Hume attempted to show that our idea of causality was derived from experience. And although his analysis has been very influential, the rigour with which he attempted this seemed to drain the concept of causality of all its content, so that causality in the objects is reduced to constant conjunction. This has led some people to conclude that the empiricist project is impossible and to embrace the opposite position, namely that certain very general notions like causality are innate and do not arise from experience. Famously, this was Kant's reaction to Hume. And this reaction has persisted; according to P.F. Strawson ‘the concept of causal efficacy is not derived from experience of a world of objects, but is a presupposition of it; perhaps better, is already with us when anything which could be called ‘experience’ begins.'Footnote ¹ However, a recent trend in interpretationFootnote ² has argued that, contrary to the traditional interpretation, Hume had a full-blown realist notion of causality, that he believed that causes have force and energy; but this interpretation puts Hume's empiricism in doubt. Hume's account of causality, and the various responses to it, suggest that there is an incompatibility between an empiricist account of causality and a realist account of causality. I will argue that it is possible to give an empiricist account of causality which is also realist.

1

Although the traditional interpretation of Hume's account of causality is that causality in the objects is just constant conjunction, his starting position is quite different. Hume starts with a realist notion of causality, with the notion that a cause has two defining characteristics: ‘tis commonly supposed, that there is a necessary connexion betwixt the cause and effect, and that the cause possesses something, which we call a power, or force, or energy.’ (T656)Footnote ³ So, first, a cause possesses a quality whereby it brings about its effect. As well as referring to this quality as power, force and energy, Hume also sometimes refers to it as efficacy, agency and productive quality. He uses expressions like: ‘the efficacy of causes, or that quality which makes them be followed by their effect’ (T156), ‘that energy in the cause, by which it operates on its effects’ (T266), ‘when we know a power, we know that very circumstance in the cause, by which it is enabled to produce the effect: For these are supposed to be synonymous.’ (E67-8) Second, a cause has a necessary connection to its effect; central to this relationship is the idea that given the cause, the effect has to happen. As well as referring to this relationship as a necessary connection, Hume also sometimes refers to it as simply a connection or tie. And this defining quality and defining relationship are related because it is the quality in the cause which determines this relationship between cause and effect: ‘we speak of a necessary connexion between objects, and suppose, that this connexion depends upon an efficacy or energy, with which any of these objects are endow'd’ (T162); ‘the very force or energy of the cause, by which it is connected with its effects, and is for ever infallible in its operation.’ (E69)

However, Hume then produces two arguments which undermine this realist notion of causality. Hume's main argument centres on the defining quality of a cause. To better understand the notion of causality, Hume attempts to discover the experience on which the concept of the defining quality is based, or, in Hume's terminology, attempts to discover the impression on which the idea of power, energy, etc is derived. However, there is a problem straightaway.

The point of this demonstration is that if power is a quality in the cause, we should be able to perceive it. But, Hume says, we cannot, and this puts into question the original notion we have of power being a quality in the cause.

Hume's secondary argument centres on the relationship of necessary connexion between cause and effect. It can be set out as follows:

1. (1) The energy of a cause necessarily brings about its particular effect. Hume understands this ‘necessarily’ as meaning that there is an a priori relationship between cause and effect: ‘were the power or energy of any cause discoverable by the mind, we could forsee the effect, even without experience.’ (E63)

2. (2) However, as Hume has argued elsewhere, for example (E27–30), from ‘the first appearance of an object, we never can conjecture what effect will result from it.’ (E63) We can only know what the effect will be from experience.

3. (3) Therefore, Hume concludes, we must be mistaken with our original idea of a cause having energy which necessarily brings about its effect.

