A guide to reading “The Reality of the Past.”

Michael Dummett’s 1969 article on the reality of the past has helped to shape on-going philosophical debates about realism and anti-realism. The aim of this guide is to help students who are just beginning their studies of metaphysics to understand and appreciate this article (specifically, it is intended for students taking my class in Philosophy In Literature). All page references refer to the article as published in Dummett’s 1978 collection, Truth and Other Enigmas. I do not hope that this will make understanding the article easy, only to make it possible.

I divide the article into different sections. I suggest that you read each section of the article, then read my notes for that section, and only then proceed to read the next section of the article.

Section 1: Anti-realism is not the same as reductionism. (Top of p.358 to end of first paragraph of p. 362 (“…independently of the use of statements of the disputed class).

A reductionist is someone who claims that when we seem to be talking about one thing, we are really talking about something else. For example, if I claim that when we talk about thoughts we are really talking about physical events that take place in the brain, that would be a form of reductionism. I could express this by saying ‘thoughts can be reduced to events in the brain and nothing more.’
At one time, many metaphysicians hoped to produce a definitive list of the things that really exist, and explain how any talk that succeeds in describing the way things are is really a way of talking about these ultimate objects of existence. For example, a materialist might say that everything that is real is made up of sub-atomic particles: ultimately any talk about anything real could be reduced to talk about subatomic particles. Or an idealist might say that what is ultimately real is the immediate objects of my senses – the feeling of the keyboard under my fingers, and the coolness of the draft of air from my fan on my skin (that is, says the idealist, the feeling and the coolness are real, not necessarily the keyboard, the fan or the skin). For the idealist, if we talk about anything real we are really talking about sensations such as these. Notice that the idealist says that all talk about material objects is really talk about sensations, and the materialist says that all talk about sensations is really talk about physical objects. Each is a reductionist about the things that the other takes to be real.
By the time Dummett wrote this article, philosophers were abandoning this kind of quest for the ultimate constituents of reality: why can’t reality include different kinds of objects – numbers, sensations, physical objects, aesthetic qualities such as beauty: couldn’t they all be real? Why should we think that one, and only one type of language is specially fitted to describe reality?
Even when we move away from this quest for the ultimate constituents of reality, questions about what is real still remain however: are numbers real? Is the past real (as Winston Smith claims in Nineteen Eighty Four?) Dummett is explaining, in this section, that a discussion about the reality of numbers of the reality of the past need not be a discussion about whether we need a reductionist account of numbers or of the past.

First Dummett shows that a reductionist does not have to be an anti-realist. If I say that thoughts are really brain events, I am a reductionist about thoughts. However, if I am a realist about brain events, then I am still a realist about thoughts. I am saying that thoughts are not what we thought they were – they are physical events – but I am not denying that they exist.

Then Dummett shows that in some cases, a reductionist is an anti-realist. Suppose, for example, a behaviourist argues that statements about people’s character are entirely reducible to statements about their behaviour. If the behaviourist says that you are brave, she means you react bravely when in danger, and if she says that you are not brave, she means you react like a coward when in danger. Suppose you die without ever being put in a situation of danger. In that case, according to the behaviourist, it was not true that you were brave nor was it false.

Of course, many of us would say that brave behaviour is the result of being a brave person, and so is the best evidence that you are brave. If you are never put into a situation of danger, we would not find out whether you are brave or not, but still, there must be some fact of the matter: a person is either brave or not brave, even if we cannot tell which. This would be a realist view about bravery.

So, let’s look at the difference between the realist and the anti-realist about bravery. The realist says that the proposition “You are brave is” is, and must be, determinately true or false. That is to say the realist applies the principle of bivalence to the statement “You are brave”. Since the realist applies the principle of bivalence, it follows that the realist is bound to admit that “either you are brave or you or not brave” – an instance of the law of excluded middle. (It is possible to separate bivalence and excluded middle, but, as far as we are concerned at the moment, if we establish bivalence then we can take it we have excluded middle as a consequence.)

The anti-realist does not think that bivalence and excluded middle can be guaranteed to hold in every case. If your bravery is never put to the test, it is not true that you are determinately brave or not brave. Of course, if you are put to the test, then we will find out whether you are brave or not – while denying as a general principle that “You are brave” must be determinately true or false, the anti-realist can say, of a specific individual, that it is determinately true that he was brave, or of some other individual that it is determinately true that he was not brave.

So, when faced with someone who performs brave actions, both the realist and the anti-realist can agree in saying “She was brave.” Where they disagree is the case where we have no evidence to tell us whether she was brave or not. The realist believes that even in this case, there is a truth about her character that we do not know, a truth that transcends the available evidence. Of course, the realist is not saying that the truth about someone’s bravery must transcend the evidence, that it cannot possibly be known: the realist is just saying that lack of evidence does not imply lack of truth.

