Ok, let's suppose Hume is wrong. Then try to solve the following problem: A billiard ball rolls toward a second billiard ball. Try to figure out (before they meet) what will happen when the two balls meet and state what method you used to do it.

By what reasoning do you find out whether the balls will attract each other, whether they will bounce off each other and in what direction, whether they will penetrate each other, or disintegrate, or explode, or ... or ...? — Jacques

If I'd never played pool before, I could read some books, and learn how the balls ought to behave through geometrical principles, and physics, then apply deductive reasoning to say that the balls on my table will move that way. Those principles I would study would have been developed from inductive reasoning. Or someone could tell me these principles. If I've played pool before, and learned from experience rather than being taught principles, I'd use inductive reasoning more directly.

@Count Timothy von Icarus @Metaphysician Undercover

I believe that Hume meant to say that in mathematics we can gain new knowledge by mere thinking, but in the natural sciences we cannot. For example, if we want to know how two balls will behave when they collide, without having observed a single collision of any objects before, we cannot find out by calculation or thinking.

So we have to do the experiment, we let the balls collide against each other more often from different directions and speeds and note the respective behavior.

Before the next collision, we can now calculate, based on our experience, what will happen, but only based on the assumption that the balls will behave as they did in the previous collisions. The assumption that the future will be similar to the past, however, cannot be justified by any calculation but only by experience.

Before the next collision, we can now calculate, based on our experience, what will happen, but only based on the assumption that the balls will behave as they did in the previous collisions. The assumption that the future will be similar to the past, however, cannot be justified by any calculation but only by experience. — Jacques

I don't think you have this quite right. The point Hume makes is that the assumption that the future will be similar to the past cannot be justified by experience, because the future is always ahead of us, and never properly experienced. Experience is always, all in the past. Therefore, that the future will be similar to the past is a principle derived from something other than experience.

The prediction, for what will happen in the future, is as you say, based on an assumption that the future will be similar to the past, but the issue between you and I is the question of what this assumption is based in. We could say that this assumption is somehow derived from experience, but we cannot say that experience justifies it, for the reasons explained above, and it is the way that it is derived which is at issue here.

So this is where inductive reasoning comes into play. When we take past experience, and produce a general principle like 'the balls behaved in such and such a way in the past, therefore the balls will behave in such and such a way in the future", this is an inductive principle. So the issue is what grounds, or justifies the inductive principle. That things behaved in such and such a way in the past, is not sufficient to produce the necessity to imply that they will necessarily behave this way in the future. What is needed is another premise which states that the future will be similar to the past. But this again appears to be just a more general form of the same inductive principle, How things have been in the past, will continue to be how they are in the future. So we do not escape the trap of relying on induction, and this does not give us the desired necessity, or certainty. However, Hume and you as well it seems, want to say that this principle (that the future will be similar to the past) is not actually derived from reason, but simply some sort of predisposition which we have toward looking at the future. We just naturally assume that things will be the same, rather than having derived this idea from experience and inductive reasoning.

We just naturally assume that things will be the same, rather than having derived this idea from experience and inductive reasoning. — Metaphysician Undercover

I'm sorry if this is a bit off-topic and I promise not to pursue it. What you say is true. But I can't resist pointing out that Hume makes a lot more of his similar point.

In Enquiry V, he repeats again that there is "no secret power or process of reasoning that leads us to expects similar results" from similar initial conditions and then points out that we are still "determined" to draw the same conclusion, even though we are "....convinced that .. understanding has no part in the operation". He concludes that "There is some other principle which determines (sc. us) to form such a conclusion. This principle is Custom or Habit." Then he says "By employing that word, ..... we only point out a principle of human nature, which is universally acknowledged, and which is well known by its effects". So it is essentially a causal explanation of why we continue in the same way despite the sceptical arguments.

So far, so good. But he goes further in footnote 1 to Enquiry VI.

"Mr. Locke divides all arguments into demonstrative and probable. In this view, we must say, that it is only probable all men must die, or that the sun will rise to-morrow. But to conform our language more to common use, we ought to divide arguments into demonstrations, proofs, and probabilities. By proofs meaning such arguments from experience as leave no room for doubt or opposition"

And then in X.1."... it is a miracle, that a dead man should come to life ; because that has never been observed in any age or country. There must, therefore, be a uniform experience against every miraculous event, otherwise the event would not merit that appellation. And as a uniform experience amounts to a proof, there is here a direct and full proof, from the nature of the fact, against the existence of any miracle; … Enquiry X. 1

So Hume really ought to be classified with G.E. Moore and Wittgenstein in as an opponent of sceptical conclusions.

So Hume really ought to be classified with G.E. Moore and Wittgenstein in as an opponent of sceptical conclusions. — Ludwig V

I don't think you can draw this conclusion so readily, because it's very unclear as to what Hume means by "uniform experience". And, Hume tends to be a little inconsistent in the way that he represents the relationship between reasoning and sense experience, so his claim that "uniform experience" can produce a proof which is beyond doubt, is in itself highly doubtful.

The problem is that by "uniform experience", Hume obviously means a number of distinct, or separate events, or experiences, compared with each other. This is how he generally represents sense experience, as distinct instances of sensation. But, it is through the use of memory, comparison, and inductive reasoning that we identify consistency through distinct events, to conclude uniformity. Of course memory and inductive reasoning are fallible, so Hume's statement that "uniform experience" provides a proof which leads no room for doubt is very unsound.

Hume's statement that "uniform experience" provides a proof which leads no room for doubt is very unsound. — Metaphysician Undercover

Yes. One of my points was precisely that he ends up ignoring the sceptical argument that he has so vigorously defended. However, his actual policy is to ignore it. He says, for example, that scepticism cannot be refuted and recommends a return to normal life and everyday occupations as a cure for it.

Hume response to scepticism is what I would call robust. (I think he would have liked Dr. Johnson's response, but he doesn't mention it.) Moore and Wittgenstein are similarly robust as well.

But, it is through the use of memory, comparison, and inductive reasoning that we identify consistency through distinct events, to conclude uniformity. — Metaphysician Undercover

You outline a standard account. But I don't accept that it is Hume's. But he is very clear a) that he accepts the sceptical argument (on the grounds that our experience provides no basis for rejecting it) and b) that we make our predictions because of association of ideas and custom or habit. He is careful to say that our understanding plays no part in this, which I think means that no process of reasoning is involved. I think his account is best classified as a causal one.

However, I have to admit that interpreting what he says is not straightforward because he never uses the word "induction". (I think it was introduced by J.S. Mill long after Hume wrote.)

You outline a standard account. But I don't accept that it is Hume's. But he is very clear a) that he accepts the sceptical argument (on the grounds that our experience provides no basis for rejecting it) and b) that we make our predictions because of association of ideas and custom or habit. He is careful to say that our understanding plays no part in this, which I think means that no process of reasoning is involved. I think his account is best classified as a causal one. — Ludwig V

This is simply a cop out by Hume. Reasoning is association of ideas, it is habit, and it is custom. And reasoning is the process of understanding. So if Hume wants to say that there is some type of association of ideas, which is not a form of reasoning, but some other type of mental habit, or custom, which is not conducive toward "understanding" like reasoning is, then he needs to explain what he's talking about. For him to assert that we pretend our predictions are based in reason, when this mental custom where predictions are based, is really something other than reasoning, requires justification. He needs to explain what other types of mental customs we have, which are other than reasoning, and how those other customs might result in successful predictions.

