Xref: utzoo comp.ai:3133 talk.philosophy.misc:1863 sci.lang:3966 Path: utzoo!attcan!uunet!lll-winken!ncis.llnl.gov!ncis!helios.ee.lbl.gov!pasteur!agate!saturn!ucscb.UCSC.EDU!hermit From: hermit@ucscb.UCSC.EDU (William R. Ward) Newsgroups: comp.ai,talk.philosophy.misc,sci.lang Subject: Biological classifications and AI (was Re: Categorization) Summary: Quantitative analysis insufficient for qualitative attributes Keywords: qualitative vs. quantitative analysys, pattern matching Message-ID: <6017@saturn.ucsc.edu> Date: 19 Jan 89 03:05:57 GMT References: <681@cogsci.ucsd.EDU> <2959@uhccux.uhcc.hawaii.edu> <684@cogsci.ucsd.EDU> <965@husc6.harvard.edu> Sender: usenet@saturn.ucsc.edu Reply-To: hermit@ucscb.UCSC.EDU (William R. Ward) Organization: University of California, Santa Cruz; CATS Lines: 23 To be brief: The difference between a penguin and a robin, or for that matter, a human being, is a qualitative one. The robin is "more of a bird" by some people's definitions, and conceivably the penguin could be classified along with humans in that both walk on two legs on land and cannot fly (barring such tricks such as jumping, hanggliding, or the Space Shuttle). Using quantitative definitions for things such as "bird," "mammal," et al., is convenient for quick and easy reference, but does not describe very well things classified into said definitions. Not being an expert on AI systems, I cannot hazard a guess toward an algorithm for processing the image of a penguin. I can only project my ideas about biological classifications. I would like to say, however, that any form of pattern matching system should not deal with details such as length of wings as opposed to length of legs, or whatever, but instead with the image as a whole, comparing it to previously stored images and imagining changes needed to change the perceived image into one it knows already. -- If creative thoughts could cause sparks, dynamite could safely be stored in this establishment. William R. Ward