The article to which you refer presents a convincing case, but I think it’s probably inconsistant with a Tarskian semantic theory of truth. (ETA: assuming it aims at defining truth, or at laying out a criterion for the identity of facts). We would have to infer from Tom’s carrying the bucket to the bucket’s being carried by Tom, since we couldn’t offer the Tarskian sentence “‘Tom is carrying the bucket’ iff the bucket is being carried by Tom” up as a definition of the truth of “Tom is carrying the bucket.”
I can see Eliezer’s point on an epistemological level, but what theory of truth do we need in order to understand anticipations as bearing on the identity of facts themselves?
Suppose we say simply that an identical set of anticipations makes two facts identical. Now suppose that I’m working in a factory in which I must crack red and blue eggs open to discover the color of the yolk (orange in the case of red eggs, green in the case of blue. But suppose also that all red, orange yolked eggs are rough to the touch, and all blue, green yolked eggs are smooth. The redness and the roughness of an egg will lead to an identical set of anticipations (the orangeness of the yolk). But we certainly can’t say that the redness and the roughness of an egg are the same fact, since they don’t even refer to the same material state of affairs.
Apparently we’re speaking across a large inferential distance. I don’t know about Tarskian sentences, so I can’t comment on those, but I can clarify the ‘anticipation controller’ idea.
Basically, you’re defining ‘anticipation’ more narrowly than what Eliezer meant by the term.
If you tell me that an egg is rough, I will anticipate that, if I rub my fingers over it, my skin will feel the sensations I associate with rough surfaces.
If you tell me that an egg is red, I will anticipate that when I look at it, the cells in my retina that are sensitive to long-wavelength radiation will be excited more than the other cells in my retina.
Clearly, these are different anticipations, so we say that redness and roughness are two different facts.
If you say to me, ‘Tom is carrying a bucket’, I anticipate that if I were to look in Tom’s direction, I would see him carrying a bucket. If you say to me ‘a bucket is carried by Tom’, I anticipate that if I were to look in Tom’s direction, I would see… him carrying a bucket. In other words, whether you phrase it as (1) or (2), my anticipations are exactly the same, and so I claim they’re the same fact.
But you seem to be telling me that not only are they different facts, but somehow one is more fundamental than the other, and I have no idea what you mean by that.
Thanks for clarifying the point about anticipations, that was very helpful and I’ll have to give it more thought. I read Eliezer’s article again, and while I don’t think his intention was to give an account of the identity of facts, he does mention that if we’re arguing over facts with identical anticipations, we may be arguing over a merely semantic point. That’s very possibly what’s going on here, but let me try to defend the idea that these are distinct facts one last time. If I cannot persuade you at all, I’ll reconsider the worth of my argument.
In my comment to Alejandro1, I mentioned three sets of facts. I’ll pare down that point here to its simplest form: the relationship between ‘X is taller than Y’ and ‘Y is shorter than X’ is different than the relationship between ‘X carries Y’ and ‘Y is carried by X’. This difference is in the priority of the former and the latter fact in each set. In the case of taller and shorter, there is no priority of one fact over the other. They really are just different ways of saying the same thing.
In the case of carrying and being carried, there is a priority. Y’s being carried is explained by X’s carrying. Y is being carried, but because X is carrying it. It is not true that X is carrying because Y is being carried. In other words, X is related to Y as agent to patient (I don’t mean agency in an intentional sense, this would apply to fire and what it burns). If we try to treat ‘X carries Y’ and ‘Y is carried by X’ as involving no explanatory priority (if we try to treat them as the same fact), the loose the explanatory priority, in this case, of agent over patient.
An example of this kind of explanatory priority (in the other direction) might be this set: ‘A falling tree kills a deer’ and ‘a deer is killed by a falling tree’. Here, I think the explanatory priority is with the patient. It is only because a deer is such as to be killed that a tree could be a killer. We have to explain the tree’s killing by reference to the deer’s being killed. If the tree fell on a deer statue, there would be no explanatory priority.
But maybe my confusion is deeper, and maybe I’m just getting something wrong about the idea of a cause. Thanks for taking the time.
Apparently you’re working in something that’s akin to a mathematical system… you start with a few facts (the ones with high ‘explanatory priority’) and then you derive other facts (the ones with lower ‘explanatory priority’). Which is nice and all, but this system doesn’t really seem to reflect anything in reality. In reality, a deer getting killed by a tree is a tree killing a deer is a deer getting killed by a tree.
Well, I’m not intentionally trying to work with anything like a mathematical system. My claim was just that if by ‘in reality’ we mean ‘referring to basic material objects and their motions’ then we loose the ability to claim any explanatory priority between facts like ‘X carries Y’ and ‘Y is carried by X’. Y didn’t just get itself carried, X had to come along and carry it. X is the cause of Y’s being carried.
But all that hinges on convincing you that there is some such explanatory priority, which I haven’t done. I think perhaps my argument isn’t very good. Thanks for the discussion, at any rate.