He needs to explain what other types of mental customs we have, which are other than reasoning, and how those other customs might result in successful predictions. — Metaphysician Undercover

I'm not trying to defend Hume, just to understand him. All we've got is what he wrote and I don't think those texts have the answers to your challenges, except that I don't think he ever claims that there is any guarantee that our predictions are always successful. That would be inconsistent.

Look at it this way. He argues 1) that all our ideas are drawn from experience 2) that experience provides no justification for making predictions based on past experience and 3) that we are going to go on doing just that. He also says that we have found this practice useful. Whether this counts as a justification or merely a cause is debateable.

So far, I don't think that's inconsistent. What is odd is that he changes the definition of "proof" (in a footnote, perhaps thinking that no-one will notice) and then ends up saying that "uniform experience amounts to a proof" - which, to be fair, is not quite the same as saying that it is a proof. The only defence I can think of for this move is a Wittgensteinian move along the lines "This is what we call proof in this context."

Does that make sense?

I think this is interesting as a response to scepticism that does not attempt to refute it.

I'm not trying to defend Hume, just to understand him. All we've got is what he wrote and I don't think those texts have the answers to your challenges, except that I don't think he ever claims that there is any guarantee that our predictions are always successful. That would be inconsistent. — Ludwig V

The point is not whether our predictions are guaranteed, or one hundred percent certain, but that we can have success in a consistent way. We can predict that the sun will rise day after day for example, with a great deal of success. Success is not "guaranteed", but such a prediction has proven successful in the past, and will likely continue to be so. The question is whether such a prediction, which has proven to be successful, and will likely continue to be so, is produced by reasoning, or some other form of habitual or customary mental activity.

I believe that the critical issue here is the matter of successfulness, reliability. We know from experiential evidence that conclusions made through reasoning are reliable. We also know from experience that predictions made without any form of reasoning at all, are highly unreliable. So, we can make random predictions without reasoning, but they are unreliable, and the question for Hume is how does he think we can produce any degree of reliability without reasoning.

Here's an example. Suppose two events occur in succession which are completely coincidental, and I wrongly conclude that one caused the other. Now I'll predict that if I want the later one to occur again, I can initiate the first, and predict the second. But that would be unreasonable because the concept of causation requires more than a simple temporal succession. There must be another premise which establishes the relationship between the two events as more than just a temporal succession, to validate "causation".

So when we look at a simple prediction, like the sun will rise tomorrow, the prediction may based solely on the continuity of what has occurred in the past, into the future, the consistency of nature which Hume is talking about. But notice that there is no "causation" invoked by this prediction. It's a sort of statistical analysis which produces a prediction based on probability, without any need to appeal to causation whatsoever.

Now, we can see that Hume tends to conflate these two types of successful prediction, the one based in statistical analysis, requiring no concept of causation, and the one based in causation. We should also recognize, that the latter, the successful prediction based in causation, is derived from a true understanding, reasoning, because it requires that further premise which establishes a true and necessary relationship between the thing considered as cause, and the thing considered as effect. But the other type of successful prediction, the one based in statistical analysis does not require that form of understanding, just recollection of past memories, and perhaps a method of applying mathematics in a more complicated prediction. We might inquire whether this type of prediction based in simple memory, and developed into an application of mathematics in statistical analysis, is a form of reasoning, or another type of habit or custom. And I think this would be a valid inquiry because such predictions of events are made without producing a "reason" for the occurrence of the predicted event. But this question is removed from the question of causation, and ought not be confused with it. And Hume seems to conflate these two types of prediction.

Look at it this way. He argues 1) that all our ideas are drawn from experience 2) that experience provides no justification for making predictions based on past experience and 3) that we are going to go on doing just that. He also says that we have found this practice useful. Whether this counts as a justification or merely a cause is debateable. — Ludwig V

So the matter which Hume needs to address, if he were here, is the extra premise required to establish a relationship between the predicted event, and the thing which is supposed to be the cause of that event. We can make predictions through statistical analysis, and we can say that these predictions are solely derived from "past experience". The sun has come up every day, I think it will come up tomorrow. And, Hume can make all sorts of claims about this type of prediction, but those claims would be irrelevant to the subject of "causation", because no cause is implied by such a prediction.

But if we want to address the type of prediction which is based on causation, and this is a type of prediction which relates two events in a necessary way, we can't simply take what is true about the other type of prediction, and apply it to this type of prediction, because they are completely different. And I really don't think we can relate two types of events as cause and effect, in the true and necessary way required to produce consistently successful predictions, without some form of reasoning. And this is why it is necessary to understand "the reason" why they are related as cause and effect, in order that the relationship proposed be the true and necessary relation required for consistently successful predictions.

The point is not whether our predictions are guaranteed, or one hundred percent certain, but that we can have success in a consistent way. — Metaphysician Undercover

I think Hume would agree with you. But he does say something to the effect that we discover from experience that there is a "pre-established harmony" between our predictions and what happens in the world. He doesn't explain himself. I think I see the finger of God there, but I'm speculating. (Hume (in the chapter on miracles) says that he does believe in God, but on faith, not reason, and he says that this faith is a miracle (i.e. caused by God).) If I'm right, this would probably be taken as a guarantee by believers.

Now, we can see that Hume tends to conflate these two types of successful prediction, the one based in statistical analysis, requiring no concept of causation, and the one based in causation. — Metaphysician Undercover

I would say that's exactly right.

We might inquire whether this type of prediction based in simple memory, and developed into an application of mathematics in statistical analysis, is a form of reasoning, or another type of habit or custom. — Metaphysician Undercover

It all turns on the question of justification. Statistics can identify correlations, but cannot justify them. I believe that statisticians do recognize the difference between correlation and causation, but I don't know how they deal with it. Certainly, statistics can't provide what you are asking for.

And I really don't think we can relate two types of events as cause and effect, in the true and necessary way required to produce consistently successful predictions, without some form of reasoning. And this is why it is necessary to understand "the reason" why they are related as cause and effect, in order that the relationship proposed be the true and necessary relation required for consistently successful predictions. — Metaphysician Undercover

I would agree that there is a valid question why correlations hold, when they do. Much science provides answers, in the form of explanations, which are defined (in philosophy) as deductive-nomological arguments. (That is, a syllogism that deduces the phenomenon to be explained from a law, or generalization) I don't find that particularly helpful, and it walks straight back into the arms of the argument against induction. I prefer to think of theories as mechanisms, showing how the effect is produced. In any case, theories put a given correlation into a larger context and so get round Hume's atomistic approach - taking each correlation on its own.

I have to say that I don't understand what necessity means here. I assume you don't mean the "true in all possible worlds" kind of necessity. That would be ambitious for an explanation of empirical phenomena.

I have to say that I don't understand what necessity means here. I assume you don't mean the "true in all possible worlds" kind of necessity. That would be ambitious for an explanation of empirical phenomena. — Ludwig V

This is how I understand "necessity" in the context of cause and effect. The cause is said to necessitate the effect, so we can say that when the cause occurs, the effect must occur. So for example, if a temperature of lower than zero Celsius is said to cause water to freeze, then we can say that whenever this temperature occurs, water will freeze necessarily. It is this "necessity" which validates the normal concept of causation, and which is very effective in prediction.

The difference between prediction by statistical analysis, and prediction by causation, is that the statistics alone cannot provide the required necessity. We could watch water freeze, always at the very same temperature, numerous times over and over again, and no matter how many times we do, we do not get the required necessity, even though we could use this to produce accurate predictions. Having it happen one hundred percent of the times, is not sufficient for necessity, because that might be one time, or a thousand times, in both cases it's one hundred percent. What provides the necessity is the understanding of how the molecules move, and this gives us the reason why water freezes at that temperature. It is this understanding of "the reason why" the two events are related, which validates the necessity of causation.

