Wednesday, November 3, 2010

The causal theory of knowledge

According to the causal theory of knowledge, I only know p if my belief in p, or my evidence for that belief, is caused by the worldly state of affairs that p reports. Thus, I only know that I won the lottery if my belief that I won the lottery, or my evidence for that belief, is caused by my lottery win.

If the causal theory of knowledge is true, then I don't know that I now have two hands. For even though I can see my hands, my evidence of my having two hands and my belief in my presently having two hands isn't caused by my present having of two hands. Rather, it is caused by my very near past having of two hands. Information travels from my hands to my brain at most at the speed of light, after all.

But I know that I now have two hands. Hence the causal theory of knowledge is false.

The argument extends to any case of our knowledge that some external physical event is now occurring.

Objection 1: A temporally extended present event can be causally observed, because its past parts can cause us to perceive it.

Response: If E is such an event, then the causal theory of knowledge may allow us to know that E occurs tenselessly, but not that E is now occurring.

Objection 2: In ordinary language, "now" refers to a temporally extended period of time that reaches somewhat back. So the very close past existence of my hands is enough to make "I now have two hands" true.

Response: If it turned out—as in some world it does—that my hands ceased to exist in the interval between the light bouncing off the hands and my forming the belief that I have hands, we would say that my belief was mistaken. Hence the content of my belief that "I now have two hands" requires the literal present existence of two hands, not just a very close past one.

31 comments:

David said...

I don't think that your response to Objection 2 works. It's of course true that if your hands ceased to exist in the interval between the light bouncing off your hands and your forming the belief that you had hands, your belief would be mistaken. But according to many versions of the causal theory, e.g., Nozick's, your belief must track the truth only in possible worlds close to the actual world. Unless, then, the world where your hands cease to exist is close to the actual world, the situation that you describe wouldn't show that you don't know that you now have hands. Your response presupposes that Nozick's account, or ones like it, is false in the relevant respect.

Alexander R Pruss said...

The objection I was considering was that <I now have two hands> requires for its truth only that I have two hands shortly earlier. The response to the objection was that, on the contrary, it logically requires the exact present existence of the two hands. I wasn't claiming that the existence of worlds where the hands cease to exist shows I don't know. What I was claiming is that the existence of those worlds shows that the content of "I now have two hands" was such as to require the presence of two hands exactly there and then.

David said...

>>What I was claiming is that the existence of those worlds shows that the content of "I now have two hands" was such as to require the presence of two hands exactly there and then.

But why do you think the existence of those worlds shows that? On your account of the causal theory, you know that "I now have two hands" only if the state of affairs that I now have two hands causes the belief, or the evidence for that belief, that you do. Objection 2 says that in ordinary language "now" refers to a temporally extended period of time that extends somewhat back, so the very close past existence of my hands is enough to make "I now have two hands" true. Suppose that in the actual world and in all possible worlds close to the actual world, my hands don't cease to exist between the time the light waves leave my hands and the time I form the belief that I now have hands. Why isn't this sufficient, i.e., why does the existence of worlds in which my hands cease to exist in the way you indicate show that the literal presence of two hands is required in order to know that I now have two hands? Why does the existence of these worlds change in the content of "I now have two hands" in a way that defeats Objection2?

Alexander R Pruss said...

The case of those worlds shows that Objection 2's claim that all that's required for the truth of "I now have two hands" is that I had hands shortly earlier is false. I need to have two hands at the very time at which I am believing this.

Alexander said...

You are at some distance x from an object emitting a steady pattern of sound waves, and are awaiting the object to stop emitting sound waves. You seem like you would say that, if you presently are still hearing the sound and if x = 50 cm, you know that the sound hasn't stopped as of now. But it seems wrong to say that you know this if x = 1 km. But there is a continuum between the two cases. Why not just say that, strictly speaking, you don't know that the sound hasn't stopped as of *now* when x = 50 cm?

Alexander said...

