III: Knowledge and Probability in Hume's Treatise

Lecture 14

Hume's Philosophical Relations—Relations of Ideas and Matters of Fact—rational and empirical science

(Concerning A Treatise of Human Nature, Book I, Part III, “Of Knowledge and Probability” .)

In this part Hume takes up the question of knowledge as contrasted with that of the constituents of knowledge, just as Locke in Book IV of An Essay Concerning Human Understanding discusses knowledge, having discussed ideas in the earlier part; but Hume has the same difficulty as Locke and Berkeley in showing how, out of the constituents of knowledge we can build up systematic knowledge; and, of course, if the ideas that these thinkers begin with are

  ― 78 ―
not themselves objects of knowledge, then knowledge cannot be built up out of them, and on the other hand, if they are objects of knowledge, then the question of knowledge is being considered from the start and does not arise merely when we come onto the question of the connection of ideas. Hume, of course, does not discuss that difficulty; he simply assumes that we can have connections of ideas although, of course, he admits in the Appendix conflicts between the two principles he has employed (T., p. 636). But here he asserts dogmatically that there are certain types of relation of importance for science or, as he calls them, types of “philosophical relation”. This is connected with the problem of substance in Locke and with Berkeley's theory of different kinds of significance and leads on to Kant's doctrine of categories.

The seven philosophical relations of Hume are: “resemblance”, “identity”, “relations of time and place”, “proportion or number” “degrees in any quality”, “contrariety” and “causation”. These relations fall into two classes, those which depend entirely on the ideas related, so that they cannot be changed without a change in the ideas, and those which can be changed without a change in the ideas. Now the first he calls “relations of ideas” and the latter, “matters of fact”, the relations of ideas being resemblance, contrariety, degree in quality, and proportions in quantity or number. For example, if we all asserted a resemblance between two ideas and then asserted an absence of resemblance, this means that in the second

  ― 79 ―
case we have changed one of the ideas or, in effect, that we have taken a different idea, so that it is something in the ideas themselves that determines each of those four relations. But this, Hume contends, does not hold in the case of matters of fact, that is, in the case of relations of identity, causation or relations of time and place. These do not follow from the nature of the ideas themselves and can be changed without any change in the ideas; that is, there is nothing in the ideas themselves to tell us if two objects are identical as contrasted with merely being very like one another, or whether one idea is the cause of another or is, say, above or below another. We could first place A above B and then B above A without this implying any alteration in A and B themselves.

But here, just as in the case of the distinction between impressions and ideas, the suggested distinction cannot be maintained; that is, we are concerned all the time with matters of fact and not with any distinct relations which could be called relations of ideas; and secondly, Hume is all the time implying a reference to things; in fact, in terms of ideas alone he cannot make the distinctions he requires.

We may note incidentally that it is by this distinction that Hume supports the conception of mathematics as a rational science as contrasted with empirical science, a view that was held by all the predecessors of Berkeley although it was rejected by Berkeley himself. Now it is quite certain that Berkeley is right here and that however we might otherwise argue on

  ― 80 ―
the matter it is impossible, in terms of the theory of simple and separate ideas, to have rational science and to develop one simple idea out of another or to show that one simple idea has a necessary relation to any other.

Now Hume also had difficulties concerning the relations of ideas; he could account for negative propositions, or what he calls contrariety, from the fact that any simple idea must exclude any other, but he could not account for the affirmative proposition without introducing considerations of complexity which alone account for the affirmative proposition. He says that two ideas go together and this means that they are both to be found in the same substance; but firstly, he cannot distinguish substance from the various ideas said to be in it, and secondly, the nature of “going together” is still unexplained. That is, the idea of inherence is as difficult as the notion of coherence and in any case, it does not explain the latter; and, of course, even if the substance were an additional idea we should then simply have three ideas and similarly the relation of “going together” would simply be a third idea, and thus we should have no solution of our problem.

Again, “going together”, whatever that means, is clearly a symmetrical relation and so although this may explain the particular affirmative proposition, it could not explain the A proposition. But apart from the working out of such objections, the fundamental point is that if simple ideas are those which must be taken one at a time if confusion is to be avoided, then there can be no way of taking them together; and we can say that a negative relation cannot be found between two simple ideas unless

  ― 81 ―
there is some way of taking the ideas together.

Now considering the relations of ideas Hume refers to, in the first place, we can see that the relation of resemblance implies that the ideas considered are complex and not simple; that is, it implies what must be allowed if any science, even mathematics, is to proceed, namely, that the same thing may have a number of different characters, and this is also implied in making distinctions of degree; and even in enumerating things we imply that they exist in complex situations; so that in asserting these relations we have an admission of complexity and interrelation contrasted with the simplicity and isolation of the objects of which science treats.

Now if that is so, then it follows that mathematics as Berkeley saw, is no more certain than any other science, and Hume himself admits that in the matter of enumeration we can make mistakes, that, except in the case of the very simplest numbers, we cannot estimate proportions by immediate inspection; and the same will be seen to apply to the other relations he mentions when it is observed that they involve complexity. If we compare two things each with a variety of characters, it is quite possible for us to overlook some of the characters which either or both possess, once we have given up

  ― 21 ―
view facsimile

the theory of unit characters. Since no relations can be discovered among unit characters, we can no longer regard each thing as having a definite enumerable set of characters all of which must be apparent to us immediately, and that means that we cannot regard ourselves as being concerned with ideas but must recognise that the relations in question are relations between

  ― 82 ―

Now if it is said that an idea is just what we know it as and so that we cannot speak of overlooking some of its characters and thus of overlooking points of resemblance and points of difference because to notice these other characters is to have a different idea, then not only are we denying the fact that we do discover resemblances and differences that we had not noticed at first, but also we make it impossible to discover the simple relations of ideas because these relations will be also just what they are known as. That is, they also will be ideas that are immediately given or else we will not know them at all. Hume, then, in his theory of relations of ideas, admits that we can discover something about ideas that we did not know at first, he admits that ideas are not just what they are known as, and thus if there are any entities entitled to be called ideas at all, then these are definite things existing in historical situations, whether they are mental or non-mental.

Now it is quite clear that in rational science we are concerned with such things, things about which we can make discoveries, and things which can have various historical relations to other things. Thus, there are many characters of triangles which we do not discover by simple inspection of them and we require not only to make various comparisons which involve a region within which the various objects can exist together, but we must also go through various operations upon these objects, to see characters which would otherwise elude our observation; and of course, it is implied that we can make mistakes about these things, this being

  ― 83 ―
another indication of the fact that they are not just what they are known as, and so it appears that the relations which Hume calls relations of ideas are really matters of fact, that we may become aware of a resemblance between things, a resemblance of which we were not previously aware, though the things themselves have not changed at all.

