George Kampis





 Lecture One
 Representation, Causality and Complexity

 This lecture deals with the problem of mental representation and its relationship to knowledge embedded in the context of evolutionary biology and embodiment. Besides, I discuss the problem of cognitive modeling from the point of view of complexity. This lecture serves as some kind of a summary of my current standpoint on cognitive science, philosophy of science and complex systems, as well as an introduction to several more specific issues, to be touched upon in a next lecture.

 Mental representation is often conceptualized in the same form as external knowledge, as text or language. However, recent studies in cognitive science and some novel philosophical developments reveal that this is a misleading picture about the mind. This symbolic or propositional model of the mind is based on the metaphor of writing. I will briefly discuss the development of writing to understand this point and to see how propositions and other categorical concepts relate to mental structure. We will then characterize mental structure with the aid of biology. Studies of pre-linguistic cognitive agents such as human infants and animals suggest that the primary medium of mental representation is based mainly not on words but pictures.

 I show some recent efforts in developing a picture theory of representation and reasoning. I am extending this towards what I call a causal picture of representation and reasoning. I do this in two steps. First, I will argue that pictures are not mental representations by themselves, but are parts of integrated physical representations called mental models; pictures are only appearances of the mental models for subjective experience.

Second, I will discuss the dynamics of physical entities (that is, causality) from a general point of view. I show that causal relations always involve an essential parallelism or “depth”. I will argue that it is this depth that lends mental models and mental processes a changing or transitory, and the the same  time unsharp character, a feature not accounted for in propositional or even in entirely picture-based representations. At the same time, the notion of “depth” allows us to intruduce the notion of complexity in a precise sense, understood as a property of dynamical systems.

The lecture concludes by suggesting that multiple meanings, hidden dynamical variables and complex (causal) processes characterize mental representation. By contrast, the prevailing approach to mental representation and to knowledge is “one-dimensional” in the sense that one single thread of meaning and one single set of variables are assumed in what ends up to be a simple system, in the sense discussed here. I discuss some of the consequences for modeling and conceptualization.



 Lecture Two
 Causality, Logic and Dynamical Systems

 This lecture embarks on the remarks of the first lecture. I will discuss in detail how biological existence presents certain preconditions for knowledge by means of embodiment; how language and meaning in particular, depend on physical structure, and how bodily actions form the basis for the understanding of causality and logic.

 Then I will show that the most fundamental forms of reasoning are what philosophy of science calls “mechanisms”. I illustrate on some simple examples that mechanisms as explanation structures involve some inexplicated or “opaque” elements – from the formal point of view of logic, enthymemes (in the sense of Aristotle). I discuss how mechanisms form a basic material form on which all causality, all logic, and all science (including explanations) can be built.

 On the basis of this notion of mechanism I will re-discuss the problem of mental representation, embodiment, and dynamical systems. In the context of the latter, I will discuss a special class of complex systems, called “chaotic itinerancies” in the sense of Tsuda and Kaneko, as possible examples of systems than can support the emergence of causality and mechanisms.

 I conlcude by offering a perspective of my ongoing research at JAIST on further related questions.