A comment on equating information with symbol strings.

Abstract

Symbol strings are advanced as the informational basis for many biological, physiological, and psychological phenomena. The role ascribed to them is that of indicating or directing states of affairs. Pattee has suggested that nature exploits information in this quasi-linguistic sense sparingly, that symbol strings are limited in detail, and that their relation to dynamics is one of complementation. A different, nonsymbolic view of information that addresses how animals can guide their locomotion in cluttered surroundings has been pursued by Gibson. It has considerable generality: information is low-dimensional qualitative properties of low-energy fields, lawfully generated by properties of systems and surround. It is argued that in the absence of information in Gibson's specificational sense, information in the indicational-injunctional sense is ineffective, and it is suggested that perplexities about the selective content of symbol strings may be resolved by a thoroughgoing understanding of Gibsonian information.