Our capacity to perform these causal inferences does not depend upon our having a scientific theory of how the cause produces the effect. In most cases we are simply ignorant of the inner workings of nature. (In fact, Hume believes that we never can have any ultimate explanation of why things happen in nature as they do.)

It is also clear, he says, that we could never infer the effect of some cause we experience unless we had previously had experience of a relevant kind. Locke tells a story of an Oriental prince being told by a traveller from Europe that in his country, when it becomes very cold, the water in the canals becomes solid, so that it could bear the weight of an elephant. The prince replies that until then he had believed what the traveller told him of Europe, but now he knows that he is a liar. Hume, recalling the story, says, ‘The Indian prince, who refused to believe the first relations concerning the effects of frost, reasoned justly.’ Since the prince had no experience of the relevant kind of phenomena, he had no reason to believe that the effect of cooling water beyond a certain point is that it becomes solid. In no case, Hume argues, can we establish causal connections a priori.

It is previous experience of what Hume calls the ‘constant conjunction’ of causes and effects, and such experience alone, which enables us to ‘reason justly’ from causes or effects which we perceive to effects or causes which we do not perceive. This form of reasoning is quite distinct from a priori reasoning in the domain of ‘knowledge’, which is based simply upon relations between ideas. It is based on experience, and, being in the domain of ‘probability’, Hume calls it probable reasoning.14 What the theory of human nature must do is explain how it is that experience of the conjunctions of causes and effects in our past experience leads to our being able to reason from a perceived phenomenon to its unperceived cause or effect. The essential feature of such reasoning, for which an explanation is required, is that what we infer is always a phenomenon of the same kind as those phenomena of which we have had previous experience. We always make inferences which presume that causal connections in nature are uniform. Why?

One possibility to be considered is that we make this presumption because we have a reason to do so. Hume thinks that this hypothesis cannot be correct. That causal connections between phenomena are uniform (that the same kind of cause always produces the same kind of effect) is not a necessary truth which we could have an a priori reason to believe. We can without contradiction imagine the negation being true – we can conceive a change in the course of nature. However, the only other sort of reason we could have would be to infer it from the evidence of past experience. We would then be inferring it by probable reasoning. But we were seeking an explanation of why probable reasoning has the character it has; and so it is circular to explain this by assuming probable reasoning already.

An alternative explanation, and the one which Hume favours, is that probable reasoning is not dependent on any other prior ‘operation of the mind’; in fact, it is a kind of instinct. In the section of the Treatise called ‘Of the Reason of Animals’, he emphasizes the link he thus makes between this form of human reason and the instincts of other animals: ‘To consider the matter aright, reason is nothing but a wonderful and unintelligible instinct in our souls.’15 Although he here says that reason is an ‘unintelligible’ instinct, he does propose an account of the mechanism in the mind by which it operates. For the moment we can postpone outlining this. For the purpose of considering the Dialogues some other comments are needed.

As we have seen, Hume intended his theory of human nature to provide a foundation for other sciences:

Even Mathematics, Natural Philosophy, and Natural Religion, are in some measure dependent on the science of MAN;… ’Tis impossible to tell what changes and improvements we might make in these sciences were we thoroughly acquainted with the extent and force of human understanding, and cou’d explain the nature of the ideas we employ, and of the operations we perform in our reasonings.16

His account of the ‘operation’ of probable reasoning should, therefore, contribute to ‘changes and improvements’ in natural science and natural religion. The theory as summarized above applies initially to those inferences we make from causes and effects in everyday life, and these are characteristically immediate and unreflective. When I see the child put her hand in the fire, I infer at once that she will be burned. Although the inference arises from my past experience of the effects of fire, I do not even call this experience to mind. Hume recognizes, however, that we also make inferences of a more reflective nature. For example, when our experience is limited in extent, we proceed with more caution. We try to review what experience suggests. We begin to form what he calls ‘general rules’ to guide our more reflective reasonings. These form the basis of the even more elaborately reflective methodological principles of the natural scientist. One thing we then do is deliberately to seek experience of the conjunctions of phenomena, in experimentation.