It is this understanding of "the reason why" the two events are related, which validates the necessity of causation. — Metaphysician Undercover

I'm afraid I do have a problem here. I don't disagree with this, but I don't understand what "validates the necessity" means.

we can say that when the cause occurs, the effect must occur. — Metaphysician Undercover

That's fine, except that I want to ask why "must". What if it doesn't?

If you say "Oxygen is necessary for life (except for anaerobic bacteria)", I understand that if there is no oxygen, most living things die. So I understand that most living things must live in an atmosphere that contains a certain percentage of oxygen.

So, for example, if a temperature of lower than zero Celsius is said to cause water to freeze, then we can say that whenever this temperature occurs, water will freeze necessarily. — Metaphysician Undercover

I'm not clear what the last word adds to the bald statement "water will freeze."

The best that I can offer is that if the prediction fails, I will not abandon the generalization, but treat it as a problem that demands an explanation that will preserve as much as possible of what I thought I knew. So if a sample doesn't freeze at that expected temperature, I will research until I find an answer - such as that the water contains too much salt to freeze at the normal temperature. Again, having learnt that fire causes burns, when I find burns occurring in the absence of fire, I will research until I realize that it is heat, not fire, that causes burns and amend my causal law accordingly. Admittedly, my belief that when a causal law fails, there must be an explanation, and my treatment of such failures as not just a fact, but a problem, is a matter of faith, (this may not be the right expression, but something along those lines is needed). Strictly speaking, when what we think is a causal law fails, that disproves the law (cf. Popper). But I can postpone abandoning the law until I'm convinced that there is no explanation for the exceptional case. There is no time limit on the postponement, so I am never compelled to abandon it and if my law is useful, I will classify the falsification as an unexplained event and continue to rely on it. Necessity is a matter of the status of "water will freeze", and not a straightforward question of truth or falsity.

Does that make any sense?

I'm afraid I do have a problem here. I don't disagree with this, but I don't understand what "validates the necessity" means. — Ludwig V

Do you see that "necessary" is a judgement, a claim, or assertion? And, this type of judgement is one which requires justification or else it is meaningless. if I assert "Y follows X, of necessity", or "it is necessary that Y follows X", those terms "necessity" and "necessary" are meaningless without justification. So when I referred to something which "validates the necessity', this type of justification is what I was referring to.

That's fine, except that I want to ask why "must". What if it doesn't? — Ludwig V

The word "must" is justified in the same way the necessity is. If the necessity is validated by justification, so that we can use "must", yet what was deemed as a necessity by that word 'must", does not actually occur, then we can infer a flaw in the justification (the reasoning).

If you say "Oxygen is necessary for life (except for anaerobic bacteria)", I understand that if there is no oxygen, most living things die. So I understand that most living things must live in an atmosphere that contains a certain percentage of oxygen. — Ludwig V

So you have given a false necessity here. Oxygen is not necessary for life, as your exception of anaerobic bacteria shows. This is evident from the fact that you proceed from "oxygen is necessary for life", through the stated exception, to your later assertion of "most life". A more appropriate understanding would determine the types of life forms which require oxygen, and then the true statement of necessity, "oxygen is necessary for these life forms..." could be made.

I'm not clear what the last word adds to the bald statement "water will freeze." — Ludwig V

This is exactly the point, the word "necessarily' adds absolutely nothing, unless it is accompanied by the justification (reasons). We can, and often do, facilitate communication by using 'necessary' in its various forms, without the explicit justification, because the justification is implicit. We simply assume when speaking, or writing, that the other person knows the reason for the claim of "necessary". But I do not think that we can go to the extreme position which Hume seems to be proposing, to say that the reasoning is not there at all, and the use of "necessary' would just be a custom or habit, with no underlying support of reasoning.

The best that I can offer is that if the prediction fails, I will not abandon the generalization, but treat it as a problem that demands an explanation that will preserve as much as possible of what I thought I knew. So if a sample doesn't freeze at that expected temperature, I will research until I find an answer - such as that the water contains too much salt to freeze at the normal temperature. Again, having learnt that fire causes burns, when I find burns occurring in the absence of fire, I will research until I realize that it is heat, not fire, that causes burns and amend my causal law accordingly. Admittedly, my belief that when a causal law fails, there must be an explanation, and my treatment of such failures as not just a fact, but a problem, is a matter of faith, (this may not be the right expression, but something along those lines is needed). Strictly speaking, when what we think is a causal law fails, that disproves the law (cf. Popper). But I can postpone abandoning the law until I'm convinced that there is no explanation for the exceptional case. There is no time limit on the postponement, so I am never compelled to abandon it and if my law is useful, I will classify the falsification as an unexplained event and continue to rely on it. Necessity is a matter of the status of "water will freeze", and not a straightforward question of truth or falsity. — Ludwig V

So what you present here is the issue, and that is the reliability of the underlying reasoning, the justification. And this is what fuels skepticism. Suppose it is customary for us to use "necessary' in its various forms, quite often, because it facilitates efficient and rapid communication. Each time we use one of those forms, there is an implied reasoning or justification. But each implied justification is unique in its reliability, depending on the different sorts of understandings which comprise the various implied justifications. So the simple word "necessary" refers to all sorts of different types and different degrees of understanding (reasoning). Now the skeptic will insist that we must analyze each one, each time that we habitually think of something as "necessary", or use that term, to ensure that there is a reliable understanding which supports it's use. Our customary ways of speaking hide misunderstanding.

The 'exceptions' that you refer to, which pop up, will appear at first, to be random. That is because there are so few of them that there is not enough to produce any sort of pattern which can be analyzed. However, the random exception demonstrates a deficiency in the necessity (the underlying understanding). The method we use to approach the underlying understanding, with the evidence of exceptions as ammunition, is the critical decision. I believe this requires a different type of thinking, maybe what they call thinking outside the box, and this might point to a type of intuition. What is important is the art of identifying relevant factors. Things which seem to be irrelevant, and which may be treated by the customary understanding as irrelevant, may actually be relevant. Also the skeptic apprehends the potential for significant misunderstanding within the things which are taken for granted.

A very good example of attacking an underlying understanding, which was riddled with exceptions is the geocentric model of the cosmos. I believe that the multitude of exceptions (retrogrades) made most intelligent people believe that the model was fundamentally wrong, for many hundreds of years. I think that even further back than Thales and ancient Greece, many believed that a heliocentric model was required. The problem was that mapping the planets as perfect circles could not produce something consistent with the observations, so it could not be supported by predictions. Therefore the desired heliocentric model could not be produced. So, just like you describe here, the use of the extremely complicated geocentric model, which employed all sorts of exceptions persisted. I interpret Aristotle as producing the metaphysical principles which denied the reality of perfectly uniform circular motion. This is the critical point, and it enabled Copernicus with the ammunition to explore various possibilities for non-circular orbits of the planets around the sun, eventually resolving the problem.

I don't think you have this quite right. The point Hume makes is that the assumption that the future will be similar to the past cannot be justified by experience, — Metaphysician Undercover

You are right, the assumption about the future cannot be justified by experience either, it cannot be justified by anything.

I think this depends on what counts as justification; justification meaning the reasons given for the claim about the future. And, we have different sorts of reasons for making such claims. A random assertion about the future would not qualify as being justifiable. Nor would the claim of "intuition tells me so", or "I have a feeling that such and such is about to occur', qualify as justification.

