Perceptual Control Theory as a Framework for Computer Modelling Across the Social Sciences
Dr Warren Mansell, Reader in Psychology, University of Manchester
Perceptual Control Theory is a psychological framework developed from
control system engineering during the 1950s and 60s (Powers et al.,
1960; Powers, 1973). The theory is supported by a range of empirical and
theoretical papers (see www.pctweb.org)
and yet its central premise - 'Behaviour is the control of perception' -
remains dominated by the same assumptions of linear 'Stimulus-Response'
or 'Input-Compute-Output' psychology that primed its development. In
this talk I will describe examples of computer modelling using PCT
across a wide range of domains (e.g. robotics, economics, sociology,
speech and language). I will then describe and demonstrate the PCT
simulator I have developed to model actively learning multi-layered
control systems. One hypothesis we are testing is that psychological
distress (i.e. mental health problems) represents chronic loss of
control due to internal goal conflict, and that this is relieved by
promoting changes in higher level, or deeper, systems. This would
explain why psychotherapies that access 'deeper meanings', 'longer term
goals' and 'broader values' are effective. To date, we have demonstrated
that learning (termed 'reorganisation' in PCT) of higher level systems
allows a more enduring optimisation of control than learning within the
lower order, conflicted, systems. Related research will be discussed, as
well as a discussion of how PCT could be utilised to improve existing
computer models of psychological systems.
Mansell, W. (2011). Control of Perception Should be Operationalised as a
Fundamental Property of the Nervous System. Topics in Cognitive
Science, 3, 257-261.
4pm Neils Bohr Common room, Schuster Building.
Last call for abstracts (Arp 7): Beyond Schelling and Axelrod: Computational Models of Ethnocentrism and Diversity - #cfpm_org Any work with simulation models that touch on these issues, critiques of existing models or relevant empirical work welcome. Just email them to d...