Clark: Coupling, Emergence, and Explanation

Clark, Andy. (2007). "Coupling, Emergence, and Explanation". In M. K. D. Schouten & H. Looren de Jong (Eds.), The matter of the mind : philosophical essays on psychology, neuroscience, and reduction (pp. x, 330 p.). Malden, MA: Blackwell Pub.

This compilation of articles is aimed at the science of cognition. In this selection, Clark presents emergence as a new tool for understanding issues surrounding cognition. “I shall argue that emergent phenomena do require new modes of explanation and understanding. But these modes do not displace more familiar project such as homuncular decomposition and representational/computational description. Instead, we must use a variety of tools to understand the multiple aspects of real-time, embodied, embedded cognition” (228).

The homuncular explanation is the first to be explored. Clark claims that it is a fully reductionist account of cognitive functioning. Some have posited that the concept of emergence is also reductionist. “...to contrast emergent explanation with reductionist explanation would be to invite a common misunderstanding of the notion of emergence, that is to suggest that emergentist accounts do not explain how higher level properties arise as a result of more basic structures and interactions…emergentist hypotheses are by no means silent on such matters. Rather, the contrast lies in the ways in which the lower-level properties and features combine to yield the target phenomena. This kind of emergentist explanation is really a special case of reductionist explanation, at least as intuitively construed, since the explanations aim to render the presence of the higher-level properties unmysterious by reference to a multitude of lower-level organization” (229).

The next target is interactive explanation. “…interactive explanation takes very seriously the role of the environment in promoting successful problem-solving activity. It seeks to display the ways in which crucial problem-solving moves may actively exploit the opportunities which the real world presents to embodied, mobile agents” (230).

“Emergent explanation is at once the most radical and most elusive member of our trinity” (232). Some people understand emergent behavior to be properties that are unexpected outcomes from the perspective of the observer. However, this produces a set of emergent properties that are objective-relative. Others have attempt to see emergence as being a property associated with successful evolution of a subject with their enviornemnt, “…and in which the patterns of results which this interaction yields require description in a vocabulary which differs from the one we use to characterize the powers and properties of the inner components themselves” (232).

Clark then provides examples from the field of artificial intelligence, one dealing with termites, the other robots. Throughout these examples, the theme of a distinction between controlled and uncontrolled variables rings out. However, according to Clark, these are not the instances of emergence that are most interesting: the most interesting cases are, “…cases in which the uncontrolled variable tracks some process…involving ‘continuous reciprocal causation’” (234).

Continuous reciprocal causation is, “causation that involves multiple simultaneous interactions and complex dynamic feedback loops, such that (a) the causal contribution of each component in the system of interest is determined by, and helps to determine, the causal contributions of…other components, and (b) those contributions may, as a result, change quite radically as the process evolves” (234).

Clark then claims that it is not the continuity of a system that allows it to be prone for emergent behavior, but rather the complexity of the system.