpressed), middle (agent-patient tend to be the same) and
reflexive (agent-patient tend to be the same, but different 'parts'
can be distinguished, as in John('s hand) shaved himself (his face).).
Other constructions in the 'middle' space include cognate objects
(fight the good fight) and implicit objects (I ate [a meal]). It is
the latter, primarily, that are discussed in our paper. (Comments and
Greek implicit object data especially welcomed!)


This archive was generated by hypermail 2.1.4 : Sat Apr 20 2002 - 15:38:17 EDT