Quick Thoughts on Some Remarks by Dummett

So I decided to read Michael Dummett's Truth and the Past to see if there was anything relevant for my dissertation. I suppose some of it may be, but I want to focus on some of the things he says in his penultimate chapter, "The Metaphysics of Time". Some of his arguments or considerations offered in this chapter are particularly weak or bordering on the question-begging (although, to be fair, these were just a series of lectures that have now been published in book form, so the brevity of some of the discussions is perhaps a little understandable). Of the four-dimensional model of the universe (by which he seems to mean a strange version of a tenseless, perdurantist view), for instance, Dummett says:

The four-dimensional model [...] deprives the world we observe of genuine change; there is only that of our awareness as we travel into the future. The model is grounded on the conception of our consciousness as moving through the static four-dimensional reality along the temporal dimension.

Now, first of all, no one that I know of holds a view like this. This reads like some unholy combination of a tenseless physical world coupled with a tensed mental world or else a timeless physical world coupled with a temporal mental world. Either way, no tenseless, perdurantist view is going to own up to anything like this. But Dummet continues:

A proponent of the four-dimensional model may deny this. We are, he says, irregular four-dimensional tubes (or hypertubes), with the peculiarity that consciousness attaches to our temporal cross-sections. Nothing changes: it is just that our different temporal cross-sections are aware of different things.

This is better, but it is still question-begging - the variation of an object along its temporal dimension, on a tenseless, perdurantist view just is the changing of the object. The perdurantist does not repudiate change, they give an account of it. Dummett misses the point here entirely. He continues:

This image is misconceived. Consider a description of other hypertubes, whose axes lie along a spatial dimension. To us these would appear long, very short-lived objects; if we learned that a different consciousness attached to each segment of one of the tubes, we should regard them as strings of distinct creatures. But if we were told that a different consciousness attached to each cross-section of such a tube at an angle orthogonal to its axis, and that the different consciousnesses varied continuously, we could make nothing of this at all.

It's not clear who the "we" is here (Van Inwagen and Dummet perhaps?), for quite a few people seem to be able to make sense of such things. Such a thing may be impossible, but it certainly seems intelligible or conceivable in a pretty strong sense. As we will see elsewhere, Dummett seems to like to think that if some view is contrary to a deeply entrenched belief of his, it must be unintelligible. Forget the fact that lots of other people seem to find it the opposite - or even to find that the denial of the perdurantist view is unintelligible itself!

Another less than stellar paragraph comes a few pages later, where Dummett writes:

Why should truth be explained in terms of knowledge? The question is whether it is possible to swallow the conception of a reality existing in utter independence of its being apprehended. [...] My question is whether it is intelligible to suppose that the universe might have been devoid of sentient creatures throughout its existence. What would be the difference between the existence of such a universe and there being no universe at all? To express the question theologically, could God have created a universe devoid of sentient creatures throughout its existence? What would be the difference between God's creating such a universe and his merely conceiving of such a universe without bringing it into existence? What difference would its existence make? It seems to me that the existence of a universe from which sentience was perpetually absent is an unintelligible fantasy. What exists is what can be known to exist. What is true is what can be known to be true. Reality is the totality of what can be experienced by sentient creatures and what can be known by intelligent ones.

It's not really clear here why we should take any of this seriously at all. This is all not so much argument as much as dogmatic assertion of Dummett's own crazy views. Of course, if one is already completely convinced of an antirealist view and think that such a view is necessarily true, one will likely find the scenario discussed here unintelligible. But that shows absolutely nothing. I could also hold crazy views about other things such that a very plausible view will then seem to me to be unintelligible. But that doesn't make the latter fact any evidence for my view - rather, it presupposes it. This is a particular example of how, unchecked, some badly formed intuitions and a lot of stubborness can snowball and lead one into incredibly implausible views. After all, many people will find the impossibility of Dummett's scenario unintelligible. After all, we normally do not think that the existence of stars or the wider universe is somehow dependent on us. That seems just as crazy (more, in fact) as the denial of the reality of the past that Dummett is so eager to escape from.