So although people ‘suppose that … they perceive the very force or energy of the cause, by which it is connected with its effect, and is for ever in fallible in its operation’ (E69), these two arguments seem to suggest that ‘we have no idea of connexion or power at all, and that these words are absolutely without any meaning, when employed either in philosophical reasonings or common life.’ (E74)

However, this is not Hume's final conclusion. Going back to his main argument, he notes that after one instance of an event being followed by another, we never say they are causally linked; rather it is when two sets of similar events are constantly conjoined that we say one is the cause and the other is the effect. Having the experience of two events being constantly conjoined will not reveal anything new about the two sets of events. The only thing that is new is

And if the idea of power and necessary connexion is derived from an inner impression, a feeling of expectation, then power and necessary connexion refer to this inner impression and not to anything that is in the objects or events themselves: ‘When we say, therefore, that one object is connected with another, we mean only that they have acquired a connexion in our thought.’ (E76)

The same point is made more emphatically in the Treatise:

Hume is aware that people will regard his conclusion a ‘gross absurdity’ (T168) or, in the more restrained language of the Enquiry ‘somewhat extraordinary’ (E76), since the normal view is that force and energy are qualities in causes. He explains away this ‘contrary biass’ (T167) by saying

The same point is made in the Enquiry at the end of E78n.

At the end of his account of causality Hume gives two definitions of causality. These two definitions should be considered as a summary of his account rather than definitions in the ordinary sense. There are two definitions because there are two aspects to his account. First, in terms of our experience of the objects, two objects are causality related if they are constantly conjoined. But as Hume's discussion shows, he is well aware that we mean something more than this when we talk of causality. We think a cause is a cause because it has force and energy which necessarily brings about a particular effect. The origin of these ideas, says Hume, is in the determination of the mind to pass from the idea of one object to the other after the past experience of their constant conjunction.

So Hume has not abandoned his initial notion that a cause has a force and energy which necessarily brings about its effect; only he undermines it. This initial notion of causality is realist because force and energy are thought of as qualities in the cause. But according to Hume's analysis, the ideas of force and energy are not derived from any sensory impression of the objects and so do not refer to any qualities in the objects. Therefore Hume subverts the ordinary notion of causality from a realist one to a non-realist one.

2

In the last few decades there has been a challenge to the traditional interpretation of Hume that all there is to causality in the objects is constant conjunction.Footnote ⁴ The new interpretation argues that Hume believed that energy, power, etc reside in causes. The way it is argued is by referring to Hume's realistic language, but the full argument can be presented as follows:

1. (1) Hume has shown that we do not have any sensory experience of energy in causes.

2. (2) Yet throughout the Treatise and the Enquiry Hume uses terms like energy in a realist way; he uses these terms in a way which clearly assumes that energy resides in the cause. And even in the many passages where Hume says we are ignorant of any energy in the cause, the argument is that: ‘One cannot be ignorant of that which cannot exist!’Footnote ⁵

3. (3) Hume admits that objects may have qualities which are beyond human experience. In the Treatise he says: ‘I am, indeed, ready to allow, that there may be several qualities both in material and immaterial objects, with which we are utterly unacquainted.’ (T168) And in the Enquiry he talks about ‘the weakness and narrow limits of human reason and capacity.’ (E76)

4. (4) Therefore, Hume must be identifying the energy of a cause with a quality in the cause which is beyond our experience. In this way Hume's view that we have no sensory experience of energy in causes is not contradicted by his continual referral to energy in a realist way. According to Wright, Hume ‘consistently maintained the point of view that there are real powers and forces in nature which are not directly accessible to our senses.’Footnote ⁶

The most impressive part of the realist interpreter's case is where they point to Hume's consistent use of terms like energy in a realist way. For example, when Hume talks of ‘the power, by which one object produces another’ (T69), the assumption has to be that Hume is talking about a power which is in the cause, and therefore he believes that power exists in a cause. Winkler, a critic of the realistic interpretation, argues that these passages in Hume should be retrospectively re-interpreted in light of Hume's completed analysis.Footnote ⁷ However I do not think this captures what is going on here; and even after Hume has completed his analysis, his use of terms like power does not change.

