Did Kant appreciate Hume?

Perception and repetition as separate aspects of experience

Ilya Bernstein

1. Introduction

Although Kants objections to Hume were ostensibly a response to Humes argument about causality, at bottom it was not the concept of causality that they focused on, but the concept of experience. As is well known, Hume argued that claims about causal relations were «not, in any instance, attained by reasonings a priori; but [arose] entirely from experience» (Hume [1758], 4.6: 27). From this argument, Kant inferred a specific conception of experience, which he attributed to Hume. He then took issue with this conception of experience and proposed a different, enriched conception, which at the same time nullified Humes claims about causality. In the debate between Kant and Hume, then, we can distinguish between three potentially different notions of experience: the enriched conception of experience proposed by Kant (E1), the conception of experience imputed by Kant to Hume (E2), and the conception of experience which Hume actually held (E3).

My aim in this paper will be, first, to define E2 and E3 and to argue that there is indeed a significant difference between them. Second, I will explore E3 itself in greater depth by looking at a related conception of experience that was developed in British legal theorya conception of experience that has strong links to Humes from a conceptual point of view, if not necessarily from a genealogical one. Third, I will revisit the difference between E2 and E3 and consider what it was that Kant failed to appreciate when he attributed E2 to Hume, and what it was that he still failed to appreciate when he proposed E1 as a replacement for E2. My goal throughout this paper both in setting up this opposition between Hume and Kant, and in drawing parallels between Hume and the common law jurists will be to delineate Humes conception of experience by highlighting what I take to be its most important and original component.

2. What Kant thought Hume thought about experience and what Hume really thought

«Experience», to Kant, to Hume, and to the philosophical traditions on which they drew, denoted a complex phenomenon consisting of a bundle of different attributes. Kant added to this complexity by arguing that experience had another feature an a priori structure, which shaped all incoming information which Hume and indeed every other philosopher had failed to take into account. But even before Kant superadded this feature to it, the concept of experience was already sufficiently complex. Let me begin by unpacking this complexity into what I take to be its two most salient components. First, «experience» refers to a form of knowledge acquisition that is grounded in perception. Such perception may be oriented outward or inward as in Lockes distinction between «sensation» and «internal sense» (Locke [1690], 1.2, 1.4: 79) but in either case the knowledge that is obtained through it is obtained by observation rather than deduction. That is, «experience» requires that we employ our perceptual apparatus, rather than some other part of our cognitive machinery. Second, «experience» implies a form of knowledge acquisition that takes place over time through exposure to repetition, through a process of trial and error, through the gradual accumulation of memories. In other words, it is not knowledge that can be had suddenly or all at once: the passage of time is an essential precondition of it.

There is a natural complementarity between these two features of experience knowledge acquired through perception is necessarily acquired over time and the distinction that I make here was not made explicitly either by Kant or Hume or the philosophical traditions on which they drew. Both aspects of experience roughly, experience-as-perception and experience-as-repetition are present in E1, E2, and E3. Indeed, both of these dimensions are still present in the term «experience» when we employ it in common usage: experience is a knowledge acquisition process that involves perception and experience is a knowledge acquisition process that takes place over time. What I would like to do in this paper, however, is precisely to differentiate between these two sides of the concept of experience and to argue thatalthough neither Kant nor Hume distinguish between them explicitly, let alone set up a conceptual opposition between themthe two authors nevertheless emphasize two different aspects of experience in their philosophies, with Kant putting experience-as-perception at the foundation of E1 and Hume putting experience-as-repetition at the foundation of E3. Furthermore, I will argue that Kant mistakenly imputed his own emphasis on experience-as-perception to Hume, making it the foundation of E2, and that the objections that Kant raised against Hume therefore ignore what is arguably Humes most original contribution to epistemology.

The flavor of what Kant had in mind when he wrote about the concept of experience is well conveyed by the opening paragraph of the Critique of Pure Reason:

There is no doubt whatever that all our cognition begins with experience; for how else should the cognitive faculty be awakened into exercise if not through objects that stimulate our senses and in part themselves produce representations, in part bring the activity of our understanding into motion to compare these, to connect or separate them, and thus to work up the raw material of sensible impressions into a cognition of objects that is called experience? (Kant [1781]: 136)

«Objects that stimulate our senses», «produce representations», «the raw material of our sensible impressions»: everything here points to an image of experience as a kind of outgrowth of perception. The senses are prominently foregrounded. Even the reference to the non-purely perceptual side of experience its capacity to «bring the activity of our understanding into motion to compare these [representations], to connect or separate them» posits as the aim of experience «to work up the raw material of our sensible impressions into a cognition of objects». Objects, representations, impressions, the senses: this is the conceptual frame of reference for Kants definition of experience. Indeed, it is the traditional frame of reference used for talking about experience in epistemology1. We are in the world of «perception vs. reason», and we are interested in experience insofar as it relates to the distinction between these phenomena. 

