Why would our habit not help us anymore on such a day? The moment a heavy object turns into a flying pig we will note such a change, presume it will happen again next time we throw a heavy object into the air and ride the pig. If rather than a permanent change in reality we are now in a world where nothing remains the same, then the heavy object might just as easily fall to the ground again next time as do any other thing, in which case maintaining our belief that it will do so is at least as likely as any other belief to yield fruitful results.

We can supplement Humean causation with a Darwinian one where constant conjunction produces reproductive success in organisms that expect the conjunction to continue in the future.

So then we animals are inherently pragmatic.

What's interesting is that Hume felt the need to explain our concept of causation with a causal psychological explanation. Perhaps because past constant conjunction alone wasn't enough to get the concept into our heads.

Turning to Darwin next, we can further explain our habituation to causation with an adaptive explanation. Animals who came to expect constantly conjoined events to continue their conjoining were better at predicting when and where there would be food, mates or danger, and thus had more reproductive success, passing that psychological tendency on. — Marchesk

The use of the word "explanation" sounds misleading, since Hume accepted the skeptical premise that the concept of causal necessity isn't reducible to "relations of ideas" (the analytic a priori) or to "matters of fact" (the synthetic a posteriori).

Hence the concept of causal necessity cannot be "explained" to the skeptic in terms of what is thinkable, and this must include the idea of psychological association in terms of an analytic or synthetic explanatory hypothesis.

If Hume is to be consistent, his notion of "psychological habit" or "custom" can at most refer to the empirical observation that historical experience of repeated trials with identical outcomes is often observed to co-occur with present expectations for similar outcomes in the future - expectations which we might call "judgements of necessity".

But to avoid contradiction the empirical co-occurrence of identically repeated empirical trials with judgements of necessity cannot itself be considered to be a "causally necessary" relation, and therefore "judgements of necessity" cannot be synonymous to either "observed matters of habit" or to histories of identical trials.

To conclude, "psychological association" can only refer to the empirical background conditions in which judgements of causal-necessity are typically observed to follow, but cannot constitute a theory of causal judgements. Causal-necessity isn't so much of a thinkable idea but a feeling of compulsion that one sometimes feels subjected to.


Secondly, although Darwinian selection provides a naturalistic explanation of our inferential capacities, it cannot augment Hume's classical skeptical argument since naturalistic explanations beg the very notion of causal necessity that is in question. Indeed it isn't even the case after granting the notion of causal-necessity that evolution is guaranteed to improve any of our empirical inferences.

Hume explained our tendency to say one thing caused another when we notice constant conjunction between the two events because of habit. — Marchesk

You can say that, but the big problem with it is that the concept of cause is entirely different from the concept of constant conjunction. They are not the same.

Kant explains: "The conception of a cause so plainly involves the conception of a necessity of connection with an effect, and of a strict universality of the law, that the very notion of a cause would entirely disappear, were we to derive it, like Hume, from a frequent association of what happens with what precedes"

Turning to Darwin next, we can further explain our habituation to causation with an adaptive explanation. Animals who came to expect constantly conjoined events to continue their conjoining were better at predicting when and where there would be food, mates or danger, and thus had more reproductive success, passing that psychological tendency on. — Marchesk

This is obfuscation now. You're thinking you solved the problem the same way the man who thinks he solved the problem by saying opium causes sleep because it has sleep-inducing properties.

The question is why, in the first place, do those animals who expect constantly conjoined events to continue to be conjoined in the future are better at predicting when and where the food, mates or dangers would be? Oh, it might be because causality really is a thing out there in the world, hmmm, I see... :-d

Oh, it might be because causality really is a thing out there in the world, hmmm, I see... — Agustino

Might just be.

Might just be. — Marchesk

I think philosophy has already settled this matter. The interesting question now is whether the causality is a priori (presupposed by our experience, and provided by our understanding) or a posteriori, derived from experience. Kant would claim that Hume definitely proved that it's not the latter, while he himself proved that it is the former. I think some Scholastics though would argue that Hume at least didn't understand causality as it had been traditionally understood, and as such was left with an impoverished notion of causality.

what was the tradition understanding?

