Classification: Logician GeneralÕs Warning

Donald Byrd, School of Informatics and School of Music, Indiana University

rev. late October 2010

 

¥       Classification is dangerous to your understanding.

      Classification ordinarily means hierarchic classification

      That means partitioning = subdividing cleanly into categories (& often subcategories). But...

      (1) Almost everything in the real world is messy; very little is well-defined

       (2) Absolute correlations between characteristics are rare

       Example: some mammals lay eggs

       Example: is the piano a keyboard, a string, or a percussion instrument?

       Example: Was the first real piano Cristofori's (ca. 1700), Broadwood's (ca. 1790), or another?

       Example: Cream member Ginger Baker says Cream was Ònot really a rock groupÓ!

       Example: Should e-mail from Jean about teaching go in the "Jean" folder or the "Teaching" folder?

¥         People often say Òan X has characteristics A, B, C, DÉÓ

¥         They nearly always mean Òan X has characteristic A [or B or C orÉ], and usually also B, C, DÉÓ

       NB: most fundamental one may not be the one mentioned first

¥         Important special case: partitions (ÒflatÓ classifications, without subcategories)

       Example: ÒAre charter schools havens for innovation, or are they undermining public schools?Ó

       Example: A famous (ex)physicist says arguments against faster-than-light travel are based on supposed partitions that aren't really partitions

¥         Messiness of the real world leads to:

       People who know better claiming absolute correlations

       Arguments among experts over which characteristic is most fundamental

       ÒIs it this or that or that?Ó questions that donÕt have an answer

       Don changing his mind

¥         But lack of classification is also dangerous to your understanding!

¥         So should we abandon hierarchic (or all) classifications?

       Of course not; they're much too useful

       Besides, it's impossible to avoid them; they're built into our ways of thinking

       Best policy is (1) be on guard for misleading classifications; (2) be aware of alternatives

       Alternative worth considering: "extended" hierarchies (e.g., Directed Acyclic Graphs instead of trees)

¥         See Hayakawa, S.I., & Hayakawa, Alan R. (1990). Language in Thought and Action, 5th ed.

 

 


Comments to: donbyrd(at)indiana.edu
Copyright 2006-10, Donald Byrd