You are at some distance x from an object emitting a steady pattern of sound waves, and are awaiting the object to stop emitting sound waves. You seem like you would say that, if you presently are still hearing the sound and if x = 50 cm, you know that the sound hasn't stopped as of now. But it seems wrong to say that you know this if x = 1 km. But there is a continuum between the two cases. Why not just say that, strictly speaking, you don't know that the sound hasn't stopped as of *now* when x = 50 cm?

Alexander R Pruss said...

Because you have inductive data that certain kinds of sounds don't stop that suddenly? But this is non-causal.

David said...

Your argument, if it is correct, shows that my seeing in the immediate past that I have hands isn't an instance of the present state of affairs that I now have hands causing my belief that I now have hands. But this doesn't suffice to show that the causal theory of knowledge is false. What if the state of affairs that I now have hands directly brings about my belief that I now have hands? It isn't logically required by your account of the causal theory that my beliefs about my hands be dependent on the physiology of how I see that I have hands, or even that my belief that I have hands be dependent on evidence that I do.

Michael said...

"According to the causal theory of knowledge, I only know p if my belief in p, or my evidence for that belief, is caused by the worldly state of affairs that p reports. Thus, I only know that I won the lottery if my belief that I won the lottery, or my evidence for that belief, is caused by my lottery win."

A state of affairs (I don't know what "worldly" adds) is itself an abstract entity, and if (as is generally supposed) they have no causal powers, the causal theory of knowledge is false.

A fact is also an abstract entity. The causal theorist will also have problems with knowledge of the future. That (assuming it exists) is caused by, if anything, something past. It is certainly not caused by anything future.

"Your lottery win" might be an event particular. The causal theory of knowledge might be:
S knows that p only if:
There is an event particular E such that:
- p entails that E existed
- E caused S to know that p
But this is wrong many ways. I know that if Socrates is human, and if all humans are mortal, then Socrates is mortal. This proposition does not entail the existence of any event particular.

Alexander R Pruss said...

"Worldly" was meant to hint that I was talking of Armstrong-style concrete states of affairs--ones that exist iff they obtain--rather than Plantinga-style abstract ones.

Maybe in your mortality case, the human nature is an immanent universal that causes beliefs about it?

Michael said...

We know there are no unicorns. "The non-existence of unicorns" is not the name of anything with causal powers, like a rock or an atom.

We know that either God exists or it is not the case that God exists. What item causes this knowledge ?

If "God exists or it is not the case that God exists" names something, I'm sure it's not a concrete particular with causal powers.

I'm afraid I don't understand what an immanent universal or a trope is (I don't know if that indicates success or failure on my part), but I don't see how human nature (humanity, as opposed to human beings) could cause anything. Fire can cause wax to melt. The nature of fire (its essence, what it is to be fire) is a property that, like humanity, causes nothing. It would be curious in any case if the causal theory of knowledge forced a metaphysics of immanent universals or tropes. It also seems irreconcilable with any of the common analyses of causation, like Humean constant conjunction, or Lewisean counterfactual dependence.

Back to the conditional, Socrates, humans, and mortals at least do not have to exist for me to know it. I gather that if there are tropes, Socrateity, Socrates' humanity, and Socrates' mortality would be tropes. They would presumably exist only if Socrates did, and since he doesn't have to exist for me to know it, they don't either. On tropes, if the ratio of the circumference to the diameter of golden circle A is Pi, and the ratio of the circumference to the diameter of wooden circle B is Pi, do they each have their own ratio ? There are two Pi's ? Do they each have their own number (one) of centers ? There are two ones ?

Alexander R Pruss said...

Quick response to one thing: "The nature of fire (its essence, what it is to be fire) is a property that, like humanity, causes nothing."

Have a look at my piece on Aristotelian forms. :-)

Michael said...

