Interactivist Summer Institute

July 22 - 26, 2003


New Trends in Cognitive Science – Lessons from the Past

Rossmanith, Nicole Reichelt, Andreas Roemmer, Brigitte
Institute for Philosophy Science and Social Studies of Science
Sensengasse 8/10
1090 Vienna, Austria


Major Theme: Analyzing the commonalities and differences between cybernetics and current developments in cognitive science (Keywords: Emergence, Social processes and realities, Robotic and computational models of interaction and cognition)

Cognitive science started out as an interdisciplinary enterprise trying to bring together various strands of cognition research. A unifying framework was provided by the symbol-manipulation-paradigm. Today, almost fifty years later, the classical paradigm has given way to a multitude of approaches, which are discussed in different communities. While some of them, as diverse as machine learning, evolutionary psychology, and cognitive neuroscience, are still compatible with the original paradigm, others explicitly reject the computer metaphor altogether, such as dynamical systems theory, situated cognition, embodied AI, and evolutionary robotics.

While each of them might provide new insights to overcome some of the problems inherent in symbol processing, they all follow their paths more or less independently, posing different questions and applying different methods, some of them moving away from the original goal, the understanding of human cognition.

As diverse as these new developments may seem, a closer look reveals many commonalities. Yet another look reveals that some of them are not so new after all, but date back to the time before the cognitive turn. We suggest that they can be better understood and related to each other by looking at them in the light of history. There are striking similarities to research carried out in the interdisciplinary field of cybernetics.

In many respects, cybernetics was ahead of its time. In addition to the development of the General Purpose Computer and information theory by mathematicians and physicists, neurophysiologists were involved, Neural Net models were designed, and autonomous robots made their first appearance. There was a strong emphasis on information and communication. Social scientists, notably Gregory Bateson and Margaret Mead, were involved from the beginning, hoping that the new theories and tools could be applied to model social systems. Concepts that could connect the diverse domains of research were at the heart of cybernetics, like feedback loops and circular causality. These were part of a larger enterprise of constructing a general theory of artificial and natural systems.

While some of the ideas developed by the cyberneticians were spectacularly successful, the field itself was eventually marginalized by cognitive science under the aegis of classical AI. Ironically, as Gardner pointed out as early as 1987, the period just before the cognitive revolution is much more similar to current developments in cognitive science than the classical computationalism that followed it. Often quoted, a deeper analysis of cybernetics rarely found.

It is worth investigating the ideas of the cybernetics movement regarding the question which ideas were retained and built upon by cognitive science and which were dismissed and didn’t resurface until much later. Interestingly, many of the arguments for and against those approaches have remained essentially the same. Leaving aside external factors for a moment, whether a particular idea or methodology is considered a fruitful approach, is partly due to what is accepted as an explanation. Was early cognitive science simply more receptive to the reductionist concepts while the more holistic ones, such as emergence, self-organization, and circular causality are gaining ground today? In what ways do the new approaches substantially differ in terms of concepts and scope? Are there still lessons to be learned from cybernetics?

Reference: Gardner, Howard (1987). The Mind's New Science. New York, Basic Books.


ISI 2003 Home Page