I'm not exactly a philosopher but I am some kind of materialist atheist. But it seems to me there's a lot of missing the point going on here. I think the cleanest answer is that some concepts exist in a different way than material things, whatever the dominant philosophy today is I hardly think denies concepts such as number or color. For instance, I don't think Mary's room means anything. Obviously, studying everything about redness without seeing it is nonsensical. (Also we could attack the concept of redness which is said to exist here, if you have only seen one shade of red you whole life you've experienced some kind of redness, but there is a lot of red stuff you aren't experiencing.) Also, the properties of water would seem to me to be laws of motion/physics here, not things. And the laws say the things change properties when joined.
Also, in the start of your essay with saying mehanics are still trying to figure stuff out, I fail to see the point of the agitation to resist it (of course the point is religion).
There is really no point to this, but I noted my disagreement I guess.
To be sure, neither am I. I submitted one paper to a journal that I know was sympathetic with the viewpoint and it was summarily rejected. However, both my parents are professional philosophy professors, and I minored in it in college, so I rate myself an excitable amateur.
> "but I am some kind of materialist atheist."
I admit, I do have misgivings about continuing to publish these letters here, since I know a chunk of the De Civ subscriber base shares your view, and committed materialist atheists are explicitly *not* the target audience for these letters.
The primary audience is my children, who have no investment in materialist atheism except what they've picked up unconsciously from the culture. The secondary audience is other parents who already believe basically the things I believe and who are trying to figure out how to pass it on to their kids in a culture where the Nones are winning bigly on both conversion and retention. (SOURCE: "Religious affiliation in youth and currently" here: https://www.pillarcatholic.com/p/the-faith-of-the-next-generation) It's that secondary audience that I'm publishing for, really.
But this leaves you in a very annoying position: this letter is kind of *about* you, but it is not designed to *persuade* you. (This letter would be three times longer and begin from a very different starting point if it were designed to persuade someone with your commitments.)
I don't know that there's anything I can do about this except acknowledge that I get where you're coming from, I get your desire to pipe up in defense of your position, and I get it if you're annoyed about this series of letters! (I'll try to balance it out, as always, with other topics.)
> "I think the cleanest answer is that some concepts exist in a different way than material things"
Then you aren't a materialist!
One of the key jobs of ontology is to characterize what exactly that "different way" is.
Materialists deny that there is a "different way," because the materialist project is to reduce *all* concepts, including number and color, to strictly and exclusively material terms. Of course, there are different approaches to materialism (reductive materialism, eliminative materialism, functionalism, epiphenomenalism, and so on), but they all share this goal. If you do not, then you aren't a materialist!
(Of course, there are plenty of non-materialist atheists, so you are in fine company with them.)
> "For instance, I don't think Mary's room means anything. Obviously, studying everything about redness without seeing it is nonsensical. (Also we could attack the concept of redness which is said to exist here...)"
The purpose of Mary's Room is not to defend any specific notion of redness, but simply to demonstrate that the subjective experience of sense data ("qualia," in the jargon) is not reducible to objective physical facts.
I agree that this is obvious! To say that one knows a color without access to its qualia is, as you say, nonsensical! But the dominant philosophy today is quite determined to either deny qualia outright or completely explain qualia in terms of objective physical facts. (Neither approach is, in my opinion, at all promising.)
> "Also, the properties of water would seem to me to be laws of motion/physics here, not things. And the laws say the things change properties when joined."
I think this is the most defensible position you stake out, and I wish that space had permitted me a longer examination of properties, but I will leave it at this: if you build out the "laws of motion" in this way, you end up with something very close to interchangeable with Scholastic ontology (with its "powers" and "potencies"), but with bigger explanatory problems.
"Also, in the start of your essay with saying mehanics are still trying to figure stuff out, I fail to see the point of the agitation to resist it."
The point is that the Scholastics already had this same "stuff" figured out. We had a (reasonably) complete, (reasonably) successful ontology, fully compatible with modern scientific methods and discoveries... and we threw it in the trash bin for what turned out to be historically contingent, not-very-good reasons. Not only did we bin it; we replaced it with a much worse theory with much bigger problems that seem intractable on their face. The anti-Scholastic temper tantrum Descartes and his lot threw at the dawn of Modern Philosophy is like if scientists got frustrated looking for the Higgs Boson in 2012, threw out the entire Standard Model, and reverted to Luminiferous Aether theory.
If the Mechanics do, in fact, eventually (miraculously) plug the most gaping holes in their ontology, we should definitely give them a second look! In the meantime, we should go with more successful ontological theories, such as Scotism, Eleatic Monism, and (my obvious fave) the Aristotelian-Thomist model.
" (of course the point is religion)"
You're certainly correct that religion is where I'm ultimately driving this car. However, my motive for swerving away from the Mechanical Philosophy is that it's not a working ontological theory, but rather a very old I.O.U. for maybe someday becoming one, and that I.O.U. is never getting redeemed. If I lost my religion (for example, because the Problem of Evil overwhelms me, or because Jesus's corpse is discovered in Jerusalem), I still would not become a Mechanic.
"There is really no point to this, but I noted my disagreement I guess."
I disagree that there was no point to it! If we never talk with one another about the Big Questions, what's the point of talking about the little ones?
I guess I would disagree that the dominant philosophy in the culture is materialism in your understanding. (However, I'm not American (I'm from Croatia!) so take this with a truckload of salt.) I think your not-mattering problem just stems from a more political philosophy of religious liberty and debating religion being really annoying to most people (don't underestimate the second one). Again, I'm not exactly well-versed in philosophy, my version of materialism is mostly related to not believing in god. And even there, I would probably limit myself to not believing in a god that wants us to follow his specific rules and punish us. (I'm not sure if you would consider that atheism, but I would say that's what most atheism is.) I guess you could probably try to shove math into the laws of motion somehow. Something I wanted to say in the previous post is, depending on how different individual brains are, you probably could in the future simulate experiences (including redness!). Also colors are social constructs at least there are many colors we consider red and so on. And the final point is, there is a difference between explaining something and explaining something well (and/or correctly), some things are just hard to explain and we should recognize that.
As a chemical engineer, I think the problem with your discussion of water vs hydrogen and oxygen is that atoms are not fundamental particles. They are composed of quarks and electrons. So when hydrogen and oxygen react to form water, it's just a rearrangement of the components of the atoms. There's nothing metaphysical about it, and the widespread success of quantum mechanics in predicting the properties of atoms and molecules, to a very high degree of accuracy, is in my view extremely strong evidence of this.
Also worth noting:
Hydrogen is not inherently flammable, as evidenced by the fact that the Sun is not on fire. Hydrogen is flammable in Earth's atmosphere because it reacts exothermically with oxygen.
Water is in fact quite a reactive molecule, although the surface of the Earth has been so saturated with water for billions of years that there's not much, if anything, left in nature that water reacts with strongly enough to generate flames.
There are plenty of manmade substances that burn with water, though. An increasingly common example is lithium metal, which can form inside of a malfunctioning lithium ion battery. Don't ever pour water on a lithium battery fire!
The boiling points that you give are for molecular hydrogen and oxygen, H2 and O2, and are not inherent properties of the atoms. An interesting illustration of this fact is that ozone, O3, is also composed only of oxygen atoms, but has a boiling point over 100 F higher than that of O2.
Much of this is very interesting! I am pleased to have these additional facts on hand from a professional chemical engineer!
In terms of the point my letter was trying to make, though, I think this comment misses it.
Whatever you set as the level of "fundamental particles," those particles have actual properties. When they interact with other fundamental particles, some or all of their actual properties are suppressed -- that is, reduced to potentials -- and the combined substance has properties of its own that are not shared by its constituents. The properties of water superseding the properties of its constituents is a great example of this for teens, both because many people intuitively think of atoms as fundamental particles, and because the properties of (molecular) hydrogen, (molecular) oxygen, and water, are all common knowledge, easily grasped by anyone past 6th grade.
This supercession cannot be explained in strictly material terms -- at least not without falling (eventually) into (a) Eleatic monism, which denies all change, (b) a really unattractive Mechanical model that treats all these behaviors as brute laws of nature, or (c) reconstructing something that looks an awful lot like Aristotelian-Scholastic metaphysics. The widespread success of quantum mechanics, far from refuting this framework, implicitly *depends* on it.
