Under this heading I shall be concerned with a part of Professor Ducasse's essay and with the whole of those by Professors Flew and Mundle.

(A) TIME IN GENERAL.

      Professor Mundle's researches show that I have "boxed the compass" about time, and in the course of doing so, have written some things which now make me blush. It may be interesting and possibly illuminating to mention very briefly the main influences under which the three accounts of time considered by Professor Mundle were written. At the back of all of them is McTaggart's paper "The Unreality of Time," published in Mind in 1908. I felt from the first, and I still feel, that the difficulty which he raises is

1. embarrassing enough prima facie to demand the serious attention of anyone who philosophises about time; but
2. almost certainly due to some purely linguistic source (common, and perhaps peculiar, to the Indo-European verb-system), which it ought to be possible to indicate and make harmless.

     At the period when I wrote the Encyclopaedia article (which, I must confess, I had wholly forgotten until Professor Mundle's essay reminded me of it) I was almost completely under the influence of Bertrand Russell in his extreme realist phase, and of Meinong as I understood him. By the time I wrote Scientific Thought I was greatly influenced by books recently published by Alexander and by Whitehead. The talk in Scientific Thought about "the sum-total of existence continually increasing by Becoming," and in the Examination of McTaggart's Philosophy about "Absolute Becoming," goes back to this source. To the influence of Whitehead was due the shocking remarks in Scientific Thought about a thing or a person being "a long event."

     By the time I wrote the Examination I had got free from the worst of that kind of crudity, largely through the careful work which had been done in the meanwhile by Moore and others on the notion of "logical constructions." What I was putting in a terribly slovenly way in Scientific Thought on this topic could be stated with more polish as follows. A sentence, whose grammatical subject is the name or a description of a thing or a person, and whose grammatical predicate is appropriate to such a substantive word or phrase, can be replaced without loss, gain, or distortion of meaning, by a complex of sentences, in each of which there occur only names or descriptions of processes, with grammatical predicates appropriate to process-words. (This is certainly not "snappy," and it may not be true; but at any rate it is not "sickmaking," like: "A thing or a person is a long event.")

     In writing each later account of time I started afresh, and was not concerned with its consistency or inconsistency with earlier accounts. What I have said on this topic in the EMcP was meant to supersede what I had said on former occasions, wherever there was a conflict between the two. The reader may assume that I continue to hold (though with much hesitation in view of the difficulty of the subject) any opinion which I expressed in the EMcP, unless I explicitly question or withdraw it in what follows below. I hope that this statement will justify me in confining my attention here, as I intend to do, to the third account and to Professor Mundle's comments on it.

(1) Qualitative Change and "Absolute Becoming."

      It seems to me that there is an irreducibly characteristic feature of time, which I have called "Absolute Becoming." It must be sharply distinguished from qualitative change, though there is no doubt a connexion between the two. In the experience of a conscious being Absolute Becoming manifests itself as the continual supersession of what was the latest phase by a new phase, which will in turn be superseded by another new one. This seems to me to be the rock-bottom peculiarity of time, distinguishing temporal sequence from all other instances of one-dimensional order, such as that of points on a line, numbers in order of magnitude, and so on.

     It is plain that Absolute Becoming is different from qualitative change. An example of the latter would be the gradual melting of a lump of ice in the sunshine, the sudden alteration in the pitch of the sound heard when a whistling locomotive rushes by one, and so on. The contrary opposite to qualitative change is qualitative invariance. An example would be the sound heard when a whistling noise of constant pitch, loudness, and tone-quality is made by a locomotive in presence of a hearer who is at rest relatively to it. Now Absolute Becoming is indifferent to whether there be qualitative variation or qualitative invariance. A superseding phase may be qualtitatively indistinguishable from that which it supersedes and from that which supersedes it. Again, in the case of a qualitative variation it is sensible to ask: At what speed is it taking place? We know that the speed of some such changes is greater than that of others, and the speed of any particular qualitative change is a matter for empirical investigation. But there is no sense in asking: At what speed does a certain phase, which was present, retreat into the past? And there is no sense in the suggestion that some might do this faster than others.

     Nevertheless, there is undoubtedly a very strong temptation to talk of Absolute Becoming in terms of qualitative change, and particularly in terms of some kind of motion. I am quite sure that all such ways of talking are misleading for the reasons given above. Moreover, if offered as an analysis of Absolute Becoming, they involve a kind of vicious circle. For the notions both of qualitative change and of qualitative invariance plainly presuppose that of Absolute Becoming, in the sense of that phrase which I have indicated. This circularity is the fundamental objection to all such metaphors. Particular forms of the metaphor have in addition, particular defects characteristic of each. The "policeman's bulls-eye" metaphor, e.g., if taken seriously, presupposes that what has not yet supervened and what has already been superseded in some sense "coexist" with each other and with what is now occurring. Again the metaphor of the history of the world "growing continually longer in duration by the addition of new slices," which I took seriously in Scientific Thought, presupposes that phases, which have already supervened and been superseded, in some sense "co-exist" with each other and with that which is now happening.

      Let us, then, avoid metaphors and similies and concentrate on the following very simple example, viz., a prolonged sound, continuing for a minute without any variation in pitch, loudness, or tone-quality. Here there is the minimum temptation to imagine that the phases which have been superseded, e.g., the first 30 seconds of this sound, "continue somehow to exist," or that the phases which have not yet supervened, e.g., the last 30 seconds of the sound, "already somehow exist." And, since we have explicitly excluded all variation in quality, there is no temptation to confuse Absolute Becoming, i.e., the supersession of earlier phases by later ones, with qualitative variation.

