These three volumes of Professor Anscombe’s collected papers encompass everything of importance that she has published, apart from her work as literary executor and translator of Ludwig Wittgenstein and her three books: Intention, An Introduction to Wittgenstein’s ‘Tractatus’ and Three Philosophers, written in collaboration with Professor Peter Geach, and containing studies of Aristotle, Aquinas and Frege. Her interest in the topic of intention and the teachings of Aristotle reappears in these papers, but they have little overtly to do with either Aquinas or Frege, and the influence of Wittgenstein is much less obtrusive than one might have been led to expect. Only in one paper, ‘The Reality of the Past’, which first appeared in 1950, during Wittgenstein’s lifetime, do we find the submissive footnote: ‘The best that I have written is a weak copy of some features of the original, and its value depends only on my capacity to understand and use Dr Wittgenstein’s work.’
‘The Reality of the Past’ is bracketed in the second volume with an article entitled ‘Memory, “Experience” and Causation’ which Professor Anscombe contributed in 1974 to the fourth series of Contemporary British Philosophy. Both pieces contain many points of interest. In the first of them, she is concerned to explore the sense of statements about the past and account for our ability to understand them. She musters a series of arguments which lead to what I believe to be a true conclusion: ‘In general we must fail if we try to explain the sense of statements about the past by means of present memory, consciousness of meaning, quality of images, or anything else of the kind. For either we have left out all reference to the actual past, or we have surreptitiously introduced it into an explanation that proposes to do without it.’ Her own solution, which she may perhaps owe to Wittgenstein, consists in subordinating what we call knowledge of the past to grammatical training in the use of the past tense. I find this unacceptable in that it runs into a dilemma: if ‘the past tense’ is construed syntactically, nothing is implied concerning the circumstances of its use: if it is construed semantically, the question is begged. For my part, I can see no objection to holding that the relation of temporal precedence is ‘given’, as spatial relations are, in sensory experience, in which case the past can result from a temporal projection: it can be taken as comprising anything earlier than the earliest slice of the specious present.
In the second article, Professor Anscombe addresses herself to the example, familiar in the literature, of Goethe’s believing, perhaps as the result of being told about them, that such-and-such episodes occurred during his childhood, but being uncertain whether he remembers them. This example has been used to support the view that a causal condition enters into the analysis of memory-judgments. To say that Goethe did remember the episodes in question is supposed to entail that the proper causal relations obtained between his witnessing of the episodes and his subsequent recollection of them. Professor Anscombe quite rightly rejects this view, without even having to dwell on the difficulties of specifying the ‘proper’ causal relation. As she truly says, ‘in a memory experience, the “past” reference, and indeed a quite particular “past” reference, must already be contained independently of the causal relation which we may postulate.’ There is, however, a minor point here on which I disagree with her. Noting that the word ‘remember’, in its ordinary acceptation, carries the implication of knowledge, Professor Anscombe coins the word ‘rember’ to stand for the residue of remembering, when this implication is cancelled. Professor Broad, in his unduly neglected book The Mind and Its Place and Nature, had used the phrase ‘ostensibly remembering’ in the same way, its purpose being to leave it open whether the episode which was ostensibly remembered had actually occurred. In Professor Anscombe’s case at least, it is assumed that if there are such experiences they are self-intimating. What she seems to find questionable is whether they exist. However that may be, she assumes that, in the Goethe example, the only points at issue are whether he witnessed the events in question and did not need to be reminded of them. If these conditions are satisfied, it follows that he remembered the events. But if his not needing to be reminded does not bring back the causal condition, which we are at one in rejecting, it appears superfluous. All that is required of Goethe is that his belief that he remembered the past events be true. But Professor Anscombe herself would admit that this is insufficient. The question is what is lacking. I contend that it is just the absence of the distinctive experience which gives such a true belief the stamp of memory. Ex hypothesi Goethe had no doubt that his account of his experiences was true. What he doubted, to borrow Professor Anscombe’s usage, was whether he genuinely ‘rembered’ them.
A more serious point on which I find that I disagree with Professor Anscombe concerns her attempt to establish a particular kind of asymmetry between the future and the past. This arises also out of one of her historical essays entitled ‘Aristotle and the Sea Battle’ which first appeared in Mind in 1956. Following Aristotle, Professor Anscombe wishes to maintain that in the case of a proposition which refers to a past or present event, its truth, if it is true, or its falsity, if it is false, is fixed once for all, or necessary, in Aristotle’s special sense of the term; whereas this does not apply to propositions concerning the future. From this she infers also that we can have no knowledge of future events. She argues that even if someone were always right in predicting her actions, he could not properly be said to know what they were going to be because, whatever his prediction, ‘the facts could go against it.’ I am not concerned at this point with the rules for the use of the verb ‘to know’, but only with the force of the argument that the possibility of the facts going against it differentiates a true statement about the future from one about the present or past. I contend that it makes no relevant difference. In the sense in which the past and present cannot be changed, neither can the future. What has been has been. What is is, and what will be will be. It may be true of a future event that it would not occur if previous events had been different, but it is equally true that if some future event were to be different some present event would not be occurring or some past event would not have occurred. The subjunctive conditional contained in ‘the facts could go against it’ works both ways in time. This is independent of the issue of determinism, with which I shall deal later on.
