July 10, 2011

Semantic theories

This is one for the linguists... A naive question I'd like to put out there:

I really enjoy semantics: thinking and writing about word meanings, cross-linguistic translatability and all the other juicy stuff you get when you really try and pin down meanings of words, morphemes, phrases etc. which are usually tantalisingly slippery and have lovely fuzzy boundaries.


When it comes to semantic theories, I just haven't come across one that I've liked. I've looked a little bit at NSM and am now learning a bit about frame semantics, but theories like this bother me. They all seem to have quite a lot of limitations, so much so that I wonder how useful they really are. Why can't we just write detailed semantic descriptions of words/morphemes/phrases etc. and leave it at that?

This really is an issue I'm grappling with because my PhD thesis will hopefully be a lot about semantics, but I have this chip on my shoulder about semantic theories.



Lauren said...

I tutored an undergraduate semantics subject last year and had a great time introducing students to a variety of semantic theories. The course included NSM, binary feature, componential, cognitive semantics and things like prototype theory - not to mention an introduction to issues in dictionaries.

I really liked showing them a little bit form each area, showing them the upsides and pit-falls and then moving on. For me, the important thing is that different semantic theories are like different tools in your tool kit. NSM is great for emotions, prototype for concrete entities, BFA for kin terms or pronouns and cognitive stuff for metaphors. No one theory is going to solve everything, and those that try and preach that theirs' can often fall over at some point.

Try implementing a few and see what works best for you and your problems!

Jangari said...

I used a bit of watered-down Jackendovian semantics for my honours thesis, which I found very useful as a tool to link semantic and morphological/syntactic levels of grammar. It was nowhere near as formal, complex or powerful as the syllogistic logic-based semantics that's normally practised in the US though.

As for NSM, I second Lauren; good for giving explications of emotions, speech acts, and so forth, but useless as a semantics-syntax interface.

nginarra said...

Thanks for the comments.

I think my issue is that I'm not convinced that semantic theories are "tools". They feel more akin to "games" to me. :-(