(2) The notion of "Successive Phases."

      We seem now to be faced with a serious difficulty. I have spoken of one "phase" of a process "superseding" another, and of its being in turn "superseded by" another. But what is a single phase? Is it supposed to have duration, or is it supposed to be quite literally momentary?

     (i) Suppose we ascribe any duration, however short, to a phase which has supervened and has not yet been superseded. Then it seems plain that it must consist of an earlier sub-phase adjoined to a later one, and that either the earlier one has already been superseded by the later one, or the later one has not yet supervened on the earlier one. On either alternative only one of the two actually exists now. Obviously the same argument applies to each sub-phase itself, and so on without end.

      If that is denied, it would seem that the denier is committed to some such view as the following. He must suppose that the sequence of successive moments is discrete (like the sequence of integers); that there is an intrinsically indivisible unit of duration (viz., the interval between one moment and the next); and that each phase supervenes at one moment and is superseded at the next, and therefore has the intrinsically indivisible unit duration. Now I find this quite unintelligible. I can write the words "phase of finite, but intrinsically indivisible, duration," but I can attach no clear idea to what I have written. So I cannot regard this as a genuine alternative.

      (ii) Suppose, then, that we say that each phase is literally momentary and has literally no duration. Then, assuming the continuity of time and therefore that the phrase "next moment" is meaningless, we shall have to say that at one and the same moment a phase supervenes and is superseded. To many this may sound palpably absurd, but I am not sure that it is so.

      Let us, for once and for this special purpose, do what I have been warning the reader against, and compare Absolute Becoming with motion. Everyone must admit that a moving particle leaves each point which it traverses at literally the same moment at which it enters it. "Entering" refers essentially back to positions occupied before, and "leaving" refers essentially forward to positions occupied after, the moment and the point in question. Might not similar remarks apply mutatis mutandis to "supervening" and "being superseded?" These refer respectively backwards to phases which have been, and forward to phases which will be; but any momentary phase just momentarily is.

      (iii) Even if this answer to the alleged difficulty in question be accepted, I think that one tends to feel dissatisfied with the notion of literally momentary phases on another count. Surely the notion of a literally momentary phase (like that of a geometrical point or line or surface) is the notion either of a boundary between successive adjoined phases, each of finite duration, or of a limit to an endless sequence of shorter and shorter durations, one inside another, like an endless nest of Chinese boxes? If so, it presupposes the existence of phases of finite duration. And surely (it might be added) the latter alone could be actual existents. The literally momentary, like the literally punctiform, bears all the marks of an abstraction, incapable of actual concrete existence, as opposed to an existent particular.

      As a preliminary comment on this last objection I will ask the reader to consider for a moment the following geometrical analogue, viz., points without any spatial magnitude, lines with length but no area or volume, and surfaces with area but no thickness. We, whose spatial experiences are of the 3-dimensional kind, consider all these to be abstractions, of the nature of boundaries or limits. We regard objects extended in three dimensions as the only possible kind of extended particular existents. But a creature whose spatial experiences were of the 4-dimensional kind would presumably think of what we call a "solid" in the sort of way in which we think of a 2-dimensional surface. He would think of it as a boundary or limit with reference to objects extended in four dimensions, and he would regard the latter as the only possible kind of extended particular existents. Conversely, a creature whose spatial experiences were of the 2-dimensional kind would presumably think of what we call a "surface," not as a mere boundary or limit with reference to objects extended in three dimensions, but as the only possible kind of extended particular existent.

      These reflexions seem to show that the question whether a person will regard a spatial entity of a given number of dimensions as a particular existent or as a mere boundary or limit, depends on the number of dimensions characteristic of his spatial experience. If the entity is of that number of dimensions (e.g., 3 in the case of human beings), he will regard it as a particular existent (e.g., as a cubical block, a spherical globe, and so on). If it is of less than that number of dimensions, he will regard it as a mere boundary or limit (e.g., as a face of a cube, the surface of a sphere, and so on). If it is of more than that number of dimensions (e.g., 4 or more in the case of a human being), he cannot perceive it as such. He can perceive only what a 4-dimensional being would regard as various 3-dimensional boundaries of it, and he will take these to be particular existents. This at least enables one to see that the question whether a given spatially extended entity is a particular existent or a mere boundary or limit, is not so simple and unambiguous as it might seem at first sight.

      But I doubt if this really helps us in the present case. The question is whether we could regard literally momentary phases as actual existents, or whether we must regard them as limits or boundaries of phases of finite duration. Now our temporal experience is at least l-dimensional, whilst a momentary phase would be an entity of zero temporal dimension. So, on the principles laid down in the preceding paragraph for spatially extended entities, it would seem that we could not help regarding a literally momentary phase as a mere boundary or limit, and not as a particular existent.

(3) The theory of 2-dimensional Time.

      The only solution that I can think of is to allege that Time is of at least two dimensions, and that a phase which has zero duration in the dimension which we commonly recognise has a finite "duration" in the other dimension.

      A theory on these lines has been put forward and argued in detail by my friend and former pupil, Mr. H. A. C. Dobbs, in the British Journal for the Philosophy of Science for August 1951. His object was primarily to deal with

1. the facts which are summarised under the phrase "the specious present," and
2. certain notions of quantum physics.

