things which we would like to say exist but that are not material objects. As Max Bennett and Peter Hacker note, ‘Laws and legal systems, numbers and theorems, games and plays are neither material objects or stuffs.’ Bennett and Hacker point out that even when it comes to material objects we often explain their behaviour, perfectly legitimately, in terms other than what they are made of. We explain some things in terms of their function (e.g. human organs), others in terms of their goals, reasons, or motives (the behaviour of animals and human beings).28 Historical events, such as the Russian revolution, are not explained in terms of what they are made of, ‘since they are not made of anything’.29 So, materialism cannot be used in support of reductionism.30
Another problem with attempts to reduce social sciences to natural ones is that social sciences often involve reference to the psychological attributes of human beings but psychological attributes of human beings cannot be reduced to any of the usual candidates that reductionist philosophers refer to – cells, molecules, brain states, or sense data. In the Philosophical Investigations one case that Wittgenstein brings our attention to is the case of knowledge. He carefully examines the grammar of ‘know’ and ‘understand’ and helps us to recognize that knowing cannot be a physical state, a mental state, or a disposition. If it were a physical state then there would be (at least) two different criteria for knowing – (i) the correct application of a relevant rule (e.g. a criterion for someone knowing the alphabet is that they can write or say ‘A, B, C, D, E,’ etc.) and (ii) the criteria for identifying the corresponding physical state or disposition. But it seems that the second criterion is not the one we would use, since even if the brain were in a particular physical state whenever someone recited the alphabet we would not take the presence of the state to indicate knowledge if someone wrote ‘A, D, F, Z, 3’ when asked to write the alphabet.31 Rather than being reducible to a physical state or disposition, knowledge is akin to an ability,32 and an ability is categorially distinct from the usual candidates that reductionists refer to (cells, molecules, brain states, physical things, or sense data). Following Wittgenstein, Bennett and Hacker note, ‘The criteria of identity for mental states, events and processes differ from the criteria of identity for neural states, events and processes.’33 This should be clear from the fact that psychological attributes are attributable to a person or to animals but neurophysiological attributes are attributable to their brains.34 So, for example, someone might believe that voting to leave the European Union (EU) was the right thing to do in the recent referendum in the United Kingdom. I attribute that belief to them (not to their brain) on the basis of their behaviour, particularly their linguistic behaviour. I attribute that belief to them, most likely, because they say that they believe that voting to leave the EU was right and I have no reason to doubt that they believe that. However, I do not attribute brain states or processes to them on the basis of their linguistic behaviour and those brain states or processes are states of that person’s brain and not of the person. The person’s beliefs cannot be neural states or events because their neural states and events have a location but their beliefs cannot be said to have a location (at least not in the same way). It makes no sense to ask, ‘Where do you believe it was wrong to leave the EU?’ Some questions sharing this form do make sense but they are not answered in a way that suggests that beliefs are neural states. So, for example, it does make sense to ask ‘where do you believe the football game between Sporting Lisbon and Benfica will take place?’ but this question is not answered appropriately by saying ‘in my head’, but by something like ‘at the Stadium of Light’.
It is also worth noting that not only are social sciences not reducible to natural sciences but natural sciences themselves cannot all be reduced to physics. John Dupré has argued convincingly that ecology is not reducible to any level below biology,35 and that there are various problems with reductionist projects in genetics.36 There have been successful reductionist projects but these successes have been very local. Biological science has not been shown to be reducible to physics and we have good reason to think that it cannot be reduced to physics, namely that categorization in biology and much of the rest of science is driven by changing human interests and there is no single privileged taxonomic scheme in biology in terms of which it could be reduced to physics.
Wittgenstein thought that the temptation to reduce phenomena in one area to phenomena in another was one of the causes of philosophical confusion. In the Blue Book Wittgenstein says that his worry about philosophers’ preoccupation with the method of science is, at least in part, a worry about ‘the method of reducing the explanation of natural phenomena to the smallest possible number of primitive natural laws’ and that ‘it can never be our [i.e. philosophers’] job to reduce anything to anything’.37 Philosophy is descriptive, that is, it describes norms of representation with the aim of getting clear about the meaning of problematic terms in order to get rid of the confusion at the root of philosophical problems.38
1.3Reasons and Causes
As mentioned in the introduction to this chapter, one of the debates that is relevant to the question of whether the social sciences are continuous with the natural sciences is the debate about whether explanations in terms of reasons are causal explanations. One approach is to claim that human actions are distinct from behaviour resulting from habits (which influence our behaviour causally). A way of bringing out this distinction is to compare human activity with the activities of animals. Peter Winch, a Wittgensteinian philosopher, uses the example of a dog learning to balance sugar on its nose and holding it there until its owner issues a command to eat it. In this case the dog has been trained into a habitual response and cannot be said to be reflectively following a rule. Like rule-following cases the dog might be said to have done something correctly or incorrectly but this is only because we are applying human norms analogically to animals, according to Winch.39 This is unlike the case of a human being continuing the series of natural numbers beyond 100 upon being ordered to do so because ‘the dog has been conditioned to respond in a certain way, whereas I know the right way to go on on the basis of what I have been taught’.40
The debates in philosophy about the distinctions that Winch makes between rule-governed human behaviour and habitual animal behaviour, and between reasons, motives, and causes, have moved on since the time of The Idea of a Social Science. A seminal anti-Wittgensteinian paper, in opposition to the kind of view that Winch presents, is Donald Davidson’s ‘Actions, Reasons, and Causes’ published in 1963.41 The arguments between Davidsonians, Wittgensteinians, and others continue to this day.42
1.3.1Social Studies and Natural Science
The considerations about differences between causal and rule-governed behaviour suggest that human activity cannot be understood in terms of the causal generalizations favoured by natural scientists. However, Winch thinks that explanations of human behaviour in terms of institutions and rules might still be defended by followers of philosophers like John Stuart Mill as being scientific because:
1.‘an institution is, a kind of uniformity’.
2.‘a