Resemblance is, then, a matter of fact, and so with the other relations; and there is no distinction between rational and empirical science. And, of course, just as Hume cannot distinguish relations of ideas from matters of fact so in dealing with those relations he calls matters of fact, he likewise introduces a reference not to ideas as things dependent on the knower of them but to independent things. In the case of identity especially we have something which cannot be accounted for in terms of a mind's impressions and ideas, and all our distinct perceptions being distinct existences, to say that a second perception was of the same thing as a previous perception had been, would be quite meaningless, and therefore to understand it we should have to think we were confronted at one and the same time with ideas of things and with things themselves. And, of course, if we are acquainted with things themselves, it is quite easy to believe that we can be confronted with the same thing twice, to believe such a proposition as “This is the same man as I met yesterday”, but if we are concerned solely with impressions, then it must be acknowledged that in the second case we have a different impression and so there is no identity in the case at all.

Thus in terms of ideas, Hume cannot distinguish, as

  ― 84 ―
he desires to do, between identity and very great resemblance and similarly with the relations of causation and of time and place; in fact, apart from the difficulties connected with Hume's theory of units of space and time, we can say that the question of objects changing their spatial relations without otherwise changing their characters, if this were possible, is a question not of perceptions but of independent things.

Lecture 15

Causal relations

We find, in connection with this theory of matters of fact, that Hume, just like Berkeley, is concerned with the possibility of prediction; in regard to matters of time and place the question is, how we can expect that one thing will appear in the vicinity of another thing, and in regard to identity, how we can expect that the same thing that appeared before will appear on a subsequent occasion.

“There is nothing in any objects to perswade us that they are either always remote or always contiguous” (T., p. 74.)

and similarly, nothing in any idea to persuade us that it is a reappearance of something that has gone before. In order to be able to pass these judgments and make these predictions, we require, Hume considers, to introduce the notion of causation.

If two things are always found together we take them to be causally related; and again, when we identify things, it is because we think there is no cause, short of perfect identity, which could produce so close a resemblance. It appears, then, that whenever we pass judgments regarding identity or contiguity

  ― 85 ―
we are implicitly passing causal judgments so that causation enters into all judgments of matters of fact, and thus cause, in Hume's theory, takes the place of significance in Berkeley's theory; and it is on this account that the question of causation is so closely connected with the question of belief.

But actually it is not correct to take cause in this way because if we say that a cause is required to explain the contiguity of two phenomena, then we are taking this cause to account for both phenomena and thus to have a complex character which cannot itself be explained in terms of causation. If we said that it could, then apart from the difficulty about complexity, we should have an infinite regress.

The reduction of predication to causation, then, brings Hume as near as he can come to a justification of the theory of separate ideas because, as he points out, cause and effect are always two different things having different places and times however close they may be to one another. Accordingly, we might conceive causation, although even this would be incorrect,

  ― 22 ―
view facsimile

as a relation between simple and isolated units whereas predication, for example, obviously introduces complexity and connection.

And again, by taking certain characters of the real relation of causation, Hume makes his theory more positive than that of Berkeley whose conception of significance is bound to be rather obscure. At the same time, we cannot say that any of the fundamental difficulties are solved in this way for if cause and effect are two different things, we must be able to think of the former both as having its peculiar qualities and as causing the

  ― 86 ―
latter. Hume quite correctly points out that we cannot take any quality of the cause and say that this is its causal quality, that this, as Locke puts it, is its power to produce the effect; but if we cannot find the causal relation in any one of the related terms and if, in the second place, these terms are separate and are not found from the beginning existing in a system of relations, then the causal relation must fall between the two, that is, it must simply be a third idea, not enabling us to say that “A causes B”, not enabling us to have before our minds that whole situation as a single object of thought. And Hume has in the end to admit that the relation does fall between the two objects and quite fails to unite them in the way which would be required if we are to formulate the complete proposition “A causes B”.

The most that Hume can make of the situation is that we first have one idea, the cause, that then there is a certain mental passage or transition and that then, finally, we have the other idea, the one we call the effect, and this does not give the required proposition; that is, it is not that we have one idea following another but that we think that some thing does follow another.

And Hume's general theory of belief fails in the same way; the statement that a certain idea becomes lively on the occasion of a certain impression taking place cannot account for the proposition that a particular subject has a particular predicate, or again, the fact that when we are given a certain idea we expect a certain other is not the same as the fact that we

  ― 87 ―
believe there is a connection between the two. In the first case, what is before our minds is first the one idea and then the other; in the second case, what is before our minds, and this is what is required for belief, is the connection of ideas. And, again, Hume's theory of the transference of liveliness from an impression to an idea fails to account for the fact that we hold and state beliefs, of no term of which we have any impression at the moment in Hume's sense of the term; we make judgments as to the past and the future and the spatially distant and our belief can be just as strong as our belief in events which are now taking place in front of us.

The main point is, then, that it is impossible to break up the proposition into a number of separate impressions and even to break up causal relations in that way because even although the cause precedes the effect, we require to be aware of the continuity or passage between the two before we can be said to believe that there is a causal relation; that is, the passage has to be part of the objective situation. It is not merely that one thought passes into another but that we think that one thing passes into another; and, of course, even to maintain the former view, Hume would have to imply that he is aware of real causal relations among thoughts just as he has implied that he is aware of complexity and thus of objective and independent situations among thoughts or perceptions.

Now while such criticisms apply to his theory, it is still possible for Hume to make many true statements about causality simply because he has a direct acquaintance with causal relations and because, however inconsistent it may be with his doctrine of

  ― 88 ―
ideas, he uses the ordinary language which implies that direct acquaintance; and in general he speaks about objects and qualities of these objects, about relations among them and, in particular, about their sequence and contiguity, all these being expressions not supported by the theory of ideas but readily understood in terms of people's experience of things.

That being understood, we can accept many of Hume's statements about causes, for example, his statement that there is no quality by means of which we can distinguish what is capable of being a cause from what is not, so that anything whatever that we perceive may be considered either as a cause or as an effect. But while this is true, Hume has no means of proving it; it implies what he has not admitted, that things exist and can be known to exist in a system of relations; and there are similar difficulties about his saying that cause and effect are contiguous and that cause precedes effect in time.

