Schyns Comms 04-06 burgund charles cleeremans

From: Willoughby Sarah (sew295@soton.ac.uk)
Date: Tue Mar 03 1998 - 18:15:35 GMT


Charles suggests that,

> "definition of cateogories is a seperate issue from recognition of
> cateogory members"

so we need not try to provide an approach that deals with both issues.
However, it is hard to see how the two can be seperated when they are
so fundamentally linked. Surely the development of categories is
determined by features of the category members. Having said this,
category members are not always obvious and clear cut. The boundaries
of peoples categories vary according to their experiences of labelling.
These boundaries may just be cognitive and not sensory. This would
appear to back up the flexible feature theory. Fixed features would not
allow for such adaption.

> "attempts to specify what the featural primitives of recognition might
> be, or to constrain the features. have been singularly unsuccessful"

and yet Cariani says that classifying objects involves feature
selection and feature mapping. The fact the identification of
primitives has been so unsuccessful raises doubt over the theory that
'feature selection is the task of finding a set of feature-primitives'
to match objects.

> "A distinction needs to be drawn between the definition of a category,
> the defining features, and the features used in recognition of category
> members, what might be called the distinctive or characteristic
> features."

It is hard to see how these features can be distinguished or how it is
beneficial to do so. If a category is not based on the features used to
recognise a category member, then what is it based on? This seems to be
an unclear distinction. It may be true that we form arbitrary category
labels which do not have much to do with the way we categorise. It
makes the concept of categories a seperate storage units questionnable.
How useful is it to categorise artificially if it is not representative
of classes of objects in the real world? It may be useful to us to
enable us to try and describe categories but doesn't necessarily give
us much insight into the way that the brain processes features.

> "What is being learned is not definitions of categories nor sets of
> invariant features, but rather the features that best characterise the
> members of a category and differentiate them from members of
> contrasting categories."

Charles appears to be advocating the prototypical model of
categorisation, a more dimensional approach. This does seem to allow
greater flexibilty.
Could it be that some features may be more flexible because they are less
concrete, while well rehearsed features become more fixed.
Burgund & Chad seem to be suggesting this..

> "alternative materials may elicit feature creation more readily than
> traditional materials because they are more likely to be processed in a
> specific subsystem. In contrast, stimuli that are readily decomposable
> into familiar parts may be processed more effectively in an abstract
> subsystem."

The use of novel items in experiments allows experimenters to look at
how people deal with 'raw' materials and categorise. Cariani allows
for a system that defines categories and creates new ones while using a
decision rule that maps pre-existing features to classes. This is
contrary to Charles who claims that such ideas about features are
restrictive and not explanitory.
It seems that there is no satisfactory explaination at present and the
issues are still open. Cariani believes that the brain forms new
categories for novel stimuli. Burgund and Chad say that it is still
unclear whether fixed or flexible features are used for category
recognition and Charles thinks that trying to tie ourselves down to
defining category features is fruitless.

> "If it is in principle impossible to distinguish between them, perhaps
> it is best to acknowledge that there may not be a set of invariant
> primitive features which underlie visual recogntion."

this dimensional approach seems to be the most plausible at present
since attempts at finding fixed features appears to be so
unsuccessful.



This archive was generated by hypermail 2b30 : Tue Feb 13 2001 - 16:23:20 GMT