Wednesday, May 14, 2008

Knowledge of the future

It is sometimes claimed that our knowledge of the future is foggy and lacking in detail, that we know very little about the future, except in a general way. I will argue that if we discount eschatological considerations, we know a lot about the future, and with a lot of specificity. So, for now, let us discount eschatology. Then, I know that American history textbooks in the year 2040 covering the first decade of the 21st century will contain a statement entailing that GWB was elected in 2004.

I know that a record of this post will be available encoded in an operating or recoverable electronic form in 2015. Anything on this blog is currently found on at least dozens of hard drives, including yours, dear reader, if you're reading on a hard-drive based computer, and including hard drives in blogspot's servers, on my home server (both the main and the backup hard drive), etc. Granted, it may be that all this data will be overwritten, but it will still be recoverable if enough effort is put into recovery (see this story on recovery of data on a hard drive in the Columbia disaster). I also know this for a lot of other information on the Internet. I also know that at least one copy of the complete works of Shakespeare will survive to 2040.

In fact, if eschatology can be discounted, I submit that I know just about as many pieces of information about how things will be in the future as I do about how things were in the past. Of course there are many things I don't know about the future, just as there are many things I don't know about the past.

So why is it that we feel that the future is so much less known than the past? I want to offer two hypotheses. The first is that we have an awareness of eschatological possibilities, of the fact that we cannot really estimate the probability that, say, in one year God will step in and really change things heavily. I am not sure this hypothesis accounts for the phenomena, though. Non-religious people are also prone to feel that the future is less well known than the past. And I can probably modify at least some of my predictions so they take the eschaton into account. Instead of talking about history books, let's talk about knowledge: I know that some people will know in 2020 that GWB was elected in 2004 (these people might be in heaven, but that doesn't affect the claim). In fact, I know quite a lot about what people will know in 2020—just about every major widely known fact about the world as it is now will be known by somebody.[note 1]

My second hypothesis as to why we feel there is an asymmetry is that this is simply because so many of the things about the future which we want to know are unknown, while comparatively fewer things about the past which we want to know are unknown. The asymmetry, then, perhaps isn't in what we know about the future or past, but in what we want to know. There is a lot more unsatisfied desire in us to know the future than there is unsatisfied desire to know the past. Or at least so it is in our culture.


Anonymous said...

A likely prediction of the future is different from knowledge of the past. Knowing that it rained yesterday is different from predicting that it will rain tomorrow. You can't know about something that hasn't happened yet, you can only make a prediction. These are different things.

To take your example, you claim that you know that an archive will exist to record that post. It is a reasonable prediction, but it is different from knowing that you made the post. You know that you made the post because you remember doing it. You could even ask people who read your post to verify that it was posted, and they can make a judgment that you really did it. They could claim that as knowledge.

However, your prediction doesn't even approach knowledge. Even if it is correct, that doesn't make it knowledge. It just makes it a correct prediction.

Note that your prediction could easily be mistaken. You could realize tomorrow how silly your post was and delete it before it was archived. Unless you think you know what you will do tomorrow, but seriously, you can only predict it.

Tanasije Gjorgoski said...

We certainly use words like 'know' ('I know she will do it'), 'see' ('I see where this is going') and other words to speak about the future.

I' inclined even to think that we literally see what will happen in the future in some cases. For example when somebody throws a ball to us, and we see where the ball is going.

I guess some unease with talking about knowledge and seeing of what will happen, might come from taking some kind of causal account of knowledge of contingent things. That is, if we assume that knowledge requires that there is causal relation between the fact and our mental state where we say that we know that fact, it won't work for those situations.

The other thing that comes into mind, is the issue of determinism. For, if the state of the universe tomorrow isn't fully determined by the state of the universe today, and if today we say 'it will rain tomorrow', the sentence will lack (definite) truth value, and hence can't be neither true nor false. So in view of nondeterminism + JTB, I think one won't agree that knowledge of future is possible.

Alexander R Pruss said...


Are you assuming that the probability of a future-directed beliefs being right is always lower than the probability of a past-directed piece of knowledge? That seems implausible. Take my knowledge of the disjunction that either in the next ten years the eschaton will occur or a copy of the prologue of the Gospel of John will survive. The probability there seems at least as high as the probability that Napoleon was vanquished at Waterloo (it is always possible there was a cover-up).


A causal requirement for knowledge makes mathematical and moral knowledge problematic. (Of course maybe the grounds of mathematical and moral truths are causally effective. But that's very controversial.)

Moreover, one might relax the causal requirement for knowledge as follows. Instead of requiring that the belief be caused by its object, one requires that the belief is causally connected to its object. Events A and B are causally connected provided that there is a chain of basic causal connections between them. A basic causal connection is the relation of causing or being-caused-by. Thus, events with a common cause, or with a common effect, count as causally connected.