As to contiguity, Hume says that “nothing can operate in a time or place which is ever so little remov'd from those of its existence” (T., p. 75); but this does not imply that cause and effect are contiguous. To say that a thing acts where it is is not to say that what it produces is in that place or in the next place; and if we do adopt the theory of next points, as is implied in Hume's theory of space, then however close these next points may be, there must be some distance between them or they would be the same point, and Hume is certainly right in saying that if a thing cannot operate at a great distance it cannot operate at a small, since “great” and “small” are relative terms.

  ― 89 ―
Thus in regard to contiguity and priority, Hume falls into difficulty on account of his theory of spatial and temporal units, so that he cannot be regarded as proving that these relations hold even though we may agree that they do hold. However, he defends himself here by saying that whatever weaknesses there are in his argument the reader will find to be of no importance. It is, however, important to observe that on the theory of units there cannot be contiguity or priority.

  ― 23 ―
view facsimile

Hume goes on to discuss the third essential of causal relations, namely, a necessary connection between cause and effect. As he puts it:

“An object may be contiguous and prior to another, without being considered as its cause. There is a necessary connection to be taken into consideration; and that relation is of much greater importance, than any of the other two above-mentioned.” (T., p. 77)

Now when we ask what this necessary connection amounts to, we find that it involves for Hume the possibility of inference; and that is really all it does involve in his theory. That is, he is really referring only to the fact that in some cases where we have found an object A to be prior and contiguous to an object B, we infer on a subsequent appearance of A that B will also appear and that in other cases we do not draw this inference. Hume's argument only shows that there is that difference between our procedure in the two cases but it does not enable us to discover any criterion of causality over and above priority and contiguity.

In some cases we predict and in other cases we do not; in some cases we become accustomed to expect a certain consequence while in other cases we have no such expectation; but Hume can give

  ― 90 ―
no account of the difference between the two sets of cases except that we may have become accustomed or not become accustomed. Now this is obviously no explanation at all of the fact that in some cases we say there is a necessary connection and in other cases do not recognise any such connection. In other words, Hume is only saying that there is a distinction without showing what it is, and thus he fails to solve the problem he set out to solve.

Lecture 16

Causal relations continued

The only distinction, as far as Hume can show, between cases in which we assert a necessary connection and those in which we do not, is that in the former we become accustomed to expect a certain consequence and in the latter we do not have this expectation. Now if this were all that was meant by the distinction, we could not describe the supposed connection as a necessary connection; in fact, we could not find any connection at all. The reference to necessary connection implies that there is, or that we think that there is, some objective difference between the two sets of cases; and if there is no objective difference, as is implied by Hume's reference to custom and to internal impression, then that would mean that we recognised no difference at all.

And it is worth noting that it is not merely a case of inference or custom so that when we are given A we expect B, whereas when we are given X we do not expect Y, or indeed have no definite expectation. On the contrary, we say that B follows A, and that it necessarily follows A; and, again, even if we do not ourselves observe B, we do not take this as a sufficient proof that the sequence has not taken place. Now if fulfilled and

  ― 91 ―
unfulfilled expectations were all that differentiates causal connection from other connections, we should certainly cease to regard AB as a causal succession if once we had found A without subsequently finding B.

All this, of course, is connected with the difficulty of accounting for propositions and any sort of relations on the basis of the theory of ideas. If it were simply a question of the succession of two particular things there would be no possibility of prediction because these two things would have occurred and never could occur again. Prediction is possible because we observe the sequence not as that of two pure particulars or simple identities but as that of two complex occurrences of certain sorts, that is, because we find one sort of thing following another sort thing or issuing from it, as Alexander puts it; and consequently, when on a subsequent occasion we observe the same sort of thing as the earlier one, the first instance, we naturally expect a thing of the same sort as the later one; that is, prediction is possible because we observe sorts of things and connections of sorts of things or necessary connections. Of course, in observations of this kind we are capable of being mistaken; we may think that that has occurred which actually does not exist, but the possibility of error is nothing against the possibility of knowledge.

Now in admitting that there is any sort of inference or prediction, Hume is admitting that we perceive on the basis of what he calls resemblances, that is, of the possession of specific factors by specific things, and he is admitting that we regard sequence, even in the simplest case of observation, as having to do with these characters and not with the things as pure particulars.

  ― 92 ―
If observation, then, implying the possibility of recognition, thus involves the observation of characters or of things as of certain sorts, then we do not require any habit or custom, any multiplication of instances, to build up in our minds a conception of universal connection, but universality is, from the very beginning, a feature of what we know. Apart from this, there could be no such thing as a multiplication of instances because what we call another instance is the appearance of another thing of the same sort. Of course, even if Hume were right in saying that the assertion of necessary connection was the result of repetition or custom, this will not show what necessary connection is, and so it would not give any additional understanding to the proposition “A is necessarily followed by B”, even if we formulate this proposition only after we have had repeated experiences of A and B in conjunction.

Of course, the very fact that on this admission we are dealing with complex things shows that we may be mistaken in our expectation. If each of two successive things has an endless variety of characters, then we might well take, as the characters relevant to discussion, characters which are actually unconnected. Such an assumption may be corrected by subsequent experience; but it is in any case impossible to formulate the connection in any but a universal way,

  ― 24 ―
view facsimile

whether that turns out to be right or wrong; and we are not here substituting the forming of suppositional connections for the observation of things that may be connected but, while we do form hypotheses which may turn out to be wrong, it is equally the case that we observe connections and that our

  ― 93 ―
observations may turn out to be right. The assumption of a distinct knowledge, first of things or events and secondly, of connections between them, leads as Hume himself admits in connection with his two principles, to contradiction (T., p. 636).

Granted, then, that in speaking of causality we assume that there is a necessary connection, although according to Hume's later showing all that we can actually find is contiguity and priority, and then mind's distinction between what is necessary and what is not necessary is quite arbitrary, Hume goes on to ask two questions, namely, why we consider it necessary that whatever comes into existence has a cause, and secondly, why we conclude that a particular cause necessarily has certain particular effects or has a certain causal relation.

Now as regards the first question, Hume considers that the view that a cause is always necessary cannot possibly be proved and that we have no reason for adhering to it. Custom may lead us to recognise a particular causal connection but it could give rise to no recognition of a general causal principle; and Hume points out that the arguments actually used to prove that everything has a cause beg the question. Of course, the fact that this view is not proved does not show that it is false. It might be that, just as Kant contends that we cannot construct space and time out of what is not spatial and temporal because the alleged construction can only take place in space and time and that we cannot think away space and time, so we cannot construct causality out of anything that does not involve causality and yet we cannot think away causality, that is, conceive a condition of affairs in which causality does not hold.

