adjective typology

[Timo Honkela]

Dear Brian,

We have written a short paper on "Measuring adjective spaces" for which we compared 3-4 classification frameworks with lexica and found them to be mutually incompatible to such a degree that made us to conclude that "[t]here does not seem to exist a consensus among linguists on how adjectives should be divided into categories in general." To have some ground truth for our experiments, we simply compared how well the different methods were able to find synonyms and antonyms. The paper is available at

http://users.ics.aalto.fi/tho/info/honkela2010measuring.shtml

Regrettably I don't have information on the classification frameworks that we studied available here and now. One potentially interesting reference is "Adjective Classes: A Cross-Linguistic Typology" (Dixon & Aikhenvald, eds.).

Adjectives are an exciting topic of research for which many classical theoretical frameworks are necessarily not fully suitable. One potentially useful direction could be to check if researchers on fuzzy sets have produced something systematic regarding adjectives. There is a quite early paper in which we modeled fuzziness of size adjectives:

http://users.ics.aalto.fi/tho/info/honkela91a.shtml

Adjectives are also a central topic in a growing number of research projects related to sentiment analysis.

Best regards,
Timo

P.S. We have recently come into the conclusion that good statistical or probabilistic methods are able to judge semantic similarities based on contextual data at a level that is comparable to a human-level performance. In other words, the differences between human judgements are so widespread that it is not necessary sensible to consider human classifications as a ground truth. This topic is discussed in our recent short paper "Exploratory Text Analysis: Data-Driven versus Human Semantic Similarity Judgments". A journal paper with broader set of experiments is under preparation.

http://link.springer.com/chapter/10.1007%2F978-3-642-37213-1_44


________________________________________
Lähettäjä: funknet-bounces at mailman.rice.edu [funknet-bounces at mailman.rice.edu] käyttäjän Tom Bartlett [BartlettT at cardiff.ac.uk] puolesta
Lähetetty: 2. heinäkuuta 2013 21:46
Vastaanottaja: Brian MacWhinney
Kopio: Funknet
Aihe: Re: [FUNKNET] adjective typology

I don't know if you're looking for a list or a classificatory framework; if the latter, you might try Gordon Tucker's The Lexicogrammar of adjectives.

All the best,

Tom Bartlett.

CLCR
Cardiff.



-----funknet-bounces at mailman.rice.edu wrote: -----
To: Funknet <funknet at mailman.rice.edu>
From: Brian MacWhinney
Sent by: funknet-bounces at mailman.rice.edu
Date: 02/07/2013 18:09
Subject: [FUNKNET] adjective typology

Dear Funknet,
      Has anyone ever produced an adjective classification along the lines of the Levin classification system for verbs?  In other words, one that breaks adjectives into syntactic/semantic groups. Something in digital format would be best, but printed would be okay too.
Many thanks for any suggestions on this.  If anyone locates anything, I will post the result to the list.

--Brian MacWhinney (macw at cmu.edu)