The conclusion that meaningful explanations of action are causal was reached in the first chapter, based primarily on the fact that they deliver theory-driven predictive power. The meaningful explanations are those of folk psychology, but the same considerations encompass models in cognitive psychology and indeed in biology, which share the fundamental idea that functional systemic activity is regulated by information-carrying states. The main argument of the present chapter has been that while explanations of this general kind are causal, their logic is not captured by certain familiar...
The conclusion that meaningful explanations of action are causal was reached in the first chapter, based primarily on the fact that they deliver theory-driven predictive power. The meaningful explanations are those of folk psychology, but the same considerations encompass models in cognitive psychology and indeed in biology, which share the fundamental idea that functional systemic activity is regulated by information-carrying states. The main argument of the present chapter has been that while explanations of this general kind are causal, their logic is not captured by certain familiar interpretations of causality.
The traditional analyses of causality were sketched in Section 4.2. Hume proposed that causal propositions are based in observation of association between kinds of event. However, this analysis failed to capture the necessity in causal propositions, a gap which has to be made good by distinguishing mere generalizations from those which are or which are covered by natural law, in particular of physics. Essential features of causal propositions according to this kind of analysis are thus empirical correlation covered by a general physical law. Meaningful explanations do exhibit the features which are expected of causal explanation—necessity and generality—though not in the way envisaged by the views of causality already outlined.
Nevertheless, once it is granted that meaningful explanations are causal, there is great pressure to bring them into the domain of the physical sciences and the notion of causality appropriate to them. Causal semantics, discussed in Section 4.3, is one way of doing this. The proposal is that A carries information about B in case B causes A, i.e. in case there is a correlation between events of kind A and events of kind B covered by a natural law. As a theory of content this tends to be either vacuous or inadequate however, and generally causal semantics fails to capture several linked features of signs, that they can be Incorrect as well as correct, and that are typically related to the signified by convention, not natural law.
Neglect of the systemic contribution to information processing and content is the main failing of causal semantics, and is made good by so-called functional semantics, to which we turned in Section 4.4. The main idea of functional semantics is that content is to be defined with reference to the (normal) function of the information-processing system. Two versions of functional semantics were considered. In its ‘causal’ version, the notion of normal function is used to make a normative distinction among the causes of information-carrying states, and hence a normative distinction among the contents of such states, with the critical task of defining the normative distinction among causes to be performed by biological/evolutionary theory. Contrary to an argument of Fodor's, evolutionary theory can deliver intensional descriptions of functions and objects, and hence also a theory of error. It was subsequently argued that a behavioural version of functional semantics can deliver the same. Causal-functional semantics disguises the fact that the notion of normal function affords primarily a normative distinction among behavioural responses, and that it is this which grounds the distinction between true and false informational content. Functional semantics in its behavioural version makes this explicit.
The intimate connection between intentional states and functional systems is what gives rise to the special causal status of intentional explanations. The familiar account of causality in terms of generality covered by natural law is appropriate for the lower-level sciences, physics and chemistry, up to, but not including, biology. With the appearance of (the study of) functional systems, in biology and psychology, different principles of causality come into play, different approaches to necessity and generality, considered in Section 4.5. If prediction from physical theory fails, and statements of initial conditions are sound, then there is an error somewhere in the theory. Either that or there has been a miracle! By contrast, if prediction from a meaningful generalization fails in a particular case, then certainly the generalization can be abandoned, but there is another possibility, namely, that the system in question is failing to function normally. This possibility is analogous to the breakdown of law, which the physical sciences never envisage. But in the case of systems breakdown can and does occur. The ‘laws’ being broken are not general laws of nature, but are rather rules or norms which apply specifically to one or another kind of functional system. In this way the causal necessity in explanations of systemic function is based in norms of function, not in general laws of nature. This is one aspect of the difference between functional semantic causality and the kind envisaged in the standard analyses. Further, while it is possible to make generalizations concerning systemic function, and meaning in particular, they are restricted to, precisely, one or another kind of functional system. The sciences from biology upwards are concerned with specifics. A connected point is that generalization tends to be at the expense of information about particular cases. This point increases in relevance as specificity of function increases, in particular as we make generalizations about meaningful connections among higher-level cognitive-affective states and action. Implications for the concept of agency fall out of the analysis (Section 4.6.3). Causal power is attributed to what is specific to the agent to the extent that explanation cannot be given in terms of a more general nature. In the extreme case, the individual person is identified as the causal origin of the act.
Chapter. 20303 words.
Full text: subscription required