Teleology and the normal/abnormal distinction

Believers in teleology also tend to believe in a distinction between the normal and the abnormal. I think teleology can be prised apart from a normal/abnormal distinction, however, if we do something that I think we should do for independent reasons: recognize teleological directedness without a telos-to-be-attained, a target to be hit. An example of such teleological directedness is an athlete trying to run as fast as possible. There isn’t a target telos: for any speed the athlete reaches, a higher speed would fit even better with the athlete’s aims. But there is a directional telos, an aim tlos: the athlete aims in the direction of higher speed.

One might then say the human body in producing eyes has a directional telos: to see as well as possible. Whether one has 20/20 or 20/15 or 20/10 vision, more acuity would fulfill that directional telos better. On this view, there is no target telos, just a direction towards better acuity. If there were a target telos, say a specific level of acuity, we could identify non-attainment with abnormalcy and attainment with normalcy. But we need not. We could just say that this is all a matter of degree, with continuous variation between 20/0 (not humanly available) and 20/∞ (alas humanly available, i.e., total blindness).

I am not endorsing the view that there is no normal/abnormal in humans. I think there is (e.g., an immoral action is abnormal; a moral action is normal). But perhaps the distinction is less often applicable than friends of teleology think.

More on experiments

We all perform experiments very often. When I hear a noise and deliberately turn my head, I perform an experiment to find out what I will see if I turn my head. If I ask a question not knowing what answer I will hear, I am engaging in (human!) experimentation. Roughly, experiments are actions done in order to generate observations as evidence.

There are typically differences in rigor between the experiments we perform in daily life and the experiments scientists perform in the lab, but only typically so. Sometimes we are rigorous in ordinary life and sometimes scientists are sloppy.

The epistemic value to one of an experiment depends on multiple factors in a Bayesian framework.

1. The set of questions towards answers to which the experiment’s results are expected to contribute.

2. Specifications of the value of different levels of credence regarding the answers to the questions in Factor 1.

3. One’s prior levels of credence for the answers.

4. The likelihoods of different experimental outcomes given different answers.

It is easiest to think of Factor 2 in practical terms. If I am thinking of going for a recreational swim but I am not sure whether my swim goggles have sprung a leak, it may be that if the probability of the goggles being sound is at least 50%, it’s worth going to the trouble of heading out for the pool, but otherwise it’s not. So an experiment that could only yield a 45% confidence in the goggles is useless to my decision whether to go to the pool, and there is no difference in value between an experiment that yields a 55% confidence and one that yields a 95% confidence. On the other hand, if I am an astronaut and am considering performing a non-essential extravehicular task, but I am worried that the only available spacesuit might have sprung a leak, an experiment that can only yield 95% confidence in the soundness of the spacesuit is pointless—if my credence in the spacesuit’s soundness is only 95%, I won’t use the spacesuit.

Factor 3 is relevant in combination with Factor 4, because these two factors tell us how likely I am to end up with different posterior probabilities for the answers to the Factor 1 questions after the experiment. For instance, if I saw that one of my goggles is missing its gasket, my prior credence in the goggle’s soundness is so low that even a positive experimental result (say, no water in my eye after submerging my head in the sink) would not give me 50% credence that the goggle is fine, and so the experiment is pointless.

In a series of posts over the last couple of days, I explored the idea of a somewhat interest-independent comparison between the values of experiments, where one still fixes a set of questions (Factor 1), but says that one experiment is at least as good as another provided that it has at least as good an expected epistemic utility as the other for every proper scoring rule (Factor 2). This comparison criterion is equivalent to one that goes back to the 1950s. This is somewhat interest-independent, because it is still relativized to a set of questions.

A somewhat interesting question that occurred to me yesterday is what effect Factor 3 has on this somewhat interest-independent comparison of experiments. If experiment E₂ is at least as good as experiment E₁ for every scoring rule on the question algebra, is this true regardless of which consistent and regular priors one has on the question algebra?

A bit of thought showed me a somewhat interesting fact. If there is only one binary (yes/no) question under Factor 1, then it turns out that the somewhat interest-independent comparison of experiments does not depend on the prior probability for the answer to this question (assuming it’s regular, i.e., neither 0 nor 1). But if the question algebra is any larger, this is no longer true. Now, whether an experiment is at least as good as another in this somewhat interest-independent way depends on the choice of priors in Factor 3.

We might now ask: Under what circumstances is an experiment at least as good as another for every proper scoring rule and every consistent and regular assignment of priors on the answers, assuming the question algebra has more than two non-trivial members? I suspect this is a non-trivial question.

And one more post on comparing experiments

In my last couple of posts, starting here, I’ve been thinking about comparing the epistemic quality of experiments for a set of questions. I gave a complete geometric characterization for the case where the experiments are binary—each experiment has only two possible outcomes.

Now I want to finally note that there is a literature for the relevant concepts, and it gives a characterization of the comparison of the epistemic quality of experiments, at least in the case of a finite probability space (and in some infinite cases).

Suppose that Ω is our probability space with a finite number of points, and that F_Q is the algebra of subsets of Ω corresponding to the set of questions Q (a question partitions Ω into subsets and asks which partition we live in; the algebra F_Q is generated by all these partitions). Let X be the space of all probability measures on F_Q. This can be identified with an (n−1)-dimensional subset of Euclidean Rⁿ consisting of the points with non-negative coordinates summing to one, where n is the number of atoms in F_Q. An experiment E also corresponds to a partition of Ω—it answers the question where in that partition we live. The experiment has some finite number of possible outcomes A₁, ..., A_m, and in each outcome A_i our Bayesian agent will have a different posterior P_Ai = P(⋅∣A_i). The posteriors are members of X. The experiment defines an atomic measure μ_E on X where μ_E(ν) is the probability that E will generate an outcome whose posterior matches ν on F_Q. Thus:

• μ_E(ν) = P(⋃{A_i:P_Ai|_FQ=ν}).

