Psychophysical Harmony and This-Any Fallacy
Source: https://apeinthecoat102771.substack.com/p/psychophysical-harmony-and-this-any · Published: 2026-10-23 · Author: Ape in the Coat (Postphilosophy) · Language: English
Note: Substack essay captured in full; image embeds and subscription widgets removed, text otherwise unchanged. The article responds to Silas Abrahamsen’s defence of the psychophysical harmony argument.
Psychophysical Harmony Argument recently caught traction on PhilStack.
On one side of the ring, we have Silas Abrahamsen who in his recent post argues that it is a strong argument in favor of theism, which works for epiphenomenalists, interactionist dualists, Russellian monists and type-B physicalists alike, and that, despite popular misconception, evolution fails to dissolve it. Other prominent substackers such as Bentham’s Bulldog and Amos Wollen have previously shared the same sentiment, though as of now they refuse to join the recent iteration of the battle.
In the other side of the ring, we have people like Stan Patton, Both Sides Brigade and Eugene Earnshaw who argue the opposite, that Psychophysical Harmony Argument is bad and is obviously dissolved by evolution, for every non-epiphenomenal case. Tyler Seacrest even goes as far as to claim that even epiphenomenalists should not be affected by the argument.
Personally, I’ve always been annoyed by the Psychophysical Harmony Argument. I consider it one of the worst theistic arguments that has ever got significant traction - which is quite an achievement because none of those are particularly good - rivaled in its badness only by Nomological Harmony and Anthropic Arguments. For a long time, I’ve wanted to properly engage with it and multiple epistemological confusions at its core.
But one of the reasons this argument is so annoying is that dealing with it is not unlike fighting a hydra - the moment you cut a head two more take its place. There are just so many alternatives to be explored so many ground to be covered before it becomes clear that there is no more place to bounce off the insights and there is no other way but to accept that the argument is faulty. So, I never quite got to it.
Until now, that is. Thanks to all the posts written by other people arguing against Psychophysical Harmony Argument to whom I can now link, saving myself the whole branches of discussion. And so, in this post we are going to see that for non-epiphenomenalist case evolution does reduce the improbability of psychophysical harmony to the point where it is completely expected. And, of course, we will also learn a very important epistemological lesson, which will hopefully inoculate us from the confusion of similar arguments in the future.
Trains of the Multiverse
While trying to show why evolution allegedly fails to account for psychophysical harmony Silas appeals to a train analogy:
Let’s say that trains that work well get copied in new designs, and trains that don’t work well are scrapped. After a while, it turns out that all the trains that do the best have pink smoke coming out of their chimneys. In this analogy, the trains are of course the physical world, and the replication and scrapping are evolution; while the smoke is the mental.
Then he distinguishes between two questions:
-
Why, given the actual facts about fuel, engines, colors, etc. do the best trains have pink smoke?
-
Why these facts are such to begin with?
And notices that while the first question is answered by evolution, the second isn’t.
If such-and-such fuel and engines (the best ones) had instead happened to produce green smoke, then train-evolution would have run its course and given us trains that produce green smoke. So train-evolution can’t be the answer to our second question.
This is correct. Indeed, evolution only works within the existent laws and therefore can’t explain them. However, to explain psychophysical harmony, neither it has to!
The best way to see it would be to resolve the core epistemological confusion. And to do it, let’s stick with this train analogy for a little while and consider our two alternative universes: the one where the best trains produce pink smoke and the one where they produce green smoke.
Let’s suppose that in the first world people noticed the incredible coincidence that among all possible colors the smoke of best trains is indeed pink. Therefore, they decided to treat it as a strong argument in favor of a Pink God who really likes the color pink.
Meanwhile, in the green smoke world people did the similar observation and made a similar update in favor of a Green God. And the same in every imaginable alternative universe, where laws are such and such that the fastest trains produce the smoke of a particular color.
As we can see, the belief updating algorithm here is to embrace a theistic hypothesis regardless of the color of the smoke. This cannot be rational behavior as it contradicts conservation of expected evidence. Probabilistic updates decrease improbability of theories only insofar as they eradicate some iterations of the experiment, redistributing their probability mass between the rest. The rarity of event on which you are updating has to be directly proportional to the strength of the update. If an event is realized in every trial of the experiment, then the update is zero.