      I am going to call the two temporal dimensions the "Θ dimension" and the "T-dimension." A completely instantaneous "phase-particle," as I will call it, would be represented by a point in the diagram, whose co-ordinates are Θ = θ and T = t. It might be denoted by the symbol p(θ,t). What we have been calling a "momentary phase" occurring at the instant t is represented in the diagram, not by a point, but by a straight line of finite length and no thickness parallel to the Θ-axis. It may therefore be described as "T-instantaneous," but it has a certain extension, which we will call "Θ-duration" in the Θ-dimension. Suppose that such a phase occurs at T = t in the T-dimension, and that it extends from Θ = θ_p to Θ = θ_q in the Θ-dimension. Then we can denote it by φ(t, θ_p -> θ_q). In the diagram the line PQ represents such a phase. ON represents T = t, OL represents Θ = θ_p, and OM represents Θ = θ_q.

      So much by way of notation and diagrammatic representation. We can now formulate the details of the theory as follows: --

      (i) We assume that there is a certain fixed direction in the 2-dimensional time-field, represented in the diagram by a fixed straight line OU, making an angle α with the axis OΘ. (It does not matter for our present purpose what the magniture of α may be, provided it is between O and π/2.)

      (ii) Every T-instantaneous phase stretches in the Θ-dimension from a phase-particle represented by a point, such as P, on the line OU, to a phase-particle represented by a point, such as Q, on a line O'U' parallel to OU and at a fixed distance from it along the Θ-axis. (For the present purpose it does not matter what may be the magnitude of the Induration represented by the distance OO'.)

      (iii) For each successive T-instantaneous phase, as the value of T continuously increases, the initial phase-particle is further along the line OU.

      (iv) Between any two T-instantaneous phases, no matter how near together be the respective values of T, there is a third T-instantaneous phase.

      It will be noted that we have secured by these suppositions a consistent combination of (a) continuity of transition, (b) the finite Θ-duration of each T-successive term, and (c) the instantaneity of each T-successive term. This is secured by the fact that T-instantaneous phases, though completely successive in the T-dimension, partially overlap in the Θ-dimension provided that the difference in their T-dates does not excede a certain maximum, and that the nearer their T-dates are to each other the more nearly complete is this overlap.

      A glance at the diagram will show that there must be a kind of "natural unit" of T-time-lapse, correlated with the "natural unit" of Θ-duration represented in the diagram by the distance OO'. (This might be compared with the natural unit of 4 right-angles in the case of angles.) In the diagram let the straight line MQ be produced upwards until it cuts the fixed line OU at P'. Then the phase P'Q' is the first successor to the phase PQ which does not overlap PQ at all. Thus the line QP' represents a kind of natural unit of T-time-lapse. This is obviously connected with the natural unit of Θ-duration and the fixed angle α by the relations P'Q/PQ = P'Q/OO' = tanα. Let us denote the natural unit of T-time-lapse by τ, and the natural unit of Θ-duration by σ. Then τ = σ tan α.

      It is plaint that τ, the natural unit of T-time-lapse, can belong only to a T-sequence of phase-particles, all of which have the same value of Θ. Such a sequence begins with a phase-particle (such as Q) which is at the terminal end of a complete T-instantaneous phase (such as PQ), and it ends with a phase-particle (such as P') which is at the initial end of a certain later complete T-instantaneous phase (such as P'Q').

      It will be of interest to consider next the T-time-lapse belonging to a sequence of sub-phases, all of which have the same initial value and the same terminal value of Θ. For this purpose we can consider the sub-phase represented by the segment pQ of the line PQ (which represents a complete phase). Through p draw a line parallel to OT. Let it cut the fixed line OU at Π, and the axis OΘ at λ. Then it is evident that there will be a sub-phase extending from Θ = Oλ to Θ = OM in every successive complete phase from PQ to Πq, both inclusive, and in no others. So the T-time-lapse belonging to this sequence of sub-phases is represented by Πp. Now Πp = Pp tan α. But Pp = (PQ - pQ) = (σ - pQ). And we have already shown that tan α = τ/σ. So Πp = (σ - pQ) τ/σ = τ(1 - pQ/σ). If we put Πp = t, and pQ = s, we can write this in the form t = τ(1 - s/σ). So the T-time-lapse belonging to such a sequence of sub-phases varies between the limits 0 (when s = σ, and the "sub-phase" is supposed to swell into a complete phase of natural unit Θ-duration) and τ (when s = 0, and the "sub-phase" is supposed to shrink into a mere phase-particle.)

(4) The Specious Present.

      This brings me to the question of the Specious Present. What I have to say about this is in principle the same as what I said in Vol. II, Part I, of the Examination (Pp. 281-288). But on the one hand it becomes considerably clearer when stated in terms of 2-dimensional Time, and on the other hand it provides a concrete illustration of the abstract account of the latter given above.

      It is evident that the 2-dimensional diagram at the top of P. 285 in my account of the specious present in the EMcP would have to be replaced by a 3-dimensional diagram. For we have now to represent two temporal dimensions (instead of one only, as in the EMcP), and in addition (as there) the magnitude which I called "degree of presentedness." The modifications needed will be understood without difficulty, if the reader will refer back to the diagram given above in expounding the general theory of 2-dimensional Time.

      Suppose now that the lines PQ, Πq, P'Q', etc., in that diagram represent the Θ-durations of T-successive specious presents. Then we should have to represent degree of presentedness by distances along a third axis, sticking out at right-angles from the plane of the paper. We must regard each of these lines as the base of a right-angled triangle, e.g.,

      It may be of interest, however, to add a diagram representing in terms of this theory the hearing of the sound of a short word, e.g., "RAG," which falls well within the Θ-duration of a human specious present. The diagram is in principle the same as that given above, but it now must be regarded as the plan in the T-Θ-plane of the complete 3-dimensional diagram.

