One sense per collocation--AGREE?

[xujiajin]

http://acl.ldc.upenn.edu/H/H93/H93-1052.pdf
One sense per collocation
Do you agree?

One sense per collocation and genre/topic varations
http://acl.ldc.upenn.edu/W/W00/W00-1326.pdf

 

[xusun575]

Very enlightening research papers which seek to find the semantic reqularities for the purpose of eventually/potentially setting up models for computer aided understanding or interpreting natural language, but pls our dear colleagues in the linguistic field look at what we have done so far in our language rearch for so many years where we could only follow the others' reluctantly abandoned shoes to do some seemingly significant studies mostly by using the SFG grammar as a main stream of linguistic study in China which will, I shall boldly predict, turn out in a few years later to be self entertained three ring circus. There are different schools of FG and why should we use Halliday's only as one and single theoretical frame in our thesis work? Cogling will be spreading across China but my concern is what our Chinese main stream linguists have done or can do to contribute to computer aided NLP, especially the task related to our mother tongue?

 

[刘语料]

According to Sinclair's study and the COBUILD project, there is strong association between meaning and collocation of a word.But one sense per collocation or
one sense per collocation and genre/topic varations may be arbitary. Obviously, one sense may have several collocations.