[B-Greek] 2nd year NTG on my own / followup

James Tauber jtauber at jtauber.com
Mon Nov 5 21:38:06 EST 2007


On Nov 5, 2007, at 9:27 AM, Eric Inman wrote:

> I'm not surprised to see that you've already considered this.  
> Similarly, I
> think affixes such as -IZW, -MA, etc. would also need to be  
> addressed before
> the analysis could be considered very accurate. I think readers find  
> them
> very helpful when initially guessing the meanings of words. In a  
> different
> language (Indonesian) I found that even when speaking it was  
> possible guess
> not only the meanings but even the existence of words I didn't know  
> yet by
> forming them with roots and a system of affixes.

Yes, the starting point is an analysis of the derivational morphology  
of each lexeme in the NT corpus. I'm working on that now.

> While ideally it would be nice to have the database loaded with  
> information
> for each word regarding whether or not its meaning can be inferred  
> from the
> meanings of its parts, I wonder if it might be sufficient just to  
> sum all
> that up in one or a small number of numerical coefficients, derived  
> through
> appropriate sampling. I don't know if that's what you were  
> indicating in
> your comment below about an empirical measure.

Yes, basically some measure of how likely it is that someone could  
guess the meaning from the components. Or more formally: the  
compositionality of semantics in derivational morphology. To be  
honest, it's not my area of expertise (my PhD is on inflectional  
morphology, not lexical semantics) but having some numerical rating in  
my database would be very useful for pedagogical reasons. It could  
even just be a three-valued field: non-compositional, weakly  
compositional, strongly compositional.

James





More information about the B-Greek mailing list