Predictability and explanation

Johansson, U., Sonstrod, C., Konig, R. and Niklasson, L. (2003) Neural networks and rule extraction for prediction and explanation in the marketing domain. The International Joint Conference on Neural Networks, IEEE Press, USA, Portland, OR, pp. 2866—2871. 

Vidal, B., M. Bouzige, R. Laroche, G. Figueres, and J. Biesse TSNAs in tobacco and smoke PLS regression for prediction and explanation of smoke contents according to chemical and physical characteristics 56th Tobacco Science Research Conference, Program Booklet and Abstracts, Vol. 56, Paper No. 55, 2002, p. 53. 

Many of our best predictions and explanations of human behavior come at the higher level of beliefs, desires, and intentions. If we take our cue from philosophy of science, then we want to say that these irreducible higher-level properties are causal properties. We do not need to reduce these mental states to neurological states. Just as real scientific explanations occur at multiple levels of explanation, and just as irreducible higher-level properties are causal in scientific practice, such is often the case with mental causation. Mental causation is just another case of higher-level causation. And just as the causal mechanism often occurs at higher-levels in science, so it often occurs at the level of irreducible mental properties. 

It seems to me that one of the most (if not the most) important things we need from a theory of mind is a good account of mental causation. If we think that the science of psychology, whose job it is to predict and explain human behavior, is a legitimate science, then we need it to be the case that mental properties are causal, fri addition, a great part of our survival as social creatures in this world depends on our ordinary assessment of one another s mental states and our assumption that those mental states are causally responsible for behavior. We regularly make predictions about what others are going to do and offer explanations of why others have behaved as they have based on our belief that mental states cause behaviors. If we want to retain the independently plausible idea that predictions and explanations of behavior in terms of mental states work because mental states are causal, then we need to account for how it is that mental states can be causal. 

Our foundational theories in physics are generally credited as being the very best science that we have got. Their universality of scope, their exactness of prediction, their depth of explanatory force are all unmatched by the theoretical accomplishments of the various special sciences. But careful reflection on just how these theories function to provide us with descriptions, predictions, and explanations appropriate to the world in which we live shows us that there are profound puzzles lurking in the way these theories function in our scientific account of the world, puzzles that beg for the detailed, careful, and thorough exploration of the methodologist of science. 

What is the scope of a theory How large a domain of happenings in the world fall within its purview A natural first response is to think of a theory as ranging over those processes in the world for which it can actually provide useful descriptions that lead to applicable predictions and explanations. But that natural response is, 1 think, wrong. 

How does idealization function in our theoretical description of the world If theories only work because of idealizations, can we reasonably take theories as even aiming at true descriptions and explanations If all theoretical description and explanation works only in the context of idealizations, shouldn t we think of all such theoretical accounts as equally false (or equally true ) of the world Shouldn t that lead us to a kind of instrumentalism that views all useful, idealized accounts as on a par, with all of them equally serving as members of a collection of instruments we can draw on as needed for description, prediction, and explanation and with none of those within this set of useful instruments as privileged over the others as truer or closer to the truth ... 

At this point I want to take up a related claim. Suppose we admit that all of our predictions and explanations concerning real systems that rest on foundational or nonfoundational physical theories require substantial... 

In addition to the various experimental techniques used to determine and verify the electrical and electronic properties of conductive acetylene copolymers, a number of theoretical calculations and models have also been adopted for the prediction and explanation of the observed electrical properties. The density of... 

In an effective media theory of a composite, a spherical or ellipsoidal grain is considered to be surrounded by a mixture, which has the effective conductivity of the composite medium. It is mainly used for composite materials with well-separated subphases for the prediction and explanation of large volume average values of electrical properties. An excellent overview has been provided by Landauer (1977). 

Design an experiment examining the effect of acid rain on snail shells. Imagine that you have access to a large supply of snail shells and common acids and laboratory apparatus. Include a research question, methodology including the control of variables, and a prediction and explanation for your expected results. 

---