In other words, such belief updating algorithm commits the This-Any Fallacy - what I previously called the worst probability theoretic confusion - it updates with such strength as if avery specificand low probableevent (This) is realized but does so every time a generic and highly probable event (Any) is realized.
Wait a second! But don’t we have prior reasons to think that God would want us to have harmonious relations between mental and physical? We update specifically because we consider the kind of relation we have to be harmonious not because it at least some relation!
And how do we know that we do not consider whatever relations we have to be harmonious in the first place? It’s not that we’ve actually precommited to some specific state of affairs while being a pure reasoner in the void between worlds before being instantiated in a universe and its laws just so happened to satisfy our precommitment. No, us and all our reasoning faculties have evolved in this universe.
Therefore, all that evolution has to do to account for psychophysical harmony is to provide an improbability reduction for why we consider our state of affairs to be harmonious and why a person who evolved in a world with different psychophysical laws would most likely consider their state of affairs harmonious as well.
And while it’s problematic to do so in case of epiphenomenalism, it’s quite easy if there are causal effects from mental to physical.
Sleight of Hand with a Camera
The train example was explicitly epiphenomenal. Silas then tries to argue that psychophysical harmony argument also works for interactive dualism but uses a completely different example.
I have a camera and a TV. On the TV I see a nice forest landscape.
From the TV I see a wire going to a camera, and the camera is pointing at a piece of paper. I cannot see what is on the paper. But I notice that if I push the paper to the right, the image of the screen is also offset to the right; when I tilt the paper, the image on the screen appears to tilt; and so on.
From this I may of course infer that the piece of paper simply contains an image of the same scene I see on the TV, and the image on the TV is caused by the picture.
But here is an alternate hypothesis. It might be that the paper looks like this:
Then it would still be true that the image on the screen is caused by the paper—the camera would simply be inverting the image and displaying it on the screen.
It could also be that the paper is simply blank, except for a single dot. Then the camera is tracking the dot, and is itself providing the rest of the information of the image, yet what we see on the screen is still caused by the paper. Et cetera, et cetera.
Experienced readers should be able to notice their confusion at this point. If the argument truly worked for non-epiphenomenalism, then we wouldn’t need to jump between metaphors. This is something one has to do if the goal is to prompt various different intuitions separately and then use some sleight of hand to arrive to the conclusion that doesn’t actually follow from the premises 1. Let’s see whether you can spot it.
In the camera example, we would probably infer that the image on the paper is like that on the screen, given what we know about cameras and screens. But if the rest of the psychophysical harmony argument is to be trusted, then merely adding in the supposition that our behavior is caused by our mental states does not ensure a priori that we get harmonious pairings.
With the train analogy Silas shows that evolution can’t account for harmony in epiphenomenalist world.
With the camera analogy Silas shows that causality doesn’t itself a priori imply harmonious pairing.
And therefore, he concludes that Psychophysical Harmony Argument works for interactive dualism as well as for epiphenomenalism.
So, did you notice the issue in this reasoning? If not, here is your last chance to solve the mystery yourself.
.
.
.
.
.
.
.
And the answer is…
.
.
.
.
.
.
…that we don’t need the pairings to be a priori harmonious! We only need them to very highly likely to be harmonious conditional on evolution. And even though evolution is dismissed as a possible explanation for epiphenomenalist case, it doesn’t mean that it also fails in the interactive dualist one. And indeed, it doesn’t. The moment mental properties have effect on physical world they also become part of the optimization process of evolution through natural selection.
Eugene Earnshaw wrote a good explanation of how evolution naturally produces harmonious creatures regardless of psychophysical laws. Here is the core insight:
Given any arbitrary set of psychophysical pairings, evolution will construct creatures with harmonious psychophysical relations. That is because, so long as the mental is not epiphenomenal, that is, the mental plays a role in causing behaviour, only harmonious relations between mind and behaviour are adaptive. The set of relations that hold between mental and physical states provides the raw material to work with, and the outcome of the evolutionary process is a harmoniously organized creature. The psychophysical harmony argument imagines changing the relations between matter and mind while holding the output of evolution constant. This is illegitimate, because whatever the relations are, evolution will work with them to create mental events that are caused by states in the world that they are appropriate reactions to.
However, Silas quickly constructed a counterexample:
You imagine that satan scrambles the psychophysical pairings, and then we let evolution play out. Your claim is that in this case evolution might produce creatures that are different on some level—but it’d still make harmonious creatures.