As a realist, I would say that if we observe constant conjunction, we're observing causality. But we often have to infer the nature of it, and it's always possible to infer incorrectly or incompletely. Something is causing everything to fall to the Earth at a rate of 9.81 m/s, otherwise why would a cannon ball fall at the same rate as a feather, once you factor out air resistance? That's totally unexpected.

what was the tradition understanding? — Marchesk

Aristotle's 4 causes.

Basically, the 4 causes leave no gaps. Efficient cause and effect are understood to be temporarily simultaneous, so the Humean notion that we could imagine A happening first without being followed by B is false. Since A (the cause) is simultaneous with B (the effect), they cannot but be linked. Like drawing a line on the paper. The line that is drawn (the effect) is simultaneous with the movement of the pencil (the cause).

As a realist, I would say that if we observe constant conjunction, we're observing causality. — Marchesk

Constant conjunction (also known as correlation) isn't causality. They are two different concepts.

Interesting. So objects on the Earth cannot but be pulled at the rate of 9.81 m/s, because that is simultaneous with the Earth's gravitational field.

Basically, the 4 causes leave no gaps. Efficient cause and effect are understood to be temporarily simultaneous, so the Humean notion that we could imagine A happening first without being followed by B is false. Since A (the cause) is simultaneous with B (the effect), they cannot but be linked. Like drawing a line on the paper. The line that is drawn (the effect) is simultaneous with the movement of the pencil (the cause). — Agustino

This doesn't seem right. Efficient cause is necessarily temporally prior to the effect. If they were simultaneous, then there would be no temporal progression between the thing which is said to be the cause, and the thing which is said to be the effect, and one cannot be claimed to be the cause of the other without a temporal progression.

For instance, the line on the paper is not simultaneous with the moving of the pencil, it follows from it, as the moving of the pencil is necessary for the existence of the line, but the existence of a line is not necessary for the moving of the pencil. The pencil might move without a line occurring.

Without movement, the pencil makes a point. After moving, and not until after moving because the pencil must get out of the way, is there a line. If the pencil continues to move, then any part of the line which already exists is simultaneous with the movement now, but any new part of the line only follows after that new movement. So it is clear that there is only a line after the pencil has moved. If the pencil moving, and the line occurring were simultaneous, then one could not be said to be the cause of the other. It would be just as likely that the line occurring would be the cause of the pencil moving as vise versa.

Constant conjunction (also known as correlation) isn't causality. They are two different concepts. — Agustino

Right, although judging by how many people wrongly equivocate Hume's skeptical conclusion with "correlation doesn't equal causation", we should be careful to avoid the potential conflation of constant conjunction with the 'inferred' notion of essential correlation.

Perhaps we should say:

1. The constant-conjunction of A and B refers to the sampled correlation of A and B over a finite history of observations, that happens to equal 1.

2. Hume's problem of induction also implies that a correlation parameter cannot be rationally deduced from a sampled correlation, regardless of the value of the sampled correlation.

For just as we cannot rationally infer causation, we cannot rationally infer correlation.

For just as we cannot rationally infer causation, we cannot rationally infer correlation. — sime

The problem is that we do both all the time. Our technology and science is based on being able to make predictions from what we consider to be correlation or causation.

It's not just that we observe past Bs following As, it's that we infer C will follow B given AB. If we know that all objects fall to the Earth at 9.81 m/s, then we can propel something into outer space with an acceleration greater than 9.81 m/s. Then given other Bs following As in chemistry, we can create the fuel to make this propulsion work. And so forth to the point that we can land spacecraft on other planets we've never been to.

That's why it's hard to take Humean skepticism seriously outside of a philosophy discussion.

The problem is that we do both all the time. — Marchesk

Why is it a problem? Hume confesses that he does it himself, and by no means demands that one does not. He points out the limits of logical deduction. You can't get a will-be from a was, any more than you can get an ought from an is. The gaps are bridged by habit and sentiment. It's only a problem for the philosopher who has a false image of himself as purely rational.