  ― 94 ―
But the point of immediate importance, as far as Hume's argument is concerned, is that if there is no general principle of causality, then all our inferences in particular cases are unfounded. Unless we can always inquire into the conditions under which a certain sort of thing happens, then the mere fact that two phenomena have been associated any number of times is not the slightest reason for expecting them to be associated again. When we say, even although we may be mistaken, that there is a constant connection between two phenomena, we imply that we know what sort of thing this constant connection is. We imply that there is an absolute order of phenomena, that there are laws of nature; and if there were no laws of nature, assuming that to be conceivable, then custom could never make us imagine that there were.

The point is, then, that to suppose that any connection whatever is causal is to suppose a general principle of causality. The attempt to derive causality from repetition corresponds to the attempt to derive universality from enumeration. But in enumerating we already recognise universality; unless we could recognise characters of things, we could not begin to enumerate, we could not say that this has the same character as that or even that this has any character at all. Thus the particulars with which we are said to begin are themselves found to involve generality; and when we take this initial recognition of generality along with the recognition of sequence, then we have the recognition in all phenomena of regular sequence which is at least the basis of the causal principle, though causality involves the further fact of interaction.

  ― 95 ―
Hume certainly finds it easy to prove that the supposed demonstrations of a causal principle, which had previously been advanced, do not really prove anything at all. The assertion that “if a thing were not caused by something it would be caused by nothing, which is inconceivable”, that conception depends on the assumption that everything must be caused; and the similar contention that “if a thing wanted a cause it would cause itself which is inconceivable, in that case the thing existing before itself exists”, is also unsound because it fails to take seriously the supposition that has been made that the thing might have no cause of any description.

Nevertheless, Hume admits that when we say a thing begins to exist, we imply that it begins to exist at a particular place or time and we cannot consider that nothing whatever existed before that time. But in that case it must be said that what did exist earlier was the condition under which the thing came into existence. It is, in fact, only if we hold to a doctrine of simple entities that we can suppose that there may be unconditional existence; but when we recognise complexity and the occurrence of things in space and time, then we are bound to think of them as conditioned; we are bound to think of things as coming to be under certain conditions which would be distinct from the conditions under which the things did not come to be. It is not a matter of something starting its existence, but of a thing's acquiring certain characters or of the appearance of a certain tension within a system; and so we do have certain universal conditions of existence and can recognise the causal principle.

  ― 96 ―

Lecture 17

Causal relations continued

  ― 25 ―
view facsimile

When we think of complex situations and not of simple ideas, then we do not have any question of the sudden appearance of a certain object, but we have the question of the conditions of existence of the situation, in general, of a thing going on in a certain way, and of it acquiring that mode of behaviour. It is quite possible that in setting down the conditions of some particular type of occurrence we should make a mistake, but that is nothing against the recognition of causal relations, because we can make mistakes in the recognition of any situation because there is no possible distinction, just as there would be no possible connection, between a type of knowledge that cannot be mistaken and one which can be. But what Hume does not, and cannot, account for is that, when we discover that we have been mistaken as to a causal relation, when our expectations have not been fulfilled, we go on to look for a more precise criterion, and do not come to the conclusion that this particular type of event is unconditioned and that there is no regular sequence to be discovered in the case.

It is not, then, that we simply recognise regular sequence in some cases leaving it possible that this may be so in other cases, but that we assume in every case that there is some general condition that could be determined whether we have determined it or not; and if we did not make that assumption, then there would be no question of our making predictions in particular cases, or of having settled expectations, because a kind of event that was said to have a certain effect in one case might quite well, in a subsequent case, be followed by a different consequence; and there would be no question of custom or a multiplication of instances making us more certain, because there would be nothing to

  ― 97 ―
be certain about. That is, on the one hand, we have the possibility of ordered activities, although as these activities are infinitely complex and interact with one another there is no question of finding a single formula or a single order to cover all of them, and against this, we have the suggestion of a mere manifold of simple ideas, and even in calling them a manifold we should be attributing to them an order incompatible with their simplicity. And again, even granted that we could arrange them all in a temporal series (and this is incompatible not only with the notion of a primitive simplicity but also with Hume's theory of time) there would be no question of prediction. We might be able to refer after the event, to a remarkable series of coincidences of two particular kinds of ideas, but that would give us no ground for expecting a single subsequent conjunction of the same kind.

We cannot, then, on the basis of Hume's assumptions account for our actual beliefs about causal relations and for our actual predictions; and, of course, we noted as a fundamental objection to Hume's position that what is to be accounted for is not the fact that when we have one idea we expect another idea, but that there is a certain objective connection that we assert, namely, that one thing causes another or is regularly followed by another.

Now similar considerations apply to Hume's theory of belief. Hume begins here by saying that:

“There is no object, which implies the existence of any other if we consider these objects in themselves, and never look beyond the ideas which we form of them.” (T., pp. 86–87.)

It might quite truly be said that prior to experience anything might have any character whatever and that anything might be

  ― 98 ―
the cause of anything else, the only difficulty being that prior to experience we could not speak of the things at all, so that there is no question of our starting with separate things and then finding by experience that they have certain connections with one another. We certainly do experience distinguishable things, but we also experience things as connected, even if we can be ignorant of, or make mistakes about, certain particular connections. Of course, the main point as regards experience as a basis of knowledge is that knowledge and experience are the same thing, and, when we do have experience, we experience or are aware of complex and interrelated activities.

There is also, as Hume urges, no question of any line of argument whereby we could show that the things we are going to experience in the future must be similar to those experienced in the past (T., p. 89). But the point is not correctly expressed by saying, as Hume does, that the instances of which we have not had experience cannot be proved to resemble past instances (T., p. 89), because unless they do resemble them they are not instances of the same sort of thing at all; and whatever is of the same sort has not merely a simple but a complex resemblance to the previous thing, in other words, it resembles the previous thing in structure and in ways of behaving even if it also differs in certain respects.