NOTE: while writing this, I discovered that Oderberg's book is freely available on the Internet Archive, and he does discuss the "quarks objection", so perhaps he will do a better, clearer, or at least more correct job of making the point than me (pp62-76; water shows up on p75; quarks have their big moment on p64 and then scattered throughout): https://archive.org/details/david-s-oderberg-real-essentialism_202305/page/62/mode/1up
It is obvious why I won't give this reading to a 13-year-old, but you're 35, so you don't need me interpreting it when you can just go to the source. :)
Thanks for the link. Unfortunately I don't agree with his arguments against atomism, on pg 64 where he starts talking about quarks. (Again with inaccurate physics, by the way, because quark flavor transitions are mediated by the weak force, not the strong force.)
His first argument, that no one has ever proven that all of the multiplicity and diversity in the universe is a result of mere recombination, is increasingly outdated. In the 24 years since his book was published, the capacity of computers to run atomistic simulations of molecules using quantum mechanics has dramatically increased, to the point that it is now common for these simulations to complement laboratory experiments not only in academia but in industry. Ruthless capitalists would not invest in the supercomputing resources necessary to run these simulations unless they worked.
Perhaps he would counter that unless we can simulate the entire universe atomistically, we cannot definitively conclude that atomism is an appropriate philosophical understanding of physical matter. However, in my view, if atomism works on the scale of a molecule, or of dozens or hundreds of molecules, this is at least sufficient to rebut his proposition that metaphysics is inherently necessary to understand the behavior of a multicomponent physical system.
His second argument is that the ability of quarks to undergo flavor change (i.e., up to down, charm to strange, etc.) is evidence of potentiality in fundamental particles. I'm certainly not an expert on quantum field theory, but my understanding is that the fields are actually fundamental, and individual particles are just excitations of their corresponding fields. https://en.m.wikipedia.org/wiki/Quantum_field_theory
For example: There is a Higgs field, and a corresponding Higgs boson which is an excitation of that field. There is an electron field, and a corresponding electron particle which is an excitation of that field. In this conceptualization of quantum physics, there are not an uncountable number of unconnected quarks constantly transforming into other kinds of quarks, but rather a handful of quantum fields permeating all of time and space which are continually interacting and exchanging energy with each other at an uncountable number of locations.
To me, energy is the closest thing in physics to this metaphysical concept of potentiality. In thermodynamics, for example, energy is considered to be the capacity to change a physical system from one state to another. And when we really drill down on "what energy is", it's admittedly somewhat mysterious: https://readfeynman.blogspot.com/2017/04/section-41-what-is-energy.html
If we know one thing about energy, though, it's that it is physical. It's governed by mechanical laws and conservation laws and famously was shown to be interchangeable with matter by Einstein (and then later by the atom bombs). So it seems to me that there is no need for metaphysical concepts like potentiality anymore, to answer the question "How do physical systems change from one state to another?", and really no room for them either.
I'm not sure you're as far apart from Oderberg as you think you are. I know it was a lot, but did you read to the end, where Oderberg considers (and agrees with) the possibility that what hylemorphists call "prime matter" might very well be identical with what physicists call "energy"?
In any case, your account of the current best scientific theories about elementary entites has moved some of the goalposts from your previous account (previously, you put forward quarks as fundamental; now you make fields out to be fundamental). That's actually not a problem for me, because my whole point here is that it *does not matter* what entity or entities one identifies as fundamental, because, whatever class of entities you select, you cannot account for the visible universe without developing an account of potency, actuality, causation, and, ultimately, substance.
Your revised account confirms this rather wonderfully! You describe the Higgs boson as an excitation of the Higgs field. Unless I've entirely misunderstood you, that, right there, shows that your model has both potency (the Higgs field encompassing all of time and space has the potential to excite itself into a Higgs boson at any given point) and actuality (since some points ARE actually Higgs bosons). Eleatic monism is avoided insofar as the Higgs field's potential is not *necessarily* actualized at each and every point, so reality is not composed *solely* of Higgs bosons.
In other words (to borrow Oderberg's language on p64), your model, as I understand it from your description, still includes substantial transformations at the most fundamental level, and pure potency therefore is still needed. Far from showing that physics has "no room" for metaphysical concepts like potency anymore, you seem to confirm that physics *depends* on those concepts. (It just insists very strongly, in your model, on not calling them "metaphysical" or making them direct objects of study. But this prejudice seems to me philosophical, not scientific, arising from the sources I briefly mentioned in the OP.)
Physics, then, can certainly help us characterize states such as: the Higgs potency, the Higgs actuality, and the material, efficient, and perhaps final causes that precipitate transitions between those states. Physics may also tell us eventually that the Higgs field (and other fields) are *not* fundamental, forcing us to look elsewhere for the most basic expressions of potency, actuality, and causation. However, what Physics *can't* help us do is deny that potentiality, actuality, and cause *exist*. (I think Physics *might* offer some insights into whether substances are real or not, perhaps pushing us toward some species of monism instead, but I'm not sure and it doesn't seem relevant to your argument, so I will leave that question bracketed.)
Fwiw, I don't find Oderberg's much-too-brief treatment of multiplicity and diversity in the universe persuasive, either. It could be sketched out into something beefier, and perhaps since this is a technical book written to fellow professionals he assumes we are familiar with that work, but he doesn't give that argument any legs here. So we agree there, and I suspect even Oderberg would if you asked him.
To repeat my thesis in a way that sounds like a good conclusion: I don't think you're as far from Oderberg as you think you are.
Well the Higgs actuality and Higgs potency, as you have defined them, are physical phenomena. They are quantitative properties of the Higgs field which can be calculated based on physical observations. If that's metaphysics, then what do you consider to be physics?
If you don't mind me answering a question with a question: aren't those quantitative properties of the Higgs field calculated with and described by mathematics? If that's math, what do you consider to be physics?
I don't see why the actualization of the Higgs boson* cannot be simultaneously an ontological process, a physical process, and a mathematical process.
*NOTE: I am not committing myself to this physical model of the universe; I'm accepting your description at face value.
As a Catholic mechanist, I suppose it’s on me to defend the obviously best way of thinking about the physical world? (I’m not really a mechanist, more of an occasionalist who thinks mechanism is the best way of gaining a useful comprehension of the world, but close enough, I suspect).
This one doesn’t really work, although I think it’s reasonably close to doing so. Two possible approaches would be:
-Convert it from a criticism of mechanism to a criticism of materialism. The advantage of this is that your three supposed critiques of mechanism fail as such, but succeed (mostly) as critiques of materialism. You would have to be less performatively mad at Descartes, and I understand that this is a pretty-close-to-terminal-value for many people, so maybe this isn’t the best option.
-Acknowledge the incredible successes of mechanism in explaining the physical world, but critique our culture’s totalitarian attitude toward it; i.e., criticize the viewpoint that mechanistic explanations are somehow in tension with theistic ones (currently, I think your essay actually implies agreement with this Culturally Prominent Bad Opinion), and point out that non-material things are not made of parts (or the parts are inaccessible to us), and that the appropriate response to our best tool not functioning in that context is probably not to pretend that such things don’t exist.
I’ll address your three specific critiques of mechanism in reverse order.
#3: Minds
There is one, key, central fact about models—and by “model” I just mean something like “proposed explanation for some facet of reality”, not anything necessarily scientific or quantitative. That fact is that there is a tradeoff between prediction and causation: models can be designed to predict, or they can be designed to show causes. They can even do both, but they won’t do both well. This tradeoff makes us (both mechanists and scholastics) very uncomfortable, so we don’t talk about it enough—certainly the mechanists have plenty of blame on our shoulders now that we are in the second century of pretending that physical models involving fields are somehow causal. For one, the tradeoff presumably doesn’t exist in the mind of God, so it isn’t necessarily a fundamental feature of reality in the grandest sense, but it is certainly the case that at the level of human reason, we are stuck with this tradeoff. This is not a new idea: Bellarmine’s responses to Galileo are based on this tradeoff, and Galileo comes off like a philosophical dummy because he ignores it. But advances in physics and chemistry in the last 150 years or so have heightened the tension, and what really smacks us in the face these days are big data models, neural nets and so forth that really explore the far edges of the tradeoff, being astoundingly predictive and completely useless in sorting out causation.
I think our responses to this tradeoff are representative of scholasticism and mechanism more broadly: you want to find explanations that reveal the causes of things; I want to find explanations that are predictive. You have very little patience with predictive explanations that are not causal, and I vice versa.
I don’t think this is just a matter of taste, though (although it probably is to some degree). Predictive explanations have a number of useful properties, such that I argue we should prefer them a priori:
-Prediction, by forcing your theory frequently to interact with the real world, has a grounding effect that causal explanations do not. This effect becomes more valuable the more biased and confused you believe human reasoning is, and I am extremely cynical about human reason.