Given the correspondence between convex functions and proper scoring rules, we can see that experiment E₂ is at least as good as E₁ for Q just in case for every convex function c on X we have:

• ∫_Xcdμ_E2 ≥ ∫_Xcdμ_E1.

There is an accepted name for this relation: μ_E2 convexly dominates μ_E1. Thus, we have it that experiment E₂ is at least as good as experiment E₁ for Q provided that there is a convex domination relation between the distributions the experiments induce on the possible posteriors for the questions in Q. And it turns out that there is a known mathematical characterization of when this happens, and it includes some infinite cases as well.

In fact, the work on this epistemic comparison of experiments turns out to go back to a 1953 paper by Blackwell. The only difference is that Blackwell (following 1950 work by Bohnenblust, Karlin and Sherman) uses non-epistemic utility while my focus is on scoring rules and epistemic utility. But the mathematics is the same, given that non-epistemic decision problems correspond to proper scoring rules and vice versa.

Comparing binary experiments for non-binary questions

In my last two posts (here and here), I introduced the notion of an experiment being epistemically at least as good as another for a set of questions. I then announced a characterization of when this happens in the special case where the set of questions consists of a single binary (yes/no) question and the experiments are themselves binary.

The characterization was as follows. A binary experiment will result in one of two posterior probabilities for the hypothesis that our yes/no question concerns, and we can form the “posterior interval” between them. It turns out that one experiment is at least as good as another provided that the first one’s posterior interval contains the second one’s.

I then noted that I didn’t know what to say for non-binary questions (e.g., “How many mountains are there on Mars?”) but still binary experiments. Well, with a bit of thought, I think I now have it, and it’s almost exactly the same. A binary experiment now defines a “posterior line segment” in the space of probabilities, joining the two possible credence outcomes. (In the case of a probability space with a finite number n of points, the space of probabilities can be identified as the set of points in n-dimensional Euclidean space all of whose coordinates are non-negative and add up to 1.) A bit of thought about convex functions makes it pretty obvious that E₂ is at least as good as E₁ if and only if E₂’s posterior line segment contains E₁’s posterior line segment. (The necessity of this geometric condition is easy to see: consider a convex function that is zero everywhere on E₂’s posterior line segment but non-zero on one of E₁’s two possible posteriors, and use that convex function to generate the scoring rule.)

This is a pretty hard to satisfy condition. The two experiments have to be pretty carefully gerrymandered to make their posterior line segments be parallel, much less to make one a subset of the other. I conclude that when one’s interest is in more than just one binary question, one binary experiment will not be overall better than another except in very special cases.

Recall that my notion of “better” quantified over all proper scoring rules. I guess the upshot of this is that interesting comparisons of scoring rules are not only relative to a set of questions but to a specific proper scoring rule.

Comparing binary experiments for binary questions

In my previous post I introduced the notion of an experiment being better than another experiment for a set of questions, and gave a definition in terms of strictly proper (or strictly open-minded, which yields the same definition) scoring rules. I gave a sufficient condition for E₂ to be at least as good as E₁: E₂’s associated partition is essentially at least as fine as that of E₁.

I then ended with an open question as to what the necessary and sufficient conditions for a binary (yes/no) experiment to be at least as good as another binary one for a binary question.

I think I now have an answer. For a binary experiment E and a hypothesis H, say that E’s posterior interval for H is the closed interval joining P(H∣E) with P(H∣∼E). Then, I think:

• Given the binary question whether a hypothesis H is true, and binary experiments E₁ and E₂, experiment E₂ is at least as good as E₁ if and only if its posterior interval for H contains the E₁’s posterior interval for H.

Let’s imagine that you want to be confident of H, because H is nice. Then the above condition says that an experiment that’s better than another will have at least as big potential benefit (i.e., confidence in H) and at least as big potential risk (i.e., confidence in  ∼ H). No benefits without risks in the epistemic game!

The proof (which I only have a sketch of) follows from expressing the expected score after an experiment using formula (4) here, and using convexity considerations.

The above answer doesn’t work for non-binary experiments. The natural analogue to the posterior interval is the convex hull of the set of possible posteriors. But now imagine two experiments to determine whether a coin is fair or double-headed. The first experiment just tosses the coin and looks at the answer. The second experiment tosses an auxiliary independent and fair coin, and if that one comes out heads, then the coin that we are interested in is tossed. The second experiment is worse, because there is probability 1/2 that the auxiliary coin is tails in which case we get no information. But the posterior interval is the same for both experiments.

I don’t know what to say about binary experiments and non-binary questions. A necessary condition is containment of posterior intervals for all possible answers to the question. I don’t know if that’s sufficient.

Comparing experiments

When you’re investigating reality as a scientist (and often as an ordinary person) you perform experiments. Epistemologists and philosophers of science have spent a lot of time thinking about how to evaluate what you should do with the results of the experiments—how they should affect your beliefs or credences—but relatively little on the important question of which experiments you should perform epistemologically speaking. (Of course, ethicists have spent a good deal of time thinking about which experiments you should not perform morally speaking.) Here I understand “experiment” in a broad sense that includes such things as pulling out a telescope and looking in a particular direction.

One might think there is not much to say. After all, it all depends on messy questions of research priorities and costs of time and material. But we can at least abstract from the costs and quantify over epistemically reasonable research priorities, and define:

1. E₂ is epistemically at least as good an experiment as E₁ provided that for every epistemically reasonable research priority, E₂ would serve the priority at least as well as E₁ would.

That’s not quite right, however. For we don’t know how well an experiment would serve a research priority unless we know the result of the experiment. So a better version is:

2. E₂ is epistemically at least as good an experiment as E₁ provided that for every epistemically reasonable research priority, the expected degree to which E₂ would serve the priority is at least as high as the expected degree to which E₁ would.