Hume's language is clearly realist. He talks of a cause having a force and energy by which is necessarily brings about its effect. This is the ordinary language of causality; it is part of what we mean by causality. Strawson and Wright are correct in emphasising Hume's use of this language. However, Hume's analysis undermines this concept of causality. According to his analysis, the idea of energy does not arise from anything within the cause but from the mind of the observer. Therefore we are mistaken in thinking that energy resides in the cause. If Hume's analysis was applied consistently it would lead to a withering away of the concept of causality; if you accept that energy does not reside in the cause, you would not use the language of causality which implies that energy does reside in the cause. You would be able to talk about regularities but not of causality.

Some philosophers have taken this position. For example, it was strongly argued for by Bertrand Russell, who bolsters his argument by referring to the practice of science. In Russell's view, when science reaches a certain level of sophistication—he is thinking of physics—the language of causality is discarded.

However, Hume does not make this step and he is not consistent in applying his analysis of causality. Even after expounding his account of causality he continues to use the language of causality and thereby implying a realist view of energy. Given his analysis of causality the only way he could rid himself of this inconsistency is by renouncing the language of causality. But Hume does not do this.

In part, Hume acknowledges this inconsistency in the final section of book 1 of the Treatise. We start off, he says, thinking of ‘that energy in the cause, by which it operates on its effect.’ (T266) But his analysis teaches us that ‘energy lies merely in ourselves.’ (T266) The notion that energy resides in the cause is ‘an illusion of the imagination.’ (T267) How far ought we to yield to this illusion, he asks:

Hume is not happy with this answer because it is so dismissive of the endeavour of philosophy. But he does not resolve this issue.

Contrary to Strawson's and Wright's argument, Hume's continuing use of the language of causality does not show that his account of causality is a realist one. Rather there is a basic inconsistency between his use of this language, which necessarily is realist, and his account of causality, which is not realist.

But the strongest argument against the realist interpretation of Hume is that the interpretation is incompatible with empiricism. Since Hume argues that we do not have any sensory experience of energy in the cause, and since the realist interpreters argue that Hume believed energy resides in the cause, it follows, for the realist interpreters, that energy has to be a quality in the cause which is beyond our experience.

This is not a position an empiricist could hold. An empiricist takes the view that all our knowledge of the world, and the concepts we use to describe the world, come from our experience of the world. And since we have a concept of energy, it must have come from our experience of the world.

Strawson is to some extent aware of this problem for the realist interpretation of Hume. He talks of a ‘meaning tension’:

The reason why a term like energy cannot manage to mean anything like energy in the cause is because we have no sensory impression of energy in the cause. In his sharpest statement of this problem, Strawson says that the realist interpretation is incompatible with Hume's theory of ideas. However, Strawson says the realist interpretation is still correct because Hume rejected his theory of ideas, strictly and literally interpreted.Footnote ¹⁰

The part of Hume's theory of ideas that the realist interpretation is incompatible with is the principle that all ideas are derived from impressions. This can be taken as Hume's statement of his empiricism. This principle is annunciated throughout the Treatise and the Enquiry. Hume calls it ‘our fundamental principle’ (T163) and it plays a crucial role in his account of causality. For the realist interpretation to be correct, Hume would have had to reject this principle. Strawson does not claim that Hume said that all ideas are not derived from impressions, or more specifically, that Hume said that the ideas of energy, force, power, etc are not derived from impressions. Rather, Strawson supports his position by citing Hume's use of terms like energy in a realist manner.

