This is the converse of my musing; I was wondering whether identity over time was sufficient, rather than necessary, for causation.
In that sense of function it will function identically; that is intended to be the case in the hypothetical. “Biological function” includes things like “representing”, “referring”. Compare the utterance “my causal history” as uttered by the two men. It seems plausible that they have different references. (Or “this body’s causal history”.)
Yes, but all such strings stand on the same footing, can be equally validly be attributed to a given clone. That is exactly where the empirical incoherence comes from!
No. Let’s take a concrete physical realization of our thought experiment: a series of beam splitters. Each beam splitter transforms the incident beam into a superposition of beams, let’s say |A> + |B>. This is thus the state after the first beam splitter. After the second stage, a pair of beam splitters now, we will have two copies of that state; the part |A> of the first copy is exactly the same as the part |A> of the second: in particular, they will combine coherently.
To say that there is a fourfold distinguishability is to presuppose the identifiability of minds in time, nothing else.
You can make that choice, but in your model, it wouldn’t be a rational choice.
The other posters don’t seem to share your problems, but well: to have experienced something is to have a certain causal connection to it. Memory is not sufficient. This is not anti-physicalist: physics comes down to kinematics and dynamics. The kinematics concerns the state, while the dynamics concerns its evolution, and thus, its causal history. You seem to want to abolish the dynamics, because you recognize that in your model, there can’t be any meaningful dynamics on the level of minds if you want to be a physicalist, because the physical dynamics is insufficient to specify that of the mental realm. The problem is then just that you’re left with too impoverished a theory to make any meaningful assertions at all.
The wiki on swampman gives a good example of why causal history is necessary to give meaning to an assertion: suppose Davidson looks at a glass marble on a shelf. His utterance the next day concerning ‘the glass marble I saw yesterday’ then refers to that glass marble. But then suppose that there is another, identical marble behind the first one. In the case in which know the glass marbles were exchanged, nothing changes in Davidson’s memory or physical state; but the referent of his utterance changes nevertheless, now to the other glass marble.
But what now about swampman, making the same utterance? Which of the marbles does it refer to? Clearly, the question cannot be decided. The clones are in a similar situation: for a clone that now has card B, he cannot decide whether he previously has had card A or B. (But unlike the case of swampman, who should in principle trust his memory, the clone has every reason to distrust it, making the problem more severe in this case—but that really is just a side thread.) His assertions and beliefs simply fail to refer. A faith in scientific reasoning is then ungrounded: there is no specific history of observations that can be associated to the clone; thus, he cannot draw conclusions based on the history.
He can maybe shrug it all away, ignore the implications, and prefer some blissful ignorance—but of course, in that case, he probably would have stuck with a ‘shut up and calculate’ approach to quantum mechanics, which has after all the same explanatory power.
You can of course interpret it that way, if you aim to construct a strawman that goes essentially against everything that I’ve said in this thread. It wouldn’t be a very good way to argue, but I can’t stop you.
Well, the argument very briefly is:
Biological concepts inherently involve a concept of “proper function.” I.e., the concept of “heart” involves, inherently, the notion that a heart “functions to pump blood.” You can’t practice biology unless you’re thinking about the “proper function” of the parts of organisms. This is because without a notion of proper function, a biologist would have to treat a heart that fails to pump blood as a perfectly fine representative of the category “heart.” This is bad because it renders the biologist unable to make good biological generalizations. He can’t make the right distinctions to come up with useful biological laws.
But how do you know what the “proper function” of an organ is? By observation of its activities? But if that’s how you figure out the proper function, then a heart with a hole in it has a function–namely, it has the function of leaking blood out into the rest of the organism, rather than functioning to pump blood. This kind of result leads to the same problem described above–an inability to practice the science of biology, due to being unable to make the distinctions necessary for useful generalizations. Because everything turns out to have a “proper function,” and that proper function is just whatever a thing happens to actually be doing. To put it simply, there’d be no way for a biologist to say “hearts function to pump blood” because some hearts don’t, in fact, pump blood. The biologist would instead have to say “hearts function to pump blood as long as they actually do pump blood” but that makes it impossible to distinguish between a heart and anything else, because anything that pumps blood now counts as a heart, and anything that doesn’t pump blood now counts as not-a-heart (even though, intuitively and according to the practice of biology, some things that don’t pump blood are nevertheless hearts.)
