euCognition Workshop
Models of Thought: Post-Cognitivist Methodologies (Munich, 20th/21st February 2008)
(For further information and details of individual presentations see the euCognition proceedings).
This recent event brought together thinking from psychology, philosophy, artificial life and related disciplines in the Municon conference facility in Munich under the banner of euCognition, the well established european society for the advancement of artificial systems.
The twin themes were post-cognitivism and artifical systems/ models. The idea arose from thinking in the 'brave new world' of post-cognitivism where consciousness arises from complex and chaotic systems and the search for general ('Newtonian') laws of human behaviour seems increasingly futile and misconceived. A number of questions were envisaged which seemed to emerge naturally from consideration of post-cognitivst theory:
- How can we best model cognition? What are the specified relationships between artificial models and animal and human behaviour?
- What role do virtual enviroments play in understanding human behaviour?
- What are the broader implications of 'post-cognitivism' for philosophy and practice in psychology and cognitive science? How do we 'do' psychology if we give up on the idea of discrete entities in the individual which are predictive of behaviour?
- What role does langauge play in all this?
Day 1
The first day speakers were Mike Wheelerfrom the University of Stirling and Max Velmans from Goldsmiths, London.
Max Velmans is interested in consciousness studies, with a particular focus on integrating work in philosophy, cognitive psychology, neuropsychology, and mind/body relationships in clinical practice.He spoke about Reflexive Monism, discussed its relation to Enactivism and described it as an attempt to resolve the so-called "hard problem(s)" of consciousness. To leave the last word to Max himself: "The reflexive model suggests that what we normally think of as the “physical world” is just the experienced world that arises from a reflexive interaction of the perceiver and perceived. First-person and third-person perspectives co-arise. Consequently there never was an explanatory gap between the physical world as-perceived and conscious experience".
His books include Understanding Consciousness (Routledge, 2000) which provides an appraisal of current theory along with a novel resolution of the "hard" problems and The Blackwell Companion to Consciousness, 2007, which provides an extensive review of current scientific and philosophical thinking in this area.
Mike Wheeleris a Reader in philosophy at the University of Stirling. He wrote Reconstructing the Cognitive World: the Next Step which was published in 2005 by MIT.
Mike spoke about themes from the book concerning such seemingly disparate aspects as Heidegger and AI-oriented robotics to articulate and defend a non-standard philosophical framework for cognitive science, saying: "It is not action-oriented representation that heads off the intra-context frame problem, but the presence of situated special-purpose adaptive couplings. Representations may figure in such mechanisms, but they may not".
Day 2
Day 2 saw further interesting and stimulating talks from speakers from a wide range of backgrounds.
Fred Cumminsgave an entertaining talk called 'Beyond the Individual: (or how we might forge links between approaches to mind which take the subjective world seriously, and those which don't)" which asked questions about the subjective "P-world" of phenomenal experience drawing from a wide range of thinking (e.g. Buddhism and neuroscience)!
Juan Escasanyspoke about "Naturalizing Cognitive Science"and argued that we must "begin by studying variables that are ecologically important rather than those that are easily manageable".
Andrej Lucnygave a talk with the title: "Can information entropy help to formalize cognition (to evaluate particular models)"?and argued that artificial cognitive systems should not only act in a reasonable way upon knowledge about its environment, but should be able to improve this knowledge and subsequent activities.
Pavel Petrovicspoke about incremental evolution in artificial systems.
Paco Calvo Garzonasked "Post-cognitivist rules: What experiment can disprove your hypothesis...."?He cited and discussed important papers from Science ranging from 1964-2002 which shed light on this question.
Participant areas of interest
Participants were asked to indicate areas of interest prior to the event on the conference 'wiki'. These helped form discussion themes. Interests and activities included:
Vincent C. Müller,American College of Thessaloniki (http://www.typos.de Working on: Basic problems of AI (book project), theory of computing, etc.] - The theme of this workshop is very timely, in my opinion. In particular, I would like to talk about and with non-cognitivist roboticists (Brooks, Pfeifer, etc.).
Juan Escasany, APEROBOT (Association for Practicing and Education in Robotics). I'm an independent researcher (email: aperobot'at'telefonica.net). We are right now moving our webpage. I agree this theme is very timely. After reading Dreyfus' article on Heideggerian AI I'm totally identified. We've been working in mobile robots trying to implement design requirements based on Heidegger and Merleau-Ponty ideas. I'll be happy to talk about problems and achievements in this approach of AI research.
Fred Cummins, University College Dublin. Post-cognitivist cognitive science is but part of the herculean task of moving beyond a psychology of the autonomous individual. The complete immersion and embedding of the person in the environment (defined both socially and physically) is beginning to be appreciated within neuroscience, ecological psychology, agent modeling, and beyond. This is not entirely new, but a rediscovery of the centrality of collective phenomena in thought and behavior. For my approach, see http://pworldrworld.com and http://pinkmonkeyfarm.com (sic)
Andrej Lucny,Comenius University Bratislava. I am dealing with control e.g. of a mobile robot which comes from interaction among set of reactive agents. (My architecture follows Brooks and Minsky).
Pavel Petrovic, Comenius University Bratislava. I recently completed a thesis at NTNU Trondheim on evolving behavior coordination mechanism for mobile robots based on finite state automata. I am also working on NXT Logo, a Lisp-like educational programming language for NXT robots, which allows building interactive projects that go beyond the standard program-and-run paradigm. Recently, I am very interested in Bayesian approaches to cognition.
Paco Calvo, Universidad de Murcia, Spain. I'd be interested in discussing: (i) The relation between embodied cogsci and classicism and connectionism (roughly, I believe that connectionism was absorbed by classicism, and the same could happen with embodied approaches, unless we have a clear idea of what it means for a system to deal with representational states), and somewhat related, (ii) does it make sense to go for situated robotics WHILE remaining an (extended) functionalist. And finally, (iii) how minimal embodied cognition can be (are invertebrates cognitive, for instance)?
Marek McGann. Mary Immaculate College, University of Limerick. I'm a psychologist, and concerned by the fact that the impact of postcognitivism in Cognitive Science more generally has left the vast majority of Psychology untouched. I'm interested in trying to address how postcognitive approaches can be "scaled up" as it were to more directly affect our understanding of "higher cognitive function" and other aspects of psychology that impact directly upon domains other than cognitive psychology. My own views are strongly enactive (in the Varela, Thompson & Rosch Embodied Mind; and Thompson Mind in Life kind of way), and affected also by the ecological psychologies of Gibson and Barker.
Marcin Miłkowski.Institute of Philosophy & Sociology, Polish Academy of Science. I'm mostly interested in computationalism, modularity and situated cognition, as well as meaning in the cognitive systems and the architecture of mind. These are broad topics, but basically I think that classical computationalism is much more open to post-cognitivist approaches and it can actually try to absorb them - which is what Paco would find probably quite repelling :)
The workshop organisers were Brendan Wallace from the University of Glasgow and Alastair Ross from the University of Strathclyde.