I have always hoped for an explanation of a substantial form :-) I'm afraid I don't have one yet, but it may be more my fault than yours. It sounds to me like there is nothing (no thing) that causes what horses do. Rather, they run because they are agitated, they eat because they are hungry, they are colored because they reflect light, they die because of disease, they shiver because of cold etc. If "it is the forms that are the causes of what the things do" then the form of the horse is not one, but many: the forms are being agitated, being hungry, reflecting light, being diseased, being cold etc. The form of the horse is not "the realization of the potentiality for being a horse". The realization of that potentiality is either the horse itself or the acquisition of the property of being a horse by something else. But neither horse nor acquisition is the cause of all the horse does.

Alexander R Pruss said...

Agitation by itself doesn't cause running. Agitation given a horse causes running, given a skunk causes an emission of noxious odor, etc. So, think of the horse's form as causing running when the horse is agitated, and of the skunk's form as what causes the emission of noxious odor when the skunk is agitated.

Michael said...

Substantial forms would be a nice sort of Swiss Army Knife. I get a little metaphysical vertigo here, but when I try to "think of the horse's form as causing running when the horse is agitated" I think of jockeys, urges to run, horse-flies etc.

Alexander R Pruss said...

Well, keep the "jockeys, urges to run, horse-flies" and everything of that sort fixed, and switch the horse to a slug. It won't run. Something causes that. :-)

Michael said...

Back to topic, let p = the proposition that my cup is neither a number nor a narwhale. If I know that p, my believing that p must be caused by something. The Causal Theory of Knowledge does not just say it must be caused, but that it must be caused by something like:
a) the fact that my cup is neither a number nor a narwhale
b) the state of affairs that my cup is neither a number nor a narwhale
c) my cup
d) my cup's being neither a number nor a narwhale
e) my cup's being concrete and inanimate
f) the immanent universals cup, number, and narwhale
g) the truth-maker for p (causal)
h) the truth-maker for p (logical)
If (d) and (e) name facts, they name different facts. (e) entails (d), but not vice versa. I might infer (d) from (e).

I see several problems for the Causal Theorist in:
1) Specifying the formula for the cause. What exactly must cause the belief that p ?
2) Knowledge of the future
3) Modal knowledge (of possibilities, necessities, impossibilities, might/would counterfactuals)
4) Denying the possibility there is a world with knowledge but no causality (ala Leibniz, RUssell)
5) Negative existentials (negative propositions)
6) Empty singular terms
7) Knowledge of mathematics
8) Knowledge of universal generalizations (like all gold melts at 1945 degrees F)
9) knowledge of existential generalizations (there are metals that melt at 1945 degrees F vs. these samples have etc.)

Pick knowledge of the future. I know I will not win the lottery because I haven't bought a ticket. Not winning does not cause my belief (believing). Not having bought a ticket does that, or belief about my doings in the recent past (if anything causes
it). But that is not the cause that we're supposed to be able to read off from the content of the belief. I think it would be a weakness in The Causal Theory if it precludes knowledge of the future, forces a particular account of causality, or forces a particular account of the metaphysics of causal relata. We need an analysis that yields the "proper" cause, from the belief. It won't do to say just that the belief must be caused, or that the belief is true.

Pick knowledge of universal generalizations. I know that all gold melts at 1945, and that some gold melts at 1945. Believing that some gold melts at 1945 is the cause the first, that these samples melt at 1945 of the second, but we can't "read it off the statement of the content". If we have to answer case by case what causes the belief, the Causal Theory reduces to the claim that if S knows that p, then something causes S to know that p. It's also not a theory of knowledge, is it. It just says that knowledge, whatever that might be, requires a cause, to be specified.

Michael said...

Formulated as an analysis, The Causal Theory might be:
S knows that p iff
S believes that p
p is true
S's believing that p is caused by the concrete state of affairs expressed by p

But there need be no concrete state of affairs expressed by p, where there is one, it need not cause the believing, and if it does there need be no knowledge. Gettier style counter-examples abound: I see a sheep in the field. I believe that there are sheep in the field. There are sheep in the field. My believing is caused by there being sheep in the field (causing the seeing, causing the believing). But I do not know there are sheep in the field: I've been given excellent reasons by the locals to believe there are no local sheep, just local sheep look-alikes, but I fail to resist.