Now we have a question we can address formally.

Let’s try.

3. A reasonable epistemic research priority is a strictly proper scoring rule or epistemic utility, and the expected degree to which an experiment would serve that priority is equal to the expected value of the score after Bayesian update on the result of the experiment.

(Since we’re only interested in expected values of scores, we can replace “strictly proper” with “strictly open-minded”.)

And we can identify an experiment with a partition of the probability space: the experiment tells us where we are in that partition. (E.g., if you are measuring some quantity to some number of significant digits, the cells of the partition are equivalence classes under equality of the quantity up to those many significant digits.) The following is then easy to prove:

Proposition 1: On definitions (2) and (3), an experiment E₂ is epistemically at least as good as experiment E₁ if and only if the partition associated with E₂ is essentially at least as fine as the partition associated with E₁.

A partition R₂ is essentially at least as fine as a partition R₁ provided that for every event A in R₁ there is an event B in R₂ such that with probability one B happens if and only if A happens. The definition is relative to the current credences which are assumed to be probabilistic. If the current credences are regular—all non-empty events have non-zero probability—then “essentially” can be dropped.

However, Proposition 1 suggests that our choice of definitions isn’t that helpful. Consider two experiments. On E₁, all the faculty members from your Geology Department have their weight measured to the nearest hundred kilograms. On E₂, a thousand randomly chosen individiduals around the world have their weight measured to the nearest kilogram. Intuitively, E₁ is better. But Proposition 1 shows that in the above sense neither experiment is better than the other, since they generate partitions neither of which is essentially finer than the other (the event of there being a member of the Geology Department with weight at least 150 kilograms is in the partition of E₂ but nothing coinciding with that event up to probability zero is in the partition of E₁). And this is to be expected. For suppose that our research priority is to know whether any members of your Geology Department are at least than 150 kilograms in weight, because we need to know if for a departmental cave exploring trip the current selection of harnesses all of which are rated for users under 150 kilograms are sufficient. Then E₁ is better. On the other hand, if our research priority is to know the average weight of a human being to the nearest ten kilograms, then E₂ is better.

The problem with our definitions is that the range of possible research priorities is just too broad. Here is one interesting way to narrow it down. When we are talking about an experiment’s epistemic value, we mean the value of the experiment towards a set of questions. If the set of questions is a scientifically typical set of questions about human population weight distribution, then E₁ seems better than E₂. But if it is an atypical set of questions about the Geology Department members’ weight distribution, then E₂ might be better. We can formalize this, too. We can identify a set Q of questions with a partition of probability space representing the possible answers. This partition then generates an algebra F_Q on the probability space, which we can call the “question algebra”. Now we can relativize our definitions to a set of questions.

4. E₂ is epistemically at least as good an experiment as E₁ for a set of questions Q provided that for every epistemically reasonable research priority on Q, the expected degree to which E₂ would serve the priority is at least as high as the expected degree to which E₁ would.

5. A reasonable epistemic research priority on a set of questions Q is a strictly proper scoring rule or epistemic utility on F_Q, and the expected degree to which an experiment would serve Q is equal to the expected value of the score after Bayesian update on the result of the experiment.

We recover the old definitions by being omnicurious, namely letting Q be all possible questions.

What about Proposition 1? Well, one direction remains: if E₂’s partition is essentially at least as fine as E₁’s, then E₂ is better with regard any set of questions, an in particular better with regard to Q. But what about the other direction? Now the answer is negative. Suppose the question is what the average weight of the six members of the Geology Department is up to the nearest 100 kg. Consider two experiments: on the first, the members are ordered alphabetically by first name, and a fair die is rolled to choose one (if you roll 1, you choose the first, etc.), and their height is measured. On the second, the same is done but with the ordering being by last name. Assuming the two orderings are different, neither experiment’s partition is essentially at least as fine as the other’s, but the expected contributions of both experiments towards our question is equal.

Is there a nice characterization in terms of partitions of when E₂ is at least as good as E₁ with regard to a set of questions Q? I don’t know. It wouldn’t surprise me if there was something in the literature. A nice start would be to see if we can answer the question in the special case where Q is a single binary question and where E₁ and E₂ are binary experiments. But I need to go for a dental appointment now.

Recollection and two types of "Aha!" experiences

On some argumentatively central occasions, Plato refers to an intellectual “aha!” experience of seeing some point (say, something philosophical or mathematical). This is supposed to be evidence for the theory of recollection, because the experience is similar to remembering a nearly forgotten thing.

After insightful comments from students in my philosophy of mathematics seminar today, I think “aha!” experiencess come in two varieties. We might express paradigm instances of the two varieties like this:

1. Aha! I’ve always thought this, but never quite put it into words!

2. Aha! Now that I think about this, I see it’s got to be true!

An example of the first variety might be someone who hears about the Golden Rule, and realizes that whenever they were at their best, they were acting in accordance with it. I had a case of the second variety when I was introduced to the distributive law in arithmetic in grade three: I had never thought about whether a ⋅ (b+c) = a ⋅ b + a ⋅ c, but as soon as the question came up, with some sort of an illustrating mental picture, it was clear that it was true.

The two experiences are phenomenologically quite distinct. Type (i) experiences fit better with the Platonic picture of innate knowledge, since type (ii) experiences feel like a new acquisition rather than the recovery of something one already had. Another difference between type (i) and type (ii) experiences is that in type (ii) experiences, we not only take ourselves to have evidence for the thing being true, but the thing becomes quite unmysterious: we see how it has to be true. But type (i) experiences need not have this explanatory feature. When I have the vision of the truth of the distributive law of arithmetic, I see why it’s got to be true though I may not be able to put it into words. Not so with the Golden Rule. I can continue to be mystified by the incumbent obligations, but cannot deny them.