-Predictive models can function in environments in which causes are fundamentally incomprehensible to the human intellect. As members of a religion that holds as dogma that various aspects of reality are fundamentally mysterious, this should be appealing to both of us—more appealing the more things you think are truly incomprehensible.
-A weird benefit of predictive models that is really only useful in this exact context is that they are sufficient in themselves to demonstrate mechanism, and your example of the mind is an excellent one.
Thought experiment: say that we get really good at imaging the brain at a sub-cellular level, tracking neurotransmitters, etc. (You seem a bit grumpy that this hasn’t been done already by Leibnitz. Our understanding of how neurons operate in any detail at all is only decades old, and there are 10^11 neurons, so I don’t really understand you on this point, but … sorry, I guess? But let’s say we crack it in the future some time.)
This lets us describe the brain in terms of networks, firing cycles, whatever. Let’s posit that for every quale, we are able to describe an associated network of neurons with a particular structure and set of firing patterns. Let’s further posit that we can manipulate the brain such that we observe:
-Whenever the firing pattern appears, people report experience of the quale.
-The firing pattern does not appear at times when people report no experience of the quale.
-Causing the firing pattern to appear produces experience of the quale even in unusual circumstances.
-Preventing the firing pattern from occurring prevents experience of the quale even in circumstances in which it would normally be experienced.
These findings—assuming they are sufficiently universal, etc—would:
1) Demonstrate the mechanistic character of qualia, i.e., that they can be explained (in the predictive sense of ‘explanation’) in terms of the behavior of their parts, in this case, the neurons of the brain.
2) Not provide ANY PROGRESS WHATSOEVER on the hard problem of consciousness, or any support for materialism, for that matter. Since we wouldn’t have any understanding of what causes qualia, it might just be that God tells your guardian angel to be on the lookout for certain brain patterns, and then give an immaterial poke to your immaterial soul to produce the associated qualia. This wouldn’t be a materialist explanation, but it would be a mechanistic one.
I think that your current attitude toward the mind is an extraordinarily dangerous one, for your daughter. I think it is extremely possible that during her lifetime we will make an enormous amount of progress in putting qualia on a mechanistic footing. It won’t explain what causes them, or do anything to harm non-materialist accounts of consciousness, but we can demonstrate mechanism without doing that. I don’t see the benefit in tying the rest of your apologetics to an unnecessary claim that is likely to be demonstrated false, and risks delegitimizing everything else you’ve said here. You don’t need to stake this claim AT ALL. Why gamble with the soul of your daughter?
#2: Water
I don’t understand why you think any of the claims you’ve made represent a problem for mechanism, so I won’t engage for now (and David seems to be anyway). If you want to explain your thoughts, I’m happy to respond!
#1: Circles
Again, this seems like something that is aimed at materialism, not mechanism. Mechanists have no problem invoking mathematical objects to describe the behavior of the world, and I’m not sure why you think that they do—is there a mechanist in the modern lineage who thinks that the components of the world are actually inert billiard balls? I’m happy to read up on him, if so, but I am skeptical he exists. Based on some of your replies to comments here, I get the impression that you want to say something like “aha, you have said that electrons have certain properties, but isn’t that just the scholastic ideas about properties and natures being smuggled in!?”
To which a mechanist would reply: Sure? Whatever? We aren’t approaching the world as if everything is lacking in properties. We’re trying to treat those properties responsibly, by which I mean that the properties invoked are mathematically describable and resulting from the interactions of the parts of the thing in question. And if there are things with “atomic” properties, we’d like to push that to as low a level as possible. But if we can do that, we mechanists are very comfortable having you jump in at the last minute to gloss all the stuff we’ve learned about physics and how it shows how Aristotle was right all along about action and potency or whatever. We are not insecure about this, because we remember which of the two approaches to properties results in nuclear reactors—and discovers entirely new realms of reality to explore and explain.
I realized that, if I sat down and just started typing, I was going to turn it into a fisking (not an unfriendly one, but lots of paragraph-by-paragraph quotes). I decided that it would be wiser for me to sit with your comment for a while, think about my reply, and finally condense it down to its essentials.
Unfortunately, I let it sit too long, and now I've completely forgotten many of my early thoughts. I suppose that's still a form of condensation.
First, rereading this letter, I agree it is rough and needs revision. I lost my own past train of thought once or twice on the reread, and had to go back to remind myself what I was on about. This was perhaps the most ambitious letter in the series (so far), and its aim outreached its grasp at times. The whole framing of your comment highlights this, because...
(Second,) This letter is *intended* as an critique of materialism. I intentionally defined "Mechanic" very narrowly here: a Mechanic (for the purposes of this post) is someone who believes that material corpuscles (and, *maybe*, their positions) are the only things that exist, period. If I weren't writing this to an adolescent, I might have avoided some confusion by using more common labels: physicalism, reductionism, eliminative materialism, with maybe a side order of skeptical empiricism in cases of doubt.
This narrow focus made me do something very painful, which you may have noticed: I let Descartes off the hook! I think his philosophy was mistaken, that it *led* to what I here termed the "Mechanical Philosophy," and, yes, you're right, performative outrage about that IS something like a terminal value! But Descartes himself was not, by any stretch of the imagination, a Mechanic within the meaning I have given the term here.
But that's all according to plan. My sole goal in this article is to persuade my daughter that the materialist model she will have absorbed through our culture (and which I found *extremely* convincing at her age) cannot possibly be correct, that there must be *some* immaterial things. This is pure bootstrapping for the next letter(s), where I will argue for some form of the First Cause and make the case that it cannot be material. I find that the way people receive those arguments depends a great deal on their priors. (At least, it made a huge difference to me when I was 14, read Russell's "Why I Am Not A Christian" in the STA library at study hall, and thought "turtles all the way down" was a brilliant refutation of Aquinas!)
As such, I think the *key* criticism you make has already been incorporated into the article, but needs to be drawn out to be much clearer in the next revision. (This is good. The reason I post these online is to gather useful feedback for the revision before the kids read them.)
But, as you identify as something-like-an-occasionalist who has no problem with certain immaterial things, this letter is, quite simply, not aimed at you, and I obviously agree with your criticism that some of my arguments do not work against (It may still have struck you with friendly fire.)
The most important part of your comment for my revision, I think, will be your notes on what I had to say about the mind. I think you are likely correct that we will, within the lifespan of human civilization and quite possibly within our lifetimes, develop the kind of neuron-to-quale mapping you describe, and I don't think anything makes it in principle impossible. It is very much *not* my intention to hang my whole argument on the certainty of that *not* happening, and I tried rather hard to avoid that specific trap at the time. Clearly, I failed! I will have to try again, and I will keep a close eye on your notes as I do so.
Three questions for you, if you're not averse to answering questions nine months later:
(1) What is the definition of "mechanist" you use in describing yourself as one? Like so many terms in philosophy, "mechanist" is overloaded. I could probably discern your definition pretty well from what you said in your comment, but it would be better to let you define your own terms.
(2) How'd you get there? Without bearing any ill will toward your metaphysics, I confess I was surprised to hear you use that label, and I'm sure you followed an interesting path to embracing it.
(3) This tradeoff between causation and prediction is a really interesting idea that has stuck in my craw without my knowing quite what to do with it. At the risk of revealing myself as out-of-touch with the literature, this comment is the first time I'd heard of the idea. Can you suggest any further reading on it? I'm not sure it will have much bearing on my revision of this letter, but I think it could be a useful way of thinking about fault line between the Moderns and the Medievals.
(Okay, that was still kinda long, but not nearly as long as it probably would've been if I'd replied right away.)
Ah, Marvin Minksy, who has variously been quoted as saying that human beings are "machines made of meat" and that our brains are "computers made of meat." I wouldn't want to go out to lunch with this guy.
But my main initial reaction to this post was "Oh no, not Star Trek!" IMHO it is far from the worst offender. It would seem that objectively and throughout the universe, racism is wrong (even if it depends on which half of your face is black), even very powerful alien beings are nothing like God ("We find the one quite sufficient"), and it is wrong to disturb the time line so that Nazis rule the world, even if the alternative makes mincemeat of your emotions. I speak, of course, of Star Trek TOS. When Star Trek: Discovery was being filmed, I read an account of one scene in which the actor playing the captain ad-libbed a comment like "Oh my God!" or "Goddamnit!" and the director sternly told him to redo the scene without that, because in the future nobody will believe in God. Saved me some wasted hours trying out the new show.
My intro to philosophy prof (one Dr. Greg Coulter) pulled a wonderful trick on us in what was, for most of us, our first month of undergrad.
