so
(Conclusion) ‘understanding social institutions is still a matter of grasping empirical generalizations which are logically on a footing with natural science’.
However, this argument is defective according to Winch because where we speak of uniformities we must have some kind of criteria of sameness. To characterize something as going on in a uniform manner is to characterize it as being the same in certain respects throughout time. However, what is characterized as being the same by one criterion might not be characterized as being the same by another. For example, someone looking at two pictures (one picture of an African elephant and one of an Indian elephant) might say that both depict the same creature, an elephant; however, we might say that they depict different species: one is an African elephant and another is an Indian elephant. Someone who is asked whether the two pictures are the same would likely be confused until they are told something further about the criteria they are supposed to apply in deciding. They might respond that they are not the same because the pose of the animal is different in each, or they might refer to the dimensions of the pictures and say the second is larger than the first.
As Wittgenstein says, ‘The use of the word “rule” and the use of the word “same” are interwoven.’43 What this means is that if we are to decide whether two things are the same or whether something counts as ‘going on in the same way’ (as in cases when we are asked to continue a series of numbers) we must do so by reference to a definition or a criterion – a rule of one sort or another. And, as Winch says, ‘rules […] rest on a social context of common activity’44 and so to decide the nature of a particular field of study we must look at the kind of activities which it involves and also at the rules embedded in those activities which tell us whether the objects of the study are of the same kind or not, or whether they continue to be the same throughout time.
If we look at the kinds of activities engaged in by natural scientists and by those engaged in fields concerned with human activity (psychology, history, sociology, literature, etc.) then we find that the things studied differ in each case. The rules which we consider in thinking about natural sciences are, for example, the grammatical rules, which constitute scientific concepts, and the rules governing the procedures of the scientists. However, in the case of those studying human activity we must consider not only the rules of the activities of the sociologists but also the rules governing the behaviour of those that the sociologist studies. It is the second set of rules, according to Winch, that tell us about the nature of sociology. It is those rules ‘which specify what is to count as “doing the same kind of thing” in relation to that kind of activity’.45
The significance of this in thinking about the relation between social fields and the natural sciences is that the two kinds of activities are quite different. John Stuart Mill had argued that studying human society is like studying a complicated mechanism. However, if Winch is correct then the sociologist’s ‘understanding of social phenomena is more like the engineer’s understanding of his colleague’s activities than it is like the engineer’s understanding of the mechanical systems which he studies’.46 Explanation in sociology is often not like the causal explanations of natural science. However, that does not imply that it is not scientific at all.
1.3.2Is Winch Correct? – Davidson’s Argument That Reasons Are Causes
Winch distinguished explanations in terms of habituation, which he said were causal, from explanations in terms of rules, which he said were non-causal. Donald Davidson, in his 1963 paper ‘Actions, Reasons, and Causes’, argued, pace Winch, that explanation of human action citing the agent’s reason for their action (i.e. the kind of action that Winch said was rule-governed) ‘is a species of ordinary causal explanation’.47 Davidson argues for this first of all by pointing out that the division between explanations in terms of reasons and explanations in terms of causes is not obviously mutually exclusive. It may be that nonteleological causal explanations do not have features that explanations in terms of reasons do, namely that explanations in terms of reasons have a justificatory element; nonetheless, ‘it does not follow that the explanation is not also – and necessarily – causal’.48
Davidson also goes further. He doesn’t rest satisfied with the claim that it is not obvious that explanations in terms of reasons are not causal. He gives an argument in favour of thinking that explanations in terms of reasons are causal. Davidson’s argument for this is that people can have a reason to do something and yet that reason was not the reason why they did it. Several different reasons in a particular case could serve to make an action intelligible. For example, somebody might raise their arm and wave it around outside of their car window in order to greet a friend or in order to signal a turn or in order to cool their hand. We might ask why somebody raised their arm and waved it around outside of their car as they drove around a bend and they might respond, ‘I saw my friend on the corner and waved at him’ or ‘my hand was hot having been on the warm steering wheel and so I wanted to cool it down’ or ‘I wanted to signal that I was turning’. How do we pick out the agent’s reason from among the reasons that they had, which might have served to make the action intelligible? – Davidson’s answer is that ‘central to the relation between a reason and an action it explains is the idea that the agent performed the action because he had the reason’. And Davidson thinks that in order to ‘account for the force of that “because”’ we should think of the relation between reason and action as causal.49
Davidson argues that his opponents, the Wittgensteinians (including people like Winch), have not accounted for this relation between reason and action by talking about patterns and contexts because ‘the relevant pattern or context contains both reason and action’.50 Davidson might not have produced a conclusive argument in favour of construing the relation between reason and action in causal terms but it seems as though he has nonetheless provided some reason for thinking that explanation in terms of reasons is a kind of causal explanation. If his opponents are to dispute that, he says that they must identify an alternative pattern of explanation.51
Davidson’s anti-Wittgensteinian arguments are formidable and have been enormously influential in terms of the way that many philosophers nowadays think about explanations of action in terms of reasons. What this demonstrates is that anyone who wants to defend a position along the lines that Winch wanted to defend must now deal with Davidson’s arguments. The debate has moved on since Winch published The Idea of a Social Science and non-Wittgensteinian thought now predominates in philosophy departments around the world.
1.3.3Is Winch Correct? – Tanney’s Response to Davidson
However, that is not to say that Davidson is correct and that a defence of ideas in the spirit of Winch cannot be given. Over the course of the past two decades, Julia Tanney has built up a powerful case against Davidson’s conception of explanations in terms of reasons and she has defended the Wittgensteinian view that Davidson attacked. She has written a series of articles about reasons and rule-following that are collected in the recent volume, Rules, Reason, and Self-Knowledge.52
In her article ‘Why Reasons May Not Be Causes’,53 Tanney examines various cases where somebody had a reason but did not act for that reason. This is the kind of case that Davidson suggested calls for thinking