If realist interpretation is correct, there would be a massive inconsistency within Hume's philosophy. This would be between Hume's fundamental principle that all ideas are derived from impressions and the view that energy resides in the cause even though we do not have any sensory impression of energy in the cause. If energy resides in the cause and we do not have any sensory impression of energy, then the idea of energy could not be derived from an impression: energy would be ‘not directly accessible to our senses.’Footnote ¹¹ Such a view would undermine everything Hume said about causality. His discussion of causality takes the form of seeking the impression for which the idea of causality and the associated ideas of energy, force, power, etc are derived. The reason why it is wrong to say that Hume rejected his theory of ideas in relation to causality is that Hume claimed to have found the impression on which the idea of energy is derived; only it is not an impression of sensation but an impression of reflection, a feeling of expectation in the observer. And if the idea of energy is derived from an impression of reflection, then energy resides not in the cause but in the observer.

I do not think it is possible to give an interpretation of Hume's account of causality without revealing an inconsistency. I have argued that the inconsistency is between Hume's analysis of causality, which is not realist, and his continuing use of the language of causality, which necessarily is realist. This inconsistency is less damaging to Hume's philosophy than the one revealed in the realist interpretation, and also, I think, gives a more accurate account of Hume's text.

Hume briefly discusses the view that the realist interpretators argue that Hume held, namely that the energy of cause is an unknown quality in the cause. Hume rejects this, because we already have an idea of energy, and this idea cannot be derived from an unknown quality.

Strawson makes a distinction between the epistemological statement:

1. (E) all we can ever know of causation, in the objects, is regular succession,

which Strawson regards as arguably true, and the ontological statement:

1. (O) all that causation actually is, in the objects, is regular succession,

which Strawson regards as wildly implausible.Footnote ¹² Strawson thinks that the traditional interpretation of Hume arose from confusing the epistemological and ontological statements. According to Strawson, Hume believes the epistemological statement but not the ontological statement.

However, although it can be accepted that the two statements have different meanings, no one should accept such drastically different assessments of their truth. To say that the ontological statement is wildly implausible is to say that you have strong reasons for believing that there is more to causation, in the objects, than just regular succession. And this is incompatible with believing that all we can know of causation is regular succession. If it is true that the ontological statement is wildly implausible, then there must be grounds for not accepting the epistemological statement.

Realist interpreters are another example of philosophers responding to Hume's account of causality by rejecting empiricism. Normally philosophers respond by putting forward a different account of causality which involves a break with empiricism. But in the case of the realist interpreters, they have re-interpreted Hume so that Hume is no longer an empiricist!

This motivation is clearly seen in the case of Strawson. He makes clear his own anti-empiricist stance. He declares: ‘our actual concept of Causation has essentially sensory-experience-transcending content,’ and ‘could not possibly be fully revealed in sensory experience.’Footnote ¹³ The problem with Hume's account of causality, as traditionally understood, that Strawson is most concerned with is that if there is no quality in the cause by virtue of which it brings about its particular effect, then

Strawson gives this as a reason against accepting the traditional non-realist interpretation; it is such a compelling argument against Hume, as traditionally understood, that Hume could not possibly have held such a view. But, however highly a philosopher may be held, it does not mean that he is capable of foreseeing all the objections that are subsequently made against his theory. Perhaps, instead of re-interrupting the theory, it should be re-examined.

3

The problem for Hume occurs right at the beginning of his enquiry into causality. He looks at a single instance of causality. He gives the example of one billiard ball banging into another causing it to move. Hume looks at the cause and looks at the effect and can find nothing on which the idea of causality could be based.

But suppose he had taken a different example. Instead of passively contemplating a billiard table, he took an example where he was actively involved, say, in pushing a cart. He would then have a candidate for the origin of the idea of causality, namely the bodily sensation of pushing. This bodily sensation would be his experience of causing or forcing the cart to move. This experience is private to him. But from similar experiences and watching other people perform similar actions, he could readily acknowledge that the other people have similar experiences to him. Also we have experiences of physical bodies acting on us so we know physical bodies can apply force. Say the cart is slowly beginning to roll down an incline, the people in front of the cart can feel the force the cart is applying to them. And even when the physical object is not applying force to us, but to another physical object, we generalize from our own experiences and accept the physical object is applying force to other physical objects, for example, when the run away cart bangs into and knocks over a fence in its way.