But consider Satan making this simple scramble: He makes it such that harmful stimuli cause pleasure. Pleasure of course causes a desire for more of the thing. However, he also scrambles it such that a desire for something causes you to avoid it. Perhaps he can’t just do that—maybe he would have to make some other adjustments elsewhere to make this coherent, but we can assume he makes the minimal adjustments to make this work.
There are several responses that can be made here. First of all, that at this point we are not talking about pure mapping between brain states and mental states but also about causality between mental states and behaviors, which seem to be outside of the realm of the psychophysical laws as initially stated.
Or that the idea of desire causing avoidance behavior is incoherent, because “desire” is just the name we give to a mental state that that causes us to search more of a corresponding stimulus 2 or because some mental states are inherently motivating. Both Side Brigade’s recent post peruses this direction of the argument.
But still, even if we need to move the goalpost a bit and even if the idea of desire causing avoidance really sounds incoherent, what if we are wrong? There still seem to be more than tiny possibility that desire is just some quale, disentangled from causation by itself and which could be arranged in any sort of cognitive machinery, including avoidance behavior.
Let’s explore this alternative then!
Mind Body Disconnection
If we imagine life in this satanically scrambled universe, what exactly happens when a person touches a hot stove? One possibility that immediately comes to mind, is that the body stops following commands from consciousness. The body executes some automatic harm avoiding reflex, despite the conscious desire, while the person keeps thinking to themself:
“If only I could touch the hot stove more! Too bad that the moment I start desiring something, my consciousness is overridden. Oh, the woe of living in such inharmonious universe!”
This would indeed be something that feels as non-harmony for the inhabitant of such universe. But unless we are assuming epiphenomenalism, such scenario is extremely implausible on evolution.
Consider, what would happen the moment the person regains the control of their body? They would just try to do the maladaptive behavior that gives them pleasure once again. And again. And again. Stuck in the cycle of mind-body conflict for an indefinite amount of time. Which is itself not very adaptive. Something that evolution would’ve selected against.
The only stable equilibrium here is that consciousness is never in control of the actions of the body. But then there is no causal connection from mind and body, which would contradict the premise of interactionism.
Now, technically it’s not impossible to have metaphysically interactionist but functionally epiphenomenalist world, where there is, in principle, a possibility of causation between mental and physical, but evolution just so happens to construct the minds of people without it. Say, first evolution created consciousness for some reason, but then started overriding it more and more until conscious experience became completely disconnected from the behavior, just an atavism like a tailbone.
But such scenario is quite improbable. Not only it would not be compatible with our evolutional history, but it would also be quite inefficient for natural selection to leave the whole cognitive machinery required for conscious experience waste resources without any purpose. With all likelihood it would either be repurposed for something, or it would be disassembled.
Motivation Reversal
The other alternative is that the person consciously chooses not to touch the hot stove because of their desire to do it.
Just like we would think:
“Ouch! That’s painful! I don’t like it at all! Therefore, I’ll avoid this kind of actions in the future.”
A person from the scrambled universe touches a hot stove and think to themself:
“Mmm… that’s pleasant. I like it a lot! Therefore, I’ll avoid this kind of actions in the future.”
Here they supposedly experience the same qualia of pleasure and desire that we do. But they exist in the functional context, which associate them with behavior to be avoided. Just like we exist in a functional context where qualia of pain and fear serve the same role.
Of course, from our perspective, evolved in the non-scrambled universe, this sounds bizarre and quite horrible. We immediately grieve the predicament of the person from the scrambled world who is doomed to be attracted to bad experience s and evade the good once. Immediately we think:
“If only they were motivated to seek what they desire and not vice versa! Oh, the woe of living in such inharmonious universe as theirs!”
But people from such world would not think so about themselves. Why would they? Having such thoughts would be maladaptive in their universe, it would potentially promote dangerous behavior therefore evolution would select against it. For them, whose intuitions and judgements evolved in the world of reversed motivations, their situation would appear coherent and harmonic, while ours - bizarre and backwards. They would grieve our predicament, in fact.
If we could communicate between universes, we may have fascinating ethical and meta-ethical arguments about who is right and whether this is even a question where one can be right and wrong. But that would be beside the point. What’s important for the sake of the question of psychophysical harmony is that both our people would think their own situation to be harmonic.