There is here, then, a possibility of error in that we may take things to resemble where they actually differ or to differ where they actually resemble, but still we can say that there are various characters of a certain class of things and that each member of the class will have all these characters. While, then, there is no question of demonstration in Hume's sense of the

  ― 99 ―
term, namely, as something of which we have absolute certainty because we know all about the things concerned, there is demonstration in the sense of correct inference from true beliefs; and as we have already noted, Hume cannot maintain his distinction between relations of ideas and matters of fact and thus between demonstrative or rational science and empirical science.

Now Hume's theory of belief is the most important of his applications of the original distinction between impressions and ideas and if the criticisms passed on that distinction are sound it will follow that Hume's theory of belief cannot be accepted.

  ― 26 ―
view facsimile

The position is that if we are to have natural science we must be able to predict, although prediction is here only a special case of predication, of finding objective truth, or truth independent of our particular experience. We must be able to say that A occurs or is followed by B, without meaning that we perceive A and B at the same time or in succession. But such objective assertions, while they are a necessary condition of science are incompatible with the strict theory of ideas, according to which the order of things and the order of our perceptions could not be distinguished. And yet we quite frequently judge that things which we perceive at the same time were not in existence at the same time and that things which were in existence at the same time were not perceived by us at the same time, as when we realise now that something happened in the past that we overlooked.

  ― 100 ―
Now if we take simply the order of ideas, then we cannot give an account of the proposition in which we make the statement that something is true independently of when the judgment is passed or of who passes it and similarly in the special case of prediction when we say that something which is present now implies something else which is not present. And in addition to the fact that we are making an assertion of objective truth, there is the special difficulty in the latter case of explaining, on the theory of ideas, how what is not present now can nevertheless be now before our minds in such a way that we are said to expect it; and in the same way, there is a difficulty about our assertions of propositions in general as when, for example, we say that “Grass is green”, without meaning that we are perceiving grass or greenness at the present time in the ordinary sense in which the term “perceiving” is used. Another point in the same connection, that is, in connection with the question of distinguishing an order of fact from the order of our ideas, is the question of existence.

Hume, as we have seen, points out that there is no difference between thinking of a thing and thinking of it as existing. But on that showing the term “existing” would seem to be quite meaningless and Hume would still have to show how people came to use it and think that they meant something by it.

Now if we adopt the theory of the proposition, then we can distinguish between the affirmative and negative proposition and thus get a meaning for the expressions “existence” and “non-existence”, even if we do not, on that account, think that there

  ― 101 ―
are two classes of things, the existent and the non-existent. But according to the theory of ideas, on the other hand, existence would at most be one idea among others, and while Hume shows that existence is not such a separate idea he fails to show how we can come to say that certain things do exist or, in the case of prediction, that certain things will exist, and if he cannot show that, then he cannot show how we can have beliefs at all.

Now Hume unsuccessfully endeavours to get over this difficulty by again introducing the criterion of force or liveliness. He begins by saying that “the idea of existence is nothing different from the idea of any object” (T., p. 94), and consequently he considers that if we have presented to our minds a proposition which we regard as false, for example, the proposition that “Caesar died in his bed” (T., p. 95), in understanding that proposition, we attribute as much or as little existence to it as if we regarded it as true. And thus we have the problem of discovering what is the difference between belief and disbelief, if it is not a difference that could be expressed in terms of existence.

Lecture 18

Causal relations continued—Hume's Theory of Belief—Rationalist views of causality: as creation and as comprehension

As regards our reception of an assertion such as the assertion that “Caesar died in his bed” Hume points out that our disbelief in this assertion is not explained by saying that after we have conceived the situation in one way

  ― 102 ―
we go on to conceive it in the opposite way because that would simply mean that we had two different beliefs one after another, which is not an explanation of disbelief; and we could also believe that “Caesar died in his bed” and then go on to disbelieve it, whereas the fact is that we never believe it at all. The position is that we understand equally the two propositions that “Caesar died in his bed” and that “Caesar did not die in his bed”, but we believe the second and not the first. In fact, we disbelieve the first, and Hume explains the difference by saying that the second has a peculiar liveliness.

Now if we apply Hume's definition of belief, namely “a lively idea related to or associated with a present impression” (T., p. 96) to the two cases, we should say that in the first case the idea that was first presented to our minds failed to gain liveliness and in the second case, that it gained liveliness. But apart from other objections, it would be difficult to say to what present impression the liveliness is related in either case because we have no present impression of Caesar, of his dying or of his bed, and what we do have an impression of, namely, the spoken word, is certainly not part of the belief. In fact, we can have beliefs about things that are not given or immediately present to us in the sense in which we speak of a direct perception. There is, then, in these cases, namely, of disbelieving that Caesar died in his bed and of

  ― 103 ―
believing that he did, no impression that could transfer its liveliness to the ideas.

But a more fundamental objection is that Hume expresses himself on this question by means of propositions in the ordinary form, namely, propositions whose truth is simply asserted without any question of liveliness being raised. Thus, on Hume's showing, we should believe that a certain idea was not lively and thus we should have to have a lively idea of non-liveliness, a position which Hume could not explain.

Then, again, to take the point already taken in connection with the distinction between impressions and ideas, liveliness will simply be one idea among others and will not qualify the others in the way that Hume requires. On the other hand, if we do admit this qualification, then we are concerned with certain situations, with finding something to be the case, and that is belief. And accepting knowledge of propositions in this way, on which basis we can account also for disbelief, we have to abandon the theory of ideas; on that theory we simply cannot account for belief or disbelief.

Of course, Hume, like his rationalistic predecessors, has to try to give some account of error, and

  ― 27 ―
view facsimile

his position here is similar to that of Descartes in certain respects. According to Descartes, a certain content could be before the understanding, and then will could enter in and attribute existence to that content. That is, in the case of Descartes we have the addition of existence and in the case of Hume, who recognises some of the confusions that that involves, we have Hume's addition of liveliness.

Now the objection to Descartes, apart from the fact

  ― 104 ―
that to understand anything is to think of it as existing, is that what is said to be attributed by the will must be capable of being understood, and so the description of something as existing cannot be described as the work of a separate faculty. But in just the same way, in Hume's case, the liveliness which is supposed to be added must be a content of exactly the same order as that to which it could be added, and consequently the new object we added would be just as much something to believe or disbelieve as the object we previously had; that is, the distinction would be between Caesar dying faintly in his bed and Caesar dying vividly in his bed and the problem of whether we were to accept either judgment would still be unsolved.