You can't get a will-be from a was, any more than you can get an ought from an is. The gaps are bridged by habit and sentiment. — unenlightened

I don't buy the argument that the gaps are bridged by habit and sentiment anymore than Kant did, even though unlike Kant, I think causality is real.

It's just obvious to me that we inhabit a causal universe, otherwise there would be no reason for it to have such a deep, uniform order to it over billions of years.

That's why it's hard to take Humean skepticism seriously outside of a philosophy discussion. — Marchesk

I don't think Hume's skepticism is taken that seriously even in philosophical discussions, to be honest. Or at least, it ought not to. I mean it's difficult to read and understand Platonic/Aristotelian philosophy especially in the later Scholastic synthesis OR to read and honestly study Kant's Critique of Pure Reason and still take Hume's skepticism seriously.

I can somewhat understand the appeal of Hume's skepticism today given the prevalence of postmodernism and different forms of positivism and scientism, but from the coherence of the arguments, I really think it is sorely lacking. I really can't take Hume very seriously anymore.

For me, the two competing worldviews that are worthy of attention are Kantianism as developed by Kant and Schopenhauer, and the Platonic-Aristotelian philosophy as developed by the Scholastics. Everything else seems children's play, though other philosophers do sometimes make important, but lonely (as opposed to systematic) points (Heidegger, Wittgenstein, etc.).

The main difficulty, of course being that, because of the prevailing cultural climate, even most professional philosophers do not seriously study Scholasticism, preferring instead to read summaries and history of philosophy type of books. Nor do most philosophers, even amongst the professionals, seriously study Kant's Critique of Pure Reason - most read through it sometimes, or parts of it, as part of their university course, and then forget about it. Or they read through it through certain lens, either those of their professor, etc. They don't seriously think about the arguments made, and Kant is difficult to read because he makes like 10-15 subtle distinctions on every page between different concepts, so if you're not really really concentrating and re-reading segments multiple times, you don't really get it.

I mean I can see how, if you read the wiki summary, or the SEP page, or some history of philosophy book that you cannot see how, for example, Kant refuted Hume. But if you would spend the time properly doing it, I think the refutation is as definite as anything can ever get in philosophy.

1. The constant-conjunction of A and B refers to the sampled correlation of A and B over a finite history of observations, that happens to equal 1. — sime

Is the only difference between constant-conjunction as described above and correlation the fact that correlation presumably involves the added necessity to continue into the future? I don't think that's the standard notion of correlation, but I may be wrong. So please clarify what is the difference between your notion of constant-conjunction and correlation.

He points out the limits of logical deduction. You can't get a will-be from a was, any more than you can get an ought from an is. The gaps are bridged by habit and sentiment. It's only a problem for the philosopher who has a false image of himself as purely rational. — unenlightened

This, of course, would ignore the fact that "oughts" can be factual too (per an Aristotelian worldview). Or, in the case of other so-called problems, it seems to me that Hume just shows a complete lack of awareness of synthetic a priori judgements that Kant discusses at length. So Hume shows a truncated understanding formed only of synthetic a posteriori judgements and analytical a priori ones (which are nothing but a systematisation of synthetic a posteriori ones).

Interesting. So objects on the Earth cannot but be pulled at the rate of 9.81 m/s, because that is simultaneous with the Earth's gravitational field. — Marchesk

I would say that's close, but if you want to be really accurate, you'd say that the universal constant of acceleration G is a property of all gravitational fields (why? - cause that's just the nature of gravitational fields, ie formal cause, due to the effects of mass (material cause) on spacetime curvature - if you ask another why now, it would be answered with the final cause, which directs the other causes towards their particular ranges of effects), and that objects on Earth experience a pull (effect) that is simultaneous with the efficient cause of being present within a gravitational field. Indeed, that's why it's called instantaneous acceleration :P

This doesn't seem right. Efficient cause is necessarily temporally prior to the effect. If they were simultaneous, then there would be no temporal progression between the thing which is said to be the cause, and the thing which is said to be the effect, and one cannot be claimed to be the cause of the other without a temporal progression. — Metaphysician Undercover

There doesn't need to be a temporal progression. The cause is logically, though not temporarily, prior to the effect. Why logically? Because the cause must contain the effect within it, and not the other way around.

