# discourse analysis and circular reasoning

Moon-Ryul Jung moon at sogang.ac.kr
Sat Apr 13 21:59:28 EDT 2002

```Rolf warned against the danger of circular reasoning in the method of
discourse analysis. To him, discourse analysis "tries  to confirm  the
very results on which your method build".

I would like to make some comments.

(1) We need to distinguish "circular reasoning" and "spiral reasoning".
If you travel on the circle, you cannot move forward. But if you
travel on the spiral, you make circular motion, but still move
forward. I think discourse analysis which for example Niccacci does
is more like spiral reasoning than circular reasoning. It does start
with the assumption that different forms would have
different functions. He distinguishes WAYYIQTOL, WEQATAL, WEYIQTOL,
QATAL, YIQTOL, and QOTEL. It is the safest assumption. I would be
happy to go on a spiral travel starting from this assumption. In the
early part of the spiral travel, we would detect the usage of these
forms from examples whose interpretations are clear. In the latter
part of the travel, we would attempt to interpret difficult cases
by using the patterns of usage discovered so far. As long as such an
attempt does not face a total breakdown, it is scientifically
legitimate to continue the trip. One should not criticise this trip
by saying that it builds upon the assumption and sticks to it, but
by saying that the resulting interpretations do not make sense at all.

You worry that the interpretation of a sentence is affected by many
factors, e.g. temporal adverbials, lexical aspects of the
verb,  other than the verb form. That is true. But the interpretation
of such sentences would be attempted in the latter part of the
spiral travel.

(2) Science is the refinement of everyday thinking. Physics Laws are
refined descriptions. For example, the Newton's three laws of motion
are just descriptions of the reality. Before we attempt refining
descriptions, we need to find them. In the case of hebrew, I think
that we need descriptions of the verb forms more than we need to
analyze them based on a few parameters. I have some experience
in computational linguistics, a field of Artificial Intelligence.
All the theories seemed to  lack "descriptive adequacy". Without it
any attempt to get "explanatory adequacy" would turn out to be
a meaningless game.

Moon
Moon R. Jung
Sogang Univ, Seoul, Korea

```