Our PCT discussions (including Bill Powers’ writings) refer to several kinds of data. For example, here are just three:
- Quantitative measurements of a variable that is perceived to be the CV, of behavioral outputs affecting it, and of environmental disturbances affecting it.
- Pairwise rankings. Examples: the substitution tests of linguistics; QSort assessments; the cognitive preference assessments of Ned Herrmanns.
- Statistical measures, insofar as they help to identify variables controlled by individuals in a population, subject to verification.
What are the prerequisites for something to be accepted as data?
You could say this about science in general—that every science changes as people gain and lose the upper hand in arguments. Every science does change, of course, but despite a popularized superstar gladiator narrative gaining and losing the upper hand in arguments is an impoverished description of the process. In science, argument is answerable to observational data. No disagreement here about that. No two sciences have the same data. No two subfields of a science have the same data. PCT has subfields.
How to represent the data of interactions, social relations, and social arrangements in a way that is amenable to quantitative PCT modeling is not a simple and straightforward matter. For example, Stephen Johnson’s graph-theoretic approach to language shows promise, but such data are very different from those for motor control in tracking and pursuit tasks. It would be impracticable to insist that a PCT simulation model control down the hierarchy all the way to the motor outputs of speech and writing and the auditory and visual inputs for hearing and reading language. One approach might be that modeling at higher levels must encapsulate control below some level as though controlled in imagination.