This means then that the realist and the anti-realist each follows a different set of logical laws. An argument that the realist accepts as valid might be rejected as invalid by the anti-realist, if, for example, the argument uses the law of excluded middle. Dummett assumes that the realist will follow classical logic, in which every proposition is either true, or false, never both true and false, never neither true nor false. (This is a simple form of logic, and usually students studying logic for the first time are taught it). The anti-realist will have to adopt some non-classical logic.

The idea that anti-realists should abandon bivalence and excluded middle was first advanced by a Dutch mathematician called Brouwer, who developed an alternative logical system called intuitionism. Dummett’s work on anti-realism was inspired by this approach, and he uses intuitionism as an example of a form of anti-realism that is not reductionist. In intuitionist mathematics, a mathematical statement is true if there is a mathematical proof, false if there is a mathematical disproof. There is no guarantee that we will have either a proof or a disproof for every proposition of mathematics, and so we cannot guarantee bivalence and excluded middle. However, mathematical statements are not said by the intuitionist to be really a way of talking about something non-mathematical: mathematical talk is about mathematics, it is just that mathematical objects have no existence that transcends their provability.

Although Dummett agrees with the intuitionists that a debate about realism and anti-realism is fundamentally a debate about logic, he does not in fact like Brouwer’s arguments in favour of intuitionism (so I won’t mention those arguments here). Instead, he has developed another argument for anti-realism, an argument that is presented in the next section.

Section 2: The Manifestation Argument (p. 362, second paragraph)

Dummett states briefly an argument that favours the anti-realist. To look at what we mean when we make a statement, we must look at what it is to understand it, and to understand a statement is always to manifest a practical skill. If I see a bird and say “Look, a bird!” that manifests my understanding of “Look a bird.” Or, if you point to a dog and say “Look, a bird!” and I say, “Don’t be silly,” that also manifests my understanding of “Look, a bird!” I manifest my understanding by responding to evidence that confirms or disconfirms the statement. But there does not seem to be any role in the account of how I manifest my understanding for me to recognize that a statement is true in a way that transcends evidence. I can respond to evidence for a statement by agreeing with the statement, I can respond to evidence against a statement by disagreeing with the statement. I can even respond to a complete lack of evidence by expressing agnosticism. But where will evidence-transcendent truth figure in my manifestation of understanding? Dummett argues that it won’t, and so the concept of evidence-transcendent truth, and thus of realism, is rendered suspect.

Dummett has presented more elaborate versions of this argument against realism, which is known as the manifestation argument. Notice that this could be applied to any debate about realism: we could use it to argue one should be a realist about physical objects, and about sensations, and about numbers and so on. When philosophers were searching for the ultimate constituents of reality, it was commonly thought that one would end up with one set of items about which one was a realist, and be a reductionist about everything else. But if Dummett’s manifestation argument could be applied to everything then we are left, it seems, with the possibility that we could be anti-realists about everything. Would that mean that nothing is real? Would that be coherent? Dummett himself wondered about this.

In 1992, in his valedictory lecture (final official lecture as Wykeham Professor of Logic at Oxford) he looked back over his work on realism and anti-realism. Unfortunately, I went to the wrong lecture room by mistake, and heard a lecture on ecclesiastical history instead. Fortunately, his valedictory lecture was published in his 1993 collection, The Seas of Language (Oxford University Press, 1993). In it, he states,

     I saw the matter … as the posing of a question of how far, and in what contexts, a certain generic line of argument could be pushed, where the answers “No distance at all” and “In no contexts whatever” could not be credibly entertained, and the answers “To the bitter end” and “In all conceivable contexts” were almost as unlikely to be right. (p.464)

The person who pushes this argument to the bitter end in every context is called by Dummett the Global Anti-Realist. Dummett doubts this position is coherent: the Global Anti-Realist has pushed a good argument too far. But he wants to find out at what point the Global Anti-Realist pushes the argument beyond its limits. In this article, the Global Anti-Realist is called G. Contrasted with G is someone who does not push the argument quite so far. This person is an anti-realist about the past, but not, as far as we know, about anything else. Dummett refers to him as T. G and T do not entirely agree with each other, but both of them are united in opposing the person who is a realist about the past. I will refer to the realist as R.
(T, G and R are all referred to as male in the article, as was usual in 1969. I will retain that here, since my aim is to help you understand that article). T, G and R are all engaged in a debate about what Dummett calls truth-value links.