So if it’s not present activities that constitute function, what is it? The theory held by Millikan et al is that what constitutes function is evolutionary history. Hearts function to pump blood because that’s what they were selected for. Specifically, an organ has function X when its existence is caused by the propensity of its ancestors (or anyway, ancestors’ corresponding organs) to do X.
Well, hopefully from that you can see the problem they’re trying to solve and why they solve it the way they do. They think that without this historically-constituted idea of function, biology is impossilbe.
I happen to disagree with them as well, but I don’t think they’re as easily dismissed as you think.
An outside observer might be able to look at him and say “gee, I feel bad for they guy; he refers to his body’s causal history, and yet he doesn’t realize that his causal history is not well-defined.” But from his point of view, him being identical, his statement is equivalent to the one made by the original version. From from his point of view, his reference, whatever he thinks it is, is identical. From the stand point of the scientific method it doesn’t matter whether what he thinks he is referencing is well-defined.
When each of the two say “this body’s causal history” they, being in the same physical state, must be thinking that they are referring to the same thing. By definition of “same physical state.” It is irrelevant whether, from some objective vantage point, what they think they are referring to is not well-defined.
So when you say “it seems plausible that they have different references”, I am assuming that you are not referring to anything that has any impact on their own subjective experience of reality. You are only making a statement about whether their references have some objective truth value, which is separate from the question of whether their subjective experience is the same. If we agree that their subjective experiences are the same, then what are we arguing about here? As far as I can tell, if we agree on this, then HMHW must agree that what I have said about probabilities in the clone example is correct. The probabilities are a subjective experience. It is irrelevant to point out that in some objective sense there is no meaning to each clone’s perceived experience of reality. From the point of view of each clone, his memories and subjective experiences are identical to if he had not been cloned, and his opinions about his remembered experience are no less valid. You could just as well say that the “original” was the impostor, and that his memories were faked. The fact that both are physically identical implies that it is meaningless to try to assert that one’s references are any less valid than the other. Either both are meaningful, or neither are. The fact that you can just as well mark the “original” as the impostor and there is no extra-physical rule to prevent one from doing this, clearly demonstrates this fact.
Thanks for the info. For the sake of this argument, what matters to me is that they be dismissed.
I shouldn’t do this in the morning while I’m also trying to get kids ready for school. I don’t think the above was very clear. Here let me do two things. First, let me lay out an outline of the main points of the argument. Second, let me say what Millikan et all say in response to the Swampman problem, as this is illustrative of their approach.
First, the argument I tried to give above in broad outline:
Two and three are probably pretty easy to agree with. Line one may be harder to see the justification for. The idea is, basically, that if you can’t say hearts are “supposed to” pump blood, then when trying to figure out things about hearts, you have to treat “defective” hearts on an equal footing with “healthy” hearts. But that doesn’t allow you to make good biological generalizations. (These guys, btw, are all about how biology is possible, and what assumptions a scienist must make in order to do it. The ontology is very much a secondary concern for them–or at least, as inheritors if not explicit holders to a pragmatist tradition, the ontological/practical distinction is not very live for them.)
Swampman coming up next.
What they say about Swampman:
There could never be a science that takes Swampman or entities like him as part of their subject. This is because it is impossible to make generalizations (biological ones anyway) that are developed from study of Swampman or that are developed from study of biological creatures and that cover Swampman as well. For no set of observations about Swampman allows for generalization to any other creature, precisely because of the uniqueness of his physical origin. When I study a human heart, I can generalize to other human hearts because I know something about this human heart’s origin, and the relation with other human hearts this origin implies. I know that human hearts are generally similar precisely because I know about this evolutionary history. But that assumption of history doesn’t apply to Swampman–hence whatever generalizations I may make about hearts, I can’t make about Swampman’s “heart,” and whatever observations I make of Swampman’s heart, I can’t generalize to any other heart or “heart.” There can be no science of Swampman.
But this illustrates that, in order to do biology, I must assume that evolutionary history is partially constitutive of the nature of biological entities.
Okay, well I hope this has been interesting and informative anyway.