But when we get into more reasonable attempts at justification, I would say that as described above, bare statistical analysis is at the lowest level. This would be simply a matter of following a pattern of occurrence, without knowing the reasons for the pattern. But when we know the reasons for the occurrence of the pattern, we can take justification, and reliability to the next level. This is because then we can pay attention to the features which are designated as the reasons for the pattern, to watch for any changes within those features, which could result in anomalies in the pattern. This would make prediction more reliable, so it's a higher form of justification.

@Metaphysician Undercover, @Ludwig V, @Count Timothy von Icarus

Please tell me what your rationale is for believing that the future will resemble the past.

What would be the point? If I give you my answer now, and you accept it as you read it, what grounds do you have for thinking it will still hold in a week, or tomorrow, or even five seconds after your read it? :cool:

I am aware of a few ways of attacking the problem:

1. Techniques in statistics and probability theory do not rely on induction. We have proofs for why Weibull regressions, multinomial logits, OLS, etc. work. We can use these techniques in the context of Bayesian Inference, while hewing to the principal of maximum entropy. This will never allow us to be absolutely certain of any inferences, but it does allow us to have high confidence in them. There are also combinatoric arguments along this line, see: https://plato.stanford.edu/entries/induction-problem/#BayeSolu .

2. You can attack Hume's premises. The uniformity principle (UP) that Hume invokes for his attack on induction doesn't seem to hold up. This doesn't necessarily resolve the problem, but it changes it.

Maybe inductive inferences do not even have a rule in common. What if every inductive inference is essentially unique? This can be seen as rejecting Hume’s premise P5.

P5: Any probable argument for UP presupposes UP.

Proponents of such views have attacked Hume’s claim that there is a UP on which all inductive inferences are based. There have long been complaints about the vagueness of the Uniformity Principle (Salmon 1953). The future only resembles the past in some respects, but not others. Suppose that on all my birthdays so far, I have been under 40 years old. This does not give me a reason to expect that I will be under 40 years old on my next birthday. There seems then to be a major lacuna in Hume’s account. He might have explained or described how we draw an inductive inference, on the assumption that it is one we can draw. But he leaves untouched the question of how we distinguish between cases where we extrapolate a regularity legitimately, regarding it as a law, and cases where we do not.

One way to put this point is to say that Hume’s argument rests on a quantifier shift fallacy (Sober 1988; Okasha 2005a). Hume says that there exists a general presupposition for all inductive inferences, whereas he should have said that for each inductive inference, there is some presupposition. Different inductive inferences then rest on different empirical presuppositions, and the problem of circularity is evaded.

https://plato.stanford.edu/entries/induction-problem/#NoRule

3. You can show that Hume's argument is self-undermining.

First, you can attack Hume's Fork, the distinction between relations of idea (logical truths) and matters of fact, see: https://plato.stanford.edu/entries/analytic-synthetic/#ProDis . There appear to be significant problems with the formulation. For example, it was considered an a priori fact that a triangle's angles add up to 180 degrees. This turned out to not be true under all consistent geometries, e.g., a triangle on a curved plane, as drawn on a ball. That is, there is no way to tell between an a priori analytic truth and a firmly held dogma. To be sure, some truths true by virtue of being simple tautologies, but then these do no lifting in any analysis, and in any event, many of these can be shown to be true only as regards arbitrary axioms.

If relations of ideas are actually matters of fact, and inductive inference preformed on such facts is invalid, than Hume's position reduces to the radical skepticism of the Academics. We end up with "knowledge is impossible." Why should we even trust our memories? Just because your memory has seemed to be accurate in the past is no assurance that it will be in the future. But the statement that "knowledge is impossible" pretends at being a knowledge statement; it's the equivalent of the man who says "I only tell lies," a contradiction.

If anyone said that information about the past could not convince him that something would happen in the future, I should not understand him. One might ask him: what do you expect to be told, then? What sort of information do you call a ground for such a belief? … If these are not grounds, then what are grounds?—If you say these are not grounds, then you must surely be able to state what must be the case for us to have the right to say that there are grounds for our assumption….

-Wittgenstein

Hume's argument is can also be attacked by looking at the "Paradox of Analysis" and the "Scandal of Deduction." If deduction gives us no new information, then we can learn nothing that we did not already know from it. This also implies that Hume's argument denies the possibility of knowledge, as we cannot learn what we don't already know if only deduction is valid.

Either of these routes then leaves Hume open to all the arguments against radical skepticism, my favorite being from Augustine's "Against the Academics," because they're witty.

---

The above gives me reasons to think the past will be like the future, while also undermining the credibility of Hume's attempt to undercut this claim. Additionally, if I buy into computationalist conceptions of physics, then what comes before dictates what comes after by the same sort of logical entailment Liebniz had in mind when he developed his conception of computation, then my expectation that the future is like the past is not grounded in Hume's UP. Or if I buy into Hegel's arguments from phenomenology and speculative logic, then I see the progression of events, at least in the big pictures, as part of a process of dialectical-logical unfolding, which is also not grounded in the UP. The same is probably true for other views of nature that don't jump to the top of my mind right now; they reject Hume's premises.


BTW, I also think Hume's idea of causation is nonsense and that it contributed to his error here, and I say that as someone who largely appreciates his work, especially his work on this very interesting topic.

When we say X causes Y we don't mean that X occurs before Y in all instances of Y (constant conjuction). We generally mean to imply some sort of step-wise chain of entailments between Y's becoming a state of affairs and X then becoming a state of affairs, not merely conjunction. (As an aside, Hume's conception of cause as being reducible to constant conjunction arguably collapses in the face of (mostly) reversible laws of physics.)

Combined with his view on induction, Hume's whole argument against causation ends up turning into what is possibly just a very convoluted form of begging the question.

Hume says we cannot sense that cause is a form of step-wise entailment. Why not? Because our senses can't tell us anything about the logical laws that may or may not be underpinning events. Why not? Because seeing events follow from one another is somehow not seeing howevents follow from one another. But this is true only if you don't accept that events follow from one another in the first place. This problem is obscured by the fact that Hume argued for undecidability rather than the denial of a world that progresses logically.

If the world is logical, then my throwing a rock at a window and seeing it break is my observing causation/entailment in the exact same way that my tallying 3+4 to equal 7 is my observing that the two sum together to 7 when the inputs 3 and 4 are given for the addition function.

Following the Wittgenstein quote above, it's worth asking what Hume would count as observation of causation/entailment? If we discovered a physical theory of everything, and all observations followed its predictions, and further if we could use mathematical induction to prove that this relation holds in n+1 cases, would Hume still deny we have grounds for explaining causation? It seems possible given his arguments, but then this is essentially just radical skepticism that has been dressed up.

Example: we know how video games work. They use logical computation to produce their outputs based on given inputs. Everytime Mario jumps on a Goomba, it falls off the screen. But if we're Hume, we have to think that the console running Mario only appears to instantiate computation, and that our observing the step-wise enumeration of mathematical entailments is actually not sure to "really" be the step-wise enumeration of mathematical entailments in the world, it just "appears" to be identical. This is Descartes' evil demon territory, because it implies that while 2 + 2 = 4, adding two apples to two apples might result in 5 apples at some time in the future; we can't be sure because we can never determine if mathematics is instantiated when it appears to be.

The argument reduces to "cause cannot be logical connection because you cannot sense such a thing, and you cannot sense such a thing because you could only sense such a thing if cause is logical connection." However, if cause IS logical connection, then seeing X after performing Y every time would be your sensing the logical connection.

At best Humeans can say "if cause is step-wise entailment then the world would look exactly like it does, and you can indeed observe cause, but it's possible to imagine that our world is observably indistinguishable from such a world but somehow different." This is just positing a potential bare illogical nature of reality though, radical skepticism.