Indeterminism does not of itself imply that there is no truth-value to future-tensed claims.

Anonymous said...

"Are you assuming that the probability of a future-directed beliefs being right is always lower than the probability of a past-directed piece of knowledge? "

No, not at all. But I am making a distinction between prediction and knowledge. We have two words, and I think they really are two different things. Of course you can say "I know that the sun will rise tomorrow", but there is a profound distinction between that and saying "I know that the sun rose yesterday". Even if I was locked in my room all day, I can check with someone else and ask them if the sun rose. I can even deduce that the sun rose because no one noted that the sun didn't rise. Or I could figure out that the sun rose because the sun actually has to rise or we'd all be dead. But I can't know that tomorrow the sun will rise. I can predict with extremely high probability, but that is not the same as knowing it, like I can know about a past event.

Even though you can legitimately say in English 'I know the sun will rise tomorrow', that doesn't mean it is the same thing. It's quite different.

Alexander R Pruss said...

I don't see how the mere possibility of checking with someone else has much to do with knowledge. I am alone in a room right now and I know that the sun rose yesterday. Suppose that, suddenly, all non-divine persons, except for me, go out of existence. Then I can't check with anybody else, other than God, but then if I can check about the past with God, I can also check about the future with him. But surely the disappearance of other people doesn't affect my knowledge of the fact that the sun rose yesterday.

I agree that we have the word "prediction". We also have the word "retrodiction". Most facts about the future are known by prediction. (Some are known by prophecy and I don't know if that counts as prediction.) Some facts about the past are known by retrodiction. (E.g., by retrodiction we can know the position of the planets on January 7, 399 BC.) Prediction and retrodiction are not opposed to knowledge. Rather, some instances of prediction and retrodiction--for instance, the very certain ones--are knowledge.

We say of our friends things like: "I know she will keep her promise, unless she has a good excuse."

Anonymous said...

"by retrodiction we can know the position of the planets on January 7, 399 BC."

This is different from predicting the future. It is a difference in kind. We can reason that if the planets hadn't been in that position at that time, they wouldn't be in the position they are now. We can check on where they are now. We can also reason that the sun rose yesterday, since we see that the earth still exists, as normal today. These aren't certain proofs, but good enough to qualify as knowledge.

It's a different thing to predict where the planets will be tomorrow. We can calculate carefully, we can be absolutely sure of what will happen, but that is a different sort of thing from direct knowledge of past events (through direct experience and memory of them) or deductions about past events (baesd on reports or scientific inverstigation or whatever).

It's an excellent distinction to keep intact. Verifiability is an important part of that.

As far as being in a room and asking God, I think this is a false thought experiment. Our knowledge of things is different in kind from God's knowledge of things. We use the same word, but we actually have no way to even discuss God's knowledge, to understand it on any level. it is not relevant to a discussion of human knowledge.

Tanasije Gjorgoski said...

Yeah, that's true about mathematical and other metaphysically necessary truths, that's why I specified that some might buy into causal requirement for knowledge of contingent truths. While I tend to agree with you that causal connection as you described is good enough, I was just saying that some people might find it intuitively right to include a requirement for direct causal relation between the fact and mental state in the meaning of 'knowledge' (of contingent truths). (for example. as Anon does, they might think that for those other cases we might properly use the word 'prediction')

Beancan Tatterpants said...

@ Alex

Spot on. The issue and your hypotheses seem right.

I would add that there's a biological bias - the way our brains perceive time. Perhaps if we were Tramalfadorian we'd give equal deference to the past and future, but time seems to only move in one direction for us.

Chad McIntosh said...

“A causal requirement for knowledge makes mathematical and moral knowledge problematic.”

Assuming, of course, such candidates for knowledge are external to us (platonic entities, say). If they are internal or already mind-dependent, then the causal requirement is unproblematic. It seems to me that even the nuanced version of the causal requirement you note seems problematic for mathematical and moral knowledge if platonic. Consider the following argument, where O is some object of knowledge (say, an abstract object or moral principle), the argument can be summarized:

1. If O is external to S, S can have knowledge of O only if there is some causal relation R between S and O
2. O is such that it cannot enter R
3. If platonism is true, then O is external to S
4. Therefore if platonism is true, then S cannot have knowledge of O
5. But S has knowledge of O
6. Therefore, platonism is false

Do you think the above argument is sound?

Alexander R Pruss said...

Actually, I myself deny (2). :-) But that's not a very common view. (But it makes sense if mathematical entities are ideas in the mind of God.)