Literal remembering of a forgotten thing seems less like (ii) than like (i). When I remember a forgotten phone number by some prompt, I don’t have an experience of seeing why it’s got to be that.

Plato’s theory of recollection does not account for the phenomenology of type (ii) experiences. And perhaps Plato would admit that. In the Republic, he talks of “the eye of the soul”. The context there is the abilities of this life, rather than recollection. Perhaps type (ii) experiences fit more with the activity of the eye of the soul than with recollection.

At the same time, while (i) is a bit more like remembering, it’s not exactly like it, either. Remembering need not have any “I’ve thought this all along” aspect to it, which type (i) experiences tend to have. So I think neither of our “Aha!” experiences is quite like the theory of recollection leads us to. Is there a third “Aha!” experience that does? I doubt it, but maybe.

Competent language use without knowledge

I can competently use a word without knowing what the word means. Just imagine some Gettier case, such as that my English teacher tried to teach me a falsehood about what “lynx” means, but due to themselves misremembering what the word means, they taught me the correct meaning. Justified true belief is clearly enough for competent use.

But if I then use “lynx”, even though I don’t know what the word means, I do know what I mean by it. Could one manufacture a case where I competently use a word but don’t even know what I mean by it?

Maybe. Suppose I am a student and a philosopher professor convinces me that I am so confused that don’t know what I mean when I use the word “supervenience” in a paper. I stop using the word. But then someone comments on an old online post of mine from the same period as the paper, in which post I used “supervenience”. The commenter praises how insightfully I have grasped the essence of the concept. This someone uses a false name, that of an eminent philosopher. I come to believe on the supposed authority of this person that I meant by “supervenience” what I in fact did mean by it, and I resume using it. But the authority is false. It seems that now I am using the word without knowing what I mean by it. And I could be entirely competent.

Kripke's standard meter

Back when there was a standard meter, Kripke claimed that it was contingent a priori that the standard meter is a meter in length.

This seems wrong. For anything narrowly logically entailed by something that’s a priori is also a priori. But that the standard meter is a meter in length entails that there is an extended object. And that there is an extended object is clearly a posteriori.

Kripke’s reasoning is that to know that the standard meter is a meter in length all you need to know is how “meter” is stipulated, namely as the actual length of the standard meterstick, and anything you can know From knowing how the terms are stipulated is known a priori.

There is something fishy here. We don’t know a priori that the stipulation was successful (it might have failed if, for instance, the “standard meter” never existed but with a conspiracy to pretend it exists). In fact, we don’t know a priori that any stipulations were ever made—that, too, is clearly a posteriori.

Maybe what we need here is some concept of “stipulational content”, and the idea is that something is a priori if you can derive it a priori from the stipulational content of the terms. But the stipulational content of a term needs to be defined in such a way that it’s neutral on whether the stipulation happened or succeeded. If so, then Kripke should have said that it’s a priori that if there is a standard meterstick, it is a meter long.

The unthinkable and the ineffable

Suppose that Alice right now thinks about some fact F and no other fact. Then we can stipulate that “Xyzzies” is a sentence whose content is that very fact which Alice is thinking. Thus:

1. If a linguistically identifiable person can think about some fact F to the exclusion of other facts at a linguistically identifiable time, then F can be expressed in a language.

It does not, however, follow that every fact can be expressed in a language. For it’s epistemically possible that there is a fact F such that a person can only think about F if the person is simultaneously thinking about G and H as well, and there may be no way for us to distinguish F from G and H in such a way as to stipulate a term for it.

This may seem like a pretty remote possibility, but I think it’s pretty plausible. There could be some fact F that only God can think. But presumably any fact has infinitely many logical consequences. But since God is inerrant and necessarily thinks all facts, necessarily if God thinks F, he thinks all the infinitely many logical consequences of F as well. And it could well be that we have no way of distinguishing F from some of its logical consequences in such a way that we could delineate F.

So it is possible to accept (1) while holding that some thinkable facts are ineffable.

However, plausibly any fact thinkable by a human can be thought by the human in a specifiably delineated way (the primary fact thought about at t₁, etc.). Thus our thought cannot exceed the possibilities of our language, since for anything we can think we could stipulate that “Xyzzies” means that. (Though, of course, our thought can (and sometimes does) exceed the actualities of our language.) Thus:

2. The humanly ineffable is humanly unthinkable.

Nonetheless, we might make a distinction between two ways of extending human language. A weak extension is one that can be introduced solely in terms of current human language. Stipulations in mathematics are like that: we explain what “continuous” is using prior vocabulary like “limit”. A strong extension is one that requires something extralinguistic, such as ostension to a non-linguistic reality.

3. There are things that are humanly thinkable that are only expressible using a strong extension of human language.

Beyond us

A being that does not represent the world has no conception of what representation might be like, since the being has no conceptions.

A being that lacks consciousness has no conception of what consciousness might be like. The being might have intentionality (our unconscious thoughts, after all, have intentionality), and so might have the contentful thought that there can be beings that have some crucial mental quality that goes beyond the unconscious being’s mentality.

A being that lacks will presumably has no consciousness of what rational will or responsibility might be like. Again, the being might have the concept of beings with “something more” in causation of activity by means of thought.

The distinctions between non-representing and representing, unconscious and conscious, and involuntary and voluntary involve immense qualitative and value gaps. In each of the three cases, we humans exemplify the higher of the two options. At the same time, we are not alone in all these on earth. We share representation with all living things, I suspect. We share consciousness with many animals. But responsibility, I suspect, is ours alone.