For instance, the line on the paper is not simultaneous with the moving of the pencil, it follows from it, as the moving of the pencil is necessary for the existence of the line, but the existence of a line is not necessary for the moving of the pencil. — Metaphysician Undercover

No, that doesn't tell me that it's not simultaneous, that just tells me that one is cause and the other is effect. You're talking of one being logically prior to the other one. The pencil can move without creating a line - if it doesn't move while in contact with, say, a page. A line cannot move a pencil, since it doesn't have that potency. Only a pencil has the potency of creating a line when moved on a paper. But the creation of the line and the movement of the pencil are simultaneous temporarily, though not logically, as explained above.

it seems to me that Hume just shows a complete lack of awareness of synthetic a priori judgements that Kant discusses at length. — Agustino

Yes, as far as I'm aware Hume never had wind of Kant at all. You can get a Kant from a Hume, but not vice versa. But as far as I can see "synthetic a priori judgements" are just a long-winded way of saying "sentiments".

But as far as I can see "synthetic a priori judgements" are just a long-winded way of saying "sentiments". — unenlightened

Why do you say that?

And have you read Kant's first Critique?

Let's take a simple synthetic a priori judgement. Here it is. It's Euclid's first postulate.

"A straight line can be drawn from any point to any other point"

The definitions involved are:

1) A point is that which has no part.
2) A line is a breathless length.
3) The extremities of a line are points.
4) A straight line is a line which lies evenly with the points on itself.

The definition and the postulate are from Euclid's Elements Book I (I paraphrase the first postulate to clarify its meaning).

Now, is the judgement synthetic or analytic? Since it cannot be derived from the definitions alone, it must be synthetic, as it adds something that isn't contained in the definition (as opposed to analytic, which merely explains something already contained in the definitions). And is it a priori or a posteriori? It is a priori. Why? Because it is universal and absolutely necessary, you cannot conceive it being otherwise. That means you don't have to appeal to experience to know it, and hence it must be a priori.

Where am I wrong?

But as far as I can see "synthetic a priori judgements" are just a long-winded way of saying "sentiments". — unenlightened

But since Hume was an empiricist, isn't he ceding ground to rationalism here by saying we have a sentiment toward causality? He's admitted there's something fundamental in our thought processes which we use to make sense of the world that doesn't come from sensory experience.

He's admitted there's something fundamental in our thought processes which we use to make sense of the world that doesn't come from sensory experience. — Marchesk

Indeed, and this is precisely Kant's point. Kant thinks that Hume is right about this:

"That which effects that the content of the phenomenon can be arranged under certain relations, I call its form. But that in which sensations are merely arranged, and by which they are susceptible of assuming a certain form, cannot be itself sensation. It is, then, the matter of all phenomena that is given to us a posteriori; the form must lie ready a priori for them in the mind, and consequently can be regarded separately from all sensation."

So indeed these forms, Kant would say, cannot come from sensory experience, they are the forms of sensory experience, and as such its conditions of possibility. But they are not, as Hume would say, merely psychological habit associations. Why?

Because "the conception of a cause so plainly involves the conception of a necessity of connection with an effect, and of a strict universality of the law, that the very notion of a cause would entirely disappear, were we to derive it, like Hume, from a frequent association of what happens with what precedes"

And also because without such forms existing a priori, sensory experience itself would be impossible. That follows from these forms being required for experience and these forms not originating from experience itself.

It is a priori. Why? Because it is universal and absolutely necessary, you cannot conceive it being otherwise.

Where am I wrong? — Agustino

Right there. I can perfectly well conceive of separate domains of any number of dimensions, euclidian and non-euclidian, such that one cannot draw any line between a point in one and a point in another. For example, in spherical geometry, there is a point (the centre) to which no line can be drawn, because the geometry is of the surface of the sphere or hypersphere.

For example, in spherical geometry, there is a point (the centre) to which no line can be drawn, because the geometry is of the surface of the sphere or hypersphere. — unenlightened

Hmmm...