      Consider the above sequence of seven lines all parallel to OΘ. They epresent seven T-successive specious presents out of a continuous seqence of specious presents. In No. 1 the sound "R," and nothing more of the sound "RAG," is just being presented to the hearer, at the latest end of a specious present, and therefore with maximal degree of presentdness. In No. 2 this is true of the "A" sound. The "R" sound is still being presented, but now with less than maximal degree of presentedness. So the "RA" sound as a whole is being presented. In No. 3 the "RAG" sound as a whole is for the first time presented. The "G" sound as now the maximal degree of presentedness, the "A" sound less, and the "R" sound still less. In No. 4 (and in all the innumerable successive specious presents between those represented in the diagram by No. 3 and by No. 5) the "RAG" sound as a whole continues to be presented, but with steadily diminishing degree of presentedness. In No. 5 the "R" sound is just on the point of ceasing to be presented and being at most remembered. In No. 6 only the "AG" sound is still presented; the "RA" sound has ceased to be presented. In No. 7 nothing is any Monger presented of the sound "RAG" except the ghost of the "G" sound in the act of vanishing.

(B) PRECOGNITION.

      This topic forms the theme of a part of Professor Ducasse's paper and of the whole of Professor Flew's. I will begin with Professor Ducasse's "Theory Theta." Although this is put forward primarily to deal with the problem of non-inferential precognition, it is a general theory of time, and therefore highly relevant to the topics which we have been discussing above.

(1) Professor Ducasse's "Theory Theta."

      The theory falls into three divisions, viz.,

1. Inter-relations of physical events,
2. Inter-relations of experiences, and
3. Relations between experiences and physical events.

(i) Inter-relations of Physical Events.

      (a) Purely physical events, which do not overlap each other, form a 1-dimensional quasi-temporal series ordered by an irreducibly triadic relation, which I will call "chronical betweenness." By saying that this relation is irreducibly triadic we mean that the statement that the physical event Y is chronically between the physical events X and Z is not analysable into the statement that either X is earlier than Y and Y earlier than Z, or X is later than Y and Y later than Z. In the sequence of purely physical events there is no asymmetrical dyadic relation, such as earlier-and-later, which would give an intrinsic direction to it.

      (b) If we take any two terms U and U' in such a series, we can subdivide all the remaining terms in it into two mutually exclusive and collectively exhaustive sub-classes, as follows, viz.,

(α) those which are on the same side of U' as U is, and
(β) those which are on the opposite side of U' to U (i.e., those of which it is true that U' lies between them and U.)

      (c) Let us now consider a term X, which is on the same side of U' as U is. Then there are two mutually exclusive and collectively exhaustive possibilities, viz.,

(α) that X is between U and U' (i.e., that X is "chronically nearer" to U' than U is), or
(β) that U is between X and U' (i.e., that X is "chronically further" from U' than U is.)

(α) "X is past to U from U' " means the same as "X is chronically on the same side of U' as U, and is chronically nearer to U' than is U." I will denote this by Π (X, U; U').
(β) "X is future to U from U' " means the same as "X is chronically on the same side of U' as U, and is chronically further from U' than is U." I will denote this by φ (X,U;U').

      I think it is wiser to keep to the symbols and their definitions, and not to use the phrases "past to . . . from" and "future to . . . from," when talking of the quasi-temporal inter-relations of purely physical events. For these phrases inevitably have associations which may mislead us.

      It should be noted that, although the two relationships Π (X,U;U') and φ (X,U;U') are mutually exclusive, they are not collectively exhaustive. Both of them presuppose that X is on the same side of U' as U is. Obviously there remains the possibility that X should be on the opposite side of U' to U. In that case obviously X and U' would be on the same side of U, and X would be further from U than is U'. So we should have φ (X,U';U). So it would seem that for any three non-overlapping purely physical events the three mutually exclusive and collectively exhaustive possibilities would be Π (X,U;U'), φ (X,U;U'), and φ (X,U';U).

      It is also worth noting that Π (X,U;U') is equivalent to φ (U,X;U'). For to say that X is nearer to U' than is U, is obviously equivalent to saying that U is further from U' than is X.

      I hope that the above is a complete and correct formal statement of Professor Ducasse's account of the quasi-temporal order of purely physical events. It may be remarked that it is precisely analogous to the intrinsic spatial order of points on a straight line. There is no intrinsic "sense" in the order of points on a line. When we ascribe one to it, we do so either by reference to our right and left hands, or by imagining something traversing it and so occupying certain points earlier and others later.

(ii) Inter-relations of Experiences.

      I am not at all sure that I fully understand Professor Ducasse's account of the temporal order of experiences. I think it is plainly concerned primarily with the experiences which together make up the mental history of some one conscious individual. Again, I think it is concerned both

1. with what almost everyone would call "experiences," e.g., feeling a twinge of toothache, or the popping up into consciousness of a name which one had been trying to recall, and
2. with what some (but not all) philosophers would refuse to call "experiences," but would prefer to describe as the "immediate objects" of certain experiences, and would call "sense-data," "mental images," and so on.

1. an experience, or
2. a sense-datum or a mental image or (in general) a "prehensum";

1. having that experience, or
2. sensing that sensum or imaging that image or (in general) "prehending" that prehensum.