I find it implausible to think that we are at the qualitative apex of the space of valuable possibilities. It seems quite likely to me that there could be beings that differ from us in further fundamental valuable qualities in such a way that we are on the lower end, and if we were to meet these beings, we would be unable to grasp what they have which we lack, though we might on testimony, or maybe even empirical observation of behavior, conclude that there is such a thing.

In fact, I suspect there are infinitely many such distinctions, and that God is beyond the higher side of all of them.

In heaven, might we be raised to have the further higher levels? Maybe, but maybe not. However, the mere epistemic possibility of us being gradually raised to acquire infinitely many further such irreducible values is enough to undercut any “argument from boredom” against eternal heavenly life.

Assuming there are infinitely many more such non-V and V pairs, I wonder what this infinity is. Does it have a cardinality?

Open-mindedness and epistemic thresholds

Fix a proposition p, and let T(r) and F(r) be the utilities of assigning credence r to p when p is true and false, respectively. The utilities here might be epistemic or of some other sort, like prudential, overall human, etc. We can call the pair T and F the score for p.

Say that the score T and F is open-minded provided that expected utility calculations based on T and F can never require you to ignore evidence, assuming that evidence is updated on in a Bayesian way. Assuming the technical condition that there is another logically independent event (else it doesn’t make sense to talk about updating on evidence), this turns out to be equivalent to saying that the function G(r) = rT(r) + (1−r)F(r) is convex. The function G(r) represents your expected value for your utility when your credence is r.

If G is a convex function, then it is continuous on the open interval (0,1). This implies that if one of the functions T or F has a discontinuity somewhere in (0,1), then the other function has a discontinuity at the same location. In particular, the points I made in yesterday’s post about the value of knowledge and anti-knowledge carry through for open-minded and not just proper scoring rules, assuming our technical condition.

Moreover, we can quantify this discontinuity. Given open-mindedness and our technical condiiton, if T has a jump of size δ at credence r (e.g., in the sense that the one-sided limits exist and differ by y), then F has a jump of size rδ/(1−r) at the same point. In particular, if r > 1/2, then if T has a jump of a given size at r, F has a larger jump at r.

I think this gives one some reason to deny that there are epistemically important thresholds strictly between 1/2 and 1, such as the threshold between non-belief and belief, or between non-knowledge and knowledge, even if the location of the thresholds depends on the proposition in question. For if there are such thresholds, then now imagine cases of propositions p with the property that it is very important to reach a threshold if p is true while one’s credence matters very little if p is false. In such a case, T will have a larger jump at the threshold than F, and so we will have a violation of open-mindedness.

Here are three examples of such propositions:

• There are objective norms

• God exists

• I am not a Boltzmann brain.

There are two directions to move from here. The first is to conclude that because open-mindedness is so plausible, we should deny that there are epistemically important thresholds. The second is to say that in the case of such special propositions, open-mindedness is not a requirement.

I wondered initially whether a similar argument doesn’t apply in the absence of discontinuities. Could one have T and F be openminded even though T continuously increases a lot faster than F decreases? The answer is positive. For instance the pair T(r) = e^10r and F(r) =  − r is open-minded (though not proper), even though T increases a lot faster than F decreases. (Of course, there are other things to be said against this pair. If that pair is your utility, and you find yourself with credence 1/2, you will increase your expected utility by switching your credence to 1 without any evidence.)

Knowledge and anti-knowledge

Suppose knowledge has a non-infinitesimal value. Now imagine that you continuously gain evidence for some true proposition p, until your evidence is sufficient for knowledge. If you’re rational, your credence will rise continuously with the evidence. But if knowledge has a non-infinitesimal value, your epistemic utility with respect to p will have a discontinuous jump precisely when you attain knowledge. Further, I will assume that the transition to knowledge happens at a credence strictly bigger than 1/2 (that’s obvious) and strictly less than 1 (Descartes will dispute this).

But this leads to an interesting and slightly implausible consequence. Let T(r) be the epistemic utility of assigning evidence-based credence r to p when p is true, and let F(r) be the epistemic utility of assigning evidence-based credence r to p when p is false. Plausibly, T is a strictly increasing function (being more confident in a truth is good) and F is a strictly decreasing function (being more confident in a falsehood is bad). Furthermore, the pair T and F plausibly yields a proper scoring rule: whatever one’s credence, one doesn’t have an expectation that some other credence would be epistemically better.

It is not difficult to see that these constraints imply that if T has a discontinuity at some point 1/2 < r_K < 1, so does F. The discontinuity in F implies that as we become more and more confident in the falsehood p, suddenly we have a discontinuous downward jump in utility. That jump occurs precisely at r_K, namely when we gain what we might call “anti-knowledge”: when one’s evidence for a falsehood becomes so strong that it would constitute knowledge if the proposition were true.

Now, there potentially are some points where we might plausibly think that epistemic utility of having a credence in a falsehood takes a discontinuous downward jump. These points are:

• 1, where we become certain of the falsehood

• r_B, the threshold of belief, where the credence becomes so high that we count as believing the falsehood

• 1/2, where we start to become more confident in the falsehood p than the truth not-p

• 1 − r_B, where we stop believing not-p, and

• 0, where the falsehood p becomes an epistemic possibility.

But presumably r_K is strictly between r_B and 1, and hence r_K is no one of these points. Is it plausible to think that there is a discontinuous downward jump in epistemic utility when we achieve anti-knowledge by crossing the threshold r_K in a falsehood.

I am incline to say not. But that forces me to say that there is no discontinuous upward jump in epistemic utility once we gain knowledge.

On the other hand, one might think that the worst kind of ignorance is when you’re wrong but you think you have knowledge, and that’s kind of like the anti-knowledge point.

Aristotle and Aquinas' Third Way

Aristotle seems to have thought that the earth and the species inhabiting it are eternal. This seems extremely implausible for reasons that should have been available to Aristotle.