And notice his rhetorical questions! Here's a good example of the sort of thing I tell undergraduates not to do - introduce rhetorical questions in the place of actual argumentation, particularly when your opponents may very well have an answer for you. In this case, it seems perfectly clear what the difference would be between the universe existing or not, or being created by God or merely conceived. If the universe exists, all sorts of properties are instantiated, there are events occurring, etc. You may as well ask what the difference would be between me existing and me not - obviously, if I did not exist, certain properties would not be instantiated nor would certain events occur had I not been around. But the answer for the existence or non-existence of the universe is exactly parallel. To deny this as Dummett does would make the existence of the external world dependent on us so that it is literally metaphysically impossible that the universe could have been destroyed or ended up in some state such that sentient life never happened. This seems, to say the least, rather implausible.

As for being created versus merely conceived, that also seems too plain to even deserve mention - if God merely conceives of something, it does not exist, whereas if he creates it then it does. And so the differences will be just those between existing and not existing (as for his "What difference would it make?", if it is asking something beyond this, I have no idea what it is or why it would be relevant). He says, "It seems to me that the existence of a universe from which sentience was perpetually absent is an unintelligible fantasy," but, on the contrary, it is his view that seems the unintelligible fantasy. He says, "What exists is what can be known to exist. What is true is what can be known to be true. Reality is the totality of what can be experienced by sentient creatures and what can be known by intelligent ones," but all this is perfectly compatible with realism - it is only incompatible if we make these out to be actual analyses, where the epistemic claim in each statement is analyzing the metaphysical one. But even if we accepted these statements, there seems no good reason to read them this way. After all, the right hand side contains what already appear on the left, thus making such analyses circular and hence no good in developing any kind of theory. But there seems no other option for a view like Dummett's. Dummett simply seems to be confused, like most antirealists, and to have canonized that confusion as dogma.

UPDATE (7/27/08): I've just discovered that an Anthony Rudd in a 1997 Phil Studies article entitled "Realism and Time" makes an argument against the B-theory of time very similar to that of Dummett's against 4Dism, with all the same horribly mistaken assumptions. Rudd's arguments in this piece are, to put it politely, quite weak.

Metaphysical Thoughts I: Past Notes

I finally finished a second chapter of my dissertation (chapter 3), clocking in at 34 single-spaced pages! That was quite a marathon. Anyways...

This post is yet another entry in my Past Notes series. These are just random thoughts on issues in metaphysics, encompassing times, for instance, when I've flirted with Carnap/Putnam type views and endurantism. Part II will be posted later with post-2002 stuff.

********

10/9/01

I am me. Who else could I be? To ask about a counterfactual situation where I was someone else is like entertaining the idea tha the Sun might not have been the Sun. It is meaningless.

10/11/01

The opposite of self-identity? There is none. Try to contradict it and what you say will be meaningless. A=A cannot be thought otherwise.

10/15/01

We cannot escape from metaphysics - every claim we make is saturated with ontology. To say that metaphysics is meaningless is itself a claim of metaphysics. "There is a cat on the mat" is a metaphysical claim. Even if we try to say it all formally, we are still being metaphysical. How can we avoid metaphysics and yet say or think anything? "Metaphysics is meaningless" is self-refuting.

4/18/02

Let us say it is secured through concepts that material objects exist independently of us. That will hold only in case we are right about our conceptual argument. The fact that even we could not imagine it to be any other way than right does not make it so. We must assume certain things in making any argument and so will always rest on assumptions which might be false, though perhaps invulnerable to doubt.

7/9/02

I learn what a material object is through experience. I have a continuous experience of a certain sort. I develop a sense of object permanence. Soon I have a full conception of a material object as a distinct object of experience - it is the kind of thing I can interact with in such and such a way and interacts with others in such and such ways. A famework for thought and experience thus arises. Whether the rudiments or beginnings of such are already in me is another question. Of course, perhaps I experience things from the beginning as discrete objects. But this seems odd. In any case, idealism could not be correct - the mental is, at the very least, those things we know of which are not material objects. To say material objects are mental in nature is to change the meaning of words and disregard their common usage. There might be some properties, known or unknown, in common between material and immaterial things, yet the distinction still remains. If idealism was true, I could not think it. I could only whether idealist "material objects" were idealist "mental" in nature. This is similar to Putnam's brain-in-a-vat. Realism is almost by definition true - it is a commitment of our thought and action, our language and concepts.