Now the solution of Hume's difficulty is, in the first place, the recognition of the proposition, that is, of our finding something to be the case or not to be the case; but, of course, we also recognise that there are psychological conditions of belief; that is apart from the question of the kind of thing we do believe there is the question of how a particular person comes to hold a particular belief; but, of course, in order to solve that problem we should require to have a knowledge of minds in the same propositional way, and neither the logical problem of truth or falsity nor the psychological problem of knowledge and error can be solved on the basis of the theory of ideas.

So much then, for Hume's theory of belief, and now to consider the theory of causality. In the work of modern philosophers there are certain main conceptions of causality, conceptions of a rationalistic kind, the two important conceptions both of which appear in the work of Descartes and to a varying extent in the work

  ― 105 ―
of his successors being that of creation and that of comprehension.

Now in the theory of causation as creation, when we ask for the cause of a thing we are simply referred to something else outside of it, that is, we are referred to something which had the power to produce that particular effect, and this, of course, is a metaphysical theory because we cannot be presented with that power and we cannot see any process of production going on. We simply have to accept the assertion that there was something adequate to the production of that effect, but, of course, this does not show us what we should consider to be adequate or why we should consider that the thing was produced at all. That is, we have an atomistic theory of individual entities, individual causes and individual effects, but we do not get any connection between them; and until we can establish such a connection we cannot talk about causes and effects at all.

Accordingly, we have the development of the other side of the rationalist theory, that which treats relations as forms of identity, because having begun by treating distinction as absolute disconnection by making it a matter of separate essences, rationalists have to go on to treat connection as identity, and thus the things to which a given thing is said to be related have to be taken to fall within the thing.

It is in connection with such difficulties that we have to take Hume's two principles that “all our distinct perceptions are distinct existences” and that “the mind never perceives any connection among distinct existences”; that is, we have the problem of distinction and connection and the rationalistic assumption of a kind of distinction that excludes connection and of

  ― 106 ―
a kind of connection which excludes distinction, whereas in order to have any logical theory at all we must recognise that distinction and connection do not exclude one another but rather are bound up together (T., p. 636).

Now it is in terms of this conception of connection as excluding distinction that we have the conception of causality as comprehension, of the cause as somehow containing or embodying the effect, so that, whereas on the creation theory we could not see how any particular cause could be assigned for any particular effect or even what was the meaning of the terms “cause” and “effect”, on the comprehension theory we likewise cannot see the meaning of these terms because the effect being embodied in the cause, nothing has been produced, nothing has happened.

Now in these rationalist theories we have the use of both of these conceptions. Thus even in the theory of Berkeley who tries consistently to maintain the conception of creation, a theory of separate agents and separate phenomena, the latter are said to be in the former or to be somehow comprehended in them as ideas in the mind.

Spinoza, on the other hand, worked almost entirely with the conception of comprehension, the identity of the effect with what has caused it, and the consequence is that his theory resolves itself into an account of one great cause which has no effect distinct from itself and which may therefore be said not to operate in any way, not to have any sort of history and thus, like the One of Parmenides, not to have any distinctions within it and to be quite indescribable. That is the upshot of the theory of identity.

  ― 107 ―
Leibniz, again, tries to work out a theory of comprehension and to deny the possibility of interaction, but in order to avoid Spinoza's conception of a single substance to which nothing could happen he recognises the existence of a multiplicity of substances which proceed in harmony with one another. But the theory of this universal harmonising involves him in the view that there is one supreme substance or universal monad which, like Spinoza's substance, would have no history and which could not then be related in any way to the individual substances; and in spite of his opposition to interaction and to any sort of external relation we find him in the end putting forward the view that at a certain time all the other substances have been created by the supreme substance in such a way that they work harmoniously ever after.

Now this theory of a single act of creation followed by parallel development without interaction is clearly a less logical view than that of the occasionalists with their theory of continual recreation.

Spinoza and Leibniz, then, having shown that the doctrine of comprehension cannot be logically worked out, Locke, Berkeley and Hume reject comprehension, but it is only with Hume that we come to the recognition of the fact that creation also is an inadequate account of causal relation and that when we reject the explanation by identity, as in the theory of comprehension, we are in effect bound to reject the theory of creation as well because the notion of creation is that of the cause having in it or comprehending not the effect but the power to produce the effect,

  ― 28 ―
view facsimile

and Hume points out that we can find in the cause neither the effect nor this productive power. If we find in the so-called cause

  ― 108 ―
anything that we can call a causal quality, then if we are to relate that to the particular effect it must have a further power to produce that effect; and again, we could never find out that any particular quality is a power to produce that effect unless we could find the effect itself embodied in it, that is, unless we reverted to the doctrine of comprehension.

Lecture 19

Causal relations continued—Rationalist views of causality continued—Locke on power—Hume's Constant Conjunction

In the case of Locke it is recognised that cause and effect are two different things, but still an attempt is made to explain causality by means of the conception of power. This notion of explanation is one of the fundamental conceptions of rationalist theory; that is, the question is not simply of asserting or showing that a thing is so but of showing why it should be so, of giving a rational account of its being so.

And that is connected with Descartes' theory of mathematics, a theory which is adopted by all these philosophers except Berkeley, namely, that mathematics is a rational science, that it deals with truths which are somehow clear or transparent, which are not so much empirical fact or brute fact, as we may call it, which we have simply to take as actually so, but that it carries with it a reason for its being so or that we can see why it is so. And, of course, that is connected with Locke's theory of the analytic character of all truths, that is, of the possibility of reducing the notion of the predicate to the notion of the subject or of seeing that the subject is essentially or necessarily so.

In the same way, then, all rational explanations of causality are attempts to show the necessity of a cause having a

  ― 109 ―
certain effect or the way in which an effect is essentially or in its nature an effect of that cause; and, as I have said, we have an example of this in Locke's theory of power.

Thus he considers material bodies as embodiments of so many powers to produce effects on our minds and on other things, these powers being what he calls the qualities of the thing. Certainly he says that some of these powers resemble the effects they produce, but he is quite unable to make such a comparison as will demonstrate that this is so. Hence his position is logically reduced to this, that anything that occurs does so because there has been a power capable of bringing it about but what that power may be like, we have no means of determining. But on this basis, not only have we no means of arguing from ideas in the mind to any external source whatever, but also we have no notion of what is meant by “having a source” or “being produced” if we have not actually observed one of these so-called powers bringing things about, or things being brought about in some way.