It is difficult to wipe out a species, but surely not possible: all it takes is to kill each of the finitely many individuals. Given a species s that cannot have more than n members, and given a long enough time, we would expect there to be a very high probability that all the members of s would have died out during some hour due to random events. Given any finite number of species each with a bound on how many members it can have, and given a long enough time, we would expect with very high probability that all the members would die off.

Now there is a finite limit on how many species there are on earth (as Aristotle knew, the earth is finite), and a finite limit on how many members the species can have (again, the earth is finite). So we should have expected all the species that existed some long amount of time ago to have died out.

The above provides an argument that if the world is eternal, new species can arise. For if new species can’t arise and the world is eternal, then by now there should have been no species left.

How could Aristotle have gotten out of this worry without rejecting his thesis about the eternity of the earth?

One way be to suppose a powerful protector of our ecosystem that would make sure that the species-destroying random events never happen. This protector would either itself have to be sufficiently powerful that it would not be subject to the vicissitudes of chance, or there would have to be an infinite (probably uncountably infinite!) number of such protectors.

Another option would be for Aristotle to reject his thesis that there is only one earth (which was based on theory of gravitation as attraction to the center of the universe: if there were more than one earth they would have both collapsed into the center of the universe by now).

If there were infinitely many earths, then it’s perhaps not so crazy to think that some earth would have lucked out and not had its species die out. Of course, this would not only require Aristotle to reject his thesis that there is only one earth, but also the finitist thesis that there cannot be an infinite number of co-actual things. (Interestingly, given the plausibility that any given species has probability one of dying out given infinite time, and given the countable additivity of probabilities, this way out would require not merely infinitely many earths, but an uncountable infinity of earths. Assuming an Archimedean spacetime for our universe, it would require a multiverse.)

In any case, Aristotle’s commitment to new species not coming into existence (or at least new species of interesting critters; he may be OK with worms coming into existence) is in tension with what he says about the earth’s eternity.

Change and matter

Aristotle’s positing matter is driven by trying to respond to the Parmenidean idea that things can’t come from nothing, and hence we must posit something that persists in change, and that is matter.

But there two senses of “x comes from nothing”:

1. x is uncaused

2. x is not made out of pre-existing materials.

If “x comes from nothing” in the argument means (1), the argument for matter fails. All we need is a pre-existing efficient cause, which need not be the matter of x.

Thus, for the argument to work, “x comes from nothing” must mean (2). But now here is a curious thing. From the middle ages to our time, many Aristotelians are theists, and yet still seem to be pulled by Aristotle’s argument for matter. But if “x comes from nothing” means (2), then theism implies that it is quite possible for something to come from nothing: God can create it ex nihilo.

There are at least two possible responses from a theistic Aristotelian who likes the argument for matter. The first response is that only God can make things come from nothing in sense (2), and hence things caused to exist by finite causes (even if with God’s cooperation) cannot come from nothing in sense (2). But there plainly are such things all around us. So there is matter.

Now, at least one theistic Aristotelian, Aquinas, does explicitly argue that only God can create ex nihilo. But the argument is pretty controversial and depends on heavy-duty metaphysics, about finite and infinite causes. It is not just the assertion of a seemingly obvious Parmenidean “nothing comes from nothing” principle. Thus at least on this response, the argument for matter becomes a lot more controversial. (And, to be honest, I am not convinced by it.)

The second and simpler response is to say that it’s just an empirical fact that there are things in the world that don’t come from nothing in sense (2): oak trees, for example. Thus there in fact is matter. This response is pretty plausible, but can be questioned: one might say that we have continuity of causal powers rather than any matter that survives the generation.

Finally, it’s worth noting that I suspect Aristotle misunderstands the Parmenidean argument, which is actually a very simple reductio ad absurdum:

3. x came into existence.
4. If x came into existence, then x did not exist.
5. So, x did not exist.
6. But non-existence is absurd.

The crucial step here is (6): the Parmenidean thinks the very concept of something not existing is absurd (presumably because of the Parmenidean’s acceptance of a strong truthmaker principle). The argument is very simple: becoming presupposes the truth of some past-tensed non-existence statements, while non-existence statements are always false. Aristotle’s positing matter does nothing to refute this Parmenidean argument. Even if we grant that x’s matter pre-existed, it’s still true that x did not exist, and that’s all Parmenides needs. Likewise, Aristotle’s famous actuality/potentiality distinction doesn’t solve the problem. Even if x was pre-existed by a potentiality for existence, it’s still true that x wasn’t pre-existed by x—that would be a contradiction.

To solve Parmenides’ problem, however, we do not need to posit matter or potentiality or anything like that. We just need to reject the idea that negative existential statements are nonsensical. And Aristotle expressly does reject this idea: he says that a statement is true provided it says of what is that it is or of what is not that it is not. Having done that, Aristotle should take himself as done with Parmenides’ problem of change.

More on the centrality of morality

I think we can imagine a species which have moral agency, but moral agency is a minor part of their flourishing. I assume wolves don’t have moral agency. But now imagine a species of canids that live much like wolves, but every couple of months get to make a very minor moral choice whether to inconvenience the pack in the slightest way—the rest is instinct. It seems to me that these canids are moral agents, but morality is a relatively minor part of their flourishing. The bulk of the flourishing of these canids would be the same as that of ordinary wolves.

Aristotle argued that the fact that rationality is how we differ from other species tells us that rationality is what is central to our flourishing. The above thought experiment shows that the argument is implausible. Our imaginary canids could, in fact, be the only rational species in the universe, and their moral agency or rationality (with Aristotle and Kant, I am inclined to equate the two) is the one thing that makes them different from other canids, but yet what is more important to their flourishing is what they have in common with other canids.