(Sorry, you got me going on something I was recently very interested in.)
Possibly a related concern follows.
In the A/B card experiment, there’s going to be a guy whose “memories” are of having always received the A card. Based on these “memories” he understandably predicts that “he”'ll get the A card next time he goes through the beam splitter, with 100% certainty or nearly so. But of course that’s wrong–only half of his continuers will get the A card.
Well, we know that guy exists and he’s in an epistemically unfortunate position. But isn’t it the case that all of the people in this network of continuers are in substantially the same unfortunate position? For all of them, for all they know, could be on some highly unlikely branch of continuations, rendering all of their predictions completely unreliable.
It may be best for each person in this network to ignore this possibility and simply hope for the best. (Really we do that in the real world as well–sure we could be in a bizarre pocket of reality where science has seemed to work for a while but it actually doesn’t, but we can’t worry about that possibility since it yields absolutely no actionable predictions in the first place.) But isn’t the upshot that you get prediction and probability despite MWI, not because of it?
It’s entirely possible I’ve completely misunderstood what’s at issue in the conversation but hopefully what I just said has some relevance at least.
Further evidence I’m not grasping the issue–I was sympathetic with iamnotbatman’s view until I actually started typing this post and after several paragraphs changed my tune completely and rewrote everything.
Not from the clones’ perspective. They each have their memories, and have a perfectly coherent conscious experience. And that is all that matters. They can do science. They can discover that p=0.5, and so on, regardless of your thinking that what they are doing is objectively incoherent. The proof is in the pudding.
I wish we could focus on the clone experiment. The fact that you are not, and that I have limited time, in practice amounts to a form of attrition. The example you provide is not a good analogy. The beams do not have memory. There is no anthropic selection.
This is not at all clear to me. Perhaps someone else could help explain.
See, this is just your own definition. I’ll define it differently: to have experienced something is to have the memory of having experienced it.
By your own definition.
No, I don’t abolish the dynamics. The dynamics are still there. They are there in the clone example, as clear as clear can be. Each clone has a clear causal history. Each clone also has a memory that maps perfectly well onto a causal history. The viewpoint from each clone is meaningful. Whether you as an objective observer find it meaningful is completely irrelevant to the practice of science from the point of view of a subjective clone test participant.
A good scientist would say “I saw such a thing with such and such characteristics.”
How is this relevant to science? The swampman says “I saw such a thing with such and such characteristics,” and he is correct.
The clone, by definition, cannot not know that he is a clone. And yes he can decide: he uses his memory. This memory that maps perfectly well to a causal history that actually existed.
How can a faith in science be ungrounded when it is experimentally verifiable that it works? You are confusing the objective reality with the reality of a conscious observer. From the point of view of any conscious observer, science works. It is irrelevant from the view point of science whether or not in some greater philosophical sense there is any meaning to his beliefs.
As a scientist, he should prefer an approach to QM that is the most parsimonious. He knows that the Schrodinger equation appears to be valid for all matter dsitributions, and that it is therefore not internally consistent to try to insist that the Schrodinger equation fails in the presence of an observer who is also a matter distribution. Therefore he considers the implications of the Schrodinger equation being unmodified. He discovers that, amazingly, thinking carefully about the evolution of the universal wave function, by itself naturally leads to the appearance of collapse. He also find that anthropic selection naturally predicts a probability measure that determines the probability associated with each appearance of collapse. Though the probability measure has not yet explicitly been shown to equal the Born rule, the fact that so few assumptions leads so close to coincidentally explaining everything that would otherwise require additional assumptions, leads him to naturally think that the approach is rather promising and that additional assumptions are probably unnecessary.
My aim is to understand where you are coming from. I am doing my best to take your assertions and form logical implications from them. It doesn’t help when you give replies like the above rather than explaining why you reject the meat of the argument out of which you pulled the quote that you actually responded to.
So far as he can tell it is equivalent.
I don’t understand what this means.
Why is it not relevant? Exactly the claim is that things matter that aren’t part of the internal physical structure of the subject. And it is compatible with physicalism because the things that matter are still physical facts (just ones that are external to the subject).
I took it that for Half Man Half Wit there is a sense in which experience is objective.