He started off hammering home for us the rules of logic and the structure of argument. That was three classes, I think, and several handouts saying "HERE IS HOW TO IDENTIFY AN ARGUMENT." Then he assigned various readings. We had some Aquinas, some Anselm, some Berkeley, some Descartes. I mean, it's a survey course, so it ran the gamut. In each reading, our assignment was to find the argument, write it down, and then offer some comments on it. I was an undergrad, so I'm sure I said several totally insane things. (I think I agreed that Anselm's ontological argument worked *only* in terms of extension or something, like some kind of insane fun-house version of Gaunilo.)
So then, after we've done three or four of these, Dr. Coulter assigns Marvin Minsky's "Minds Are Simply What Brains Do" with the usual assignment: find the argument, write it down, offer comments.
When we all came into class the next session, haggard and confused, he asked who found the argument. Some of us sheepishly offered up what he'd written on our homework and he shot them down one by one. Then, finally, we had no answers left, and he revealed that this was correct. The reason we'd had a hard time finding an argument in there was because there wasn't one. It was just a series of assertions, which made no attempt to proceed to a conclusion from logic. It paints a picture; it doesn't prove a case. "And yet," he commented (I paraphrase), "often compiled into introductory philosophy readers1"
That always stuck with me. People can sound very convincing. But look for the argument.
(Also, write the argument.)
You are, of course, absolutely correct about Star Trek. My kids know that it is my favorite show of all time by far and that I am deeply obsessed with it. (I have always been very active in certain parts of the fandom.) So if I am willing to admit it has bad metaphysics, they won't take my comments about My Little Pony as an attack.
You are also correct that Discovery was terrible, although somehow not nearly as bad as Picard. That complete misunderstanding of how God figured into the Roddenberry universe was only the first sign of the madness that had taken hold on set. (Did you hear how they hired the great Black American mystery novelist Walter Mosley for Season 2 and then he wasn't woke enough so they fired him?!) (And Discovery was SO MUCH BETTER than Picard! GAH!)
Star Trek Prodigy, however, is quite good. It's aimed at kids and I am watching it with my kids, but I would watch it without them, because it's a fun, pure, joyful space adventure, the likes of which you don't really see much of on TV anymore.
I have always loved the "We find the one quite sufficient" line, because it canonizes Kirk's monotheism (and, almost certainly given Bread and Circuses, his Christianity) no matter how much Roddenberry et. al. tried to make theism untenable in future shows. (And then DS9 snuck a space pope in through the back door anyway.)
1. I would say mechanism (sensu-I-am-a-mechanist) is the idea that the explanations for complex things are to be found in the interactions of their parts. In terms of the opinions you ascribe to mechanists:
-The world is a clockwork machine: ✅
-It is made up of colorless, odorless particles bouncing off one another aimlessly, governed by just a few laws of motion: ✅
-It is made up of featureless particles: I don’t know what this means
-There really is no such thing as a brick or a window: ✅
-There really is no such thing as your body: possible not; I’m prepared to believe God keeps a “ledger” of which particles are assigned to your soul.
-We wouldn’t need to reason about bricks and windows: I am with Douglas Hofstadter that limitations of human reason mean that it is usually only practical to do one level of mechanical explanation. So bricks and windows are very useful approximations for us in understanding the world. But, s/we/angels/ and ✅
-We wouldn’t need to reason about humans: humans have a soul, which might make this more complicated.
-Particles are the only things that could be known and understood: ❌
-There’s not really any such thing as a tree or a brick or the color red: ✅
-Anything that isn’t made of particles can’t really exist: ❌
-Anything that isn’t made of particles can’t really be known: ❌
-Any free will we actually experience can only be an illusion: ❌
2. Well, my journey isn’t interesting from the inside, at least. I was exposed to the scholastic worldview in college, and it was immediately evident that no one had bothered integrating the last 200 years of advances in biology into it, and that if you did, the results would be farcical. Given that this was the only field that I knew anything about, and that I’ve never been one for Gell-Mann amnesia, I didn’t really see the point in exploring it further. Nothing (including spending much of the time since in Catholic academia) has made me think I ought to reconsider that; as we’ve discussed above, I don’t find the arguments against mechanism to be at all compelling, and when the other camp’s model fits the world I know about so badly, I've never felt either a push nor a pull away from mechanism.
3. No, I can’t give you any readings. This is a view gained from the trenches, as it were. Maybe it would be useful to give a brief history of models in biology? That might be helpful in finding readings.
-The original model is the Linear Model, y=ax + b, where x is your predictor, y is the thing you are explaining, and a and b are parameters. This is the classic causal model: you do your experiment, measure x and y, and calculate values for the parameters. If a > 0, then x causes y (assuming you’ve designed your experiment properly). Once you have an estimate for a and b, you can then use measurements of x to predict y, or vice versa. So the same final equation, y=0.7x+5.3 (or whatever) is both causal and predictive.
-Of course, the problem is that in the real world, nothing is ever really caused by one thing. Things have lots of causes. So you’ll want to add more predictors to your equation. You might try y=ax1 + bx2 + cx3 + d. The problem is that you can’t actually stop there, because the world is lousy with interactions between variables. A classic example might be: having cancer is bad. Having radiation poisoning is bad. But having radiation poisoning while you have cancer isn’t double-bad, it’s just chemo, and it’s actually kinda good. So you need to account for these interactions in your model, which is now: y=ax1 + bx2 + cx3 + dx1x2 + ex1x3 + fx2x3 + gx1x2x3 + h. Every predictor you add doubles the number of parameters you need to estimate, which creates a number of problems:
A seven-predictor model requires 128 parameters to be estimated, which means you need 64 times the sample size compared to a single-predictor, two-variable model. This is a practical problem.
Even if you have a big enough sample size, there are other practical problems with the parameter estimation—for instance, the methods here are very sensitive to predictor variables being correlated with each other, and go haywire when they are. Alas, it turns out that things are often correlated with each other in reality.
Even if you solve those problems, and correctly figure out the parameter associated with the x1x2x4x7 interaction term, what does that parameter mean? You could write out a sentence that describes it, sure, but it will be hard to understand that meaning, and you will have dozens of these to account for simultaneously. You won’t be able to do this. Human brains just don’t think in more than three dimensions, sorry. So we’ve lost the nice causal meanings that simple, non-interaction-bearing models have.
For these reasons, the Best Practices for linear modelers is to curate your set of predictor variables heavily; keep them few, don’t add predictors unless you have some expectation that they are causal, and ensure that they are uncorrelated. Don’t include any interaction terms in your model, ever, except for possibly one or two which you have a strong a priori reason to believe might be important. If you follow the advice, then you get a beautiful model from which causation is easily inferable from its parameter values. But because you have (intentionally) ignored vast swaths of reality, this will limit the predictive power of your model.
Now, there is a very pleasing fact about these linear models, which is that they were very susceptible to overfitting. I won’t really explain this, because I don’t understand it. But if you ignore the Best Practices, if you add in tons of predictors and interaction terms, even if you have a large enough sample size to calculate your parameters accurately, overfitting meant that your predictive powers didn’t actually improve, even as your were wrecking your ability to draw causal conclusions from your model. So for … I dunno, half a century? … when linear modelling was the only game in town, there was no tradeoff, in practice at least, between causal explanations and predictions.
The next important type of model is the family of Principal Components Analysis (PCA) and relatives. By the time I was in graduate school in the 2010’s, PCA was already well-established. With PCA, you throw away the linear modelling best practices. Instead of curating your predictor variables, your gather every possible predictor you can think of. You don’t worry about predictors being correlated with each other, you don’t worry about them being useful, you just throw in everything you got. Then, you make a bunch of made-up variables that are combinations of your real variables, and identify which of these predict your data the best. One of these synthetic variables might be something like 53% x1 + 4% x2 + 14% x3, etc, etc. Oftentimes, systems that are resistant to linear models will cluster nicely in a PCA analysis. For instance, you might have two species that are very difficult to distinguish. You measure a whole bunch of traits (head width, leg length, body mass, wing area, spine number, etc), none of which distinguish the species. But some synthetic combination of those traits will: anything above 1.6 on that mystery madeup synthetic variable is species 1, anything below is species 2, or whatever. Once you have your PCA model calculated, making predictions is going to be more successful than if you had tried to do a linear model.