Compare this now to the conceptual world we enter when we turn to Humes discussion of experience. The flavor of his thinking emerges with particular clarity when he describes experience as the basis of our idea of causality, as in the following quote:

[I]n all single instances of the operation of bodies or minds, there is nothing that [...] can suggest any idea of [...] necessary connexion. But when many uniform instances appear, and the same object is always followed by the same event; we then begin to entertain the notion of cause and connexion. We then feel a new sentiment [...] to wit, a customary connexion in the thought or imagination between one object and its usual attendant. (Hume [1758], 7.30: 78)

The frame of reference for thinking about the concept of experience is announced at once: it is no longer the opposition between deduction and observation, but the opposition between “single instances” and “many uniform instances”. Hume goes on:

[As the] idea [of necessary connection] arises from a number of similar instances, and not from any single instance, it must arise from that circumstance, in which the number of instances differ from every individual instance. But this customary connexion or transition of the imagination is the only circumstance in which they differ. In every other particular they are alike. The first instance which we saw of motion communicated by the shock of two billiard balls [...] is exactly similar to any instance that may, at present, occur to us; except only, that we could not, at first, infer one event from the other; which we are enabled to do at present, after so long a course of uniform experience. (Hume [1758], 7.30: 78-79)

We are very far here from any emphasis on perception as the foundation of experience: perception is not mentioned at all. What Hume is interested in, rather, is the form of knowledge acquisition that becomes available to us not through perception, but through repetition. It is this form of knowledge acquisition that is specifically associated for him with experience. When we see two billiard balls collide over and over again, according to Hume, our perception of the event does not change; yet our experience does. This is because experience depends not only on our powers of observation, which remain constant, but also on our ability to register repetition to differentiate between the «individual instance» and the «number of similar instances», between the single event and the «long [...] course of experience». It is this second aspect of experience that most interests Hume as can be seen, for example, from the kinds of arguments in which he invokes the concept of experience, such as his argument about causality. For Hume, experience is fundamentally grounded in the passage of time and in our sensitivity to differentiations between degrees of repetition or in what I will characterize below as our differential capacity for knowledge acquisition over different time scales.

Let us turn to Kant again for another comparison. In the second paragraph of the Critique, he attempts to formulate a crisp criterion «by means of which we can security distinguish a pure cognition from an empirical one», and this criterion further illuminates the core of his conception of experience, a core that remains unchanged in the transition from E2 to E1:

At issue here is a mark by means of which we can security distinguish a pure cognition from an empirical one. Experience teaches us, to be sure, that something is constituted thus and so, but not that it could not be otherwise. First, then, if a proposition is thought along with its necessity, it is an a priori judgment [...]. Second: Experience never gives its judgments true or strict but only assumed and comparative universality (through induction), so properly it must be said: as far as we have yet perceived, there is no exception to this or that rule. Thus if a judgment is thought in strict universality, i.e. in such a way that no exception at all is allowed to be possible, then it is not derived from experience, but is rather valid absolutely a priori [...]. Necessity and strict universality are therefore secure indications of an a priori cognition, and also belong together inseparably. (Kant [1781]: 137-138)

To single out necessity and strict universality as the defining attributes of a priori knowledge, as Kant does, is to imply a pair of opposite attributes as the defining characteristics of empirical knowledge. If knowledge that is had a priori is necessary and universal, then knowledge that is acquired through experience must be contingent and partial. Kants emphasis on these two particular aspects of empirical knowledge, as implied by his definition of a priori knowledge, is revealing. I would argue that it points, once again, to a model of experience that is founded on perception. It is perception that is necessarily contingent, due to the fact that the point of view of the perceiving subject is necessarily bounded in time and space; and it is perception that is necessarily partial, due to the fact that it is bounded by the limitations of our sensory organs. By transferring these two hallmarks of perception contingency and partialness to his conception of experience in general, both in E1 and E2, Kant again sheds light on the inspiration behind that conception of experience: his empirical knowledge is at bottom the knowledge of the perceiving subject endowed with organs of sense.

Hume also offers something of a criterion for distinguishing between knowledge that is had a priori through reason and knowledge that is acquired through experience, and the difference between his criterion and Kants could not be more eloquent:

From causes which appear similar we expect similar effects. This is the sum of all our experimental conclusions. It seems evident, that, if this conclusion were formed by reason, it would be as perfect at first, and upon one instance, as after ever so long a course of experience. But the case is far otherwise. Nothing so like as eggs; yet no one, on account of this appearing similarity, expects the same taste and relish in all of them. It is only after a long course of uniform experiments in any kind, that we attain a firm reliance and security with regard to a particular event. Now where is that process of reasoning, which, from one instance, draws a conclusion, so different from that which it infers from a hundred instances, that are nowise different from that single one? (Hume [1758], 4.20: 36)

With this sweeping rhetorical question, Hume effectively implies that the one criterion by which we can always distinguish a conclusion formed by reason from a conclusion formed by experience is by seeing whether repeating the premise which gave rise to the conclusion adds anything to the conclusion or not: if it does, then the conclusion was formed by experience; if it does not, then the conclusion was attained through reason.

As can be seen, Hume approaches the difference between experience and reason from a completely different angle than Kant. The distinctive hallmarks of experience are no longer those of perception and the constraints of the perceiving subject; rather, they are those associated with the process of repetition and the manner in which knowledge is acquired over time. In fact, although he himself does not wish to cut the link between experience and perception, one can see how such an approach to experience as the one which he develops in this passage can suggest a conception of experience that dispenses with any mention of perception altogether and lead to a definition of experience that does not rely on perception at all, but is formulated exclusively in terms of the relation between knowledge and the passage of time. This, in my view, is the most important and original implication of Humes epistemologyand one completely overlooked by Kant, whose own notion of experience could never be severed from perception in this way.