Alexander R Pruss said...

Some causalists may have a very wide view of what concrete states of affairs there are.

Armstrong (I don't know what his epistemological views are, though) thinks that to every true proposition that p there corresponds a concrete state of affairs of its being the case that p. So, there are negative states of affairs.

A lot of people think that negative states of affairs can enter into causal explanations. Why was there a leak? Because no one plugged the hole.

I don't think any causalist would accept the analysis you propose. Minimally, you need "causes in a non-aberrant way" to rule out Gettier cases. And even then I bet it's not good enough.

Alexander R Pruss said...

Oh, and some causalists are quite happy to deny knowledge of the future.

James Bejon said...

A lot of people think that negative states of affairs can enter into causal explanations. Why was there a leak? Because no one plugged the hole.

I guess you have to hold to something like the PSR if you take this route?

Alexander R Pruss said...

I think the PSR is true, but I don't think you need the PSR to think that negative states of affairs can be causes.

Michael said...

I'm suspicious of "negative states of affairs":

not stopping for the light = driving through the light
not eating = fasting
not voting = abstaining
not moving = standing still
not talking = being silent
not deceiving = being honest
not married = single
not staying = leaving
not finite = infinite

The conditions to exclude non-aberrant ways have never to my knowledge been spelled out, and I too doubt that they can be, even to an approximation. I can't deny the Causalists their happiness over accepting no knowledge of the future, but my unhappiness with that consequence leads me to reject it. I know that something will happen tomorrow.

Michael said...

On the Causalist theory, can different beliefs have the same cause, or the same belief have different causes ? All inductive knowledge may be in jeopardy: the states of affairs that cause my beliefs that constitute my inductive base also cause my belief in the inductions over that base. Our common beliefs often (always ?) will have different causes. It's not just aberrant causal chains that are the problem, it's mapping back from common beliefs to their common causal chains "when everything goes right".

James Bejon said...

I think the PSR is true, but I don't think you need the PSR to think that negative states of affairs can be causes.

But without the PSR, it's hard to see how negative states of affairs can be causes, right? Absent the PSR, that, say, no-one moved my keys, can't causally explain why they're still where I left them, because they could have just disappeared without cause.

Separately, on such a view, isn't, for instance, my failure to build a time machine and go back in time and prevent my birth a cause of my being born? If so, we have causal loops all over the place.

Michael said...

PSRs are not required for negative states of affairs to be causes: not stopping for the redlight (which I would say = driving through it) caused the accident. That can be true in a world where some events have no causes, and some events have no explanations.

James Bejon said...

True. But the word "negative" is fairly inactive here. For, rather than making any kind of metaphysical claim, this just seems to be an example of rephrasing a given cause is negative terms.

Michael said...
This comment has been removed by the author.
Michael said...

That sounds like more or less my point above: There is no deep metaphysical distinction here between positive and negative states of affairs. There's a difference in descriptions, or in phrasing as you put it. No metaphysical distinction is going to be grounded in the aptness of English words and particles like "not" (non, un-, in-, anti-, a-, etc.). The "negative expressions" are fairly inactive: not eating (that is, fasting) for a month can cause a person to die. PSRs are many and various (see Alexander's essays for many of them), but none is needed to know that failure to pay your taxes can cause you to be subject to a hefty fine.

Michael said...

Alexander's original argument seems to be this time-gap argument:

1) I know (and hence believe) I now have hands.
2) Any cause of that belief happened prior to now.
3) Therefore the belief is not "caused by the worldly state of affairs that p reports".

p reports on now, the causes all happened earlier (even if only ever so slightly). This is right. I object it can't as stated account for knowledge about the future. He objects it can't even account as stated for knowledge about the present. The proponent might respond that we do not in fact know that we now have hands, only that we did ever so recently as not to matter. My hands might have been annihilated very slightly before I get the news. My objection gives us a little more temporal breathing room: no way something happening tomorrow causes me to believe that something will happen tomorrow.

James Bejon said...

I see. All good points. Thanks.