Section 3: What are truth-value links and why do they matter? (p. 362, “Statements about the past form a class the application to which…” to p. 365, 5th line from end of page “…(‘was going to…’, etc.) are used.”


The fact that one statement is true at one time implies that other sentences will be true at other times. This is what is meant by a truth value link. For example, Dummett is truly in his room in college at some time in 1969, it is true in 1970 that Dummett was in his room at some time during the previous year. A statement A, said at one time, implies the truth of a statement B, said at another time.

The realist needs to explain how we can grasp that it is possible for something to be true, even though there is no evidence. The realist explains that we can grasp that evidence can go missing. If in 1969, it is true that Dummett is in his college room, that implies that in 1970 it will be true that he was in his college room. It does not, however, imply that there will be evidence that he was in his college room (not in the strict logical sense). What if, in December 1969, a bomb is dropped on Oxford, and all evidence of Dummett’s movements is destroyed? Of course, if it is now true that Dummett is in his college room, it is likely that, in one year’s time, there will be evidence that he was in his room a year ago, but the link between “Dummett is in his room” in 1969 and “There is evidence that Dummett was in his room last year” in 1970 is not an example of a truth-value link: it is possible, however unlikely, that the first sentence could be true and the second could be false.


(Reading Nineteen Eighty Four will give you a good idea of how someone could deliberately go about altering evidence. Indeed, if I am in Winston Smith’s position, I can manifest my knowledge of a present tense truth by responding correctly to evidence that supports it now, and at the same time manifest my understanding of the truth that this evidence will soon be erased by responding correctly to the evidence that the Party is erasing the evidence. Smith realizes, for example, that there is evidence that Syme exists, and evidence that all evidence of Syme’s existence will be erased. Arguably, this kind of example raises even more problems for the anti-realist than the examples Dummett considers – it is not that the evidence might be lost, there is evidence that the evidence will be destroyed).


Section 4: The debate between G, T and R. (p.365, “At least two versions of anti-realism can be distinguished…” to p.374, end of the article).

First, Dummett introduces us to T, who is an anti-realist about the past but not about the present. When Dummett was writing this article, he was introducing new philosophical ideas. When introducing a new idea, it is useful to compare it with ideas that might be familiar to your audience. On p.366, Dummett begins to explain T’s position by comparing it to a position one might hold about set theory, because he imagines that many of his readers will be familiar with set theory. If you are familiar with set theory, this passage will be helpful, but if you have not studied set theory in some detail, there is no point first explaining set theory and then using that to explain T’s position: better to jump straight to an explanation of T’s position. So, if you don’t understand p.366, omit it and don’t worry about it. As explained on p.367, T’s theory is in fact easier for a beginner than set theory.

Here, I will explain T’s theory by means of an example. There are some things we know about the past, such as the fact that two princes died in the Tower of London. The two main suspects are Richard III and Henry, Earl of Richmond. There is no longer sufficient evidence to say which of these two is guilty. So, T thinks about this as follows: Possibly Richard killed the Princes. Possibly Henry killed the Princes. Certainly the Princes died. This may also be expressed by saying that there is a possible history in which it is true that Richard killed them, and a possible history in which it is true that Henry killed them. ‘Possibly’ here means ‘possible as far as we know’: what is possible is what is consistent with our current historical evidence. For the sake of simplicity, let us suppose that these are the only two possibilities. In all of the possible histories, the Princes are murdered – and we can express this by saying that certainly they were murdered, or that it is an absolute truth that they were murdered. So ‘possibly’ indicates what is true in at least one possible history, ‘certainly’ what is true in every possible history.

In that case, certainly Richard either murdered them or did not murder them. In every possible history, ‘Richard murdered the princes’ is either true or it is false. However, we cannot say that he certainly murdered them or certainly did not murder them.
What happens to bivalence for T? Remember that T distinguishes what is absolutely true and what is possibly true. Absolute truth does not follow a principle of bivalence – not every proposition is absolutely true or absolutely false. However, it is absolutely true that, within any possible history, every proposition is true in that history or false in that history. Bivalence holds within any possible history for T because for T, at any time, bivalence holds for present tense statements about that time – T, remember, is a realist about the present.

G on the other hand is not a realist about the present. G does not apply bivalence to statements about the present, and so certainly does not want to extend this principle to the past.

So, G and T both deal in a different way with the possibility, raised by R, that something did happen, but all the evidence has since been destroyed.