Thus, we have articulated the basic outlines of the distinction between E1 and E2, on the one hand, and E3, on the other. It is, of course, a distinction that is neither strict nor made explicitly by either of the two authors. And it would not be difficult to argue that Kant and Hume are simply using two different approaches to talk about what is really the same thing. But my aim here is precisely to call attention to the difference in their approaches.

That Kant imputes his own approach to Hume is made evident not only by the manner in which he characterizes E2, but also by the fact that there is no difference in this respect between E2 and Kants own, enriched and revised, version of the concept of experience, E1in both cases, experience retains a pronounced affinity with perception. «My reader», he writes in the Prolegomena, «[...] has probably been long accustomed to consider experience a mere empirical synthesis of perceptions, and hence not considered that it goes much beyond them, as it imparts to empirical judgments universal validity, and for that purpose requires a pure and a priori unity of the understanding» (Kant [1783]: 75). Here we have both E2 and E1, succinctly summed up in the span of a single sentence, with the conception of experience as a «mere empirical synthesis of perceptions» which Kant will explicitly attribute to Hume in the following paragraph rejected in favor of a conception of experience that also relies on the «pure and a priori unity of the understanding».

Roughly speaking, Kant sees experience as being positioned between perception and reason; and what he is concerned to show is that experience is not wholly on the side of perception, as it has traditionally been seen «my reader [...] is recommended to pay special attention to this distinction of experience from a mere aggregate of perceptions» but that it also has a structural component that can come from nowhere other than reason. Thus, the new conception of experience that he offers as a response to what he regards as Humes challenge straddles the opposition between perception and reason, rather than being bound to perception alone. But in neither of these alternatives (E1 or E2) is the possibility ever entertained that experience might be fruitfully viewed not in terms of the opposition between perception and reason, but in terms of the opposition between repetition and reason, as demonstrated by Hume (E3). Kant goes on:

Now we are prepared to remove Humes doubt. He justly maintains, that we cannot comprehend by reason the possibility of causality, that is, of the reference of the existence of one thing to the existence of another, which is necessitated by the former [...].But I am very far from holding [this concept] to be derived merely from experience, and the necessity represented in [it], to be imaginary and a mere illusion produced in us by long habit. On the contrary, I have amply shown, that [it] and the theorems derived from [it] are firmly established a priori, or before all experience. (Kant [1783]: 75)

As can be seen, Kant does not deny that repetition is involved in experience. He recognizes the phenomenon of habit. But this phenomenon does not appear to be of philosophical interest to him. It is neither something that needs to be explained nor something that itself has significant explanatory power. Rather, it is the phenomenon of perception that he focuses on in arguing against Humes notion of experience, and he implicitly imputes such a focus on perception to Hume as well. Yet as I have been suggesting, Humes own focus is actually and fundamentally on the phenomenon of habit, not perception. Habit, custom, repetitionthese are the counterweights to reason in Humes epistemology, all of them subsumed in his notion of experience, and it is their origins, properties, and influences that Humes epistemology seeks to explain2.

It would be misguided, I believe, to attempt to define a formal difference between E2 and E3, because neither of these approaches to the concept of experience rules out the other. What I am attempting to bring out is the difference in the spirit of Kants and Humes viewsa difference that may seem subtle enough in and of itself, but whose implications are extremely far-reaching. It is to these implications that I will now turn.

3. The Epistemology of the Common Law

My aim above has been to show that Kant was making a specific interpretative choice when he read Humes notion of experience as he did, and that it was this interpretative choice that led him to transform this notion in the way that he did. In making this choice, of course, Kant was not stepping off of any beaten path in philosophy. On the contrary, he was following the traditional epistemological interest in perception and in knowledge acquired through the sensory apparatus as opposed to knowledge acquired through the deductive faculties. Traditionally, however, the two contrasting aspects of experience which I have been highlighting were never deliberately distinguished, and authors indiscriminately emphasized both of them as the need for one or the other arose, depending on context3. As a very loose generalization, it may be possible to argue that perception traditionally tended to be foregrounded as a component of experience in epistemology, as part of an interest in the cognition of the individual, whereas repetition traditionally tended to be foregrounded as a component of experience in political theory and ethics, as part of an interest in tradition, law, custom. I do not want to make that argument here. What I want to do, rather, is to turn to the field of thought in which the term «experience» was used to mean exclusively what I have argued it meant for Hume, and in which there was a centuries-long tradition of explicitly conceiving of experience as the result of the passage of time. That field of thought is British legal theory.

The argument which I will make below will not be historical in nature. I will not claim that Hume read the specific texts from which I am going to quote, nor that it was his deliberate aim to introduce the principles formulated by common law jurists into epistemologyalthough the fact that he was thoroughly acquainted with the British legal theoretical tradition is evident from his own writings on ethics and, especially, British history (see McArthur [2005]: 67-82). My purpose in going over this background will be rather to flesh out the implications of the notion of experience espoused by Hume by drawing on that intellectual tradition which articulated and developed these ideas most fully. And though stopping short of a concrete historical argument, I will make the claim that Humes thought marked the first full-fledged penetration of these ideas into epistemology.