What I have noticed in the past, is that I can pick an imaginary time in the future, say tomorrow morning for example, and say something about it, like it will get light outside, and the sun will rise, and when that time comes, it will happen, just like before. So in my experience, I've gone to bed when it's dark, expecting that it will get light in the morning, because it has done so in the past, and it actually does as I expect. So this I believe, justifies (is the rationale for) my belief that the future will resemble the past, with respect to this feature of reality.

If, in your question, you are assuming a more general definition of "future" and "past', and you are ready to produce a definition of these things, as abstract objects, I believe that the future cannot resemble the past in such definitions. The two are in a way mutually exclusive, but not in the way of proper opposites, where there is a similarity derived from being exactly opposite, they are opposed more in the sense of dichotomous. But if we take some particular aspects of reality, particular material things, like in my example, I think there is rationale for believing that the future will be similar to the past for those particular things.

1. Techniques in statistics and probability theory do not rely on induction. — Count Timothy von Icarus

The theories themselves do not necessarily rely on induction to be produced, but a judgement of the reliability of them, in application, does rely on induction. So people might produce thousands or millions of such theories, in any random way, but we would only choose the ones proven by induction as reliable, to be used, and these would become the conventional.

One way to put this point is to say that Hume’s argument rests on a quantifier shift fallacy (Sober 1988; Okasha 2005a). Hume says that there exists a general presupposition for all inductive inferences, whereas he should have said that for each inductive inference, there is some presupposition. Different inductive inferences then rest on different empirical presuppositions, and the problem of circularity is evaded.

This is another way of putting the point I made above in my reply to Jacques. Inductive reasoning relies on particulars, and it proceeds toward making a general statement about similar particulars. And, predictions always concern particulars. So if we start with general principles, abstractions like "future" and "past" in the most general sense, the question of how they are similar is a completely different question. To show that two general abstractions like "future" and "past" are similar, would be to place them into a broader category of abstraction, the concept of "time', or "temporality", and say that they are both temporal concepts, therefore similar in that way.

So Hume really just makes an inductive conclusion about inductive conclusions, that they all employ some sort of presupposition about temporal continuity. This may be useful if we want to know something about the process of inductive reasoning, but then again it might itself be faulty induction. But to put inductive reasoning into the larger context, as to how the conclusions of induction are used by us, i.e. how induction is actually useful, we need to show how they are related to other logic, deduction.

This turned out to not be true under all consistent geometries, e.g., a triangle on a curved plane, as drawn on a ball. — Count Timothy von Icarus

This is not really a consistent geometry though. A "curved plane" is contradictory because the curve of a sphere requires three dimensions while the plane is two. To make the curved plane we need to annihilate the convention of dimensions, but in doing this we annihilate the triangle. So really, the triangle is incommensurable with the proposed "curved plane" (which is a misnomer because it's not a plane at all), and in reality a triangle's angles always add up to 180 degrees. Because the proposed triangle on a curved plane is not a triangle at all.

That is, there is no way to tell between an a priori analytic truth and a firmly held dogma. — Count Timothy von Icarus

This is why, as I said above, we could make thousands or millions of such theories, by what is called "pure mathematics" but only the ones which prove themselves to be useful (and this is itself an inductive method) are accepted into convention. The usefulness is what inspires the "firmly held dogma". So if it turns out that it's better for us (more useful) to annihilate the conventional spatial dimensions, such that we have the so-called "curved plane", then that will become conventional, or "firmly held dogma".

Additionally, if I buy into computationalist conceptions of physics, then what comes before dictates what comes after by the same sort of logical entailment Liebniz had in mind when he developed his conception of computation, then my expectation that the future is like the past is not grounded in Hume's UP. — Count Timothy von Icarus

As you point out though, your expectation is grounded in some sort of UP, or to put it more precisely, "a UP". And, as I discussed in my prior post, such principles obtain varying degrees of reliability. So we would need to isolate and analyze this specific UP as to its own peculiarities and uniqueness, in order to determine whether your expectations about particular aspects of the future are well grounded. Therefore the skeptic wins out in the end, because each such expectation is unique, and therefore must undergo examination through the skeptic's microscope, in a way unique to it.

Hi @Count Timothy von Icarus, thank you for your very interesting post.

If anyone said that information about the past could not convince him that something would happen in the future, I should not understand him. One might ask him: what do you expect to be told, then? What sort of information do you call a ground for such a belief? … If these are not grounds, then what are grounds?—If you say these are not grounds, then you must surely be able to state what must be the case for us to have the right to say that there are grounds for our assumption….

-Wittgenstein

I am not the man to say that information about the past could not convince me that something similar will happen in the future. I would be confident that it will happen again in the same way, I just couldn't say what my confidence is based on. That is Hume's point.

Do you know of any law that guarantees the future will necessarily correspond to the past? I for one currently believe there is none. You named a few in your post and I will try to understand and address them soon.

Please tell me what your rationale is for believing that the future will resemble the past. — Jacques

Wittgenstein, as so often, has it right when he says:-

If anyone said that information about the past could not convince him that something would happen in the future, I should not understand him. One might ask him: what do you expect to be told, then? What sort of information do you call a ground for such a belief? … If these are not grounds, then what are grounds?—If you say these are not grounds, then you must surely be able to state what must be the case for us to have the right to say that there are grounds for our assumption….-Wittgenstein

I would only want to add:-

First, it's not just about the future and the past. There is much about the past and the present that we do not know. What we normally do is to expect that what we do not know will resemble what we do know. It's about the known and the unknown.

Second, it's not a simple either/or. It would be quite unreasonable to expect that the future will totally resemble the past and to expect it to be totally different. We actually do is to expect that the future will resemble the past in some respects and to expect that it will be different from the past in other respects. That seems reasonable to me.

But we also have no choice but to continue to use the same language to describe the present, the past and the future. We can and do adapt our language in the light of unexpected events as they occur. What's the alternative?

Does that help?

Because seeing events follow from one another is somehow not seeing how events follow from one another. But this is true only if you don't accept that events follow from one another in the first place. — Count Timothy von Icarus

That's right. The difficulty is to see exactly what "how" means and to understand that asking such a question means rejecting Hume's idea of atomistic idea of experience (which analytic philosophy largely inherited from Hume). That requires understanding Wittgenstein's reasons for abandoning his logical atomism - that is, the colour-exclusion problem and his remark to the effect that a single proposition is never "compared to reality" but a system of propositions. Similarly, we do not experience the world as a succession of atomistic, independent events. We need to pay attention to the idea of a "Gestalt", to understand the part/whole relationship in a more complex way.

Either of these routes then leaves Hume open to all the arguments against radical skepticism, my favorite being from Augustine's "Against the Academics," because they're witty. — Count Timothy von Icarus

I haven't read the Augustine book. I'll make a note of that. Thanks.

Therefore the skeptic wins out in the end, because each such expectation is unique, and therefore must undergo examination through the skeptic's microscope, in a way unique to it. — Metaphysician Undercover

I'm afraid I disagree with both of you. You misunderstand Hume. His position is that scepticism is right if it recommends careful and judicious examination of the facts and judicious decisions based on them, wrong if it is applied excessively. I think that's about right. It's not a case of radical scepticism (Pyrrhonism according to Hume) or nothing.

