Re: Mari Broman Olsen (nee Mari Broman)

Jonathan Robie (
Wed, 11 Dec 1996 17:09:24 -0500

Hi, Don,

I'm going to try to keep this short. We agree on a bunch of things:

1. A good theory has to falsifiable
2. To be accepted, a theory must account for the phenomena
3. In Greek grammar, the Greek corpus is the most important phenomenon to
4. Mari's theory is still quite new, and in need of more testing

We seem to have been using the phrase "cut the Gordian knot" differently.
When I suggest that slashing a Gordian knot is a good thing to do, I mean
that formulating a clean theory which can be tested in finite time using
available tools is useful whenever possible. For me, the real "Gordian knot"
is the approach which says that I have to be thoroughly familiar with the
entire Greek corpus before I can know anything at all about Greek grammar -
people like Bauer or Blass or Robertson (and I suspect Carl Conrad)
certainly have this kind of familiarity with the Greek corpus, and I would
be very careful about ideas which run in the face of what they have to say.
Now that we have the ability to propose precise linguistic theories and test
them by searching the corpus, let's do it!

I don't care if the results favor traditional accounts like Blass/Debrunner
or Robertson, or more modern accounts like Mari's. The point is to go at it
systematically, and to be clear about how we formulate theories and how we
prove them. In fact, this is fundamental to scientific inquiry: our job is
to formulate theories and to prove them or disprove them. Who wins and who
loses in the process is irrelevant, in the long run we all win.

Incidentally, your statements on context make me suspect that we are talking
past each other on the role of context. Mari's use of context is *very*
different form Porter's - she uses words like NUN, which establish time
independently of the verb tense, and sees how they combine with the tense or
aspect ideas conveyed by the forms of the verbs. If the context can cancel a
time sense associated with a verb form, then that time sense is not inherent
to the meaning of the verb.

Let's take an example:

EZHTOUN SE LIQASSI "they were trying to kill you"
Time reference: past time
Aspect: imperfective

NUN + EZHTOUN SE LIQASSI "just now they were trying to kill you"

In this case, EZHTOUN retains the past time reference, which means that it
is not cancelled by NUN. I think this is a good example of how the tense of
the imperfect interacts with the context, and it tells us something about
the meaning of the imperfect. In the wider context of John 11:8, I think it
is clear that this verse *must* mean what I have said above. Suppose we come
up with a couple of examples like this, do a Gramcord search, and come up
with only with examples that must be interpreted this way and examples that
might easily be interpreted this way. We would be justified in assuming that
it probably works this way. Of course, someone with access to TLG could
enlighten us by doing a wider search, and someone like Mari can provided
evidence from other languages like English, Russian, Mandarin, Kukuyu, and

Russian, incidentally, is a very interesting language for investigation of
tense and aspect because it uses different morphemes to encode each!


Jonathan Robie
POET Software, 3207 Gibson Road, Durham, N.C., 27703
Ph: 919.598.5728 Fax: 919.598.6728
email:, <--- shockwave enabled!