      On these assumptions, I think that this part of the theory certainly includes the following propositions: --

1. What is "present to" a person at any given moment consists of a certain set of sub-phases. All of these are then present to him with some degree or other of what Professor Ducasse calls "liveness," and each different one of them is present to him with a different degree of liveness. One and only one of them (which may itself be internally complex) is then present to him with the maximal degree of liveness. I propose to call them "degree-of-liveness sub-phases."
2. It follows that what is present to a person at any given moment constitutes a finite segment, in respect of the different degrees of liveness, from maximal to minimal, with which each different sub-phase is then present to him. Let us call this a "degree-of-liveness segment."
3. What is present to a person at any moment is not merely in fact a degree-of-liveness segment. It is presented to him as such a segment, and he can formulate these facts about it if he inspects it and reflects on his findings.
4. That sub-phase of what is present to a person at any moment which is then present to him with maximal degree of liveness, is at that moment strictly present. All the other sub-phases in the segment are then strictly past. The degree of pastness of each is correlated conversely with its degree of liveness.
5. Every degree-of-liveness sub-phase of the segment which is present to a person at any moment may be called "speciously present." This serves to contrast them all with sub-phases which have been present to the person, but are no longer so, and with others which have never been present to him but may be so later.

      Now it will be noted that all the propositions which I have ascribed above with some confidence to Professor Ducasse have involved the phrase "at any one moment." I find it impossible to state the theory (or indeed any account of "specious presentness") without introducing that phrase, or some equivalent of it. But it is plain that the theory would be hopelessly inadequate unless it also referred to a plurality of successive specious presents, and it is at this point that I feel very uncertain as to Professor Ducasse's meaning.

      It seems to me that at least the following statements would need to be added, but I am not sure which (if any) of them Professor Ducasse would accept: --

1. If a sub-phase is present to a person with the maximal degree of liveness at the moment t₁, then

   (α) at no moment before t₁ was it present to him at all;
   (β) at each successive moment after t₁ (up to and including a certain moment t₂) it will be present to him with a lesser and lesser degree of liveness, and at t₂ with minimal degree; and
   (γ) after t₂ it will never be present to him again.

2. If a sub-phase is present to a person with a degree of liveness less than the maximal at the moment t, then there is a moment t₁ (earlier than t) such that it was present to him with maximal degree of liveness at t₁.

      Now one reason why I am doubtful whether I have fully understood Professor Ducasse's theory is this. On the one hand, he appears to make such statements as the following. To call any phase "strictly present" means simply and solely that it is present to one with the maximal degree of liveness. To call any phase "strictly past" means simply and solely that it is either

1. present to one with less than the maximal degree of liveness, or
2. not present to one, but capable of being attended to only by g "recalling" it in memory.

      Now I agree that the ordered degrees of liveness with which a number of different sub-phases are present to a person at any given moment of his life may well be one essential factor in the experiential basis of our notions of past and present, of earlier and later, and so of Time. But surely a no less essential factor is the experience

1. of what was just lately present to one with maximal liveness being now present to one with lesser liveness,
2. of something which just lately was not present to one at all being now present to one with maximal liveness, and
3. of what was just lately present to one with minimal liveness being now no longer present to one at all.

(iii) Relations of Experiences to Physical Events.

      I doubt if I fully understand Professor Ducasse's account of the relation between the triadic quasi-temporal order of purely physical events and the dyadic genuinely temporal order of experiences. It is this which plays an essential part in his theory of Precognition. The following points seem to be certain: --

1. To say of a physical event that it is "present" at a certain moment means simply and solely that it is then the object of a perception which is strictly present to some percipient.
2. It is logically possible that one and the same physical event should be perceived on several different occasions (either by the same person or by different persons), and therefore, by definition, that it should be present on as many different occasions.
3. In a genuine case of precognising non-inferentially a physical event e, what would happen would be this.

   (α) The person who is said to "precognise" e in fact perceives it, but in an abnormal way, viz., not by means of sensations (which Professor Ducasse describes as "vivid images caused at the time through the functioning of the sense-organs").
   (β) Later he, or someone else, perceives e in the normal way, i.e., by means of sensations.

      It seems to me that these three statements would at any rate need the following qualifications.

1. I think that the first of them would need to be amplified somewhat as follows. A physical event e is present at any moment, if either

   (α) it is then the object of a perception which is strictly present to a percipient, or
   (β) it is contemporary with another physical event e', of which this is true.

   Unfortunately Professor Ducasse has given no account of "simultaneity" between purely physical events.
2. As Professor Ducasse recognises, it would be necessary (in view of the finite velocity of light and sound and of the transmission of nervous impulses) to modify the original statement somewhat as follows. A physical event, perceived normally by the senses, which would be called "present" in accordance with the proposed definition, would always in fact be earlier (and in some cases very much earlier) than the strictly present sense-perception of it. It would therefore be really past at the moment when, if the definition were taken as it stands, it would be called "present."
3. Professor Ducasse admits the possibility that a physical event which is non-sensuously "perceived" by one person may later be perceived normally by another person. This presupposes some temporal correlation between the mental histories of different persons. Obviously there is such a correlation. But all Professor Ducasse's statements about the temporal interrelations of expriences have been confined to those which fall within the mental history of a single person.

      The above are comments on matters of detail. The two following are more general.