So we would need to isolate and analyze this specific UP as to its own peculiarities and uniqueness, in order to determine whether your expectations about particular aspects of the future are well grounded. — Metaphysician Undercover

Hume's position is that even though our inferences are not well grounded, we will continue to make them, as a result of what he calls "custom or habit". He then makes a sequence of moves, as I outlined in an earlier post, to arrive at a non-sceptical position that "uniform experience" is proof. One may or may not think that's legitimate; it's certaintly dubious. There is also the problem that experience is not uniform, unless we select among our experiences. Which, as you are indicating, we do, and in the process notice differences as well as similarities.

but only the ones which prove themselves to be useful (and this is itself an inductive method) are accepted into convention. The usefulness is what inspires the "firmly held dogma". — Metaphysician Undercover

That's exactly what Hume says, in the end.

The difficulty is to see exactly what "how" means and to understand that asking such a question means rejecting Hume's idea of atomistic idea of experience (which analytic philosophy largely inherited from Hume). — Ludwig V

Yes, this I believe is the root of the problem. Hume described the experience of sensing as a series of static states which may change as time passes. This implies a break, a divide between each state. Then he moves to address the problem of how the mind relates one state to another. The distinct states being what sensation gives us. But i think that in reality, sensation is an experience of continuous activity, which we produce breaks in through withdrawing our attention, either intentionally or unintentionally.

This is a substantial difference because on the one hand we have the perspective of sensation providing images of natural states with implied natural divisions between them (Hume), and on the other hand we have the position that sensation provides natural continuous activity, with the sensing being imposing artificial divisions onto that continuity. So from the Humean perspective, it appears necessary for the observing mind to understand the natural relation between natural distinct states, in order to understand the natural progression of these states, and this is what he thinks of as causation. But if the other perspective is right, then sensation does not provide us with any naturally distinct states, and no natural separations. So any description of causation by this means would be to describe the relations between completely artificial separations.

But from the other perspective, these separations are apprehended as completely artificial, so they may even be totally arbitrary, or at best the separations are imposed in different ways for different purposes, (analogous to Wittgenstein's boundaries in meaning). So the difference manifests as Hume looking for an independent, objective relation between cause and effect, a relation which creates a unity through a natural form of synthesis, while the other perspective looks for the subjective principles of analysis whereby we divide what is present as a natural continuity.

I'm afraid I disagree with both of you. You misunderstand Hume. His position is that scepticism is right if it recommends careful and judicious examination of the facts and judicious decisions based on them, wrong if it is applied excessively. I think that's about right. It's not a case of radical scepticism (Pyrrhonism according to Hume) or nothing. — Ludwig V

I definitely wasn't saying that Hume is anti-skeptic. I see all philosophy as fundamentally skeptical. Even the anti-skeptic would be skeptical of skepticism. if one was just stating certainties, that would not be philosophy. The issue of course is the way that a philosopher places limits to one's own skepticism. No philosopher, not even Socrates, ever seems to be skeptical in an absolute sense, they always seem to believe that they get to the bottom somehow (In the skeptical form of analysis), and here they claim to find some sort of self-evident truth, something to take for granted (eg, Descartes' "I think therefore I am."). It may be the case that the best philosophers are the ones who never seem to get to the bottom, never finding any self-evident truth, and always remaining skeptical, open minded.

Hume's position is that even though our inferences are not well grounded, we will continue to make them, as a result of what he calls "custom or habit". He then makes a sequence of moves, as I outlined in an earlier post, to arrive at a non-sceptical position that "uniform experience" is proof. One may or may not think that's legitimate; it's certaintly dubious. There is also the problem that experience is not uniform, unless we select among our experiences. Which, as you are indicating, we do, and in the process notice differences as well as similarities. — Ludwig V

So the point I make above, is that this proposed "uniform experience" is not an inference, a custom or a habit at all. It is simply a statement of description of the sense experience, an observation. Therefore there is no need to inquire about the mental activity which creates this uniformity, it is not within the realm of conscious thought, and cannot be described as habit or custom.. However, since we are prone, (by habit), in our mental activity to impose divisions on this given uniformity, and we do seek truth, we ought to inquire whether there is any true, natural separations, upon which such divisions could be based.

This requires a deeper form of skepticism. We must cast doubt on what the senses provide for us, in the way of Plato, who teaches us that the senses deceive. From this perspective we can apprehend the continuity which is given by sensation as manufactured, created by the apparatus which produces the sense experience, and therefore there is the potential that this is not a true representation. Now we would have the proper platform for inquiring into the possibility of true divisions, the true separations in time, which the experience of sensation, as a continuity, hides from us in its deceptive ways.

I'm not sure if I made my point clear. I'm not saying "I believe the world progresses based on logical entailments." I'm saying that, IF this was true, then seeing things follow these deductive laws is seeing causation in action. When I throw the rock, I experience the cause of the window's breaking if my throwing the rock does indeed cause the window to break. It's hard to see what more Hume could ask for or what he thinks experiencing cause would look like if it could be experienced.

I don't think Hume is merely a skeptic, although his point might reduce to radical skepticism. He does not seem to be saying "I don't think we can ever be sure if we are seeing cause," rather, he is saying "cause reduces to constant conjunction and we can't see one action entailing another because one action doesn't entail another."

Now I know there is a later school of Humeans that emerged in the 1980s who say Hume is only talking about epistemological limits. I just don't see it, granted I've not reread his work extensively. It seems to me like he is taking the more concrete position of denying that causation, as generally understood, exists at all. But this argument is entirely based on the fact that seeing a billiard ball hit another one "isn't actually seeing the moving ball cause the still ball to move." This is where it seems like begging the question.

The Problem of Induction is much more sophisticated, and so Hume's real argument about cause gets lost in the mix. But the Problem of Induction only says that we can't be sure if causation will work the same way in the future as it has in the past. Rejecting induction doesn't require rejecting causation. The denial of causation as popularly understood doesn't hinge on the Problem of Induction, it hinges entirely on Hume's assertion that common experiences of cause aren't actually experienced of cause... because cause can't be experienced... because it doesn't exist... which is the very point the argument sets out to prove.

The theories themselves do not necessarily rely on induction to be produced, but a judgement of the reliability of them, in application, does rely on induction. So people might produce thousands or millions of such theories, in any random way, but we would only choose the ones proven by induction as reliable, to be used, and these would become the conventional.

Right, but the selection isn't arbitrary. It's based on a principle of indifference, as further formalized by the principle of maximum entropy. You're not going for best fit, because of over fitting problems, but the least assumptions, a sort of formalized Ockham's Razor.

So Hume really just makes an inductive conclusion about inductive conclusions, that they all employ some sort of presupposition about temporal continuity.

Good point. I think the larger issue that gets buried in Hume and much modern philosophy of science is that the acceptance that:
1. Logic and our understanding of it is valid; and
2. The world is a logical place where at least some things follow from others; and
3. The logic of the world is intelligible to us,

All need to come prior to any knowledge statements. If this is not true, and one thing doesn't follow from another, then any prediction is impossible; I
we can't even trust our memories. Our theories of information and semiotics also collapse in a context where outcomes for any observation X have no relationship to any others.

Claims against the rationality of the world also need to explain how so much science can involve doing deductive work on a chalkboard. Often, experiments are only doing the work of confirming deductive arguments about how nature progresses from state to state, i.e., cause. If rules accessible by deductive reasoning don't guide state progression, why should they seem to? If deductively accessible logical laws do cause progression, then seeing the rock break a window IS seeing causation.

Hume is following the Platonist tradition in allowing some types of knowledge to side step this problem. "A thing can't be green and not green," still supposedly holds. I don't disagree that it holds, but rather maintain that this requires that we trust that our sense of logic is meaningful prior to accepting this as a true statement.

