Summary of Seminar #2
I’m posting some brief comments on the discussion led by Tom in seminar #2. I didn’t take any notes, so this will also be a test for my failing memory. Please, feel free to add/correct.
Tom’s concern was to question the role of AI in shifting paradigms in cognitive science. He briefly described the shift from cognitivism (understood broadly as a framework that sees the mind as computational/representational) towards embodied-embedded cognitive science (where cognition is understood as the engagement in appropriate coping through situated and embodied sensorimotor coupling). Tom’s first point was that developments in AI (e.g., Brooksian robotics) were principal factors in instigating a turn towards the embodied-embedded conception of cognition. During discussion on this point it emerged that it would also be proper to recognize that traditional AI was also instrumental in defining the cognitivist paradigm in the first place (eg Turing machines, von Neumann architecture; the picture of influences is blurred a bit when we remind ourselves that Turing’s machine is itself based on a cognitive theory of how a person would solve an algorithmic problem).
Tom later described the new shift that is starting to take place now, a shift towards enactivism. This perspective takes on the lessons of embodied-embedded cognitive science, but asks new questions such as the nature of agency, values, but most importantly in Tom’s view, the role of experience in informing, directing, or inspiring both theoretical and empirical research (eg first person approaches, neurophenomenology, etc).
It is unclear whether this shift is empirically-driven, eg, by the development of new methodologies in AI. It may be that stumbling blocks in embodied-embedded (and dynamical!) research in robotics may indicate the need for a new set of questions and methods. But, according to Tom, the main shift will probably be driven by the experiental element, by the need to take experience seriously in guiding research.
Tom presented a diagram where the interactions between experience/phenomenology, theoretical and empirical work were all present and bi-directional. Of course, questions emerged about the practicalities of such a view. When we talk about experience, are we talking only about humans? (for how can we know about other cases?). How does a phenomenological insight illuminate modelling research, say a new kind of robot? During discussion some people suggested that experience was not absent in current scientific methods, only very well regulated through experimental procedures and embodied practices that can be learned and replicated. The problem remains how to use more ineffable structures of experience that emerge from phenomenological (or other) methods a) in a learnable, communicable way and b) in a practical combination with theory and practice to produce new results. Tom suggested that this would be a real challenge because the structure of our scientific and academic communities do not facilitate such kinds of developments.
One general question that I have is how much of these paradigm shifts (the one that happened and the one that is coming) are real paradigm shifts or splits in the cognitive science communities. I’m always amazed at how much of what I would personally call GOFAI/cognitivism is still going on strongly throughout the world. So maybe there’s no shift happening at all? On the one hand, this is sign that we may indeed be dealing with a paradigm shift situation (because if we were talking about within-paradigm developments, then you would expect the within-paradigm community to follow the turn towards embodied-embedded and maybe even enactive perspectives). The fact that logical arguments do not seem to suffice to produce such a change in the practice of certain communities indicates that we are indeed dealing with a change of views which is more related to personal and emotional perspectives on the relative importance of some elements or some questions within cognitive science.