Two very interesting articles which Professor Anscombe reprints are ‘The Intentionality of Sensation’, which she contributed to the second series of Analytical Philosophy in 1965, and ‘The First Person’, which was one of the Wolfson College lectures for 1974, published under the title Mind and Language. She derives the concept of intentionality from that of intention, and picks out three features of intention as being relevant for her subject: first, that an intentional act is so only under some, not all, of its descriptions; secondly, that the descriptions under which an act is intentional may be vague and indeterminate; and thirdly, that what is intended may not come true. This has as a corollary that the ‘direct objects’ of intentional verbs like ‘thinking of’ and ‘worshipping’ may or may not be real. Professor Anscombe makes the brilliant suggestion that philosophers who have on the one side asserted or on the other denied the existence of objects of the order of ‘ideas’ or ‘impressions’ or ‘sense-data’ have alike overlooked the fact that verbs of sense like ‘see’ are intentional. Where sense-datum philosophers have gone astray, in her opinion, is in construing expressions like ‘what you see’ materially, with the result that they have been led to treat sense-data as genuine objects. I own to having made this mistake in the past. It seems to me, however, that Professor Anscombe grants the sense-datum theorist all that he should need when she allows that ‘to say “X saw A” where “saw” is used materially, implies some proposition “X saw – ” where “saw” is used intentionally; but the converse does not hold.’
The argument of ‘The First Person’ is extremely subtle, and there may well be a vital point that I have missed. I will do my best to summarise it fairly. The use of the word ‘I’ is usually explained by its being said to be the word that each of us uses to speak of himself, but Professor Anscombe finds this explanation inadequate. To begin with, ‘I’ does not function like a proper name. There may be circumstances – for instance, one’s having a homonym – in which one does not know that one’s name is being used to refer to oneself, but this is never true of the use of ‘I’. There could indeed be a society in which the same particular sign, perhaps marked on each person’s wrist, was used by each person to refer exclusively to himself, while he used different signs to refer to other people, but the use of such a sign would be different from that of ‘I’ because it would not carry the implication of self-consciousness. So what is self-consciousness? What is a self? It might be suggested that it is a human being in a special aspect and ‘I’ would be the name ‘used by each one only for himself and precisely in that aspect’. This suggestion is dismissed as nonsensical because ‘it would be a question what guaranteed that one got hold of the right self.’ This is not very clear, but the point seems to be that a repeated use of a proper name involves a re-identification of the object it refers to, but this is not the case with ‘I’. I may make false statements about myself, but they are not made false by my using the first-person pronoun with the wrong reference.
But why should ‘I’ not function as a demonstrative like ‘this’? If I understand her rightly, Professor Anscombe’s answer is that the reference of ‘I’ is supposed to be guaranteed in a way that the reference of ‘this’ is not. The example she gives is that of someone’s presenting an urn and saying ‘these ashes’ when in fact the urn is empty. But she herself admits that ‘I’ lends itself to a similar pretence as when someone is believed to be speaking through a medium. Another example would be its being supposed to issue from the mouth of an oracle.
A more telling point is that if ‘I’ is a name qua demonstrative, one may have trouble in saying what it names. Surely not a Cartesian ego, nor yet a body since one could still use it in the same way if one were suffering from total sensory deprivation. The best answer seems to be ‘this person’, but that is not illuminating. It should, however, be noted that this is rather because of the difficulty of analysing the concept of a person than because of an inability to explain the use of ‘I’.
Professor Anscombe cuts the knot by concluding that ‘ “I” is neither a name nor another kind of expression whose logical role is to make a reference, at all.’ According to her, ‘ “I am this thing here” is, then, a real proposition, but not a proposition of identity. It means: this thing here is the thing, the person ... of whose action this idea of action is an idea, of whose movements these ideas of movement are ideas,’ and so forth.
I do not find this conclusion shocking. After all, Russell maintained in his Inquiry into Meaning and Truth that no other demonstrative than ‘this’ was needed, and I argued in my book The Problem of Knowledge that ‘I’ functioned not as a name but as a signal. I have some qualms about the contention that ‘I am Elizabeth Anscombe’ is not an identity statement, since in the right context the statement would be a revelation of her identity. However, the disclaimer could perhaps be justified as a precaution against those misguided philosophers who think that statements of identity are necessarily true, if they are true at all.