If you can say “the only sort of experience that matters is subjective experience”, then it looks like this talk of swamp-man etc is an irrelevancy. I don’t know if that is true or not.
Some people think that “cannot tell the difference” doesn’t imply “there is no difference”. The fact that we cannot tell who is the original and who is the clone doesn’t matter. What matters is that there is a difference. And the difference is cashed out physically, e.g. the “original” is the thing that went in first (or the thing that was not caused to exist by the machine, or thing that corresponds to the space-time-worm that such and so, or whatever). Your opponent would say, fine, there are cases when we can’t tell what the causal histories of a thing are, but that doesn’t mean they don’t have different causal histories.
You may well disagree with this, but I think this is what you guys are arguing about.
I’m possibly abusing swampman. I only wanted the “doesn’t have the right history” part of it (so people that, in particular, have a casual theory of reference should think that swampman goes wrong, but for possibly different reasons).
Correct. This is also true in any interpretation of QM. There is always a probability that a person will be “lucky” and predict he will get card “A”, and always will. He will discover that p=1.0. This is a tiny probability, just as it is a tiny probability to find yourself to be such a clone. There are a binomial distribution of clones that discover p=0.4, 0.5, 0.6, etc, for a given number of clonings. The vast majority will discover p to be closer and closer to 0.5 as the number of clonings increase.
This is the same position anyone is in whenever they flip a coin, even in classical mechanics. You flip a coin 10 times and it’s heads every time. Is the coin fair, or are you just an unfortunate person?
The clone example shows how given a deterministic classical world in which cloning happens, a pure probability can originate from anthropic selection. You get prediction and probability because of the cloning, not in spite of it. By analogy, you get prediction and probability because of the MWI, not in spite of it.
It seems to me that by definition of experience the only sort of experience that matters is subjective experience.
But if there is no difference in subjective experience, then what does it mean to say that the difference is “cashed out physically”, when physically the clones are identical? There is a causal connection between the original at t1 and the original at t2, and there is a causal connection between the original at t1 and the clone at t2. The original continued to time evolve to t2, and the clone time evolved from t1 to t2, through the actions of the clone machine. But the laws of physics tell us – and this is experimentally tested into the ground – that physically it doesn’t matter if we take a physical ensemble and move it from (x1,y1,z1,t1) to (x2,y2,z2,t2). This is all the clone machine does. It applies a physical transformation to the original at t1, to propagate it to t2. There is still a causal connection. It is the same as though the original had simply been transported to (x1,y1,z1,t1). It could have been by car or by foot, but it happens to have been by clone machine. Is HMHW arguing that if I walk from (x1,y1,z1,t1) to (x2,y2,z2,t2), that I lose my identity in some non-subjectively noticeable but still somehow objectively important way? And if not, then why should being transported through cloning be any different? The action of Newton’s laws are equivalent to the continued infinitesimal destruction and cloning from (x1,y1,z1,t1) to (x1+dx,y1+dy,z1+dz,t1+dt) to (x2,y2,z2,t2).
It means there is some physical fact to appeal to when one makes claims about identity. And “physical fact” could be any physical fact; it doesn’t have to be internal to the subject.
The relevant physical facts in this case being “through the actions of the clone machine”.
It doesn’t make any difference to the internal structure of the thing. But, given that it makes sense to say it moved, or that it was at (x1,y1,z1,t1) and now at (x2,y2,z2,t2), we still have physical facts to appeal to; namely that it moved, and that it was at such and so, and now is at such and so. And then with cloning, it’s origin, or it’s method of motion (or whatever else is left physically) will be the difference maker. Why that difference matters I won’t attempt to answer.
My point is that the physical fact that is pointed to by HMHW has no characteristic that distinguishes it from the physical facts of ordinary existence. If I move from (x1,y1,z1,t1) to (x2,y2,z2,t2), HMHW can point to the physical fact that I am now at (x2,y2,z2,t2), and say “you have no identify.” The rules by which he decides that one physical fact merits non-identity and another does not, seem arbitrary to me.
But HMHW could just as well point to a taxi and then the relevant physical fact would be “through the actions of the taxi.”
Well, then, I am of the position that the difference does not matter, due to the translational invariance of physical law.