Now Berkeley tries to avoid this difficulty by saying that we do have experience of causes, only it is a different sort of experience from our experience of effects; that is, we have, on the one hand, experience of the active and on the other hand, experience of the passive. According to Berkeley we experience our own minds as causes, we have a notion of their activities and we are aware of the images which are the effects of these activities. But the inevitable objection is that if the experience of causes is a different kind of experience from the experience of effects, then there can be no single experience which will inform us of the fact that a cause produces an effect; and if, to get over that difficulty,

  ― 110 ―
we admit that we have the same sort of experience of our minds as of other things, then we are no longer in a position to separate the active from the passive and to say that only a mind can produce effects; and thus it is possible to consider that any effect whatever may be a cause.

In holding that only minds can cause, Berkeley takes causation to involve a sort of comprehension, the cause comprehending the effect as the mind comprehends an idea. But if we took the matter in that way, if we considered that an idea is not an independent existence but that what exists is a mind having an idea, then it will be that whole situation, the mind having that idea which is the effect of the mind's activity in forming the idea, and thus the effect will not be comprehended in the cause, nor will it be purely passive. But, of course, the fundamental objection still is that we have no ground for distinguishing between the active and the passive and that we have no knowledge of any peculiar power that agents have.

Coming now to Hume, we find that he recognises, as Locke and Berkeley had also initially done, that cause and effect are distinct; but by taking that fact quite seriously Hume is able show that there can be no such thing as this power recognised by Locke and Berkeley and appearing in some form in all animistic metaphysical thinking.

The main point is simply this, that granted that there is a cause A and an effect B, then to say that there is in A a power to produce B is to say that B itself is already somehow in A, at any rate that we already know B in knowing A, for unless we do, we do not know what power A has. But this contradicts the distinction

  ― 111 ―
between cause and effect and is manifestly untrue. As Hume points out, apart from experience we might take anything to be the cause of anything else. Now if we say that to know the power is not to know B, the effect, that the power is not defined as “that which produces B”, then this so-called power is simply a certain quality of A, say, X; and then the position is that A, which is X, or AX, produces B, and we should have to inquire all over again how AX comes to have the power to produce B.

Hume has shown, then, that if causality is not a form of identity with all the difficulties that that involves, if cause and effect are different, then the notion of power has to be abandoned; and of course, Hume is also able

  ― 29 ―
view facsimile

to reject the usual a priori arguments for causality, arguments purporting to show rationally that a cause is necessary because, as he points out, all of these arguments beg the question; they are not arguments but mere identities asserting the thing to be proved.

Having dismissed, then, the conception of causality as power, Hume puts forward his own theory of causality as constant conjunction. But this really means that he has substituted for one bond or tie between cause and effect another bond or tie which also cannot be clearly indicated, and so he reintroduces difficulties similar to those he had pointed out in the case of power. The rejected view is that we have A's power of producing B and then B; on Hume's view, we have A, a certain mental transition and then B. If we are to speak not simply of two separate pieces of knowledge but of what is known or asserted as a whole situation, then the question is not of a mental transition but of the assertion of a transition between A and B, in other words, of A passing into

  ― 112 ―
B. We find, then, comparing Berkeley and Hume, that Berkeley has asserted that cause and effect are on different levels or exist in different ways; and on that basis there could be no relation between them. Hume asserts that causes and effects are on the same level, but what is required to complete that theory is that the relation itself should be at the same level and that we are capable of knowing the relation at the same time and in the same way as we know the things themselves, and this substantially is what Kant demonstrates, namely, that apprehending objects is apprehending them as in space and time and hence as having various relations, including the causal relation; similarly, James develops Hume's position by saying that we know connections and distinctions between things along with the things themselves.

Hume's argument is incomplete, then, because he gives no account of the relation on the level of the things related; in fact, he avoids doing so by making the distinction between internal and external impressions, a distinction which he is not entitled to make and which, logically speaking, makes no difference to the case because an internal impression would be known or would be an object just as much as an external one, and so it would not be that we were aware of A, then a mental transition took place and then we were aware of B; on the contrary, we should be aware of A passing to B.

That is the real outcome of Hume's argument although it is not in accordance with his original assumptions, and it implies that frequency has nothing to do with the case. The question is not, under what conditions we have come to associate A and B but what we have before our minds when we do associate them, that is,

  ― 113 ―
what is the nature of the complex, A and B associated, or A and B conjoined in that particular way. And if it is admitted that that whole complex can be known by a single act of thought, and otherwise there can be no causal judgments, then it is implied that that complex can be known even if repetition has not taken place, in other words, that causality is a matter of observation, granting that here, as elsewhere, our observation may be mistaken; the view that any observation may be mistaken is, of course, contrary to the Cartesian theory of rational knowledge and particularly to the theory of ideas, to the view that we can be absolutely certain of having an idea even if we cannot know that that idea corresponds to reality.

Now what has been indicated in consideration of the various theories referred to is that starting with elements or with certain supposedly indubitable pieces of knowledge we cannot work out any theory of complex things, or of things of which there may be a doubt; and in particular we cannot, in terms of the essence theory, show what correspondence is.

Now according to Berkeley causality is a species of correspondence, namely, that between spirits and ideas, and one of the things that Hume has shown is that that position is untenable. But Hume's own doctrine of impressions and ideas is also a doctrine of correspondence and so his whole position can be undermined by arguments of a similar character. Indeed, we can say that his theory of causality is also a correspondence theory, namely, one in which it is said that there is some relation between cause and effect but in which it cannot in the least be indicated what that relation is.

  ― 114 ―

Lecture 20

Causal relations continued—Rationalist views of causality continued—Hume's criticisms—Abstract Agency

We have been considering causality as rationalistically viewed by Hume's predecessors and as reformulated by Hume. In criticism of some of these rationalistic conceptions, Hume shows that the treatment of causality as identity, that is, the theory of causality as comprehension, is unsound, and likewise the theory of causality as creation. He brings out the fact that the theory of creation although it purports to avoid reducing causality to identity and to distinguish cause from effect does not really do so because the power attributed to the cause must either embody the effect or else it must itself have a power to produce the effect and so on indefinitely. And Hume indicates clearly enough that there cannot be a tie or bond uniting cause and effect, making them one or making them any less distinct than they were originally taken to be, that is, that we cannot find any essence or nature which makes the one the cause of the other or whereby their notions are bound up together. But the consequence of Hume's treatment of cause and effect not merely as different things but as distinct existences in his sense, that is, simple and separate natures, is that we cannot give any account of their connection at all; and this is substantially admitted by Hume when he speaks of the customary transition from the one to the other.