Considerable attention is paid throughout these volumes to the work of Hume, for whose insights Professor Anscombe has a pronounced respect, though she disagrees with him more often than not. For instance, she convicts him of carelessness in arguing that it is the impossibility of carrying on our inferences ad infinitum that accounts for our allowing the series which begins with immediate perceptions, say of printed letters, to terminate with a belief in some historical event, such as Julius Caesar’s murder. It is the present impression that stops the regress: the historical event is the first link of the chain. The argument should be that eye-witnesses of Caesar’s murder passed the information on to others, who accepted it on the strength of their visual or auditory impressions, and, remembering these impressions, passed the information on in their turn until finally it came to anchor in our own sense-experiences. But Professor Anscombe argues that ‘it is not like that. If the written records that we now see are grounds of our belief, they are first and foremost grounds for belief in Caesar’s killing, belief that the assassination is a solid bit of history.’ Such a belief may be ‘a belief that there has been a chain of tradition and reports going back to contemporary knowledge; it is not a belief in the historical facts by an inference that passes through the links of such a chain. At most, that can very seldom be the case.’
This might seem a rather trivial point, but a deeper issue lies behind it. Professor Anscombe wants to suggest that pieces of solid history, like the death of Caesar, are no longer open to question, short of our engaging in total Cartesian doubt. Wittgenstein did indeed allow in his Philosophical Remarks that one might find written evidence that no such man as Caesar ever lived, but Professor Anscombe brushes this aside as ‘one of the rare pieces of stupidity’ in his writings and anyhow at variance with his later work On Certainty. If we came upon such apparent evidence we should explain it away. I think we probably should, though I do not rule out the possibility of our coming across a mass of evidence that would at least make us hesitate. I am, however, inclined to concede Professor Anscombe’s point that many of our historical beliefs are not open to serious doubt.
Professor Anscombe tangles with Hume also on the question of causality. In her inaugural lecture, delivered at Cambridge in 1971 and entitled ‘Causality and Determination’, she gives Hume due credit for seeing that the connection between cause and effect is not a logical connection, but reproaches him for lending support to the prevalent mistake of equating causality with necessitation. I think that she may be doing him a slight injustice here. Where Hume ‘got into a mess’, as F.P. Ramsey said, ‘was by taking an “idea” of necessity and looking for an “impression” ’. But I agree with Ramsey that Hume understood very well, even when he spoke of feelings of determination, that there was nothing more to it than regularity and that ‘he gave his readers credit for more intelligence than they display in their literal interpretations.’
Nevertheless Hume did represent our attributions of causality as founded on our experience of ‘constant conjunction’, and here, too, Professor Anscombe finds him at fault. She denies that particular causal statements are covertly universal. I think that she is right to the extent that such statements, especially when human behaviour is in question, do not claim the backing of universal laws. As I put it in The Central Questions of Philosophy,
for the most part, when we speak of the causes of human behaviour ... we are claiming that the behaviour in question would not have been forthcoming if such-and-such an event had not occurred. The idea behind this is that there is a finite number of ways in which such behaviour comes about. That is to say, the behaviour is linked with different events by different generalisations of tendency. If one of these generalisations is exemplified on a particular occasion, and the others not, we say that the event which enters into the generalisation is the cause.
I am not suggesting that this relaxation of the requirements of causality goes far enough to satisfy Professor Anscombe. If I understand her rightly, she maintains that a causal concept is contained in many observational terms like ‘push’, ‘carry’, ‘eat’, ‘burn’, ‘cut’, ‘hurt’, without bringing with it any whiff of generality. I think that the Humean answer to this should be that a word like ‘cutting’ or ‘burning’ stands for a process which can be broken down into logically independent stages. The important point for Hume is that the fact that the process has taken a particular course on one occasion is no guarantee that it will do so on another. If we assume it will, we are generalising from experience.
Professor Anscombe has a nice example, suggested by the physicist Professor Feynman, of what she calls a non-necessitating cause: ‘a bomb is connected with a Geiger counter, so that it will go off if the Geiger counter registers a certain reading, whether it will or not is not determined, for it is so placed near some radio-active material that it may or may not register that reading.’ Since Professor Anscombe seems willing to follow C.S. Peirce in assigning a role to chance in nature, I find it odd that she disputes Hume’s contention that it is possible for an object to come into existence without a cause. Her tracing of a number of objects to their origins is hardly a sufficient rebuttal.
In her well-known article ‘Modern Moral Philosophy’, first published in 1958, she argues that the concepts of moral obligation and duty, right and wrong, make no sense when severed from ‘a law conception of ethics’. I should go further and maintain that the concept of a lawgiver is out of place in moral philosophy. I share her unease, however, over the standing of any moral theory from which it can be deduced that ‘in such-and-such circumstances one ought to procure the judicial condemnation of the innocent.’
Professor Anscombe is philosophically robust. She is not a slave to fashion and however intricate her play of the cards may be, at least she lays them all upon the table.