The salient difference between knowledge acquired through experience and knowledge acquired through reason for the common law jurists was that reason belonged to one person, while experience belonged to many people in many generations. In other words, knowledge acquired through reason could be acquired by one person, and no matter how many other people acquired this knowledge, and over what span of time, it would remain the same; but knowledge acquired by experience could not by definition be acquired by a single person, but was defined as knowledge that could be acquired only by many people over time. To get a sense of how the term «experience» was used in the British legal tradition, we can begin by turning to some representative passages from the writings of Edward Coke.

We are but of yesterday (and therefore had need of the wisdom of those that were before us) and had been ignorant (if we had not received light and knowledge from our forefathers) and our dayes upon the earth are but as a shadow, in respect of the old ancient dayes and times past, wherein the Laws have been by the wisdom of the most excellent men, in many successions of ages, by long and continual experience (the trial of right and truth) fined and refined, which no one man (being of so short a time) albeit he had in his head the wisdom of all the men in the world, in any one age could ever have effected or attained unto. (Coke [2003], vol. 1: 173)

The opposition here is between «any one age» and «many successions of ages» the short run and the long run and the different kinds of knowledge that can be acquired in each case. Experience is tellingly characterized as «long and continual», and this contrast between a knowledge acquisition process that is long and continual, on the one hand, and a knowledge acquisition process that is sudden, on the other, is a theme that runs through all of Cokes thought:

[I]n these dayes of many it may be justly said, Quod statim sapiunt, statim sciunt omnia, neminem verentur, imitantur neminem, ipsi sibi exempla sunt [that they are instantly wise, instantly know everything, respect no one, imitate no one, set their own precedents]. But it is safe for the Client and for the Councellor also [...] to follow Presidents formerly approved and allowed, and not to trust to any new frame carved out of his owne invention, for Nihil simul inventum & perfectum est [nothing is invented and perfected at the same time]. (Coke [2003], vol. 2: 561)

Conceptually, there is a direct correspondence between Cokes differentiation of knowledge that is «instant» and knowledge that is «not instant», and Humes temporal approach to reason and experience, with the former defined as that form of knowledge acquisition whose conclusions are «as perfect at first, and upon one instance, as after ever so long a course of experience» (Hume [1758], 4.20: 36), and the latter defined as that form of knowledge acquisition whose conclusions can be enriched over time.

Of course, Coke and the other jurists had a much narrower range of knowledge in mind when they wrote about reason and experience than did Hume, Kant, and other philosophers: the jurists were interested exclusively in legal knowledge, i.e. knowledge of laws, precedents, legal procedure, and the concrete facts surrounding the cases that came before them. But what «experience» meant to them, when it was opposed to reason, had little to do with the specific cognitive faculty that people relied on in making legal decisions; rather, it had to do with relying on knowledge gathered over many generations, as opposed to relying on knowledge accessible to an individual mind. To quote Coke again:

And therefore if all the reason that is dispersed into so many severall heads were united into one, yet could he not make such a Law as the Law of England is, because by many successions of ages it hath been fined and refined by an infinite number of grave and learned men, and by long experience growne to such a perfection, for the government of this Realme, as the old rule may be justly verified of it, Neminem oportet esse sapientiorem legibus [no man ought to be wiser than the laws]: No man (out of his owne private reason) ought to be wiser than the Law, which is the perfection of reason. (Coke [2003], vol. 2: 701)4

Note that the distinction being made by Coke is ultimately not between knowledge whose origins are more recent and knowledge whose origins are more remote in time, but between knowledge acquired over a large time scale and knowledge acquired over a small time scale. Thus, tradition, in British legal theory, is conceived of as deriving its authority not so much from the fact that it conveys knowledge handed down from the past, as from the fact that it conveys knowledge that has endured over a long period of time. This orientation toward thinking about knowledge in terms of different time scales becomes even more explicit in Matthew Hale:

Again it is a reason for me to preferre a Law by which a Kingdome hath been happily governed four or five hundred yeares then to adventure the happiness and Peace of a Kingdome upon Some new Theory of my owne tho I am better acquainted with the reasonableness of my owne theory then with that Law. Again I have reason to assure myselfe that Long Experience makes more discoveries touching conveniences or Inconveniences of Laws then is possible for the wisest Councill of Men at first to foresee. And that those amendments and supplements that through the various Experiences of wise and knowing men have been applyed to any Law must needs be better suited to the Convenience of Laws then the best Invention of the most pregnant witts not ayded by such a series of tract of Experience. (Pocock [1987]: 172)

J.G.A. Pococks The Ancient Constitution and the Feudal Law, in which this passage from Hale is quoted, contains the classic exposition of the common law conception of experience5. The jurists whom Pocock cites invariably regard experience as a knowledge acquisition process that depends fundamentally on the passage of time. Perception a key component of experience for the epistemological tradition in philosophy and for Kant in particular is not mentioned as an aspect of experience at all. Indeed, since legal knowledge is abstract in character, actual perception necessarily plays only a subordinate role in its acquisition. The opposition between experience and reason thus necessarily boils down not to an opposition between perception and deduction, but to an opposition between «theories» to use Hales word that are of short standing and those that are of long standing:

As it is said of every Art or Science, which is brought to perfection, Per varios usus Artes experientia fecit [experience, through many trials, made the art; so may it properly be said of our Law, Per varios usus Legem experientia fecit. Long experience, and many trials of what was best for the common good, did make the Common Law6.