The beautiful thing about PCA is that it is basically immune to overfitting, so you don’t need to curate your variables at all, and you can include any number and quality of them, which is especially nice in a world of Big Data. One drawback to it is that it is kinda just linear-modelling-with-extra-steps, so it also has the same difficulties that linear modelling has when interactions between variables are important. Another drawback is that it is much harder to pull out causation. It’s not totally impossible—you can go back and look at your mystery made-up synthetic variable that you are actually using to make predictions, and you can decompose them into the real variables that make them up. That will at least give you some clues about what distinguishes your two species (“well, body length and leg length are both big contributors here, so maybe it’s something to do with the ratio of the two?”). But it’s quite difficult, and the only situations where you are going to get clean answers out of it are simple scenarios where linear modelling would have worked anyway.
By the late 2010s, more state-of-the-art machine learning models like neural nets were beginning to make their way into biology. As a CS guy, you probably know more about neural nets than I do (which is almost nothing). For anybody reading along who is interested, I recommend looking up the Random Forest Algorithm. This is a competitor of neural nets, and I don’t think it’s very popular these days, but it has the virtue that it is relatively easy to understand what is going on under the hood, and it is also easy to understand why nowhere in the model is there any number that would even approximate causation. These models were often big improvements on PCA approaches, in terms of prediction, because they were well suited for dealing with interactions. Alas (and maybe, "Because of this..."), they are even more inscrutable than PCA in terms of causation.
All three approaches are very much used to this day. So if you are modelling in biology, you are probably going to pick one of these three approaches to make sense of your data, and which one you use will depend on what you want—do you want to find causal factors? Then you’ll do some linear modelling and pray the whole time that the real world is simple enough to capture that way. Do you want to make predictions and damn the rest? Out come the neural nets.
I'm not exactly a philosopher but I am some kind of materialist atheist. But it seems to me there's a lot of missing the point going on here. I think the cleanest answer is that some concepts exist in a different way than material things, whatever the dominant philosophy today is I hardly think denies concepts such as number or color. For instance, I don't think Mary's room means anything. Obviously, studying everything about redness without seeing it is nonsensical. (Also we could attack the concept of redness which is said to exist here, if you have only seen one shade of red you whole life you've experienced some kind of redness, but there is a lot of red stuff you aren't experiencing.) Also, the properties of water would seem to me to be laws of motion/physics here, not things. And the laws say the things change properties when joined.
Also, in the start of your essay with saying mehanics are still trying to figure stuff out, I fail to see the point of the agitation to resist it (of course the point is religion).
There is really no point to this, but I noted my disagreement I guess.
> "I'm not exactly a philosopher"
To be sure, neither am I. I submitted one paper to a journal that I know was sympathetic with the viewpoint and it was summarily rejected. However, both my parents are professional philosophy professors, and I minored in it in college, so I rate myself an excitable amateur.
> "but I am some kind of materialist atheist."
I admit, I do have misgivings about continuing to publish these letters here, since I know a chunk of the De Civ subscriber base shares your view, and committed materialist atheists are explicitly *not* the target audience for these letters.
The primary audience is my children, who have no investment in materialist atheism except what they've picked up unconsciously from the culture. The secondary audience is other parents who already believe basically the things I believe and who are trying to figure out how to pass it on to their kids in a culture where the Nones are winning bigly on both conversion and retention. (SOURCE: "Religious affiliation in youth and currently" here: https://www.pillarcatholic.com/p/the-faith-of-the-next-generation) It's that secondary audience that I'm publishing for, really.
But this leaves you in a very annoying position: this letter is kind of *about* you, but it is not designed to *persuade* you. (This letter would be three times longer and begin from a very different starting point if it were designed to persuade someone with your commitments.)
I don't know that there's anything I can do about this except acknowledge that I get where you're coming from, I get your desire to pipe up in defense of your position, and I get it if you're annoyed about this series of letters! (I'll try to balance it out, as always, with other topics.)
> "I think the cleanest answer is that some concepts exist in a different way than material things"
Then you aren't a materialist!
One of the key jobs of ontology is to characterize what exactly that "different way" is.
Materialists deny that there is a "different way," because the materialist project is to reduce *all* concepts, including number and color, to strictly and exclusively material terms. Of course, there are different approaches to materialism (reductive materialism, eliminative materialism, functionalism, epiphenomenalism, and so on), but they all share this goal. If you do not, then you aren't a materialist!
(Of course, there are plenty of non-materialist atheists, so you are in fine company with them.)
> "For instance, I don't think Mary's room means anything. Obviously, studying everything about redness without seeing it is nonsensical. (Also we could attack the concept of redness which is said to exist here...)"
The purpose of Mary's Room is not to defend any specific notion of redness, but simply to demonstrate that the subjective experience of sense data ("qualia," in the jargon) is not reducible to objective physical facts.
I agree that this is obvious! To say that one knows a color without access to its qualia is, as you say, nonsensical! But the dominant philosophy today is quite determined to either deny qualia outright or completely explain qualia in terms of objective physical facts. (Neither approach is, in my opinion, at all promising.)
> "Also, the properties of water would seem to me to be laws of motion/physics here, not things. And the laws say the things change properties when joined."
I think this is the most defensible position you stake out, and I wish that space had permitted me a longer examination of properties, but I will leave it at this: if you build out the "laws of motion" in this way, you end up with something very close to interchangeable with Scholastic ontology (with its "powers" and "potencies"), but with bigger explanatory problems.
"Also, in the start of your essay with saying mehanics are still trying to figure stuff out, I fail to see the point of the agitation to resist it."
The point is that the Scholastics already had this same "stuff" figured out. We had a (reasonably) complete, (reasonably) successful ontology, fully compatible with modern scientific methods and discoveries... and we threw it in the trash bin for what turned out to be historically contingent, not-very-good reasons. Not only did we bin it; we replaced it with a much worse theory with much bigger problems that seem intractable on their face. The anti-Scholastic temper tantrum Descartes and his lot threw at the dawn of Modern Philosophy is like if scientists got frustrated looking for the Higgs Boson in 2012, threw out the entire Standard Model, and reverted to Luminiferous Aether theory.
If the Mechanics do, in fact, eventually (miraculously) plug the most gaping holes in their ontology, we should definitely give them a second look! In the meantime, we should go with more successful ontological theories, such as Scotism, Eleatic Monism, and (my obvious fave) the Aristotelian-Thomist model.
" (of course the point is religion)"
You're certainly correct that religion is where I'm ultimately driving this car. However, my motive for swerving away from the Mechanical Philosophy is that it's not a working ontological theory, but rather a very old I.O.U. for maybe someday becoming one, and that I.O.U. is never getting redeemed. If I lost my religion (for example, because the Problem of Evil overwhelms me, or because Jesus's corpse is discovered in Jerusalem), I still would not become a Mechanic.
"There is really no point to this, but I noted my disagreement I guess."
I disagree that there was no point to it! If we never talk with one another about the Big Questions, what's the point of talking about the little ones?
So thank you for your comment (as always).
I guess I would disagree that the dominant philosophy in the culture is materialism in your understanding. (However, I'm not American (I'm from Croatia!) so take this with a truckload of salt.) I think your not-mattering problem just stems from a more political philosophy of religious liberty and debating religion being really annoying to most people (don't underestimate the second one). Again, I'm not exactly well-versed in philosophy, my version of materialism is mostly related to not believing in god. And even there, I would probably limit myself to not believing in a god that wants us to follow his specific rules and punish us. (I'm not sure if you would consider that atheism, but I would say that's what most atheism is.) I guess you could probably try to shove math into the laws of motion somehow. Something I wanted to say in the previous post is, depending on how different individual brains are, you probably could in the future simulate experiences (including redness!). Also colors are social constructs at least there are many colors we consider red and so on. And the final point is, there is a difference between explaining something and explaining something well (and/or correctly), some things are just hard to explain and we should recognize that.
As a chemical engineer, I think the problem with your discussion of water vs hydrogen and oxygen is that atoms are not fundamental particles. They are composed of quarks and electrons. So when hydrogen and oxygen react to form water, it's just a rearrangement of the components of the atoms. There's nothing metaphysical about it, and the widespread success of quantum mechanics in predicting the properties of atoms and molecules, to a very high degree of accuracy, is in my view extremely strong evidence of this.
Also worth noting:
Hydrogen is not inherently flammable, as evidenced by the fact that the Sun is not on fire. Hydrogen is flammable in Earth's atmosphere because it reacts exothermically with oxygen.
Water is in fact quite a reactive molecule, although the surface of the Earth has been so saturated with water for billions of years that there's not much, if anything, left in nature that water reacts with strongly enough to generate flames.
There are plenty of manmade substances that burn with water, though. An increasingly common example is lithium metal, which can form inside of a malfunctioning lithium ion battery. Don't ever pour water on a lithium battery fire!
