Representation and Process in Defining Holism Vs Partism: Contributions from General Recognition Theory and Stochastic Cognitive Process Theory

Our target in this brief paper is the issue of holism vs. partism (to coin a term) considered from the points of view of representation, usually thought of as static and either consisting of a feature list or engaging properties of geometry, and process by which we infer properties like dynamism and mechanism. We approach these topics later from the perspectives of general recognition theory (GRT, Ashby & Townsend, 1986) and stochastic cognitive processing theory (e.g., Townsend & Nozawa, 1995). Ultimately, we believe that dynamics, involving how the original percept develops as well as subsequent, and perhaps more cognitive or motor processes operate, must become an integral part of what representation means (e.g., O’Toole, Wenger, & Townsend, 2001; Spencer-Smith et al., 2001). The notion of holism is itself challenging and its definition continues to evolve through discussion and experimental manipulations. A closely related concept is Garner’s integrality. Each is in turn related to a ‘good form’, configurality, or Gestalt, all intimating a unity that goes beyond any mere description of its parts. We shall use the relatively unencumbered name of Whole to refer to the complete unity of a thing. One aspect of Whole is that all aspects should be present. As we see below, “all” may be finite or infinite in nature. The opposite of Whole should entail some listing of parts that are in some sense independent. A substantial impediment arises because “independence” has meant a myriad of things to many investigators. We shall relate our view based on our mathematical theories below. We note in passing that an even more opposite notion to a Whole would be an object with parts that are actually antagonistic to one another. It has been thought for decades that the famous Stroop effect, with literally hundreds of studies to its credit (Stroop, 1935) incorporated mutually inhibitory dimensions, that were highly automatic. However, it now appears that the effect appears only with differential salience of the dimensions, or the existence of certain correlations in the dimensions across stimulus presentations (Algom, Dekel & Pansky, 1996). Anyhow, eventually, objects with antagonistic features or dimensions should be brought into the theoretical fold that encompasses Wholes and Part-sets. All parties who have dealt with the topic of holism have tended to give it their own flavor either through theoretical definition (somewhat rare), operational-experimental definition (more common–see Appendix) or both (occasionally). One other important avenue is to study the structure of good vs. not-so-good vs. bad Gestalts. This is an important question but we are more concerned here, in our ventures with the general aspects of