«Long experience and many trials»: that is the cognitive alternative to reason or to «natural reason» as the British legal theorists referred to it rather than any use of the sensory apparatus.

What we see here is something that might be described as the epistemology of the common law. It is an epistemology that categorizes different types of knowledge in terms of the speeds at whichor the time scales over whichknowledge is acquired. The distinction between gradual knowledge acquisition and sudden knowledge acquisition becomes elevated in it to a fundamental principle. And although, as noted above, the actual knowledge that the common law jurists were concerned with was very narrow in scope, pertaining exclusively to the law, the model of knowledge that they formulated was remarkable for the prominence that it gave to the role played in cognition by the passage of time.

4. The Common Law Flavor of Humes Epistemology

It is this model of knowledge, I would like to argue, that became generalized in Humes philosophy into an explicit epistemology. What the common law jurists claimed about knowledge of the law, Hume claimed about knowledge in general. In both cases, knowledge was seen as being acquired through experience, and in both cases, experience was understood to mean principally not knowledge acquisition through perception, but knowledge acquisition over time. Between Daviess emphasis on «many trials» as an aspect of experience, cited above, and Humes emphasis on «repetition» as an aspect of experience, there is a direct conceptual link. And insofar as Humes philosophy constitutes a form of empiricism, it should be regarded as an empiricism founded on a notion of experience that has less to do with the senses than it does with the passage of time.

Nothing expresses the affinity between the epistemology developed by the common law jurists and the epistemology formulated by Hume more succinctly than the fact that the term which Hume used to denote the fundamental principle of his epistemology had a long history of usage in British legal theory, in which it referred to the cornerstone of the British legal theorists model of the law. That term, of course, was «custom»7. Only where the common law jurists had used it literally, to refer to customs, traditions, and the cultural and institutional heritage of a people, Hume employed the term «custom» abstractly, and thus far more broadly, to encompass not only everything that it encompassed for the common law jurists, but also the habits of individuals, and not merely those habits which were observable in their conduct, but also those that were present in their cognitive activities. The definition of custom in the Enquiry is formulated in the most abstract manner:

[W]herever the repetition of any particular act or operation produces a propensity to renew the same act or operation, without being impelled by any reasoning or process of the understanding, we always say, that this propensity is the effect of Custom. By employing that word, we pretend not to have given the ultimate reason of such a propensity. We only point out a principle of human nature, which is universally acknowledged, and which is well known by its effects. Perhaps we can push our enquiries no farther, or pretend to give the cause of this cause; but must rest contented with it as the ultimate principle, which we can assign, of all our conclusions from experience. (Hume [1758], 5.5: 43)

Everything that we know from experience, Hume concludes, we know, effectively, through custom. Let us note parenthetically once again that perception, although undoubtedly a component of experience, has no role to play in this crucial discussion: as for the theorists of the common law, the mechanism behind experience the mechanism which gives experience access to knowledge that deductive reasoning lacks is custom. By contrast with the common law jurists, however, Hume has expanded this category of knowledge from specifically legal knowledge to all knowledge in general, and even more significantly, he has abstracted the idea of custom from the historical handing-down of traditions in a society and has related it to the workings of the individual mind.

Cokes and Hales model of a social process of knowledge acquisition is thus applied to individual cognition. And in both cases, the classification of knowledge into different types is based on essentially the same logic. The common law jurists, as we have seen, distinguish between knowledge that is accessible on the time scale of the individual and knowledge that is accessible on the time scale of «many successions of ages». The power of the common law, to their way of thinking, comes from the fact that it transcends the temporal limitations of the individual and thus has access to knowledge whose «reasonableness» an individual might not even be in a position to discern, yet which is superior to the «reasonableness of [his] owne theory» because it has stood the test of time. Hume frames the problem of cognition in much the same terms:

Even after one instance or experiment where we have observed a particular event to follow upon another, we are not entitled to form a general rule, or foretell what will happen in like cases; it being justly esteemed an unpardonable temerity to judge of the whole course of nature from one single experiment, however accurate or certain. But when one particular species of event has always, in all instances, been conjoined with another, we make no longer any scruple of foretelling one upon the appearance of the other, and of employing that reasoning, which can alone assure us of any matter of fact or existence. (Hume [1758], 7.27: 74)

Just as in the common law, it is contact with different time scales that distinguishes one type of knowledge from another. An event that takes place once tells us nothing. It is only after the same event is repeated on multiple occasions, over a longer time scale, that it becomes the grounds for making generalizations and predictions about events that are similar to it. The parallel between the conceptual tools used by Hume and the common law jurists to talk about knowledge is evident: both of them are attentive above all to the relationship between knowledge and the passage of time.