7/15/02

I cannot consistently deny realism. It is implied in all our assertions. Realism is not a theory - it is the way in which we must think. It cannot be unmasked.

10/23/02

There are 3-dimensional objects. These are not mere time-slices of the "real" objects, which are the spatiotemporal series. Real change requires this - the same object to have one property at one time and a different one at the next (replacing the other). If the real objects are space-time worms, there can be no change. Each worm has each of its properties in every time. Consider a coffee-worm. At 1:00 it has the properties of being hot-at-1:00 and cold-at-2:00. At 2:00 it has the exact same properties. Objection: The worm does not at 1:00 have the property of being hot, rather it has the property of being-hot-at-1:00. The same with 2:00. And it always has exactly the same properties. Consider a poker, where one end is hot and the other cold. Point to one end and say that at that point the poker is hot and point to the other say that at that point the poker is cold. But the poker cannot be both hot and cold. What is really true to say is that the poker has the property of being hot-at-poker-end and the property of being cold-at-handle-end. Objection: But it is not the same object that changed. The properties changed, but so did the object. One temporal slice is not another. So this is not sufficient for real change. Space-time worms, then, cannot be the "real" objects, but rather the three-dimensional objects we know and love.

Moral Indexicals, Or Why Judgment Internalism Is Not Evidence Against Objectivist Moral Realism

So one argument I've come across more than once is that Judgment Internalism - on the version I am here interested in, the view that, insofar as one is rational, moral judgments will be intrinsically motivating - provides evidence against moral realism or at least against objectivist versions thereof (could a subjectivist view count as moral realist? I'd have to think about that, but it's late so I won't). The reasoning here is that objective facts are not intrinsically motivating, so when one makes a moral judgment one can't be judging that some objective fact is the case. I have two main responses to this argument, either one of which would effectively defang it:

First, I would contend that in fact not all moral judgments do motivate on their own. Consider this one: 'Ian Spencer ought to A'. That's not going to motivate me to do anything unless I know that I am Ian Spencer. Andy Egan thinks that since self-locating beliefs such as 'I' beliefs are motivating and hence that moral judgments must be self-locating beliefs ascribing to oneself the property of being such that one's ideal rational self would prescribe or proscribe such and such. But notice that, as we just saw, the only moral judgments that are in fact motivating are the ones that contain an explicit first-person reference. This has nothing to do with the fact that it is a moral judgment - it only has to do with the fact that it contains a first-person indexical! So Egan is right to find the motivating factor in a motivating moral judgment to come from self-location but he is wrong to think that this has anything to do with the relativity of morality. After all, an moral realist objectivist could perfectly well agree that self-location is doing the work here but disagree with Egan's relativism - the non-motivating third-person judgment and the motivating first-person one express the same facts and these can perfectly well be objective, morally realist facts. Similarly, 'Ian Spencer is being chased by a bear' and 'I am being chased by a bear' express the same objective, realist fact even though the latter will motivate me all on its own whereas the former will not (that requires me to know that I am Ian Spencer). Note that this also shows that there may also be non-moral judgments that are also intrinsically motivating insofar as I'm rational!

Second, suppose I am wrong about the above. Notice that Judgment Internalism says that it is only if one is rational (or insofar as one is so) that one is motivated by moral judgments. But if we view morality as in the business of dealing with reasons for action, we can view moral judgments as embodying or expressing reasons for or against different actions. Now, insofar as one is rational, one will be motivated by one's reasons. So judgment internalism follows nearly-trivially from just these two conceptual points about morality and its connection with rationality. No need for relativism or emotivism or what-have-you. The nature of rationality and morality jointly do all that work for us. So whichever of these two arguments you choose to employ, it looks like the move from Judgment Internalism to relativism or anti-realism done for.

