Coherence

[Holger Schauer]

>>>>"BM" == Bill Mann schrieb am Thu, 16 Dec 1999 12:07:46 -0500:

 BM> Coherence has to do with an impression of wholeness.  It is
 BM> expected of texts, but not text fragments.

Huh? This comes as a suprise to me. What about the following two
sentences:

1) Fred got up early.
2) The sun was shining in Paris.

Depending on the context in which these two sentences occur I can
clearly make a claim whether this fragment is coherent or not. In case
that the fragment stems from a larger text it is likely that it is
indeed coherent:

3) So Fred went to the next cafe to take his breakfast sitting in the
sun.

 BM> Coherent texts without any of the well known cohesive devices can
 BM> be constructed, but they are extremely hard to find.

I can imaine that for a smaller text, but do you have any references
for a larger such text ?

 BM> RST analysis uses cohesive links as evidence, but RST has little
 BM> or nothing to say about cohesion itself.  It is
 BM> "pre-realizational," whereas cohesion is not.

I strongly believe that, despite its great success, its (i.e. RSTs)
greatest weakness is that is says too little about the relationship
between cohesion and coherence: it leaves a lot open to speculation
about the intention of the writer (and based on the pre-realizational
nature of the structure it adresses, it does so rightfully) and gives
thereby not much guidance to a person analysing texts.

Holger

--
Holger Schauer                         CLIF - Computational Linguistic Lab
                                       Freiburg University, Germany