The manner in which Hume formulates the distinction between experience and reason is based on the same logic as well: as we have seen, they are differentiated principally in terms of the different roles played in each of them, as knowledge acquisition processes, by the passage of time:

[A]fter the constant conjunction of two objects heat and flame, for instance, weight and solidity we are determined by custom alone to expect the one from the appearance of the other. This hypothesis seems even the only one which explains the difficulty, why we draw, from a thousand instances, an inference which we are not able to draw from one instance, that is, in no respect, different from them. Reason is incapable of any such variation. The conclusions which it draws from considering one circle are the same which it would form upon surveying all the circles in the universe. But no man, having seen only one body move after being impelled by another, could infer that every other body will move after a like impulse. All inferences from experience, therefore, are effects of custom, not of reasoning. (Hume [1758], 5.5: 43).

Reason needs one instance and additional instances add nothing to it; experience requires multiple instances, and the knowledge acquired through experience changes as the number of instances grows, even if they are identical.

What Hume is at pains to capture is the effect of repetition on knowledge acquisition. He effectively poses the question: What is it that changes for knowledge when something is repeated without any changes? What is it that happens to knowledge when nothing changes, in other words, except for time passing? The only thing that changes, with repetition, is the time scale at which the knowledge endures. And in accordance with the time scale at which it endures, Hume concludes, knowledge itself changes as well.

It is this concern with the relationship between knowledge and time that, I submit, is at the heart of Humes epistemology, and it is this concern that is expressed in his conception of experience, on which his argument about causality is based. What Hume wanted to call attention to, therefore, when he argued that our knowledge of causality was based on experience, was not so much that none of this knowledge was ours innately or a priori, which was what Kant took his principal meaning to be, but rather that all of this knowledge depended on passage of time and on the repetition of events that were identical in all respects except for their number. This implication of Humes argument about experience failed to make a strong impression on Kant. But by arguing that our knowledge was based on repetition and our uncertainty on lack of repetition, Hume put time at the center of his epistemology and in this way effectively proposed a new conception of cognition.

5. Humes Idea of «Synthetic A Priori Knowledge»

As I have suggested above, Kant made a particular interpretative choice when he understood Humes notion of experience as he did, as being oriented predominantly toward perception rather than repetition. It was this interpretative choice that led Kant to respond to Hume as he did, by proposing an enriched conception of experience that was grounded not only in perception, but also in pure reason. Had Kant made a different interpretative choice in reading Humes notion of experience, and had he seen it as being oriented predominantly toward repetition rather than perception which, as I have been arguing, was Humes own approach he might not have been as satisfied with his response to Humes arguments as he was. Because the response that Kant gave to Humes arguments, while making the point that perception had to have an a priori structure that was not itself the result of perception, and that experience had to have an a priori structure that was not itself the result of experience, nevertheless failed to say anything illuminating about what I take to be the deepest implication of Humes epistemology, namely, that different kinds of knowledge emerge over different time scales. Thus, the structure that Kant envisioned experience as having was given in experience once and for all: a little bit of experience already contained the entire «synthetic a priori» apparatus, which was operating in it and shaping it, and no additional experience would change or add anything to that apparatus, as it had been given from the first. Hence, synthetic a priori judgments, to Kants way of thinking, expressed a static form of knowledge, for which the passage of time made absolutely no difference; our synthetic a priori judgments were always with us, and what they were at the beginning, they would be at the end. It seems not to have occurred to Kant that the a priori structure of cognition might operate differently, or manifest different features, over different periods of time. For him, what synthetic a priori knowledge was in the short term, it was also in the long term: no differentiation was made between the two.

But this was precisely Humes insight: the implication of his epistemology was that cognition had different properties at different time scales. Hume did not deny that cognition had a certain a priori structure. He called it «instinct». If custom or habit plays as important a role in our cognition as Hume contends it does, then where does our tendency to form habits come from? Is it itself a habit? No, says Hume, our tendency to form habits is not a habit. Rather, it is a «principle of human nature», in other words, a part of the structure of our cognition that precedes experience and cannot be explained by experience. Although Hume does not characterize our ability to form habits as a form of «synthetic a priori» knowledge, what he refers to using his more colloquial terminology is effectively the same thing: custom for him is a cognitive faculty that shapes our experience without itself being the result of experience. As he writes in the passage already quoted above:

By employing that word [custom], we […] point out a principle of human nature […] Perhaps we can push our enquiries no farther, or pretend to give the cause of this cause; but must rest contented with it as the ultimate principle, which we can assign, of all our conclusions from experience. (Hume [1758], 5.5: 43)

Elsewhere, both in the Enquiry and in the Treatise, Hume explicitly identifies this “ultimate principle” with instinct as it is observed in animals. As he writes in the Enquiry:

But though animals learn many parts of their knowledge from observation, there are also many parts of it, which they derive from the original hand of nature; which much exceed the share of capacity they possess on ordinary occasions; and in which they improve, little or nothing, by the longest practice and experience. These we denominate Instincts, and are so apt to admire as something very extraordinary, and inexplicable by all the disquisitions of human understanding. But our wonder will, perhaps, cease or diminish, when we consider, that the experimental reasoning itself, which we possess in common with beasts, and on which the whole conduct of life depends, is nothing but a species of instinct or mechanical power, that acts in us unknown to ourselves; and in its chief operations, is not directed by any such relations or comparisons of ideas, as are the proper objects of our intellectual faculties. (Hume [1758], 9.6: 108)