And of course, we can be fooled as to logical statements. Is there a mathematics who student hasn't had at least one occasion where they have argued with their teacher about how they MUST be right because of iron clad logic, only to find out they are wrong? Has there ever been a programmer who hasn't run their program, absolutely certain the logic works out, only to get an inconceivable error they want to attribute to broken logic gates?

But if Hume's take gets reduced to being skeptical of all knowledge claims in this way, then it is just the claim of the radical skeptics, Descartes' Evil Demon, the Academics, etc.

If someone wants to maintain that Hume's Fork holds, they have to counter Quine and Co's arguments against it and explain why we, as creatures in an illogical world, who can know nothing certain of that world, can still somehow access inviolable a priori truths from the ether. It seems to me that if worldly creatures can access those truths then, in at least some sense, logic is in the world. But how can logic be in just part of the world and not collapse from the Principle of Explosion? This seems to require some sort of dualism.

This is not really a consistent geometry though. A "curved plane" is contradictory because the curve of a sphere requires three dimensions while the plane is two

"Curved plane," is my sloppy, improper terminology; it's a surface with curvature. The easiest way to visualize a triangle with more than 180 degrees is to think of a triangle drawn with a ruler on a sphere, or for one with less than 180 degrees, one drawn on a saddle. While more intuitive, this is a misleading analogy because we don't need three dimensions to make the triangle have degrees unequal to 180; hyperbolic geometry, on a hyperbolic plane, accomplishes this.

The argument that all mathematics is simply invented, and selected for its usefulness, is another angle from which Hume's Fork can be attacked. However, if one accepts that abstract objects are real ontological entities, then this also seems to provide a reason for doubting the reliability of Hume's distinction. I don't tend to buy into the argument that mathematics is only selected based on usefulness. Often, investigations are based on elegance and aesthetic preferences

To better understand my point of view, I would like to give an example:

The electric attraction between an electron and a proton can be described by Coulomb's law, which is given by:

F = (k * q1 * q2) / r^2

This formula has worked brilliantly since its invention until today, but is there a formula that guarantees that Coulomb's law will still be valid tomorrow?

Do you know of any law that guarantees the future will necessarily correspond to the past? I for one currently believe there is none. You named a few in your post and I will try to understand and address them soon.

Sure, but as to whether these laws actually describe our world is another question entirely.

One good example is Max Tegmark's Mathematical Universe Hypothesis. Tegmark posits that all abstract objects exist and that our universe is one such of those objects. Tegmark's speculation requires that a number of things be true, the theory of eternal inflation in physics, the existence of a single set of natural laws that govern the universe, and the ability of abstract objects to somehow generate the first person subjective experience we are familiar with. The last of these is simply glossed over in his book because it isn't his area of expertise, so all we get is that "very complex informational patterns produce experience."

But let's look at what happens if his highly speculative theory is true. If it is true, then everything in the multiverse is determined. A description of the abstract object of which we are a part could, at least in theory, tell us, using only deduction, exactly what will happen in the future and what happened in the past. Granted that, in Tegmark's view, this would be a description of an unimaginable,although finite, number of discrete "universes," plural.

Causation, correctly understood, would look quite different in such a world. All future states are already defined, so causation would really just be the enumeration of state transitions that occur according to mathematical laws. Cause doesn't really exist as commonly conceived in this case, since the universe is a complete four dimensional object, but it can be formally described as what apparent state transitions look like for an experiencing entity within the universe. This is true even if the nature of our universe entails that what we take to be physical laws radically change in the future, since those changes would also be merely traits of the abstract object that is co-identical with the universe.

More broadly, if the universe works according to set laws that can be described purely by using deduction, then there exists laws that can define that the future will correspond to the past. Indeed, if the universe has no randomness, then the very fact that the future and the past both are determined is such a law denoting similarity.

We, as finite entities in the universe, might never fully understand these laws. We might think we understand them and get them wrong. We might mistake something for a law that is really the manifestation of the interaction of more basic laws, which could lead to "laws" we think exist changing on us. However, there is no reason to reject that such laws could be known, although arguably there is good reason to reject the idea that they could ever be perfectly predictive, because you cannot plug all the information in the universe into a function and read the output while being within the universe yourself.

My main argument would be that such laws could be known entirely deductively, a type of knowledge Hume would accept. Indeed, this is how many people see the Holy Grail of a Theory of Everything. That we use induction to test the validity our mathematical models of how the universe works, or that our deductions are informed by prior inductive findings, does not preclude a wholly deductive understanding of the universe. This is true in the same way that, if a mathematician graphs equations to look at them and get an intuitive understanding of how they work with numbers close to zero, it does not entail that, if she later develops a deductive proof of some inequality, etc. it isn't "really" deductive because she used inductive reasoning based on the graphs to inform how she went about making the proof.

Now, people can argue that we can never be truly sure that any deductively derived description of the laws of the universe actually maps to the universe, no matter how much we verify it with induction, but this just collapses into radical skepticism. You might as well argue that we can never be sure if 2 + 2 will be 4 in our world, because "what if nature is instantiating some other abstract principle and it just looks like it it instantiating the one that has predicted everything up to this point," applies for all attempts to use deduction.



Maybe. If the universe follows laws, if it is deterministic (even in a stochastic way), then it seems possible, maybe even plausible given the successes of attempts to identify such laws, to define the root rules by which the present always evolves into the future. Perhaps the universe is deterministic but follows undefinable laws though? Then such a thing isn't possible.

The elephant in the room here is "initial conditions." How much of our universe is determined by brute fact initial conditions? If the universe progresses due to laws, but it has unexplainable initial conditions, then it might not be possible to determine if "law-like" phenomena continue in the future. Tegmark's theory gets around this by positing a multiverse in which all discernible initial conditions actually exist.

If the world doesn't progress in any determined manner then it is unclear if any knowledge can be grounded. Maybe we, and all our memories, spontaneously sprang into existence a second ago and will disappear 30 seconds from now? We can't know.

On another note: Hume's assumption that people think the future will be like the past because, in prior cases the future has indeed been like the past, is also flawed. It doesn't describe why people actually have these beliefs.

People often do think the future will be like the past, but this is often because this relationship is entailed by another belief of theirs, not because of inductive inference from past resemblances between future and past.

E.g., Hegel thinks the future will be like (and in more ways unlike) the past due to everything that exists existing due to logical necessity. Logos theologians think the future will be like the past because of their faith in a particular conception of God, not due to generalizations. Such belief might be tied to a single mystical experience.

The same is true for beliefs that the future will be unlike the past. Many Patristics believed in a doctrine of "Christian historical progress," where the world gets better over time. They believed this due to a faith in God paired with an interpretation of a few specific verses in Scripture. When the Roman Empire was collapsing and things appeared to be getting worse, this didn't lead to an inductive reassessment, but rather simply caused them to take these events as small setbacks against the backdrop of a larger trend. You see the same thing with Marxists. The belief about the resemblance between past and future comes from other beliefs, which might be held due to inductive arguments, but as often are held due to deductive arguments.

Not the mention that some conclusions that appear inductive are actually tautological. That water is H2O is a necessary truth, but one only arrived at by observation. However, just because observation was involved does not mean that UP is required to say water will H2O in the future. A lot of work in science, while utilizing observation, is focused on identifying these sorts of truths. Arguably, the finding that "causation" is identical with "progressions between states based on x specific laws," could represent the same sort of identity relationship. Indeed, the way causation is generally understood, if something spontaneous occurs in the future, as a brute fact, not according to any laws that can possibly be known, we would call it "uncaused" or "a miracle."