G argues (p. 371) that it simply does not make sense to suppose that something happened that left no evidence. G admits that evidence can be destroyed, however, if I say ‘Evidence has been destroyed’ correctly, it must be because I have evidence that evidence has been destroyed. As long as Winston has a memory of destroying evidence that the Party produced a fake confession, that memory itself constitutes a piece of evidence that the Party produced a fake confession. But, suppose that Winston says “Syme’s confession was a fake produced by the party, and there is not a single scrap of evidence to support that statement,” then, argues G, he is simply manifesting his confusion. Either he does not understand “Syme’s confession was a fake produced by the party” or, perhaps, he does not understand “…and there is not a single scrap of evidence to support that statement.”

T would deal with this example in a different way. As far as T is concerned, we can say that possibly an event happened that left no traces, but we can never say that certainly an event happened that left no traces. Anything that is certainly true for T must be something that we are compelled to say by our evidence, and we can never be compelled to say by the evidence that something happened for which there is now no evidence. For T, what is certainly true, that is absolutely true, is the only real truth: what is true in one possible history but not another is nothing more than a maybe. So for T, an event that has left no trace would be a maybe that can never be true. In the words of T.S. Eliot,
“What might have been is an abstraction,
Remaining a perpetual possibility
Only in a world of speculation.” (Four Quartets: 1 Burnt Norton)

What seems to be of more interest to Dummett however is not the contrast between T and G, but what separates both of them from R. At the top of p.370, Dummett argues that the anti-realist (I take this to include both G and T) takes the reality of time more seriously than the realist:
 
 It is the anti-realist who takes time seriously, who thinks in the way McTaggart described as believing in the reality of time; it is the realist who takes the view McTaggart was advancing when he proclaimed the unreality of time.

To say that the anti-realist about the past takes time more seriously than the realist about the past may seem at first to be a deliberate paradox, the kind of thing philosophers say to confuse their readers and sound more profound. Dummett is not, however, that kind of a philosopher and, when one understands what he means here by ‘the reality of time’, his statement is not so confusing. As he states on p.370, he takes the phrase from McTaggart. So what did McTaggart mean by it?

McTaggart pointed out that we have two ways of describing the way events are ordered in time. Description (a) invokes tenses: events are past, present and future. Description (b) sets one event before or after another, but without reference to tense. The (a) series and (b) series are not equivalent. Suppose that I wake up, suffering from amnesia. I wonder where I am. I have a map, and it tells me where one building is in relation to another. However, to use the map I must pinpoint my position – I need someone to say “You are here.” I now know where I am, but where should I be. Suppose I have a copy of my schedule with me. For every day of the year, I have written down what appointments I need to keep at what time. I know that going to the dentist precedes attending a wedding, and that only after the wedding should I visit my brother. However, just as I need to know where to use the map, to use the schedule I need to know what time it is now. The schedule gives me the (b) series, I need a clock to give me the (a) series.

The fact that I am in a place makes it different, to me, from other places. I refer to the place where I am by the word ‘here’, and you can use the same word to refer to the place where you are. Of course, there is not one ‘here’ for the whole universe. Where I am, my here, is a special place for me, but not for everybody. If I say ‘I am here in Panama’ and my mother says ‘I am here in Manchester’, we cannot say that somehow the place where I am is really here, whereas where she is is really there, even though she thinks it is here.

Is ‘now’ analogous, in this respect, to ‘here’? This is McTaggart’s question. For McTaggart, to answer ‘yes’ to this question is to say that time is unreal, and Dummett, as we have seen, is using the term ‘reality of time’ in McTaggart’s sense.

(I should note that McTaggart’s question cannot be answered by referring to different time zones. Suppose I am recovering my memory. Address book in hand, I telephone my mother. In her time zone, it is six o’clock, in my time-zone, it is one o’clock. Still, if she tells me of some event that it is in the future, then that holds for both of us. If she says “The wedding you have to be at will take place in five hours time”, then I know that it is not taking place now, but that in six hours it will have taken place, and that is true for her and for me. The fact that the Sun is in one position relative to her and another position relative to me (which is what talk of time-zones amounts to) is a fact about our different spatial arrangements, not a fact about time. Being in a different time-zone has no significance for McTaggart’s question. Entirely separate from the question of time-zones is Einstein’s theory of relativity. It would however be beyond the scope of this guide to discuss relativity further. I am only trying to explain what McTaggart’s question about the reality of time is, not to determine the correct answer.)

Let us now consider why it is not surprising that denying the reality of the past is a way of affirming the reality of time. If one affirms the reality of time, in McTaggart’s sense, ‘now’ marks the division between two different metaphysical entities, the past and the future, and so one must be able to point to some metaphysical difference between the past, the present and the future. To affirm the reality of time is to affirm the reality of change. (McTaggart, incidentally, denied the reality of time, and thus argued that all apparent change is an illusion).