1. What are we to understand by the kind of non-sensuous perception of a physical event, which Professor Ducasse postulates in contrast with ordinary sense-perception? He seems to assume that there is some generally admitted definition or description of the genus or determinable "perception," which leaves the two possibilities "sensuous" and "non-sensuous" open as specifications of it. But, if so, what is it?
2. At an earlier stage of his essay, in dealing with the alleged "fatalistic objection" to the possibility of non-inferential precognition, Professor Ducasse asserts (quite rightly) that many of the instances of "veridical precognition" are not instances of cognition (i.e. knowledge) of the future event which will in course of time verify or refute the "precognition." I do not see how this fits in with the later suggestion that in cases of veridical precognition one and the same physical event is twice perceived, first non-sensuously and later sensuously, by the same person or by different persons.

(2) Professor Flew's Comments.

      I pass now to Professor Flew's essay. This is concerned with my treatment of three prima facie objections which I alleged that many people feel in connexion with the very notion of veridical non-inferential foreseeing. I called these the "Epistemological," the "Causal," and the "Fatalistic" objections.

(i) The Epistemological Objection.

      As Professor Flew agrees in the main with my statement of this and with my answer to it, I will comment only on the following point. I stressed an alleged analogy between ostensible foreseeing and ostensible remembering of incidents, persons, and things. In doing so, I said that a present image is involved in ostensible remembering, and I asserted or implied that one would be involved in ostensible foreseeing also. Professor Flew points out that in The Mind and its Place in Nature I had denied that a present image is necessarily involved in ostensible remembering, and he speculates on the cause of my "backsliding."

      In point of fact I have never seen occasion to alter the opinion on this point which I expressed in MPN. The explanation of my apparent "backsliding" in "The Philosophical Implications of Precognition" is this. I was thinking exclusively of the sporadic cases on record at the time in the books and papers which I mentioned. The experimental work in connexion with card-guessing, which has since become the most important evidence for "precognition," was not then available. Now images (in a wide sense which includes the quasi-sensa of dreams and waking hallucinations) are involved in most of the sporadic ostensibly precognitive experiences, and images are involved in many experiences of ostensibly remembering events, persons, or things. In the paper in question I was concerned to stress the resemblance to memory, and to undermine the common assumption that veridical non-inferential precognition, if it occurs, must be of the nature of perception.

      It is plain that the card-guessing results bear very little resemblance to experiences of ostensible remembering, and do not fit at all into the framework of my Aristotelian paper. This is of considerable importance in reference to Professor Flew's essay, for it is evident that the empirical data which he has in mind are correlations between a sequence of guess-values and a sequence of target-values, as in a card-guessing experiment.

      Here the only relevant property of any guess is that it is a guess that the target-card bears such and such a one of a small number of known alternative symbols, e.g., a cross, where the alternatives are known to the guesser to be, e.g., a cross, a square, a circle, a wavy line, or a triangle. No-one would say here of any particular correct guess that it is "at least a very remarkable coincidence." One would say this only of the proportion of correct guesses in a long sequence of guesses. And one would say it only if that proportion were to differ (either by excess or defect) from "the proportion most probable on the hypothesis of chance-coincidence" by several times the "standard deviation" for such a sequence on that hypothesis. At a certain point, which would differ from person to person, one would be inclined to say: "This excess (or defect) is altogether too great to be reasonably regarded as a mere freak of chance."

      Now the kind of case which I had in mind was different. Here a person is presented on a certain one occasion with an image or a quasi-sensum of a very detailed and elaborate kind, and no obvious cause can be suggested for the occurrence of that experience in him at that time. Not too long afterwards there happens an event in the external world, which could not normally have been expected by him at the time, and it corresponds in a remarkable way in its details with the experience in question. One is inclined to say of any particular pair of events of this kind that it constitutes "at least a very remarkable coincidence." And if the singularity and unexpectedness of the later event, and the degree of detailed correspondence between it and the earlier experience, surpass a certain point (which again would differ from person to person), one would be inclined to say: "This correspondence is altogether too peculiar and too detailed to be reasonably regarded as a mere freak of chance."

      Since Professor Flew confines himself to evidence of the first kind, I shall do so too. I will only remark that I think it would be very difficult, in the case of some of his arguments, to adapt them to evidence of the second kind, which is what I had in mind in writing the paper on which he is commenting.

(ii) The Causal Objection.

      I think that Professor Flew's formulation of the alleged objection is essentially correct, but I will re-state it in my own way. A certain person P makes a sequence of guesses as to which one of a small number of known alternative symbols will be on the face of the next card which is about to be turned up in a certain experiment. Here a "hit" is a case where the symbol guessed is the same as that on the face of the card next turned up after the guess has been made. What is said to be "too great to be a mere chance coincidence" is the deviation (positive or negative) between the actual proportion of hits in the whole sequence and what is called "the most probable proportion of hits, on the hypothesis of chance coincidence, in such a sequence."

      Now

1. by definition we are not to count the results as evidence for foreseeing, if the difference between the actual and the most probable proportion of hits can be dismissed as "merely a remarkable coincidence." But
2. to deny that it is a mere coincidence is to allege that there is some causal connexion between

   (α) some at least of the events described as "a guess that the next card will have such and such a symbol on its face" and
   (β) some of the events described as "the turning up, immediately after the making of that guess, of a card with such and such a symbol (the same or different) on its face."

