- From: Dylan Evans <evansd66@googlemail.com>
- Date: Wed, 7 May 2008 14:44:36 +0100
- To: "Bill Jarrold" <jarrold@ai.sri.com>
- Cc: "Marc Schroeder" <schroed@dfki.de>, EMOXG-public <public-xg-emotion@w3.org>
Dear Emoters, I think Bill's comments about the range and type of values for intensity and emotion regulation are very important. But I think he overestimates the problems associated with setting an upper bound for a scalar value. There are various scalar measures of happiness and sadness and all of them involve an upper limit (eg. say how happy you are on a measure of 1 to 10 where 10 is the happiest you've ever been). There are problems with such scales, of course (you may be happier now than you have ever been) but on the whole its surprising how valid these measurements turn out to be. They are widely used, and would be easy for researchers to convert to real values in the interval [0,1]. Nevertheless, I still agree with him in principle that it would be great if we didn't limit future researchers to assuming an upper bound. A researcher who measures happiness by the concentration of a certain neurotransmitter will not be able to map those values to the interval [0,1] without some extremely arbitrary assumptions, for example. So, maybe we need five options for specifying the intensity of an emotion: [a] qualitative values [b] qualitative values relative to a particular class [c] a partial ordering of values [d] scalar values with an upper bound [0,1] [e] scalar values with no upper bound [units to be specified by user] Or is this too complex? Dylan -- -------------------------------------------- Dr. Dylan Evans Senior Research Scientist Cork Constraint Computation Centre (4C) University College Cork, Cork, Ireland. Tel: +353-(0)21-4255408 Fax: +353-(0)21-4255424 Email: d.evans@4c.ucc.ie Web: http://4c.ucc.ie http://www.dylan.org.uk --------------------------------------------
Received on Wednesday, 7 May 2008 13:45:15 UTC