Cockburn and Dummett on Understanding Statements About the Past

So I've been reading David Cockburn's (boy, I bet with a name like that he got made fun of as a kid) book Other Times. It's very different from anything else I've read on time since it takes a weird Wittgensteinian behavioristically anti-realist "No, I'm not an anti-realist" stand that threatens to collapse into either a strong anti-realism or a confused realism. A few comments on some stuff I've read:

Cockburn's take on Dummett is that Dummett thinks that it is wrong to think that there is a common core, "A is F", which, when one understands that and also has a general understanding of the past and future tenses, one can then understand "A was F" or "A will be F". Cockburn's Dummet (CD hereafter) thinks that, instead, to understand "A was F" is to know, for instance, what counts as present evidence for that - which will depend on the kind of event in question. On CD's view, it's not enough to have the general understanding of the tenses plus an understanding of the present-tensed version of the sentence. But isn't it? If one knows what being A and being F are, one knows the kinds of causes and effects associated with them to some degree. If one (perhaps expertly) deeply understands "A is F", one needs to understand what being A and being F are. But that, combined with a general understanding of the past tense, will also yield knowledge of what counts as present evidence for the past tense claim that A was F.

On pg. 61, Cockburn says that a 'fundamental aspect of our use of a sentence' is 'the ways in which it may feature in the justification of actions and emotions. This feature...is not one we should expect to be able to derive from other feature of its use... We cannot even characterise those supposed 'other' features of the use of a sentence independently of the actions and emotions with which it is characteristically linked.' I'm not sure there's sufficient evidence for this sort of claim. It's not clear how justification can be fundamental unless we become some sort of behaviorist or something close to it. Even then, I'm not sure what justification would even mean. If the entire meaning of a sentence is captured in its inferential role and we leave no room for reference or reality or correspondence to external facts, this is an extreme anti-realism. Otherwise, either the facts expressed will themselves determine the sentences' inferential role or, more in line with Cockburn's ideas, the inferential role of a sentence will determine the facts. That is, if p justifies q then, given that facts are "chosen" from the world that match the given inferential role, the fact expressed by p will be the sort to ground such a justification or be a reason - not just any old facts, but the ones that actually fit the role. Either way, we still can ask about what in reality is playing that role. Cockburn doesn't seem to allow this - his whole focus is on sentences and our behavior but he fails to deal with what those sentences correspond to.

Notes on Ludlow: Ch. 10

**WARNING: Technical Post**

In chapter 10, the antirealism or verficationism shines through strongly. Whereas previously Ludlow noted that we tend to evaluate claims about the past or future based on current evidence, he now seems to say that claims about the past or future are really claims about this evidence (in which case, it's not really evidence except in the trivial sense where we count something as evidence for itself). This, of course, is an unargued leap that he seems to take without being aware of it - if one is to be a presentist, there are a number of ways of grounding past or future truths other than simply in the evidence for such truths. Perhaps Ludlow's linguistic discussion is supposed to provide some such evidence, but it seems to me inconclusive at best, a confusion of assertibility conditions and truth conditions that one would only expect if one were already persuaded of some kind of antirealism.

Indeed, the move seems to be completely unwarranted unless one has already ruled out the alternatives or are assuming some kind of verificationism about meaning. But then it is hard to see how we can rule out many similar, obviously bad moves (notice a pattern developing in this book?). We might, for instance, note that we evaluate claims about other people or places based on personal, spatially proximate evidence. If we follow Ludlow, we would have to conclude that claims about other people or places are just claims about evidence internal to me or where I am at. But this seems to me to be clearly false - solipsism just isn't a viable option. Perhaps someone may object that the evidence we consider can be located in other persons than me or places than the one I am at. It's just mediated by more proximate events, objects or processes. But then we could give exactly the same answer for time - we can be, for instance, in possession of temporally remote evidence about stars via current light processes now reaching the earth from these stars. The same thing goes for fossils, which interestingly, Ludlow thinks are really what sentences purportedly about dinosaurs are really about. But that's just crazy (darn it!).