   Let us call these respectively a "G-event" and its "A-correlate." Now
3. it is contrary to the notion of causation that the A-correlate to a G-event should be a factor in causing the latter. For the A-correlate does not begin until after the G-event in question has ceased. Therefore
4. the only kinds of causal connexion that are possible are the following. Either

   (α) a G-event is a cause factor in a causal ancestor of its A-correlate; or
   (β) a G-event and its A-correlate are effect factors respectively in an earlier and in a later causal descendant of some common causal ancestor; or
   (γ) the G-event was determined by the result of an inference as to the nature of its forthcoming A-correlate, either made somehow by the guesser himself or made by someone else and somehow imparted by him to the guesser. But

5. all these alternatives are so many ways of accounting for the difference between the actual and the most probable proportion of hits in the sequence without supposing that any of the guesses is an instance of veridical non-inferential precognising.
6. It would seem, therefore, that my definition of "foreseeing" rules out the possibility of any of the guesses counting as instances of foreseeing the nature of the target-card, no matter how great may be the deviation between the actual proportion of hits and the proportion which would be most probable on the hypothesis of chance-coincidence.

      In the above reasoning it is the second step to which Professor Flew takes exception, viz., the transition from denying that so great a divergence can be a mere chance coincidence to asserting that there must be some causal connexion between some at least of the G-events and their A-correlates. Professor Flew's alternative, as I understand it, may be stated as follows: --

1. It is a fact, with regard to certain persons P₁, P₂, . . . . . . P_n, who have been investigated, that on all or most occasions when any of them has made a long sequence of guesses under certain assigned conditions, the proportion of hits has differed significantly from the most probable proportion on the hypothesis of chance-coincidence.
2. It is reasonable to believe that the guess made by such a person on any occasion is causally determined, and it is reasonable to believe that the immediately subsequent turning up of a card with such and such a symbol on its face by the experimenter is causally determined. But in a properly designed experiment there is no reason to believe that there is any causal connexion, direct or indirect, between the former event and the latter, whether the guess be a hit or a miss.

      Notwithstanding this complete lack of causal connexion, the follow. ing expectations might reasonably be entertained and the following enquiries might reasonably be undertaken, according to Professor Flew.

1. It would be reasonable to expect, with regard to any of these persons, that, if he were to make a further sequence of guesses under similar conditions, the proportion of hits would still diverge significantly and in the same direction from the most probable proportion on the hypothesis of chance-coincidence.
2. It would be sensible to compare these persons with each other, and to contrast them with others who (under apparently similar conditions) consistently make scores which do not differ significantly from what might be expected on the hypothesis of chance-coincidence, in order to discover some characteristic ψ, common and peculiar to the former. If such a characteristic were found, it would be reasonable to put forward, and to test by further experiment and observation, following wider generalization: -- Most (if not all) persons having the characteristic ψ, and few (if any) who lack it, would, if making a sequence of guesses under conditions C, score a proportion of hits significantly different from the most probable proportion on the hypothesis of chance-coincidence.
3. Even if only the narrower extrapolation, labelled (a) above, were available, the following statement would be true. Suppose that an observer O accepted this extrapolation. Suppose he knew that on a certain occasion one of these persons P_r had guessed that the next card to be turned up would be of such and such a kind. Then O would be justified in conjecturing, with greater conviction than would otherwise have been warranted, that such a card would be turned up next (if P_r's previous performance had shown a significant positive deviation), or that such a card would not be turned up next (if P_r's previous performance had shown a significant negative deviation).

      Now, as regards the logic of the question, I find myself largely in agreement with Professor Flew here. Let us assume that the cards are properly randomised, that the experiments are properly conducted so as to eliminate all possibility of fraud, sensory leakages, and so on. Then the conclusion which can legitimately be drawn from a successful series of such experiments can be accurately stated in the following rather complex sentence:

"In view of the actual results obtained with the subject S, it is extremely unlikely that the probability of his assigning any particular one of the alternative symbols to the next card to be turned up is the same no matter whether that symbol or any of the alternatives to it will in fact be on the face of that card."

      That is the prima facie case in favour of Professor Flew's contention. I think it is strong, but not absolutely conclusive. Both "probability" and "causation" are extremely obscure and ambiguous notions, and one cannot be quite sure that sentences which explicitly mention only the former may not implicitly refer to the latter.

      Without entering into that question, we can ask ourselves the following one: -- Under what circumstances would one's initial impression, that the deviation of the actual proportion of hits from the proportion which would be most probable on the hypothesis of chance-coincidence is too great to be a chance-coincidence, be strengthened? And under what circumstances would it be weakened? It seems to me plain that it would be strengthened or weakened according as the answers to such questions as the following were affirmative or negative. Are the results as a whole repeatable,

1. in the sense that there are generally a few subjects who can produce them, and
2. that each such subject can go on producing them over a longish period when the known conditions are kept as constant as possible?

      Before concluding this sub-section I would add the following remark. If we consider in detail how card-guessing experiments are designed and conducted, it seems that in most of those which are said to provide evidence for "precognitive telepathy or clairvoyance" there is no necessity to postulate foreknowledge at all. The results could in fact be interpreted causally, and the causation would involve no temporal difficulties or paradoxes, though it would be in other respects extremely peculiar.