The anti-realist, as we have seen, always explains the meaning of a tensed statement in terms of evidence that is available at the time of making it, or after it is made, that would enable us to recognise that it is correct or incorrect. The difference between the realist and anti-realist is difficult to explain, because they will often both say the same thing, but what they mean by it is different in each case, because they have a different way of explaining the meaning. Dummett tries several different ways of presenting this difference to us, I find what he says on p.373 particularly helpful. He points out that both the realist and the anti-realist will agree to the following principle, which I will call P:

P: If a statement is true, there is something in virtue of which it is true.

In explaining the different ways of interpreting P, Dummett talks about different “restrictions on the range of the quantifier.” This is a piece of technical terminology, familiar to Dummett, and, he expects, familiar to his audience, but probably not familiar to you. However, it is easily explained. The phrase ‘There is a …’ can be called a quantifier. When I use a phrase containing this quantifier, I must specify a range of objects over which it applies. Suppose for example I say ‘There is a student who understands Dummett’s article.’ To know whether this is true, you would have to consider whether I am talking about students in this class, students in this campus, students in this university, or any students anywhere on the planet. Each of these is a different range of quantification.

So, when we look at P, we see a quantifier ‘there is something in virtue of which it is true.’ For the realist about the past, the range of this quantifier is the whole of time. For the anti-realist (either T or G), the range of the quantifier is the evidence that is available at the time one is making the assertion, and evidence that may come to light in the future. So, consider the sentence ‘Richard III murdered the princes in the tower.’ Both the realist and the anti-realist agree that

If ‘Richard III murdered the princes in the tower’ is true, then there is something in virtue of which it is true.

 For the realist about the past, the range of the quantifier is the same whether this assertion is made in 1483, in 1583, in 1674 or in 2050. For the anti-realist, the range of quantification is different depending on the time when we make the statement. For the anti-realist, the range of quantification is different in each case. In 1483, we might suppose, much more evidence was available than in 1583. In 1674, new evidence came to light when the skeletons of two children were discovered in the tower. Perhaps by 2050, more evidence still will be available, more things that could count as being a ‘something in virtue of which it is true.’

On p.373, Dummett points out that if the anti-realist is correct that the range of quantification differs, that means that different things can be said at different times. Here is another example to make that clear.
Suppose I meet up with my friend from the USA, Dr. Patzia, also a philosopher. I say to him proudly, “One of my students understands Dummett’s article on ‘The Reality of the Past.’” If he says “One of your students understands Dummett’s article on ‘The Reality of the Past’” what he says does not mean exactly what I meant. I made the comment with pride in the abilities of my students, he said it in awe and respect of the abilities of the students of another professor. (“My car is fast” said by you and “Your car is fast” said by me do not mean the same thing – unless you are happy to let me assert rights of ownership over your car. When you say something about what you call “My car” you do not just say something about the car that other people could say, that it is fast, you say something about your relationship to it that other people cannot truly say: you assert your ownership).  

Of course, Mike Patzia might say “That’s great Ben – but you know I have some good students too. One of my students understands Dummett’s article on ‘The Reality of the Past.’” Mike and I both speak English, and if we were teaching someone English, we would teach them to translate the sentence said by me in just the same way that they would translate the sentence said by Mike, so, if we consider meaning to be a property of a sentence, where that sentence is just any utterance or inscription of a given set of words in a given order, then what I said and what Mike said means the same. However, we did not say the same thing, in the sense that what he said and what I said could have different truth values – different truth values because of different ranges of quantification.

According to the anti-realist, just as what I say by “One of my students understands Dummett” is different from what Mike Patzia says by asserting the same words, and he cannot say exactly what I say when I say those words, (nor can I say exactly what he says), what we say now when we say ‘If ‘Richard III murdered the princes in the tower’ there is something in virtue of which it is true that he murdered them’ is different from what someone would have said by saying those same words in 1583, and we cannot now say the thing that they said by them.

So, I am saying that for all of us, there are things that can be said by one person that cannot be said by another person. Dummett’s anti-realist about the past is saying that furthermore, things that are said at one time cannot be said at another time. Dummett’s realist about the past denies that there is this difference in expressive power between different times and it is in this sense that the realist about the past is denying the reality of time – there is a form of change, change in what can be said, that the anti-realist believes in and the realist rejects.

The conclusion is typical of Dummett’s writing: after all that hard work, we have a better understanding of the strengths and weaknesses of the different positions, and we see how one philosophical issue, the reality of the past, has implications for another, the reality of time. We do not discover what the right answer is: that will require further research.


Back to PHI 3882