The main problem Ludlow deals with in this chapter has to do with inferences like the following:

(2)

I am hungry.

--------------------------------------------------

Next Tuesday it will be true that I was hungry.

Now, it might be the case that all evidence for today's hunger disappears by next Tuesday. But then, in that case, it will not be true then that I was hungry today. Ludlow discusses two ways out of this. The first is to say, for instance, that a future tense version of a sentence is true iff the sentence in the present tense is true. So even though there will be no evidence for my hungriness Tuesday, the prediction of future truth is still true. Even though at that later time it ends up not true that I was hungry. This is a pretty weird way out and not very plausible.

The second alternative is to say that the content of my sentences changes over time. So the above inference is going to work even though the evidence for the premise will be long gone because the words purportedly ascribing hungriness to myself change in meaning. But we have no idea what meaning they will take on later, nor for that matter what meaning they had previously. We may, however, not be able to get out of the first strategy here. After all, the proposition that next Tuesday it will be the case that I was hungry seems to follow from the proposition that I am hungry. Since it doesn't make any sense to speak of propositions changing their contents (since, presumably, propositions just are certain sorts of contents), explaining the inference in cases of evidence loss cannot rely on this second strategy. So it looks like Ludlow is going to have to be faced with taking the first alternative after all, which doesn't seem like a good idea.

The main issue behind all of this is whether we can "lose" facts about the past - whether past facts about people or places or events or whatever can simply disappear from reality. And that just seems implausible. The past, whatever else we may say about it, seems firm and fixed and not subject to erasure. Intuitively, this kind of change doesn't seem possible. Despite claims for presentist views that they are common sense, the issues brought up in this chapter I think show that it is indeed quite far from it (or at least Ludlow's version is), Ludlow's protests to the contrary.

Notes on Ludlow: Chs. 7-9

**WARNING: Technical Post**

In Chapter 7, Ludlow begins to construct an alternative, tensed semantics. Parts of this have a strongly antirealist, idealist, or verificationist sort of feel about them. In fact, Ludlow seems to agree with uberantirealist Michael Dummet, stating, 'As Dummet (1969) has argued,a semantic theory that accounts for an agent's semantic knowledge must show how portions of the language are learned from the evidence available to the language learner' (p.99). I'm not sure that this is really correct as it pushes us towards an untenable kind of antirealism about practically everything. And I'm not sure whether Ludlow wants to be committed to such a view. Then again, he seems to flirt with idealism throughout the book, so maybe I shouldn't be so surprised. To continue the quote,

But now consider how we learn to use past-tense expressions such as (4).
(4)
Dinosaurs roamed the Earth.
We do not evaluate this sentence by imagining some time earlier than now and determining whether at that time (4) is true. Rather, we evaluate (4) by right now conducting the sort of investigation that is appropriate for past-tense statements like (4). (For example, we might study fossil records.) Likewise for any past-tense statement. We have certain procedures for determining whether a past-tense proposition is true, and these procedures do not involve the evaluation of a proposition at some time past; rather, we simply evaluate the proposition in a particular way - a way which is independent of how we evaluate present-tense and future-tense propositions.
Consider the future-tense proposition (5).
(5)
The economy will recover in the third quarter.
Clearly we do not evaluate such a proposition by picking some time in the third quarter and determining whether it is true at that time that the economy is recovering. Rather, we evaluate it by studying the currently available economic data. Crucially, our evaluation of (5) can proceed without our ever attending to a corresponding present-tense proposition at some future time index.

It's not quite clear to me how any of this is relevant to the debate over tense. After all, the tenseless theorist can simply grant that Ludlow is correct that we look at present evidence to determine the truth of past or future tense statements. The mistake is to infer from this that when we determine the truth or falsity of past or future statements we are not thereby determining whether a certain tenseless fact holds. Using present evidence and not attending de dicto to any present-tense proposition at some future time index is perfectly compatible with this. The only way one could think otherwise would be to assume that tensed r-mirroring truth-conditions must also be m-mirroring truth-conditions. And that, as I've been arguing is clearly a mistake. After all, the same sorts of things Ludlow says about tensed statements could be said about first-person or 'here' sentences as well.