In spherical geometry, assuming we're talking about an intrinsic as opposed to an extrinsic curvature of space, there would be no point at the centre of the sphere would there? I mean, if space itself is in the shape of a sphere, then there is no other space to contain that point in the centre of it, is there? Or am I wrong?

In spherical geometry, assuming we're talking about an intrinsic as opposed to an extrinsic curvature of space, there would be no point at the centre of the sphere would there? I mean, if space itself is in the shape of a sphere, then there is no other space to contain that point in the centre of it, is there? Or am I wrong? — Agustino

We can argue about it. In a sense you are right, because the centre would be postulated to exist in 'another (higher) dimension', outside the geometric space. But I think the point is already made, that what is conceivable or inconceivable varies according to how daring one's thinking is.

it is universal and absolutely necessary, you cannot conceive it being otherwise. — Agustino

I think we have already conceived it being otherwise.

We can argue about it. In a sense you are right, because the centre would be postulated to exist in 'another (higher) dimension', outside the geometric space. But I think the point is already made, that what is conceivable or inconceivable varies according to how daring one's thinking is. — unenlightened

Okay, but then we've just moved the problem one step further no? I mean from the POV of this higher dimensional space that contains the geometric space we're talking about, any point can be connected to any other point by a line no? The only way this wouldn't be possible is if we're dealing again with a non-Euclidean space stuck in a higher dimension space. But at some point, we would obviously have to stop with adding dimensions no? Otherwise, we'd have an infinite regress. So when we do that, it seems to me that the postulate still holds absolutely true, no?

So I think Kant's point still holds. I see that the judgement does have absolute necessity, though it may lack universality.

There doesn't need to be a temporal progression. The cause is logically, though not temporarily, prior to the effect. Why logically? Because the cause must contain the effect within it, and not the other way around. — Agustino

If you analyze "logical priority" you will see that the only valid way that something can be prior to another is that it is temporally prior. The claim that there is a logical priority which is not reducible to a temporal priority, would have to be justified either by examples and demonstrations, or else by some logical argument which demonstrated such. In the case of the former, all examples of priority will be demonstrated to be actually temporal priority or else the claim of "priority" will not be justified. In the case of the latter, if you have a logical argument which demonstrates that there is a type of priority which is not a temporal priority, then produce it.

No, that doesn't tell me that it's not simultaneous, that just tells me that one is cause and the other is effect. — Agustino

It can be demonstrated quite easily. Try it yourself. There is no line until after the pencil moves. Prior to movement the pencil is at a point and there is no line. After the pencil moves there is a line. The line does not appear until after the pencil moves.

The pencil can move without creating a line - if it doesn't move while in contact with, say, a page. A line cannot move a pencil, since it doesn't have that potency. Only a pencil has the potency of creating a line when moved on a paper. But the creation of the line and the movement of the pencil are simultaneous temporarily, though not logically, as explained above. — Agustino

Yes, the creation of the line is simultaneous with the movement of the pencil, the two are the same thing. However, the creation of the line is necessarily temporally prior to the existence of the line. The existence, or "being" of a thing is completely distinct from the "becoming" of that thing. And, the becoming of the thing, by which process the thing comes to be, is temporally prior to the being of the thing.

I have never understood the idea that causation does not exist. Causation is a deeply commonsensical idea. Sure, we can mistake correlation for causation, but the ideas of causation and correlation are very different. If I throw a rock at a glass window, the window breaks. There is a transfer of momentum and energy from the rock to the glass. The energy transfer changes the motion of the components of the glass causing stresses that exceed the strength of the material.

Isn't that the common sense meaning of "cause," at least in a physical sense - a transfer of momentum and energy manifested as a change in velocity.

If you analyze "logical priority" you will see that the only valid way that something can be prior to another is that it is temporally prior. — Metaphysician Undercover

Some things can't come to exist without the priors. Babies exist because of mothers. Life exists because of chemistry. Chemistry exists because of physics. In fact, chemistry has a close mathematical relationship to physics. The physics of atoms dictate that molecules will form under the right conditions.

Some things can't come to exist without the priors. — Marchesk

Are these "priors" not temporally prior? If the "prior" is necessary for the existence of the thing, then isn't the prior necessarily temporally prior to the existence of the thing