The boiling points that you give are for molecular hydrogen and oxygen, H2 and O2, and are not inherent properties of the atoms. An interesting illustration of this fact is that ozone, O3, is also composed only of oxygen atoms, but has a boiling point over 100 F higher than that of O2.
Much of this is very interesting! I am pleased to have these additional facts on hand from a professional chemical engineer!
In terms of the point my letter was trying to make, though, I think this comment misses it.
Whatever you set as the level of "fundamental particles," those particles have actual properties. When they interact with other fundamental particles, some or all of their actual properties are suppressed -- that is, reduced to potentials -- and the combined substance has properties of its own that are not shared by its constituents. The properties of water superseding the properties of its constituents is a great example of this for teens, both because many people intuitively think of atoms as fundamental particles, and because the properties of (molecular) hydrogen, (molecular) oxygen, and water, are all common knowledge, easily grasped by anyone past 6th grade.
This supercession cannot be explained in strictly material terms -- at least not without falling (eventually) into (a) Eleatic monism, which denies all change, (b) a really unattractive Mechanical model that treats all these behaviors as brute laws of nature, or (c) reconstructing something that looks an awful lot like Aristotelian-Scholastic metaphysics. The widespread success of quantum mechanics, far from refuting this framework, implicitly *depends* on it.
NOTE: while writing this, I discovered that Oderberg's book is freely available on the Internet Archive, and he does discuss the "quarks objection", so perhaps he will do a better, clearer, or at least more correct job of making the point than me (pp62-76; water shows up on p75; quarks have their big moment on p64 and then scattered throughout): https://archive.org/details/david-s-oderberg-real-essentialism_202305/page/62/mode/1up
It is obvious why I won't give this reading to a 13-year-old, but you're 35, so you don't need me interpreting it when you can just go to the source. :)
Thanks for the link. Unfortunately I don't agree with his arguments against atomism, on pg 64 where he starts talking about quarks. (Again with inaccurate physics, by the way, because quark flavor transitions are mediated by the weak force, not the strong force.)
His first argument, that no one has ever proven that all of the multiplicity and diversity in the universe is a result of mere recombination, is increasingly outdated. In the 24 years since his book was published, the capacity of computers to run atomistic simulations of molecules using quantum mechanics has dramatically increased, to the point that it is now common for these simulations to complement laboratory experiments not only in academia but in industry. Ruthless capitalists would not invest in the supercomputing resources necessary to run these simulations unless they worked.
Perhaps he would counter that unless we can simulate the entire universe atomistically, we cannot definitively conclude that atomism is an appropriate philosophical understanding of physical matter. However, in my view, if atomism works on the scale of a molecule, or of dozens or hundreds of molecules, this is at least sufficient to rebut his proposition that metaphysics is inherently necessary to understand the behavior of a multicomponent physical system.
His second argument is that the ability of quarks to undergo flavor change (i.e., up to down, charm to strange, etc.) is evidence of potentiality in fundamental particles. I'm certainly not an expert on quantum field theory, but my understanding is that the fields are actually fundamental, and individual particles are just excitations of their corresponding fields. https://en.m.wikipedia.org/wiki/Quantum_field_theory
For example: There is a Higgs field, and a corresponding Higgs boson which is an excitation of that field. There is an electron field, and a corresponding electron particle which is an excitation of that field. In this conceptualization of quantum physics, there are not an uncountable number of unconnected quarks constantly transforming into other kinds of quarks, but rather a handful of quantum fields permeating all of time and space which are continually interacting and exchanging energy with each other at an uncountable number of locations.
To me, energy is the closest thing in physics to this metaphysical concept of potentiality. In thermodynamics, for example, energy is considered to be the capacity to change a physical system from one state to another. And when we really drill down on "what energy is", it's admittedly somewhat mysterious: https://readfeynman.blogspot.com/2017/04/section-41-what-is-energy.html
If we know one thing about energy, though, it's that it is physical. It's governed by mechanical laws and conservation laws and famously was shown to be interchangeable with matter by Einstein (and then later by the atom bombs). So it seems to me that there is no need for metaphysical concepts like potentiality anymore, to answer the question "How do physical systems change from one state to another?", and really no room for them either.
I'm not sure you're as far apart from Oderberg as you think you are. I know it was a lot, but did you read to the end, where Oderberg considers (and agrees with) the possibility that what hylemorphists call "prime matter" might very well be identical with what physicists call "energy"?
In any case, your account of the current best scientific theories about elementary entites has moved some of the goalposts from your previous account (previously, you put forward quarks as fundamental; now you make fields out to be fundamental). That's actually not a problem for me, because my whole point here is that it *does not matter* what entity or entities one identifies as fundamental, because, whatever class of entities you select, you cannot account for the visible universe without developing an account of potency, actuality, causation, and, ultimately, substance.
Your revised account confirms this rather wonderfully! You describe the Higgs boson as an excitation of the Higgs field. Unless I've entirely misunderstood you, that, right there, shows that your model has both potency (the Higgs field encompassing all of time and space has the potential to excite itself into a Higgs boson at any given point) and actuality (since some points ARE actually Higgs bosons). Eleatic monism is avoided insofar as the Higgs field's potential is not *necessarily* actualized at each and every point, so reality is not composed *solely* of Higgs bosons.
In other words (to borrow Oderberg's language on p64), your model, as I understand it from your description, still includes substantial transformations at the most fundamental level, and pure potency therefore is still needed. Far from showing that physics has "no room" for metaphysical concepts like potency anymore, you seem to confirm that physics *depends* on those concepts. (It just insists very strongly, in your model, on not calling them "metaphysical" or making them direct objects of study. But this prejudice seems to me philosophical, not scientific, arising from the sources I briefly mentioned in the OP.)
Physics, then, can certainly help us characterize states such as: the Higgs potency, the Higgs actuality, and the material, efficient, and perhaps final causes that precipitate transitions between those states. Physics may also tell us eventually that the Higgs field (and other fields) are *not* fundamental, forcing us to look elsewhere for the most basic expressions of potency, actuality, and causation. However, what Physics *can't* help us do is deny that potentiality, actuality, and cause *exist*. (I think Physics *might* offer some insights into whether substances are real or not, perhaps pushing us toward some species of monism instead, but I'm not sure and it doesn't seem relevant to your argument, so I will leave that question bracketed.)
Fwiw, I don't find Oderberg's much-too-brief treatment of multiplicity and diversity in the universe persuasive, either. It could be sketched out into something beefier, and perhaps since this is a technical book written to fellow professionals he assumes we are familiar with that work, but he doesn't give that argument any legs here. So we agree there, and I suspect even Oderberg would if you asked him.
To repeat my thesis in a way that sounds like a good conclusion: I don't think you're as far from Oderberg as you think you are.
Well the Higgs actuality and Higgs potency, as you have defined them, are physical phenomena. They are quantitative properties of the Higgs field which can be calculated based on physical observations. If that's metaphysics, then what do you consider to be physics?
If you don't mind me answering a question with a question: aren't those quantitative properties of the Higgs field calculated with and described by mathematics? If that's math, what do you consider to be physics?
I don't see why the actualization of the Higgs boson* cannot be simultaneously an ontological process, a physical process, and a mathematical process.
*NOTE: I am not committing myself to this physical model of the universe; I'm accepting your description at face value.
left all the best stuff for those last two footnotes!
As a Catholic mechanist, I suppose it’s on me to defend the obviously best way of thinking about the physical world? (I’m not really a mechanist, more of an occasionalist who thinks mechanism is the best way of gaining a useful comprehension of the world, but close enough, I suspect).
This one doesn’t really work, although I think it’s reasonably close to doing so. Two possible approaches would be:
-Convert it from a criticism of mechanism to a criticism of materialism. The advantage of this is that your three supposed critiques of mechanism fail as such, but succeed (mostly) as critiques of materialism. You would have to be less performatively mad at Descartes, and I understand that this is a pretty-close-to-terminal-value for many people, so maybe this isn’t the best option.
-Acknowledge the incredible successes of mechanism in explaining the physical world, but critique our culture’s totalitarian attitude toward it; i.e., criticize the viewpoint that mechanistic explanations are somehow in tension with theistic ones (currently, I think your essay actually implies agreement with this Culturally Prominent Bad Opinion), and point out that non-material things are not made of parts (or the parts are inaccessible to us), and that the appropriate response to our best tool not functioning in that context is probably not to pretend that such things don’t exist.
I’ll address your three specific critiques of mechanism in reverse order.