But that transition in the form of an internal impression as contrasted with an external impression, even supposing that Hume were able to make that distinction, does not account for the fact that we definitely say “A causes B”; that is, that we assert the existence of a situation of which “causing” is just as much a feature as are A and B, so that, as we

  ― 115 ―
have seen, it is not a question of a mental transition from the knowing of A to the knowing of B but of the knowing of A passing into B; so that without recognising any tie or bond of identity we still recognise complex situations, and we can get to know their further complications and their interrelations with other complexes without in any way contradicting the complex position we start from; but, of course, this also implies a historical medium of things, which is opposed to Hume's general theory of simple entities and to his

  ― 30 ―
view facsimile

particular reduction of space and time to so many units.

The general position is, then, that the difficulties brought out by Hume cannot be solved so long as the conception of an essence or distinct existence or unitary being is retained; and this theory of essence is that which dominates rationalist theory in general. We see, for example, that rationalists separate the domain of causes from that of effects, that they set up the conception of that which is a cause but not an effect and of that which is an effect but not a cause, this being paralleled in the case of predication and also of knowledge in the notion of S that cannot be P, of P that cannot be S, of a knower that cannot be known, or of a known which cannot know. But in the last case we see that it is not simply a matter of rejecting the separation of the two domains of a relation because we find it quite possible to think of known things that cannot know, in fact, that is the view that is commonly held, that is, as an empirical fact. Or again, we have the relation of ancestor and descendent where we see that although the two domains are not separate, although every ancestor is a descendent, some descendents are not ancestors. Then if we take the relation of husband and wife, we have two absolutely separate domains, a husband

  ― 116 ―
cannot be a wife nor a wife a husband; so that we cannot begin by assuming that there must be something wrong with a theory which separates causes from effects or even which holds that some causes are not effects or that some effects are not causes.

What we have to consider is, on the one hand, empirical evidence and, on the other, the rational mode of procedure. Do we find, as a matter of fact, that causes are effects and effects causes, or do we find effects which are not causes and causes which are not effects? Now the common view at least is that we find events which are both causes and effects, and the rational feature of the theories of first causes and epiphenomena is that a certain abstraction has been made, a certain feature of things has been separated off as a peculiar essence or nature. Thus because one thing causes another, it is assumed that there is a certain causal power or causal nature which enables it to do so and which is, in fact, the real cause of the effect; and again it is argued, not that we find causes which are not effects but that there must be such or such a cause in order that there may be any effects. Thus from the empirical fact that in a causal situation of the form “A causes B” the cause is not the effect, it is assumed, without any further evidence, that we can think of a cause which is not an effect; and then it is further argued that there must be a cause which is not an effect in order that causation may operate.

Now this creation of a certain feature of things into a principle is the sort of thing James calls “vicious intellectualism”,note the point being that because a certain feature is different from another feature of things it is assumed, not merely

  ― 117 ―
that the two can be separate but that they actually exclude one another, that, as James points out, because when we speak of things as distinct we are not at the same time speaking of them as connected, therefore we are speaking of them as unconnected and are setting up a pure distinctness independent of all connection. But the fact that we can say that distinction is not connection and that connection is not distinction is not the least reason for saying that distinct things are unconnected or that connected things are not distinct.

As far as causality is concerned, then, the point is that we find causes to be effects and effects to be causes and the fact that we can distinguish “being an effect” from “being a cause” is not the least reason for supposing that there can be an effect which is not a cause or a cause which is not an effect, let alone that there must be such a thing.

And just as the position is that we observe causal situations in which the relation of causing is a feature and that we observe the relations of such situations to other situations in which what was previously the cause is now the effect and vice versa, so in dealing with the question of a first cause or of nature as a thing requiring a cause we have still to consider, and can only consider, observable situations, that is, the possibility of finding a certain cause being first or of finding nature being caused; and if we do not find a cause being first or nature being caused, all the more if we do not find “nature” at all, we have no occasion to speak of a first cause or of the cause of nature.

Now, of course, these rationalistic conceptions, the separation of a certain feature of things as an abstract agency, are

  ― 118 ―
to be found in the various arguments considered in Hume's Dialogues Concerning Natural Religion and are the main ground of criticism of these arguments. It is to be noted that the design argument is connected with the general division of things into causes and effects or into the active and the passive and is substantially the form of argument put forward by Berkeley in his Principles: the theory of significance, the theory of the laws of nature as God's language, is equivalent to the design argument.note Berkeley, that is, thinks of ideas as merely passive, as having no efficacy, and as having a certain arrangement imposed upon them, so that we cannot say that one idea causes another, but only that one idea signifies another; and this means that we cannot properly speak about ideas in themselves at all. We should have to say not merely that one idea signifies another but that one idea is presented by God to our minds as a sign of another. It is only if we specify the agents and their relations that we really know what has happened; but, of course, in that case we have the problem of explaining what it is that is presented or, in general, what are these placid entities which are distinct from the agents and also of course related to them; and, as we noted, Berkeley cannot keep up the distinction, he has to admit that finite minds are passive in being acted upon by God as well as being active in their own way, and thus the distinction between the purely active and the purely passive breaks down. There again it is conceivable that besides things which are active and passive, there are things which are passive but not active and things which are active but not passive, that is, that the two domains of the active and passive intersect. Again, it is a

  ― 119 ―
matter for empirical consideration which, however, shows us not the purely active or the purely passive but interaction, and the only reason for assuming the purely active and the purely passive is this theory of natures, this separation off of a particular feature of things as being a thing in itself. In the same way, Berkeley has tried to work out a theory of the purely perceived or the idea but has found it quite impossible consistently to develop a theory of that whose nature it is to be perceived.

While, therefore, we have to drop the rationalistic theory of natures, we find that there is no basis for the conception of separate classes, of agents and

  ― 31 ―
view facsimile

things operated upon, of possible arrangers of phenomena as contrasted with phenomena to be arranged, of designers as contrasted with the designed; and thus we find it impossible to uphold the design argument; in other words, we do not require to go beyond the facts themselves, and, of course, we cannot logically do so, in order to obtain an answer to the question: how did things come to be arranged in this particular way.

In considering the conditions of any particular arrangement of things, we can find it only as issuing from another arrangement of things, and we cannot distinguish the matter arranged from the forces arranging it, but can only take the matter, in so far as we can give a definite signification to that term, as something which itself acts. That is, we cannot think of the purely passive which can be operated upon, and neither, of course, can we think of the purely disordered which can be put into order; but at any time the question is of interrelated activities.