So, to go back to Hume's original billiard ball example, it is true we can not see the white ball forcing or causing the red ball to move; the experience of forcing is not a visual experience. However, from our experiences of physical objects acting on us, we understand and accept that the white ball is exerting a force on the red ball. There is nothing mysterious about this. We could feel the force the white ball applies to the red ball by simply putting our hand between the white and the red ball. When the white ball hits your hand you feel a pressure on your hand, which is the force it would have hit the red ball with.

Hume does not consider the possibility that a bodily sensation of forcing might be the original of the idea of causality in the main body of the Treatise. But he does consider something similar first briefly in the Appendix to the Treatise (T632-3) and then at greater length in the Enquiry (E64-69)

To argue against this attack on his main argument in the Treatise, i.e. the argument that we have no idea of energy in the cause because we cannot see the energy, he relies on the secondary argument, involving necessary connection, which only briefly appears in the main body of the Treatise. (T161-2) This argument is that we think the energy of a cause necessarily brings about its effect (and Hume understands this ‘necessarily’ as implying an a priori relationship between cause and effect); but from the cause we can never know a priori the effect; therefore we are deceived in thinking we have an idea of energy residing in the object which necessarily brings about its effect.

Hume uses this argument in the appendix to the Treatise and the Enquiry against the suggestion that in the performing of actions we acquire the idea of force and energy. In most of the cases Hume considers the cause is an act of volition and the effect is the corresponding movement of the body. Hume considers the suggestion that in the act of willing we feel an energy and force and this is the source of the idea of causality. I do not believe that this experience is the source of our idea of causality but the much more straightforward experience of a person's action bringing about change in the external world. However, one example Hume considers is similar to that of pushing a cart.

The argument he uses against this, is the same argument he uses in the cases of an act of will causing the movement of a part of the body, so all the cases can be considered together.

Hume argues that the act of volition gives us no privileged knowledge of the effect. As with other cases of causality, we only know what effect a cause will have by experience. For example, if a man has a stroke it is only by trying and failing to move his arm that he discovers he has lost the use of his arm. (E66) ‘But were the power or energy of any cause discoverable by the mind, we could foresee the effect, even without experience.’ (E63) Since we cannot foresee the effect without experience, we cannot be conscious of any power or energy.

What is striking about the argument is that the premise that there is an a priori relation between cause and effect conflicts with the premise that from the cause we can never know a priori the effect. Accepting the second premise as correct, which Hume certainly does, you would think the conclusion Hume would draw is that the first premise is not true; that if the energy of a cause necessarily brings about the effect, this ‘necessarily’ should not be understood in the sense that it implies an a priori relationship between cause and effect. However, this would be a mistake in interpretation. It is clear that Hume held the view that the necessity involved in causality is that of an a priori relation between cause and effect. For example Hume says: ‘We must distinctly and particularly conceive the connexion betwixt the cause and effect, and be able to pronounce, from a simple view of the one, that it must be follow'd or preceded by the other.’ (T161) And: ‘were the power or energy of any cause discoverable by the mind, we could forsee the effect, even without experience.’ (E63)

This has to be understood historically. Before Hume there was a tradition in philosophy which took the view that there was an a priori relationship between cause and effect.Footnote ¹⁵ The demise of this tradition was largely due to Hume. When Kant said that Hume interrupted his ‘dogmatic slumber’Footnote ¹⁶ he was referring to the fact that Hume ‘indisputably proved that it is wholly impossible for reason to think such a connection [between cause and effect] a priori.’Footnote ¹⁷ Kant took from Hume that in the individual case we have no a priori knowledge of the effect from the cause; nevertheless he argued that the concept of causality was a priori and was not ‘as Hume had feared, derived from experience.’Footnote ¹⁸

However, despite arguing strongly that, from the cause, we can never a priori know the effect, Hume retained the notion that part of what we mean by causality is that there is an a priori connection between cause and effect. In doing this, Hume is being inconsistent. This is similar to his inconsistency in holding that part of what we mean by causality is that a cause has energy by which it brings about its effect and arguing that energy does not reside in the cause.