Scientific identities like:

Water = H2O

Gold = the element with atomic number 79
Hesperus = Phosphorus

Are necessarily true although they are discovered a posteriori.

The terms flanking the identity sign are rigid designators.


[Def. A rigid designator is a labelling device whose function is to pick out the same object or natural kind in every possible world, that is, in every possible counterfactual situation.]

Identity statements between rigid designators are necessarily true if they are true. Each term independently picks out the same thing in every possible world.

Although these identities cannot be known a priori, they are necessary empirical truths, discovered a posteriori, like all scientific identities.

Once we know that ‘water’ and ‘H2O’ refer to the same thing, we treat both terms as rigid designators.

They have different uses or connotations – a chemist would use the former, an ordinary speaker the latter – but they denote the same natural kind.

If water is necessarily H2O, there is no possible world (i.e . situation) in which pure water at normal pressure, if it is the natural kind we designate by that term, is not H2O, does not have the molecular structure it does, does not freeze at 0 0C.

This supports the essentialist picture. If a thing’s identity depends on what it is made of, its microstructure will necessarily determine its disposition to behave in particular ways, i.e. its causal powers.

Just replace the last paragraph, which has a reductionist view, with the more modern theory of "fundemental" parts only being definable in terms of the whole of which they are a part (fields), and you have an essentialist picture that dictates relations between the past and future that exist by necessity.

That's a lot to respond to in your posts. It's an impressive extended argument. I don't pretend this is comprehensive. This is just a series of comments.

If deductively accessible logical laws do cause progression, then seeing the rock break a window IS seeing causation. — Count Timothy von Icarus

I agree with the consequent, but I don't understand the antecedent. If the antecedent is false, then the project of understanding the world is hopeless. Or is there an alternative approach?

If the universe follows laws, if it is deterministic (even in a stochastic way), then it seems possible, maybe even plausible given the successes of attempts to identify such laws, to define the root rules by which the present always evolves into the future. — Count Timothy von Icarus

There's an ambiguity between "follows" in the sense of "comes after" and "follows" in the sense of "is constrained by". It doesn't make any sense to me to speak of the universe being constrained by natural laws. Natural laws are what the universe does given that it is not constrained. Actually, it is neither constrained, nor not constrained; it just does what it does.

There's a similarly weakness in the idea of causation. There's an idea that a cause somehow forces its effect. But that's a category mistake.

This supports the essentialist picture. If a thing’s identity depends on what it is made of, its microstructure will necessarily determine its disposition to behave in particular ways, i.e. its causal powers.

Unfortunately, although the idea that a thing's identity depends on what it is made of seems plausible, I can't accept essentialism, particularly not the variety that derives from Kripke.

The problem with that argument, for me starts from:-

Identity statements between rigid designators are necessarily true if they are true. Each term independently picks out the same thing in every possible world.

So far as I can see, Kripke's argument proves that for every statement capable of truth or falsity, if it is true, it is necessarily true. But doesn't it follow that any statement that is not true in every possible world, is not necessarily true in every possible world. I find this unhelpful.

I'm afraid I got confused by your post.

In the beginning, you say:-

Hume described the experience of sensing as a series of static states which may change as time passes. This implies a break, a divide between each state. Then he moves to address the problem of how the mind relates one state to another. The distinct states being what sensation gives us. But i think that in reality, sensation is an experience of continuous activity, which we produce breaks in through withdrawing our attention, either intentionally or unintentionally. — Metaphysician Undercover

But by the end, we have:-

From this perspective we can apprehend the continuity which is given by sensation as manufactured, created by the apparatus which produces the sense experience, and therefore there is the potential that this is not a true representation. Now we would have the proper platform for inquiring into the possibility of true divisions, the true separations in time, which the experience of sensation, as a continuity, hides from us in its deceptive ways. — Metaphysician Undercover

I agree with the first quotation, but not with the second and, although I accept that we often get things wrong, I'm not at all sure that it is because our sensations deceive us; it may be that they neither deceive nor reveal. The problem may like in our interpretations.

I agree with the first quotation, but not with the second and, although I accept that we often get things wrong, I'm not at all sure that it is because our sensations deceive us; it may be that they neither deceive nor reveal. The problem may like in our interpretations. — Ludwig V

In the first passage, I describe how my understanding of what we get from sensation differs from Hume's. There is a problem though with my perspective, and that is that for logical purposes, we describe things in terms of static states, in the same way that Hume describes sensation gives us. We say things like "this was the situation at time1, and this was the situation at time2. So the reason I believe Hume's position is incorrect, is because I think these static states are produced by us, by our minds, for the purpose of applying logic, and it is incorrect to say that this is what sensation gives to the mind. What sensation gives to the mind is a temporal continuity, and the mind breals the continuity into distinct states, in a variety of different ways, depending on the intended purposes. That is why I argued earlier that all the breaks in the continuity which constitute the divisions between t1 and t2, and such, are subjective, imposed for various purposes.

There is a fundamental incompatibility between the perception of reality as a persistently changing continuity, and as a succession of separate but contiguous discrete instances. This is an incommensurability which mathematicians have not been able to resolve. Therefore, one of the ways of representing the world must be wrong, either the way of sensation, as a continuity, or the way of logic, as a succession of discrete instances.

The inclination (intuition) is to accept the sense representation as the correct representation, and conclude that the way we use logic to represent reality is just a model, and this model is incapable of providing a true representation. But this renders that part of reality, the procession of time, as fundamentally, and necessarily unintelligible. However, as I explained in that post there is a potential way around this problem. It is completely possible that the way that our senses present the world to us is "deceptive", as explained by Plato. And from this perspective we can say that it is absolutely possible that the procession of time is truly intelligible to the human mind as a succession of discrete instances, and we just need to identify those breaks in time which constitutes the separate moments.

So, you say that the problem may lie in our interpretation. This is consistent with what I've argued. if reality is as I've said above, consisting of discrete moments, then the faulty interpretation is the assumption that how the senses present reality to us, as a continuity, is completely consistent with the way that reality actually is. If, on the other hand, we assume that the approach of our logic is incorrect, or inconsistent with the way that reality is, then we assume that reality is necessarily unintelligible. So to allow for the possibility that reality is intelligible to us, we must assume that this other assumption, that 'how the senses present reality to us is completely consistent with the way that reality actually is' is the faulty premise of interpretation. So to allow for the possibility that reality is intelligible to us, we must assume that the senses deceive us.

There is a fundamental incompatibility between the perception of reality as a persistently changing continuity, and as a succession of separate but contiguous discrete instances. This is an incommensurability which mathematicians have not been able to resolve. Therefore, one of the ways of representing the world must be wrong, either the way of sensation, as a continuity, or the way of logic, as a succession of discrete instances. — Metaphysician Undercover

This goes way beyond my criticism of Hume for his atomistic idea of experiences. Logical atomism is a different issue, and I'm not aware that anyone thinks it is viable. Your conclusion is that:-

So to allow for the possibility that reality is intelligible to us, we must assume that the senses deceive us — Metaphysician Undercover

Your conclusion has a certain paradoxical appeal. I agree that sometimes we draw the wrong conclusions from what our senses tell us (that's a bit over-simplified, but it will do for now); but surely we sometimes get it right. Similarly, reality is partially intelligible to us and partly not, and we work hard to understand the latter part. You seem very fond of comprehensive statements, but the truth is more mundane than that. For example, you say:-

We say things like "this was the situation at time1, and this was the situation at time2. — Metaphysician Undercover

And I say "Don't we also say things like "between t1 and t2 this process was going on?"

I'm not at all sure what you are getting at in this post.