[ETA]: One can say that I no longer have the same identify if I take a taxi from (x1,y1,z1,t1) to (x2,y2,z2,t2), but then the word “identity” loses all meaning, for no one ever has any identify. One might as well take a more pragmatic approach: I call myself iamnotbatman at (x1,y1,z1,t1), and I call myself iamnotbatman at (x2,y2,z2,t2), etc.
But the difference between the states of the beams is exactly the same difference between the states of the clones: they are different by virtue of having received a different card. The beams map one-to-one to the clones and their states. The fact that you see that in the beam case, there is no memory to support your thesis makes it all the more bizarre to me that you can’t grasp that the same is true for the clones.
Which is, of course, all-out brain-in-a-vat-ism: memories, whether they refer to anything or not, are treated on equal footing. With such an assumption at the base of your epistemology, all science is just a comforting bed-time story.
But they don’t! For such a history, you need the identifiability of minds across time. It’s just not the case that the clone that received card A at time t[sub]2[/sub] is in any way the same as the clone that received B at t[sub]1[/sub]; this identification is not given by the physical evolution (I thought we were clear on that!), and thus, is simply not permitted (without postulating an extra-physical rule).
But which one it maps to is arbitrary. Really, if you’re just gonna answer one point in this post, answer the following one: Why does the clone, having drawn card A, have a memory of having drawn card B previously (say)? Or, more pertinently, why doesn’t he have a memory of having drawn card A? Because both possibilities existed, and are on equal footing: a clone at a previous time did draw card A, and a clone at the same previous time did draw card B. And since the physical evolution does not—and cannot!—determine that the present clone is the successor of the clone that drew B, as opposed to the clone that drew A, there is literally not a matter of fact asserting that the ‘correct’ history is BA, as opposed to AA; and this obviously holds for all histories.
This is bizarre. How can you misunderstand my position that much? I say, keep saying and have always said, in every post that touches on the matter, that the identity is provided by the physical evolution: any map that takes your past state and spits out your present one provides an identity relation. Such a relation (between minds) is missing in your scenario. But it is only such a relation that can make a memory of having observed A, say, my memory; otherwise, it would be equally plausible that I have observed B, as my present state is compatible with both.
Let’s go through the experiment.
Step 1) HMHW steps into the clone machine, steps out, and then receives a card.
There is now a copy of HMHW who thinks he is HMHW who remembers stepping into and out of the clone machine, and remembers having received card A. Call this version of HMHW: HMHW_A.
There is now also a copy of HMHW who thinks he is HMHW who remembers stepping into and out of the clone machine, and remembers having received card B. Call this version of HMHW: HMHW_B.
Step 2) HMHW_A steps back into the clone machine, steps out, and then receives a card.
There is now a copy of HMHW who thinks he is HMHW who remembers stepping into and out of the clone machine the first time, and remembers having received card A that first time, and who then stepped again into the clone machine, and received card A again. There is therefore a version of HMHW who remembers having received card A twice. Call this version HMHW_AA.
There is now a copy of HMHW who thinks he is HMHW who remembers stepping into and out of the clone machine the first time, and remembers having received card A that first time, and who then stepped again into the clone machine, and received card B the next time. There is therefore a version of HMHW who remembers having received card A once, and then card B the next time. Call this version HMHW_AB.
Next HMHW_B steps back into the clone machine, steps out, and then receives a card.
There is now a copy of HMHW who thinks he is HMHW who remembers stepping into and out of the clone machine the first time, and remembers having received card B that first time, and who then stepped again into the clone machine, and received card A. There is therefore a version of HMHW who remembers having received card B and then card A. Call this version HMHW_BA.
There is now a copy of HMHW who thinks he is HMHW who remembers stepping into and out of the clone machine the first time, and remembers having received card B that first time, and who then stepped again into the clone machine, and received card B again. There is therefore a version of HMHW who remembers having received card B twice. Call this version HMHW_BB.
Is that clear?
This is what your position logically leads to, as I have explained numerous times now.
You could help clarify if you answer, for example, the question: do you believe that if you are erased at (x1,y1,z1,t1) and reappear in the exact same physical state at (x2,y1,z1,t1), that you are now somehow a zombie, or that you are somehow missing some je ne c’est quoi?
If you answer yes or no, I can lead through the logic.