      This is not the place to develop the point in detail, and the following general hint must suffice. Suppose we assume that the bodily action which the "telepathic agent" performs, on receipt of his cue from the experimenter that the subject has made a guess, is completely determined causally. Then a complete causal ancestor of that action must have already existed immediately before the subject made his guess. Now it is this action, together with the experimental set-up, which determines what symbol the agent will perceive on the receipt by him of his cue from the experimenter. Suppose we assume that the bodily action which the subject performs, in writing down such and such a symbol as his guess about the card which the agent is about to look at, is also completely determined causally. Then all that we need to assume in order to account causally for (say) a significant predominance of hits is this. We must suppose that that causal ancestor of the agent's future action of selecting and looking at such and such a card, which immediately precedes the subject's present action of writing down his guess as to the nature of that card, contains a factor which influences the subject to write down as his guess the symbol which it is already determined that the agent will perceive. This would be a very odd kind of causal law, but its oddity would arise from its unfamiliarity and not from its involving causal influence from future to present or later to earlier.

(iii) The Fatalistic Objection.

      I will begin by re-stating the objection. Suppose that at t₁ a person A veridically foresees an event which in due course happens at t₃; and suppose that an essential factor in causally predetermining that event was a voluntary decision, made at a certain intermediate moment t₂ either by A himself or by another person B. (The following would be an example. Mr. Jones correctly foresees at t₁ that Mr. Smith will be killed in an aeroplane-crash at t₃; and a necessary precondition of this happening was that Mr. Smith decided at t₂ to travel by a certain plane and not by another plane or by boat.)

      Then the argument runs as follows.

1. Since the event was correctly foreseen at t₁, it must have been completely predetermined causally by that time.
2. It must therefore have been completely predetermined causally at the later moment t₂.
3. Therefore the voluntary decision made by B at t₂, which was an essential factor in the causal ancestor at t₂ of the event in question, must have been completely pre-determined causally at least as early at t₁.
4. We must therefore draw the following conclusion in such a case. Either

   (α) B's voluntary decision at t₂ was not (as we had assumed it to be) a causally necessary precondition of the occurrence of the foreseen event. Or
   (β) that decision was completely predetermined causally at least as early at t₁.

      Now both alternatives are highly distasteful to the feelings of many, though that is of course no reason for holding that neither of them could be true. The first alternative is in the worse position. We have as good reason for holding that Mr. Smith, in my example, would not have travelled by the plane in question, unless he had decided at an intermediate moment to do so, as we have for almost any empirical belief as to the consequences of unfulfilled conditions.

      In commenting on this "objection," as I originally stated it, Professor Flew makes the following points: --

1. He says that he rejects the assumption that freedom to decide between alternatives entails that voluntary decisions are not completely predetermined causally. If the reader will look at my re-statement above he will see that that assumption is not involved. All that I allege is that many people find it distasteful to think that their voluntary decisions are completely predetermined causally long before they come to be made. I think that they would do this, even though they admitted that such decisions would still be free in a number of important senses, which were mentioned by Locke and have been repeated ad nauseam by other philosophers since his time.
2. The premiss in this "objection" which I myself questioned was that which comes first in the above re-statement, viz., that, if an event were correctly foreseen at a certain moment, it must have been completely predetermined causally at that moment. I do not know what attitude Professor Flew would himself take towards this premiss. But he does argue that to reject it (as I was inclined to do) is inconsistent with certain things which I said or implied about veridical foreseeing

      The point is this.

1. My definition of "veridical foreseeing" entails that an event would not be said to have been "veridically foreseen" unless there were some kind of causal connexion between the experience of a "foreseeing" and that event "foreseen." For, otherwise, according to what I said, the correspondence between the two would be no more than a remarkable coincidence.
2. My account of an event being "completely predetermined causally at t" was this. There is a set of facts about the states and dispositions at t of the various things and persons then existing, which, together with the laws of nature, logically entail that an event of precisely that kind would happen at the time and place at which that event did happen or will happen.
3. Now suppose that there occurs at t an experience answering to my definition of "veridical foreseeing." Then it must be included among "the states and dispositions at t of the various things and persons then existing." And among the "laws of nature" we must include, not only the laws of normal physics and psychology, but also those causal laws (whatever they may be), in accordance with which experiences of veridically foreseeing and the events veridically foreseen are (according to my definition) causally interconnected.
4. But, when that is done, it is not open to me (consistently with my definitions) to question that, if an event is veridically foreseen at t, it must be causally predetermined at t.

      I accept this criticism of Professor Flew's on my consistency, at any rate to the following extent. I certainly did forget to include the experience of veridically foreseeing itself among the states and dispositions of the various things and persons existing at the time when it occurred. And I certainly failed to notice that my assertion that there must be some kind of causal connexion between an earlier experience and a corresponding later event, if the former is to count as a "veridical foreseeing" of the latter, implies that among the laws of nature there are laws concerning just that kind of causal connexion. But it does not seem to me to follow, even when these factors are taken into account, that the event foreseen at t must be completely predetermined causally at t. It might at that moment be causally predetermined only within certain limits.

      This has a bearing on the last point which I will consider under this head. Consider the following modification of our previous example. Mr. Smith, hearing of the experience in which Mr. Jones ostensibly foresees his death in the crash of a certain plane (or, alternatively, having such an experience himself), cancels his booking and thus saves his life, though that plane does crash and all the passengers on it are killed. It is sometimes said that in such cases the occurrence of a veridical foreseeing causes voluntary action to be taken which prevents it from being fulfilled.

      About all such cases it seems to me that Professor Ducasse is right. In so far as it is known that a claim has been made to foresee that a certain kind of event will happen at such and such a time and place, and in so far as that knowledge leads to action which prevents an event of exactly that kind from happening there and then, the claim as it stands is mistaken. But, if the claim were re-stated in a conditional form, or in a less determinate categorical form, there may be no reason against and good reasons for admitting it.