The functional organization of self-constructing observer-actors
A framework describing the functional organization of computers, scientific predictive models, and cybernetic percept-action systems is presented (Cariani, P. (2011). The semiotics of cybernetic percept-action systems. International Journal of Signs and Semiotic Systems, 1(1), 1-17.). Cybernetic systems have internal goals that adaptively steer, switch, and/or construct informational functions.
Basic percept-action systems engage in ongoing transactions with their surrounds using several basic informational operations: measurements, computations/coordinations, and actions. Measurements alter internal states contingent on sensor-environment interactions; coordinations/computations reliably map internal states to other internal states; actions activate effectors to create changes in the external environment contingent on internal states.
Scientific models can be described operationally (following Hertz, Bridgman, Bohr, Cassirer, van Fraasen, von Glasersfeld) in terms of measurements and formal predictive computations that assumes no knowledge of any objective reality, (i.e. is compatible with radical constructivist epistemology). Systems-theoretic criteria based on observed state-transition structure (Ashby) can be used to demarcate “epistemic cuts” where measurements end and computations begin.
The semiotics of percept-action systems can be described in terms of the Morrisean triad: measurements determine the external perceptual semantics of internal states, computations the syntactic relations between them, and actions their external action-semantics. Pragmatics is the relation of internal states to system-purposes. Purposive systems have embedded goal states and mechanisms for adjustment of action contingent on sensory inputs, current internal state, and evaluative criteria. Purposive systems have agency to the extent that they can act autonomously to pursue internal goals.
Adaptive percept-actions systems are those that modify their internal structure and operation contingent on the evaluated efficacy of their actions in order to improve actions to better satisfy internal goals. Within this framework, adaptive adjustments can occur on three levels: incremental steering adjustments, switching between existing discrete alternatives (“combinatoric creativity”), and creation of entirely new alternatives (“emergent creativity”)(Cariani, P. (2012). Creating new primitives in minds and machines. In McCormack & D’Inverno (Eds.), Computers and Creativity (pp. 395-430), Springer.).
For steering, incremental adjustments of parameters are made that deform but do not qualitatively change percept-action mappings.
Adaptive switching alters mappings qualitatively, but within the confines of existing perceptual features and action-alternatives. Virtually all trainable machines fall into this category. Adaptive creation of new alternatives can arise through self-construction of new underlying hardware that enables additional, entirely new internal states, measurements (sensors) and actions (effectors) to be created. Biological evolution is filled with the emergence of new senses, new action-alternatives, and neural coordinative possibilities that successively expand the Uexkullian umwelts of lineages. Self-construction of sensors permits epistemic autonomy (creation of new observables). Pask’s electrochemical device is an artificial example of functional emergence of this sort that expanded the dimensionality of its distinction-space (Cariani, P. (1993). To evolve an ear: epistemological implications of Gordon Pask’s electrochemical devices. Systems Research, 10(3), 19-33. Pask, G. (1959). Physical analogues to the growth of a concept. Mechanization of Thought Processes, Vol II. (pp. 765-794). H.M.S.O.). These concepts bear potential implications for construction of new concepts in brains and for constructivist epistemology more generally.
TRADITIONS: EPISTEMOLOGY, NEUROBIOLOGY
- 3) Experimental epistemology; constructivism; philosophy of science
- 1) Computer science; AI; robotics