Thus, Hume is far from denying, as Kant accused him of doing, a role for what Kant would call the «pure concepts of the understanding» in shaping our experience and our cognition. We expect the future to resemble the past by instinct, like animals8. Our tendency to form habits is a «principle of human nature», an a priori feature of our cognition. Hume is even more eloquent on this point in the Treatise:

Nothing shows more the force of habit in reconciling us to any phenomenon, than this, that men are not astonished at the operations of their own reason, at the same time, that they admire the instinct of animals, and find a difficulty in explaining it, merely because it cannot be reducd to the very same principles. To consider the matter aright, reason is nothing but a wonderful and unintelligible instinct in our souls, which carries us along a certain train of ideas, and endows them with particular qualities, according to their particular situations and relations. This instinct, it is true, arises from past observation and experience; but can any one give the ultimate reason, why past experience and observation produces such an effect, any more than why nature alone should produce it? Nature may certainly produce whatever can arise from habit: Nay, habit is nothing but one of the principles of nature, and derives all its force from that origin. (Hume [1739], 178-179)

Thus, contrary to what Kant argued about Humes notion of experience, experience for Hume is not simply a blank slate that becomes filled up with perceptions. Rather, it is grounded in tendencies or instincts that is, it possesses an a priori structure which precede perceptions and shape experience as perceptions come in. Hence, the disagreement between Kant and Hume on this point may not have been as sharp as Kant took it to be.

Be that as it may, what Hume was in fact interested in was not the borderline between this a priori structure of experience and the perceptions that filled experience up and were shaped by this structure, but rather the manner in which both the a priori structure and the accumulating experience behaved over time. For Hume, it is not the difference between the a priori and the a posteriori aspects of experience that is at issue: the all-decisive difference is between a single experience and multiple similar experiences. When Kant talks about the a priori features of cognition, it makes no difference to him whether the cognition in question arises from a single experience or from an identical experience replayed a thousand times: the a priori features of cognition are the same in both cases. Not so for Hume: short-term and long-term experience are completely different in nature, and the category of causality is specifically an attribute of long-term experience, as opposed to short-term experience. As we pass from a short-term to a long-term experience, according to Hume, our cognition changes. But this change comes about not as a result of our stepping out of the world of pure reason and into the world of experience stepping out of the a priori and seeing what there is to be learned a posteriori, as Kant would have it; rather, it comes about because we step out of the domain of the single experience and pass into the domain of the repeated experience.

So it is indeed difficult to be precise about the sense in which the category of causality is a precondition or parameter of human experience, according to Hume, and in what sense it is a piece of knowledge that is acquired. It is a parameter or precondition of human experience   «the ultimate principle, which we can assign, of all our conclusions from experience» but only over the long term; it is not an attribute of human experience over the short term; thus, we must acquire experience in order to bring this attribute to light.

This complication is the reason for the confusing terminology at the end of the passage quoted directly above, in which Hume speaks of an «instinct [...] [which] arises from past observation and experience», but then asks rhetorically whether anyone «can [...] give the ultimate reason, why past experience and observation produces such an effect?». In other words, while instinct arises from past observation and experience, there is nothing about past observation and experience per se that suggests that they should necessarily give rise to such an effect. Past experience and observation evidently produce such an effect because of some a priori tendency a «principle of human nature» which is not itself the result of past experience and observation.  But this a priori tendency has to do precisely with the way in which experiences and observations accumulate over time, rather than with shaping them as they come in. It is a tendency that arises only when confronted with repeated experience, that is, with the same experience over a large time scale. For Kant, by contrast, the «pure concepts of the understanding» would seem to shape experience over all time scales indiscriminately.

Thus, though Humes argument is that causality is an impression that is formed as a result of experience, it is an impression that is formed specifically after repeated experience. Hence, it is a form of knowledge that is acquired not in the general way in which anything that comes through perception is acquired, but in the particular way in which anything that comes through repetition is acquired. This decisive but unstated distinction the distinction between E2 and E3 is the source of the confusion surrounding the epistemological argument between Kant and Hume.

6. Conclusion

My aim in this paper has been to suggest that Humes conception of experience was far more original than Kant took it to be. While Kant appreciated Humes argument about causality, he assumed that the conception of experience on which this argument was based was the conventional conception of experience found in epistemology. This conventional conception interpreted experience as the product of a particular set of cognitive faculties sensation, perception, external and internal observation and explored experience primarily as it related to those faculties. Hume, however, was operating with a different conception of experience, not one that was new to philosophy in general, and certainly not one that was new to the history of thought in general, but one that was, I would argue, virtually unprecedented in epistemology. Humes conception of experience interpreted experience as a knowledge acquisition process with a particular temporal profile referring specifically to knowledge that could be acquired only over time and explored experience primarily as it related to that temporal profile.