At the same time, it would be easy for an Aristotelian theorist to accommodate my canids. One needs to say that the form of a species defines what is central to the flourishing, and in my canids, unlike in humans, morality is not so central. And one can somehow observe this: rationality just is clearly important to the lives of humans in a way in which it’s not so much these canids.

In this way, I think, the Aristotelian may have a significant advantage over a Kantian. For a Kantian may have to prioritize rationality in all possible species.

In any case, we should not take it as a defining feature of morality that it is central to our flourishing.

One might wonder how this works in a theistic context. For humans, moral wrongdoing is also sin, an offense against a loving infinite Creator. As I’ve described the canids, they may have no concept of God and sin, and so moral wrongdoing isn’t seen as sin by them. Could you have a species which does have a concept of God and sin, but where morality (and hence sin) isn’t central to flourishing? Or does bringing God in automatically elevate morality to a higher plane? Anselm thought so. He might have been right. If so, then the discomfort that one is liable to feel at the idea of a species of moral agents where morality is not very important could be an inchoate grasp of the connection between God and morality.

The overridingness of morality and Double Effect

You’ve been imprisoned in a cell with a torture robot. The cell is locked by a combination lock, and your estimate is that you will be able to open it in a week. If the torture robot is left running, it will stimulate your pain center, causing horrible pain but no lasting damage, and not slowing down your escaping at all. An infallible oracle reveals to you that if you disable the robot, through a random confluence of events this will affect your character in such a way that in a year you will be 0.1% less patient for the rest of your life than you would otherwise be.

Now, sometimes, a small difference in the degree of a virtue could make a big difference. For instance, perhaps, you will one day be in a position where an extremely arduous task will need to be done to save someone’s life, and you just barely have enough patience for it, so that if you were 0.1% less patient, you wouldn’t do it. You ask the oracle whether something like this will happen if you turn off the robot. The oracle replies: “No, it’s just that you will be 0.1% more annoyed whenever you engage in an arduous task, but that’s never going to push you past any significant threshold—you’re not going to blow up in a big way at your child, or neglect a duty, or anything like that.”

It seems obviously reasonable to disable the robot. Thus, enormous short-term hedonic considerations can win out over tiny long-term virtue considerations. It is thus not the case that considerations of virtue always beat hedonic considerations.

What are we to make, then, of the deep insight—perhaps the most important insight in the history of Western philosophy—about the primacy of morality over other considerations?

Two things. First, moral considerations tend to be much more important than non-moral considerations.

Second, we should never do what is morally wrong, no matter what the price for avoiding it, and no matter how small the wrong. But there is a difference between doing what is morally wrong and doing something morally permissible that makes one less virtuous.

Here is a second case. You and an innocent stranger are in the cell. The robot is set to torture the stranger. The oracle now reveals to you that right after the escape, you will forget the last two weeks of your life, and your life will go the same way whether you disabled the robot or not, with exactly one morally relevant exception: if you have chosen to disable the robot, then one day, feeling peckish and having forgotten your wallet, you will culpably steal a candybar from a cornerstore.

It seems obvious that you should disable the robot, despite the fact that doing so leads to your doing a minor moral wrong. The point isn’t that disabling the robot justifies stealing the candybar—at the time that you steal it, you will have forgotten all about the robot, so there is no justification. The point is that even though you should never do wrong that a good might come of it, nonetheless sometimes for the sake of a great good it is permissible to do something that you know will lead to your later doing something impermissible.

Sometimes theologians have incautiously said things like that the smallest sin outweighs the greatest evil that is not a sin. I think this is incorrect. But what is correct is that you shouldn’t commit the smallest sin for the sake of the greatest good. However, the Principle of Double Effect applies to future sins: you can foresee but not intend that if you perform a certain action—turning off the robot, say—you will commit a future sin.

The badness of non-intentional harming

Consider a trolley problem where on both tracks there is exactly one innocent stranger. Alice is driving the trolley. If she does nothing, the trolley will head down the left track. But the right track will get Alice to her destination three minutes sooner. Alice redirects.

It seems that Alice did something wrong. Yet, why? We can say that she intended to save the person on the left track and get to her destination faster, and did not intend to kill the person on the right track. What went wrong?

One option is this. In the proportionality condition on Double Effect, we need that the outcome chosen have a significantly better consequence than the alternative, and three minutes (normally) is not significant.

But that’s probably not right. There are times when it is permissible to redirect a trolley even when the outcome is a bit worse. For instance, suppose that we have a trolley setup with one person on each track, but things are such that if the trolley hits the person on the right track, the death will be a bit more painful. The trolley is controlled by the person on the right track. It seems obvious that the person on the right track is permitted to redirect the trolley to the right even though the outcome is a bit worse.

Maybe the issue is this. Even though it’s not always wrong to become the non-intentional cause of a grave harm to someone, we have moral reason to avoid becoming such a cause. This fits with our intuitions: we feel really bad when we become such a cause. Murray Leinster’s first novel Murder Madness is all about the horror of a drug that makes one involuntarily kill people (I won’t recommend the novel because of a number of pieces of outrageous racism).

This makes sense from an Aristotelian point of view. For a social organism, helping members of the group is a part of flourishing. This is true for animals that are not moral agents. A meerkat sentinel that saves the group by warning of a danger is thereby flourishing. This is even true in the case of non-intentional cooperative activity. A slime mold that, as part of a stalk, enables reproduction by slime molds that are part of the fruiting body is thereby flourishing. It makes sense, thus, to think that for social organisms harming members of the group is contrary to flourishing whether or not one is morally responsible for the harm, and even when the harm is one that one is not intending.

Scientific realism about mass

While I’ve grown up as a scientific realist, and been trained as one as a philosophy graduate student, and I suppose I still identify as one, I’ve been finding it more difficult to say what scientific realism claims.