#3: Minds
There is one, key, central fact about models—and by “model” I just mean something like “proposed explanation for some facet of reality”, not anything necessarily scientific or quantitative. That fact is that there is a tradeoff between prediction and causation: models can be designed to predict, or they can be designed to show causes. They can even do both, but they won’t do both well. This tradeoff makes us (both mechanists and scholastics) very uncomfortable, so we don’t talk about it enough—certainly the mechanists have plenty of blame on our shoulders now that we are in the second century of pretending that physical models involving fields are somehow causal. For one, the tradeoff presumably doesn’t exist in the mind of God, so it isn’t necessarily a fundamental feature of reality in the grandest sense, but it is certainly the case that at the level of human reason, we are stuck with this tradeoff. This is not a new idea: Bellarmine’s responses to Galileo are based on this tradeoff, and Galileo comes off like a philosophical dummy because he ignores it. But advances in physics and chemistry in the last 150 years or so have heightened the tension, and what really smacks us in the face these days are big data models, neural nets and so forth that really explore the far edges of the tradeoff, being astoundingly predictive and completely useless in sorting out causation.
I think our responses to this tradeoff are representative of scholasticism and mechanism more broadly: you want to find explanations that reveal the causes of things; I want to find explanations that are predictive. You have very little patience with predictive explanations that are not causal, and I vice versa.
I don’t think this is just a matter of taste, though (although it probably is to some degree). Predictive explanations have a number of useful properties, such that I argue we should prefer them a priori:
-Prediction, by forcing your theory frequently to interact with the real world, has a grounding effect that causal explanations do not. This effect becomes more valuable the more biased and confused you believe human reasoning is, and I am extremely cynical about human reason.
-Predictive models can function in environments in which causes are fundamentally incomprehensible to the human intellect. As members of a religion that holds as dogma that various aspects of reality are fundamentally mysterious, this should be appealing to both of us—more appealing the more things you think are truly incomprehensible.
-A weird benefit of predictive models that is really only useful in this exact context is that they are sufficient in themselves to demonstrate mechanism, and your example of the mind is an excellent one.
Thought experiment: say that we get really good at imaging the brain at a sub-cellular level, tracking neurotransmitters, etc. (You seem a bit grumpy that this hasn’t been done already by Leibnitz. Our understanding of how neurons operate in any detail at all is only decades old, and there are 10^11 neurons, so I don’t really understand you on this point, but … sorry, I guess? But let’s say we crack it in the future some time.)
This lets us describe the brain in terms of networks, firing cycles, whatever. Let’s posit that for every quale, we are able to describe an associated network of neurons with a particular structure and set of firing patterns. Let’s further posit that we can manipulate the brain such that we observe:
-Whenever the firing pattern appears, people report experience of the quale.
-The firing pattern does not appear at times when people report no experience of the quale.
-Causing the firing pattern to appear produces experience of the quale even in unusual circumstances.
-Preventing the firing pattern from occurring prevents experience of the quale even in circumstances in which it would normally be experienced.
These findings—assuming they are sufficiently universal, etc—would:
1) Demonstrate the mechanistic character of qualia, i.e., that they can be explained (in the predictive sense of ‘explanation’) in terms of the behavior of their parts, in this case, the neurons of the brain.
2) Not provide ANY PROGRESS WHATSOEVER on the hard problem of consciousness, or any support for materialism, for that matter. Since we wouldn’t have any understanding of what causes qualia, it might just be that God tells your guardian angel to be on the lookout for certain brain patterns, and then give an immaterial poke to your immaterial soul to produce the associated qualia. This wouldn’t be a materialist explanation, but it would be a mechanistic one.
I think that your current attitude toward the mind is an extraordinarily dangerous one, for your daughter. I think it is extremely possible that during her lifetime we will make an enormous amount of progress in putting qualia on a mechanistic footing. It won’t explain what causes them, or do anything to harm non-materialist accounts of consciousness, but we can demonstrate mechanism without doing that. I don’t see the benefit in tying the rest of your apologetics to an unnecessary claim that is likely to be demonstrated false, and risks delegitimizing everything else you’ve said here. You don’t need to stake this claim AT ALL. Why gamble with the soul of your daughter?
#2: Water
I don’t understand why you think any of the claims you’ve made represent a problem for mechanism, so I won’t engage for now (and David seems to be anyway). If you want to explain your thoughts, I’m happy to respond!
#1: Circles
Again, this seems like something that is aimed at materialism, not mechanism. Mechanists have no problem invoking mathematical objects to describe the behavior of the world, and I’m not sure why you think that they do—is there a mechanist in the modern lineage who thinks that the components of the world are actually inert billiard balls? I’m happy to read up on him, if so, but I am skeptical he exists. Based on some of your replies to comments here, I get the impression that you want to say something like “aha, you have said that electrons have certain properties, but isn’t that just the scholastic ideas about properties and natures being smuggled in!?”
To which a mechanist would reply: Sure? Whatever? We aren’t approaching the world as if everything is lacking in properties. We’re trying to treat those properties responsibly, by which I mean that the properties invoked are mathematically describable and resulting from the interactions of the parts of the thing in question. And if there are things with “atomic” properties, we’d like to push that to as low a level as possible. But if we can do that, we mechanists are very comfortable having you jump in at the last minute to gloss all the stuff we’ve learned about physics and how it shows how Aristotle was right all along about action and potency or whatever. We are not insecure about this, because we remember which of the two approaches to properties results in nuclear reactors—and discovers entirely new realms of reality to explore and explain.
I had a genius plan to reply to this comment.
I realized that, if I sat down and just started typing, I was going to turn it into a fisking (not an unfriendly one, but lots of paragraph-by-paragraph quotes). I decided that it would be wiser for me to sit with your comment for a while, think about my reply, and finally condense it down to its essentials.
Unfortunately, I let it sit too long, and now I've completely forgotten many of my early thoughts. I suppose that's still a form of condensation.
First, rereading this letter, I agree it is rough and needs revision. I lost my own past train of thought once or twice on the reread, and had to go back to remind myself what I was on about. This was perhaps the most ambitious letter in the series (so far), and its aim outreached its grasp at times. The whole framing of your comment highlights this, because...
(Second,) This letter is *intended* as an critique of materialism. I intentionally defined "Mechanic" very narrowly here: a Mechanic (for the purposes of this post) is someone who believes that material corpuscles (and, *maybe*, their positions) are the only things that exist, period. If I weren't writing this to an adolescent, I might have avoided some confusion by using more common labels: physicalism, reductionism, eliminative materialism, with maybe a side order of skeptical empiricism in cases of doubt.
This narrow focus made me do something very painful, which you may have noticed: I let Descartes off the hook! I think his philosophy was mistaken, that it *led* to what I here termed the "Mechanical Philosophy," and, yes, you're right, performative outrage about that IS something like a terminal value! But Descartes himself was not, by any stretch of the imagination, a Mechanic within the meaning I have given the term here.
But that's all according to plan. My sole goal in this article is to persuade my daughter that the materialist model she will have absorbed through our culture (and which I found *extremely* convincing at her age) cannot possibly be correct, that there must be *some* immaterial things. This is pure bootstrapping for the next letter(s), where I will argue for some form of the First Cause and make the case that it cannot be material. I find that the way people receive those arguments depends a great deal on their priors. (At least, it made a huge difference to me when I was 14, read Russell's "Why I Am Not A Christian" in the STA library at study hall, and thought "turtles all the way down" was a brilliant refutation of Aquinas!)
As such, I think the *key* criticism you make has already been incorporated into the article, but needs to be drawn out to be much clearer in the next revision. (This is good. The reason I post these online is to gather useful feedback for the revision before the kids read them.)
But, as you identify as something-like-an-occasionalist who has no problem with certain immaterial things, this letter is, quite simply, not aimed at you, and I obviously agree with your criticism that some of my arguments do not work against (It may still have struck you with friendly fire.)
The most important part of your comment for my revision, I think, will be your notes on what I had to say about the mind. I think you are likely correct that we will, within the lifespan of human civilization and quite possibly within our lifetimes, develop the kind of neuron-to-quale mapping you describe, and I don't think anything makes it in principle impossible. It is very much *not* my intention to hang my whole argument on the certainty of that *not* happening, and I tried rather hard to avoid that specific trap at the time. Clearly, I failed! I will have to try again, and I will keep a close eye on your notes as I do so.
Three questions for you, if you're not averse to answering questions nine months later:
(1) What is the definition of "mechanist" you use in describing yourself as one? Like so many terms in philosophy, "mechanist" is overloaded. I could probably discern your definition pretty well from what you said in your comment, but it would be better to let you define your own terms.