The article to which you refer presents a convincing case, but I think it’s probably inconsistant with a Tarskian semantic theory of truth. (ETA: assuming it aims at defining truth, or at laying out a criterion for the identity of facts). We would have to infer from Tom’s carrying the bucket to the bucket’s being carried by Tom, since we couldn’t offer the Tarskian sentence “‘Tom is carrying the bucket’ iff the bucket is being carried by Tom” up as a definition of the truth of “Tom is carrying the bucket.”
I can see Eliezer’s point on an epistemological level, but what theory of truth do we need in order to understand anticipations as bearing on the identity of facts themselves?
Suppose we say simply that an identical set of anticipations makes two facts identical. Now suppose that I’m working in a factory in which I must crack red and blue eggs open to discover the color of the yolk (orange in the case of red eggs, green in the case of blue. But suppose also that all red, orange yolked eggs are rough to the touch, and all blue, green yolked eggs are smooth. The redness and the roughness of an egg will lead to an identical set of anticipations (the orangeness of the yolk). But we certainly can’t say that the redness and the roughness of an egg are the same fact, since they don’t even refer to the same material state of affairs.
Apparently we’re speaking across a large inferential distance. I don’t know about Tarskian sentences, so I can’t comment on those, but I can clarify the ‘anticipation controller’ idea.
Basically, you’re defining ‘anticipation’ more narrowly than what Eliezer meant by the term.
If you tell me that an egg is rough, I will anticipate that, if I rub my fingers over it, my skin will feel the sensations I associate with rough surfaces.
If you tell me that an egg is red, I will anticipate that when I look at it, the cells in my retina that are sensitive to long-wavelength radiation will be excited more than the other cells in my retina.
Clearly, these are different anticipations, so we say that redness and roughness are two different facts.
If you say to me, ‘Tom is carrying a bucket’, I anticipate that if I were to look in Tom’s direction, I would see him carrying a bucket. If you say to me ‘a bucket is carried by Tom’, I anticipate that if I were to look in Tom’s direction, I would see… him carrying a bucket. In other words, whether you phrase it as (1) or (2), my anticipations are exactly the same, and so I claim they’re the same fact.
But you seem to be telling me that not only are they different facts, but somehow one is more fundamental than the other, and I have no idea what you mean by that.
Thanks for clarifying the point about anticipations, that was very helpful and I’ll have to give it more thought. I read Eliezer’s article again, and while I don’t think his intention was to give an account of the identity of facts, he does mention that if we’re arguing over facts with identical anticipations, we may be arguing over a merely semantic point. That’s very possibly what’s going on here, but let me try to defend the idea that these are distinct facts one last time. If I cannot persuade you at all, I’ll reconsider the worth of my argument.
In my comment to Alejandro1, I mentioned three sets of facts. I’ll pare down that point here to its simplest form: the relationship between ‘X is taller than Y’ and ‘Y is shorter than X’ is different than the relationship between ‘X carries Y’ and ‘Y is carried by X’. This difference is in the priority of the former and the latter fact in each set. In the case of taller and shorter, there is no priority of one fact over the other. They really are just different ways of saying the same thing.
In the case of carrying and being carried, there is a priority. Y’s being carried is explained by X’s carrying. Y is being carried, but because X is carrying it. It is not true that X is carrying because Y is being carried. In other words, X is related to Y as agent to patient (I don’t mean agency in an intentional sense, this would apply to fire and what it burns). If we try to treat ‘X carries Y’ and ‘Y is carried by X’ as involving no explanatory priority (if we try to treat them as the same fact), the loose the explanatory priority, in this case, of agent over patient.
An example of this kind of explanatory priority (in the other direction) might be this set: ‘A falling tree kills a deer’ and ‘a deer is killed by a falling tree’. Here, I think the explanatory priority is with the patient. It is only because a deer is such as to be killed that a tree could be a killer. We have to explain the tree’s killing by reference to the deer’s being killed. If the tree fell on a deer statue, there would be no explanatory priority.
But maybe my confusion is deeper, and maybe I’m just getting something wrong about the idea of a cause. Thanks for taking the time.
Apparently you’re working in something that’s akin to a mathematical system… you start with a few facts (the ones with high ‘explanatory priority’) and then you derive other facts (the ones with lower ‘explanatory priority’). Which is nice and all, but this system doesn’t really seem to reflect anything in reality. In reality, a deer getting killed by a tree is a tree killing a deer is a deer getting killed by a tree.
Well, I’m not intentionally trying to work with anything like a mathematical system. My claim was just that if by ‘in reality’ we mean ‘referring to basic material objects and their motions’ then we loose the ability to claim any explanatory priority between facts like ‘X carries Y’ and ‘Y is carried by X’. Y didn’t just get itself carried, X had to come along and carry it. X is the cause of Y’s being carried.
But all that hinges on convincing you that there is some such explanatory priority, which I haven’t done. I think perhaps my argument isn’t very good. Thanks for the discussion, at any rate.