For instance, what does it mean to be a realist about mass in a Newtonian context? A naive thought is that for each physical object, there is a positive real number, the mass of the object, which mathematically enters into the laws of nature such as F = ma and F = Gm₁m₂/r². But that seems to commit one to there being some odd objective facts, such as to which objects have the property that the square of their masses is less than their mass—a property that barely seems to make any sense, since normally in physics, we don’t compare masses with squares of masses, as they are measured in different units.

A more sophisticated thought is that there is a determinable mass, and a family of determinates, with various mathematical relations between them, with the family isomorphic with the positive real numbers with respect to the relations, but without necessarily a single isomorphism being privileged. But this more sophisticated thought is much more philosophy than physics: physicists hypothesize entities like forces and particles and the like, but not such entities like determinables and determinates. Indeed, this approach commits one to the denial of nominalism, and surely realism about mass in a Newtonian context shouldn’t commit one to such a controversial metaphysical thesis.

Is there some alternative? Maybe, but I don’t know.

Hyperreal worlds

In a number of papers, I argued against using hyperreal-valued probabilities to account for zero probability but nonetheless possible events, such as a randomly thrown dart hitting the exact center of the target, by assigning such phenomena non-zero but infinitesimal probability.

But it is possible to accept all my critiques, and nonetheless hold that there is room for hyperreal-valued probabilities.

Typically, physicists model our world’s physics with a calculus centered on real numbers. Masses are real numbers, wavefunctions are functions whose values are pairs of real numbers (or, equivalently, complex numbers), and so on. This naturally fits with real-valued probabilities, for instance via the Born rule in quantum mechanics.

However, even if our world is modeled by the real numbers, perhaps there could be a world with similar laws to ours, but where hyperreal numbers figure in place of our world’s real ones. If so, then in such a world, we would expect to have hyperreal-valued probabilities. We could, then, say that whether chances are rightly modeled with real-valued probabilities or hyperreal-valued probabilities depends on the laws of nature.

This doesn’t solve the problems with zero probability issues. In fact, in such a world we would expect to have the same issues coming up for the hyperreal probabilities. In that world, a dartboard would have a richer space of possible places for the dart to hit—a space with a coordinate system defined by pairs of hyperreal numbers instead of pairs of real numbers—and the probability of hitting a single point could still be zero. And in our world, the probabilities would still be real numbers. And my published critiques of hyperreal probabilities would not apply, because they are meant to be critiques of the application of such probabilities to our world.

There is, however, a potential critique available, on the basis of causal finitism. Plausibly, our world has an infinite number of future days, but a finite past, so on any day, our world’s past has only finitely many days. The set of future days in our world can be modeled with the natural numbers. An analogous hyperreal-based world would have a set of future days that would be modeled with the hypernatural numbers. But because the hypernatural numbers include infinite numbers, that world would have days that were preceded by infinitely (though hyperfinitely) many days. And that seems to violate causal finitism. More generally, any hyperreal world will either have a future that includes a finite number of days or one that includes days that have infinitely many days prior to them.

If causal finitism is correct, then “hyperreal worlds”, ones similar to ours but where hyperreals figure where in our our world we have reals, must have a finite future, unlike our world. This is an interesting result, that for worlds like ours, having real numbers as coordinates is required in order to have both causal finitism true and yet an infinite future.

Causation and contingency

A correspondent yesterday reminded me of a classic objection to the “inductive” approach to the causal principle that all contingent things have causes in the context of cosmological arguments. As I understand the objection, it goes like this:

1. Granted, we have good reason to think that all the contingent things we observe do have causes. However, all these causes are contingent causes, and so we have equally good inductive support to think that all contingent things have contingent causes. Thus, to extend this reasoning to conclude that the cosmos—the sum total of all contingent things—has a cause is illegitimate, since the cosmos cannot have a contingent cause on pain of circularity.

An initial response is that (1) as it stands appears to rely on a false principle of inductive reasoning:

2. Suppose that all observed Fs are Gs, and that all observed Fs are also Hs. Then we have equally good inductive support for the hypothesis that all Fs are Hs as that all Fs are Gs.

But (2) is false. All observed emeralds are green and all observed emeralds are grue, where an emerald is grue if it is green and observed before 2100 or it is blue and not observed before 2100. It is reasonable to conclude that all emeralds are green but not that they are all grue. Or even more simply, from the facts that all observed electrons are charged and all observed electrons are observed, it is reasonable to conclude that all electrons are charged but not that all electrons are observed.

Nonetheless, this response to (1) does not seem entirely satisfying. The predicate “has a contingent cause” seems to be projectible, i.e., friendly to induction, in a way in which “is grue” or “is observed” are not.

Still, I think there is something more to be said for this response to (1). While “has a contingent cause” is not as obviously non-projectible as “is observed”, it has something in common with it. We are more suspicious of inductive inferences from all observed Fs being Gs to all Fs being Gs when being G includes features that are known prior to these observations to be concommitants of observation. For instance, consider the following variant of the germ theory of disease:

3. All infectious diseases are caused by germs that are at least 500 nm in size.

Until the advent of electron microscopy, all the infectious diseases whose causes were known were indeed caused by germs at least 500 nm in size, as that is the lower limit of what can be seen with visible light. But it would not be very reasonable to have concluded at the time that 500 nm is the lower limit on the size of a disease-causing germ. Now, something similar is happening in the contingent cause case. All observable things are physical. All physical things are contingent. So being contingent is a concommitant of being observed.

Finally, there is another epistemological problem with (1). The fact that some evidence gives as good support for q as for p does not mean that q is as likely to be true as p given the evidence. For the prior probability of q might be lower than that of p. And indeed that is the case in the reasoning in (1). The prior probability that everything contingent has a contingent cause is zero, precisely for the reason stated in (1): it is impossible that everything contingent have a contingent cause! But the prior probability that everything contingent has a cause is not zero.