Facing facts about 'information' in psychology

Abstract

The coherence of theorizing in terms of ʻinformationʼ is generally taken for granted in the cognitive sciences. Despite widespread employment of ʻinformationʼ as an apparently technical concept, no one knows what it means. Historical and conceptual investigations reveal that it is meaningless. A crucial distinction is drawn between ʻsignalsʼ and ʻmessagesʼ, respectively, with regards to the role of these concepts in Claude Shannonʼs (1948) information theory. Sharpening this woefully neglected distinction reveals the vastness of the gulf separating two bodies of work that are historically associated with the phrase ʻinformation theory.ʼ On the one hand, there is Shannonʼs (1948) statistical theory of signal transmission; on the other, there are the incoherent, ʻinformationʼ-laden speculations of Norbert Wiener (1950) and Warren Weaver (1949).\n \nIt is demonstrated that certain of pioneering information-processing psychologist George Millerʼs (1951, 1953, 1956) most celebrated works involve gross distortions of Shannonʼs concepts, and conflation of Shannonʼs (1948) mathematics with Wienerʼs (1950) and Weaverʼs (1949) respective metaphysics. The widespread misperception that information-processing psychology is substantively related to Shannonʼs (1948) information theory is demonstrated to be false. Subsequently, the suggestion that it is reasonable for psychologists to employ ʻinformationʼ and/or ʻinformation-processingʼ as core concepts without defining these terms of art is sharply criticized.\n \nHowever, throughout these critical investigations, care is taken to distinguish between conceptual clarity and empirical knowledge. It is demonstrated that information- processing psychologists can and have made important contributions to scientific knowledge despite the incoherence that is endemic in their metatheory. On the assumption that a helping hand is preferable to a pointed finger, these critical investigations constitute a therapy rather than an indictment.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 0 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.