Once the notion that part of what we mean by causality is that there is an a priori connection between cause and effect is rejected, then Hume's argument collapses. There can still be a necessary connection between cause and effect, but this necessary connection should not be understood as implying an a priori relationship between cause and effect. But in examining the argument, we should not ignore the nonsensical nature of Hume's conclusion. Comparing a man in perfect health to a man suffering from a stroke, he says: ‘neither in the one case nor in the other, are we ever conscious of any power.’ (E66) And to show that this is not some kind of slip on Hume's part, he repeats the point a little later. ‘How indeed can we be conscious of a power to move our limbs, when we have no such power.’(E66) Transferring his argument to the type of example I would prefer, Hume would have to say that a person pushing a cart is not conscious of exerting a force on the cart. But since part of what ‘pushing’ means is ‘exerting a force’, it is nonsense to say that a person pushing a cart, and is conscious of doing this, is not conscious of exerting a force.

At the end of his discussion, Hume relents on denying the obvious, namely that we have the experience of applying force. In discussing an example similar to my own of ‘the resistance which we meet with in bodies, obliging us frequently to exert our force and call up all our power’ (E67n)—he states:

But there are not two different ideas of power, one vulgar the other not. There is only one concept of power, which everyone shares.

Hume is correct in saying that we can only know the effects of our actions from experience; but this does not mean that we do not have the experience of applying force. A person may apply a force but it is only from experience that he discovers whether the effect on a certain type of object is that the object moves, or, for example, that it bends or it breaks.

In opposition to Hume, a similar view was put forward by Thomas Reid as to the empirical origin of our idea of causality.

However, Reid also thought that it is only with regards voluntary action that the concept of cause should be used. In relation to everything else, Reid took a very non-common sensical position and held that it was incorrect to use the language of causality. Discussing an example of holding a magnet close to a nautical compass and observing the compass needle to move, he acknowledges that people would say that the magnet caused the compass to move. Yet Reid's position is that: ‘We deceive ourselves if we conceive that we can point out the real efficient cause …’Footnote ²⁰ Reid's position was similar to Russell's in that he thought physicists discovered laws, i.e. regularities, and not causes.

However, if Reid accepts the experience of a voluntary action forcing something as the origin of the concept of causality, then he should also accept the experience we have when physical objects exert force on us, say the feeling of pressure on your hand when you stop a billiard ball. If the appeal is to experience, then the experience of acting on the external world and the experience of being acted on by the external world are equally valid.

4

Hume started with the everyday realist notion of causality in which the cause has energy residing in it. However, because he was unable to find any sensory experience we have of energy in the cause and to stay true to his empiricism, he ended up with a theory that the origin of our idea of energy lies in the feeling of expectation in the observer. In this way Hume ends up with a non-realist account of causality where causality in the objects is reduced to constant conjunction.

The realist interpreters have re-emphasised the importance of force and energy. However, because they have followed Hume in accepting his view that we have no sensory experience of force and energy, they have ended up with the notion of force and energy as being non-sensory qualities in the cause. As an interpretation of Hume this is not credible because on this account Hume would no longer be an empiricist. It is also an enfeebled account of causality since, as non-sensory qualities, force and energy cannot play any role in understanding the world. Strawson admits as much: ‘Such terms [power, energy, force, etc.] can have no role to play in the explanation or understanding of particular phenomena’ and are ‘of no use to physics’.Footnote ²¹ But in reality, force and energy are fundamental concepts in physics.

I have tried to follow Hume in giving a strictly empirical account. But by rejecting Hume's view that we have no sensory experience of force and energy, I am able to show how it is possible to give an empiricist and realist account of causality.