Therefore, not to commit the This-Any Fallacy, thinking that your situation is harmonic can’t lawfully update you in favor of theism.
The Degree of Harmony
Okay, maybe evolution does significantly reduce the improbability of psychophysical harmony conditionally on non-epiphenomenalism. But still, isn’t theism a better explanation? After all, even though improbable, non-harmonious worlds are still possible on evolution. Meanwhile, on theism they would not be possible at all! So, by the power of Bayes theorem, we should still update a bit in favor of the existence of God, shouldn’t we?
Sorry, but no. Even this consolation prize is beyond the reach of theists. In fact, if we properly account for all the relevant data, theism gets a penalty to its probability. The claim that non-harmonious worlds are impossible on theism is completely wrong and, in fact, such possibility is higher than on naturalistic evolution without any divine interventions.
Previously we were talking about universes disharmonious in a way that, from our perspective, would be unpleasant to live in. A reasonable case can be made, that omnibenevolent God would not create those. Let’s accept it for now. But consider a universe in which psychophysical laws are such that everything gives conscious beings same immense pleasure. Such universe would not be harmonious but would be completely compatible with omnibenevolent God.
One may even think that such universe is more likely on theism, than a harmonic one. After all, the layman conception of Heaven is not a place with perfectly refined psychophysical harmony. It’s a place of immense bliss, period. All this talk of how such harmony is allegedly expected on theism is quite recent. Which sounds very much like an ad hoc rationalization and not an honest prediction of what a tri-omni God would do.
Would you really give more than 50% probability that God would create a harmonious world instead of a heavenly one, if you didn’t already know which world we live in?
But isn’t it a false dilemma? Why can’t it be best of both worlds? Certainly, God can create a perfectly harmonic and perfectly blissful universe! For example, the one where every action is very pleasant, but the pleasure of harmful stimuli isn’t addictive while the pleasure of helpful stimuli is. Indeed, this should be the central example of a universe created by God!
But then the problem is that our universe is not like that at all. Even if we ignore the fact that it’s not heavenly, in order not to double count the Problem of Evil, if you look at the degree of psychophysical harmony that we have, it’s clearly less than divine. Lots of harmful behaviors are pleasant, and lots of helpful behaviors are not. Heroin feels awesome and our taste buds track fail in properly tracking nutritional value. It looks exactly like the kind of harmony that a blind optimization process of inclusive genetic fitness in the ancestral environment is expected to create and not at all like the harmony imposed by an omnibenevolent God.
At this point a theist may want to appeal to a Divine Hiddenness Model according to which God isn’t supposed to be too obvious about being our creator for some complicated reasons. Which is completely unhelpful, because we don’t know whether this model itself is true, therefore, total improbability only increases.
Moreover, such appeal would be an its own goal in a different way. Remember the reasonable case according to which God would not create unpleasantly disharmonious worlds, that we’ve accepted previously? Now it’s time to reevaluate it in the light of Divine Hiddenness!
After all, doing something like this would be a great way to hide better, wouldn’t it? This way God could be more obvious in some other way. Imagine a world in which there is a single religion with several veridical miracles and prophesies, however people have more unpleasant and disharmonic lives. Then after death they go to perfectly blissful and perfectly harmonic heaven. Is it a priori impossible that a tri-omni God would create such a world?
If we apply the same amount of rationalization that people do while constructing theodicies, nothing is impossible. If eternal torture for a majority of humans who have ever lived is compatible with omnibenevolence, then so is the scenario above. Sure, it’s not the most expected outcome, but neither is the universe we observe.
Conclusion
Well, that was quite a ride. But in the end, once again, everything adds up to normality. As soon as we solved an epistemological conclusion at the core of the argument, not only did naturalistic evolution turn out to reduce the improbability of the exact type of psychophysical harmony that we have but even did it better then theism.
However, this is not going to be the last time we engage with Psychophysical Harmony Argument or even with this particular post by Silas Abrahamsen. I’ve promised you multiple epistemological confusions, didn’t I? And we’ve so far dealt with only one. So in the future we will explore the rest of his post and the fascinating epistemological mistake that the proponents of the Psychophysical Harmony Argument need to exploit to try to sell it to type B-physicalists.
I’m not accusing Silas of bad faith arguing here, I’m pretty sure he is honestly confused on the matter himself.
Notice that this is a purely semantical issue. We don’t need to subscribe to any particular ontological theory to define “desire” this way.