But he continues,

If this picture of the underlying robust theory is correct, then it immediately leads to a second advantage for the A-theory [tensed] proposal under discussion - in fact, a striking epistemological advantage. The B-theorist is in the untenable position of asserting that there is actually reference to past and future times and/or events. However, this flies in the face of everything we know about reference. We are in neither a perceptual relation nor a causal relation with future events, and our causal connection with most past events is tenuous at best. In regard to times, the idea that there could be reference to such abstract objects surely requires major adjustments to current epistemological thinking.

This argument or set of arguments here seems to be a non-starter. I'm not sure how anything Ludlow says makes reference to or quantification over future or past events or times at all problematic. That our causal relation to past things is tenuous seems irrelevant since all that is needed for causal theories of reference is causation - not "super duper not-so-tenuous causation". And if we have a causal theory of reference, then it is reference to present things that is problematic since causation is a cross-temporal relation. That we do not have any causal relation with future things is, I think, not as clear as Ludlow seems to think, but let's give him that for the moment (I tend to think it's false, actually). But quantification or reference do not necessarily require causal relations - one can fix the reference of a name, for instance, by introducing it via an identifying description without having any clear causal contact whatsoever with the object satisfying the description. And quantification over certain entities does not seem to require being causally related to all of them and there's no clear reason why we would need to be. In addition, on most theories of time, times are not abstract but rather concrete objects. In any case, they are treated the same sort of way as places or parts of space. We seem to be able to refer to or quantify over space or regions thereof, so why not times? There seems to be no difference here. All of Ludlow's criticisms here could just as well be thrown against the view that other persons or object outside of myself exist and that we quantify or refer to them. If Ludlow were correct, his views would be pushing us towards a dangerous ontological solipsism where only I exist or an epistemic or semantic solipsism where only I can be referred to or quantified over by myself.

Ludlow ultimately comes to think that his semantics leaves presentism as one of the only plausible, consistent accounts of time. But if we accept presentism for time based on the problems outlined in the book, it seems that similar problems for first-person sentences or 'here' sentences are going to force us into the ontological solipsism mentioned above. After all, if presentism is a main way to get out of McTaggart's Paradox for time, solipsism will be an analogous way to get out similar paradoxes for persons.

Indeed, Ludlow's tensed semantics could be transformed into an analogous first person or 'here' semantics. Ludlow claims in Chapter 8, for instance, that apparent reference to times like 'June 24, 1972' can be paraphrased away as 'when standard calendars read "June 24, 1972"' and that normal tensed sentences will actually be decomposed as complex sentences composed of two tensed sentences joined by 'when', 'after' or 'before'. But we can do the same sorts of things with apparent reference to places and decompose 'here' sentences as complex sentences composed of two 'here' sentences joined by 'where', etc. So 'Paris' becomes something like 'where standard tracking systems read "Paris"'. If we do want reference to times, we can build times up as collections of when-clauses, according to Ludlow. But then if we want reference to places, we can build them up as collections of where-clauses. Perhaps we can do this sort of thing with persons as well - only I exist, but I can refer to other persons as collections of who-clauses (?).

At the end of Chapter 8, Ludlow shows that his theory can apparently get him out of one formulation of McTaggart's Paradox. But it's far from clear that it can escape a reformulation to match Ludlow's theory. Heather Dyke's formulation, suitably adjusted to face Ludlow, seems, for instance, like it would cause Ludlow particular trouble.

Chapter 9 consists in listing some psychological considerations that may or may not help the tensed theorist. I think they do not - the tenseless theorist should be at ease with all the data discussed. In fact, that's just the sort of data one would expect if the New Tenseless Theory were true - people think tensedly. In fact, some have argued that the data actually favors the tenseless theory. In addition, not all of the discussion is clear or very clearly well-motivated. Some of the discussion of and quotes from Merleau-Ponty, for instance, is metaphorical and opaque at best and of unclear relevance to the topic or the use Ludlow seems to want to put it to. So I think chapter 9 is inconclusive at best.

The last-ish notes are soon to com.