To argue that cognition was sensitive to mere repetition, as Hume did, was to make a break with the epistemological tradition, since it was to foreground a dimension of knowledge that had been virtually ignored by this tradition, from Plato and Aristotle to Descartes and his followers. In various ways, all of these thinkers were interested in defining epistemological categories and differentiating cognitive faculties; but they were not interested in exploring the fact that these different categories and faculties endured in time. Hume, by contrast, placed the passage of time, as expressed in the phenomenon of repetition, and the endurance of knowledge, as expressed in the phenomenon of custom, right at the center of his analysis of cognition. The ultimate implication of his approach was that all knowledge had a temporal dimension, that all knowledge was the result of exposure to one or another degree of repetition, and that all knowledge had to be associated with one or another time scale. For Hume, the temporal dimension of knowledge was the fundamental fact that epistemology had to confront. As he wrote, «all our experimental conclusions proceed upon the supposition that the future will be conformable to the past» (Hume [1758], 4.19: 35). What were the grounds for this supposition? What effects did it have on knowledge? These were, for Hume, problems of primary importance. For most philosophers who preceded him, they were simply not an issue.

But while Humes focus on repetition, time scales, and the resemblance between the future and the past represented a break with his predecessors in the epistemological tradition, it was by no means original in the history of thought in general. As we have seen, long before Hume, British legal theory had a highly developed conception of the relationship between knowledge acquisition and the passage to time. Without going so far as to make an argument about concrete textual influences, it is possible to view Humes approach to cognition as the entrance of the British legal way of thinking into the mainstream of epistemological thought. This entrance involved a generalization of the common law jurists ideas, which were concerned exclusively with knowledge of the law rather than with knowledge in general, and Humes conception of experience along with the accompanying notions of custom and repetition    accomplished just such a generalization.

Kant, meanwhile, overlooked this aspect of Humes thinking and failed to appreciate what it was that Hume was bringing into epistemology. He interpreted Humes argument about causality within the traditional epistemological framework using the conventional epistemological conception of experience and did not recognize that a different approach to thinking about knowledge and experience existed and had in fact been introduced into epistemology by Hume. Kants own response to Hume, then, may be seen as a response formulated within the traditional epistemological framework, which was ultimately interested in exploring cognition in order to understand the gap between the self and the world, rather than within the framework suggested by Hume, which was ultimately interested in understanding cognition as a form of evolution.


Coke, E., 2003: The Selected Writings and Speeches of Sir Edward Coke, edited by S. Sheppard, Liberty Fund, Indianapolis.

Hume, D., 1739: A Treatise of Human Nature, edited by L.A. Selby-Bigge, 2nd ed. revised by P. H. Nidditch, Clarendon Press, Oxford 1978.

Hume, D., 1758: An Enquiry Concerning Human Understanding, edited by L.A. Selby-Bigge, 3rd ed. revised by P. H. Nidditch, Clarendon Press, Oxford 1975.

Kant, I., 1781: Critique of Pure Reason, edited by P. Guyer, A.W. Wood, Cambridge University Press, Cambridge 2000.

Kant, I., 1783: Prolegomena to Any Future Metaphysics, edited by B. Logan, Routledge, London 1996.

Locke, J., 1690: An Essay Concerning Human Understanding, edited by P.H. Nidditch, Clarendon Press, Oxford 1979.

McArthur, N., 2005: David Hume and the Common Law of England, in “The Journal of Scottish Philosophy”.

Pocock, J. G. A., 1987: The Ancient Constitution and the Feudal Law, Cambridge University Press, Cambridge.

1 Kant adheres to this approach throughout the Critique. As he writes elsewhere: «Experience is an empirical cognition, i.e., a cognition that determines an object through perceptions. It is therefore a synthesis of perceptions, which is not itself contained in perception but contains the synthetic unity of the manifold of perception in one consciousness, which constitutes what is essential in a cognition of objects of the senses, i.e., of experience» (Kant [1781]: 295-296).

2 As a curious illustration of this divergence between Kant and Hume, it may be interesting to note that the words «habit», «custom», and «repetition» appear in Humes Treatise 32, 97, and 26 times, respectively, while in Kants Critique, «habit» is used six times (four of them in connection with Hume), «repetition» is used twice (both times specifically in connection with the idea of quantity), and the word «custom» does not appear in the text at all.

3 This tendency to leave the concept of experience in a complex form without analyzing it into its constituent parts was even more strongly expressed in the related concept of «experiment», as it was developed in the philosophy of the natural sciences. What defined an experiment was precisely the co-presence of both observation and repetition. In other words, an experiment involved both reliance on the senses and reliance on the passage of time.

4 Note that «reason» here is not being used in any technical sense. Rather, it refers to cognition generally.

5 See in particular Pococks discussion of Cokes, and subsequently Hales, distinction between «natural reason» and «artificial reason» (Pocock [1987]: 30-55, 148-181).

6  John Davies, also quoted in Pocock, (1987: 41).

7 On the extensive discussion of this term in British legal theory, and on the interpretation of the common law as a formalization of custom, see Pocock (1987) especially 30-55.

8 Thus, Hume is making a distinction between our tendency to form habits in general and the concrete habits that we actually form: which specific events in the future and in the past we expect to resemble each other depends on our past experience, but the general expectation that the future will be like the past is instinctual and precedes experience.


Creative Commons License

This work is licensed under a Creative Commons Attribution 4.0 International License (CC-BY- 4.0)

Firenze University Press
Via Cittadella, 7 - 50144 Firenze
Tel. (0039) 055 2757700 Fax (0039) 055 2757712