And:
>I'm just trying to get my head around why the 'morphologicon'
## morphicon?
>contains
>only roots and affixes or why it contains roots that are also stems. On one
>level the answer is obvious: the morphologicon contains all morphemes
>and nothing but morphemes. But I'm still unclear about the underlying
>rationale.
## I suppose the main rationale is that we clearly have a list of forms
which we can recycle and recognise across words. Another reason is that I
found statistical evidence for it in my paper "Inherent variability and
linguistic theory" where the figures for t/d deletion in words like "mist"
required three opportunities for it, and the morphicon provides the first
of three:
morpheme {mist}
stem /mist/
whole /mist/
(I think that's how the analysis went.)
>
>Do you still think that all morphemes are nothing more than phonological
>strings? Or do they have semantic and combinatoric properties?
## In a network analysis every morpheme is directly linked to the words
that contain it, which have semantics, word classes, etc. But what I am
claiming is that these connections are only indirect.
Richard (= Dick) Hudson
Phonetics and Linguistics, University College London,
Gower Street, London WC1E 6BT.
+44(0)20 7679 3152; fax +44(0)20 7383 4108;
http://www.phon.ucl.ac.uk/home/dick/home.htm
|