(2) How'd you get there? Without bearing any ill will toward your metaphysics, I confess I was surprised to hear you use that label, and I'm sure you followed an interesting path to embracing it.
(3) This tradeoff between causation and prediction is a really interesting idea that has stuck in my craw without my knowing quite what to do with it. At the risk of revealing myself as out-of-touch with the literature, this comment is the first time I'd heard of the idea. Can you suggest any further reading on it? I'm not sure it will have much bearing on my revision of this letter, but I think it could be a useful way of thinking about fault line between the Moderns and the Medievals.
(Okay, that was still kinda long, but not nearly as long as it probably would've been if I'd replied right away.)
Ah, Marvin Minksy, who has variously been quoted as saying that human beings are "machines made of meat" and that our brains are "computers made of meat." I wouldn't want to go out to lunch with this guy.
But my main initial reaction to this post was "Oh no, not Star Trek!" IMHO it is far from the worst offender. It would seem that objectively and throughout the universe, racism is wrong (even if it depends on which half of your face is black), even very powerful alien beings are nothing like God ("We find the one quite sufficient"), and it is wrong to disturb the time line so that Nazis rule the world, even if the alternative makes mincemeat of your emotions. I speak, of course, of Star Trek TOS. When Star Trek: Discovery was being filmed, I read an account of one scene in which the actor playing the captain ad-libbed a comment like "Oh my God!" or "Goddamnit!" and the director sternly told him to redo the scene without that, because in the future nobody will believe in God. Saved me some wasted hours trying out the new show.
My intro to philosophy prof (one Dr. Greg Coulter) pulled a wonderful trick on us in what was, for most of us, our first month of undergrad.
He started off hammering home for us the rules of logic and the structure of argument. That was three classes, I think, and several handouts saying "HERE IS HOW TO IDENTIFY AN ARGUMENT." Then he assigned various readings. We had some Aquinas, some Anselm, some Berkeley, some Descartes. I mean, it's a survey course, so it ran the gamut. In each reading, our assignment was to find the argument, write it down, and then offer some comments on it. I was an undergrad, so I'm sure I said several totally insane things. (I think I agreed that Anselm's ontological argument worked *only* in terms of extension or something, like some kind of insane fun-house version of Gaunilo.)
So then, after we've done three or four of these, Dr. Coulter assigns Marvin Minsky's "Minds Are Simply What Brains Do" with the usual assignment: find the argument, write it down, offer comments.
When we all came into class the next session, haggard and confused, he asked who found the argument. Some of us sheepishly offered up what he'd written on our homework and he shot them down one by one. Then, finally, we had no answers left, and he revealed that this was correct. The reason we'd had a hard time finding an argument in there was because there wasn't one. It was just a series of assertions, which made no attempt to proceed to a conclusion from logic. It paints a picture; it doesn't prove a case. "And yet," he commented (I paraphrase), "often compiled into introductory philosophy readers1"
That always stuck with me. People can sound very convincing. But look for the argument.
(Also, write the argument.)
You are, of course, absolutely correct about Star Trek. My kids know that it is my favorite show of all time by far and that I am deeply obsessed with it. (I have always been very active in certain parts of the fandom.) So if I am willing to admit it has bad metaphysics, they won't take my comments about My Little Pony as an attack.
You are also correct that Discovery was terrible, although somehow not nearly as bad as Picard. That complete misunderstanding of how God figured into the Roddenberry universe was only the first sign of the madness that had taken hold on set. (Did you hear how they hired the great Black American mystery novelist Walter Mosley for Season 2 and then he wasn't woke enough so they fired him?!) (And Discovery was SO MUCH BETTER than Picard! GAH!)
Star Trek Prodigy, however, is quite good. It's aimed at kids and I am watching it with my kids, but I would watch it without them, because it's a fun, pure, joyful space adventure, the likes of which you don't really see much of on TV anymore.
I have always loved the "We find the one quite sufficient" line, because it canonizes Kirk's monotheism (and, almost certainly given Bread and Circuses, his Christianity) no matter how much Roddenberry et. al. tried to make theism untenable in future shows. (And then DS9 snuck a space pope in through the back door anyway.)
BUT I DIGRESS!
1. I would say mechanism (sensu-I-am-a-mechanist) is the idea that the explanations for complex things are to be found in the interactions of their parts. In terms of the opinions you ascribe to mechanists:
-The world is a clockwork machine: ✅
-It is made up of colorless, odorless particles bouncing off one another aimlessly, governed by just a few laws of motion: ✅
-It is made up of featureless particles: I don’t know what this means
-There really is no such thing as a brick or a window: ✅
-There really is no such thing as your body: possible not; I’m prepared to believe God keeps a “ledger” of which particles are assigned to your soul.
-We wouldn’t need to reason about bricks and windows: I am with Douglas Hofstadter that limitations of human reason mean that it is usually only practical to do one level of mechanical explanation. So bricks and windows are very useful approximations for us in understanding the world. But, s/we/angels/ and ✅
-We wouldn’t need to reason about humans: humans have a soul, which might make this more complicated.
-Particles are the only things that could be known and understood: ❌
-There’s not really any such thing as a tree or a brick or the color red: ✅
-Anything that isn’t made of particles can’t really exist: ❌
-Anything that isn’t made of particles can’t really be known: ❌
-Any free will we actually experience can only be an illusion: ❌
2. Well, my journey isn’t interesting from the inside, at least. I was exposed to the scholastic worldview in college, and it was immediately evident that no one had bothered integrating the last 200 years of advances in biology into it, and that if you did, the results would be farcical. Given that this was the only field that I knew anything about, and that I’ve never been one for Gell-Mann amnesia, I didn’t really see the point in exploring it further. Nothing (including spending much of the time since in Catholic academia) has made me think I ought to reconsider that; as we’ve discussed above, I don’t find the arguments against mechanism to be at all compelling, and when the other camp’s model fits the world I know about so badly, I've never felt either a push nor a pull away from mechanism.
3. No, I can’t give you any readings. This is a view gained from the trenches, as it were. Maybe it would be useful to give a brief history of models in biology? That might be helpful in finding readings.
-The original model is the Linear Model, y=ax + b, where x is your predictor, y is the thing you are explaining, and a and b are parameters. This is the classic causal model: you do your experiment, measure x and y, and calculate values for the parameters. If a > 0, then x causes y (assuming you’ve designed your experiment properly). Once you have an estimate for a and b, you can then use measurements of x to predict y, or vice versa. So the same final equation, y=0.7x+5.3 (or whatever) is both causal and predictive.
-Of course, the problem is that in the real world, nothing is ever really caused by one thing. Things have lots of causes. So you’ll want to add more predictors to your equation. You might try y=ax1 + bx2 + cx3 + d. The problem is that you can’t actually stop there, because the world is lousy with interactions between variables. A classic example might be: having cancer is bad. Having radiation poisoning is bad. But having radiation poisoning while you have cancer isn’t double-bad, it’s just chemo, and it’s actually kinda good. So you need to account for these interactions in your model, which is now: y=ax1 + bx2 + cx3 + dx1x2 + ex1x3 + fx2x3 + gx1x2x3 + h. Every predictor you add doubles the number of parameters you need to estimate, which creates a number of problems:
A seven-predictor model requires 128 parameters to be estimated, which means you need 64 times the sample size compared to a single-predictor, two-variable model. This is a practical problem.
Even if you have a big enough sample size, there are other practical problems with the parameter estimation—for instance, the methods here are very sensitive to predictor variables being correlated with each other, and go haywire when they are. Alas, it turns out that things are often correlated with each other in reality.
Even if you solve those problems, and correctly figure out the parameter associated with the x1x2x4x7 interaction term, what does that parameter mean? You could write out a sentence that describes it, sure, but it will be hard to understand that meaning, and you will have dozens of these to account for simultaneously. You won’t be able to do this. Human brains just don’t think in more than three dimensions, sorry. So we’ve lost the nice causal meanings that simple, non-interaction-bearing models have.
For these reasons, the Best Practices for linear modelers is to curate your set of predictor variables heavily; keep them few, don’t add predictors unless you have some expectation that they are causal, and ensure that they are uncorrelated. Don’t include any interaction terms in your model, ever, except for possibly one or two which you have a strong a priori reason to believe might be important. If you follow the advice, then you get a beautiful model from which causation is easily inferable from its parameter values. But because you have (intentionally) ignored vast swaths of reality, this will limit the predictive power of your model.
Now, there is a very pleasing fact about these linear models, which is that they were very susceptible to overfitting. I won’t really explain this, because I don’t understand it. But if you ignore the Best Practices, if you add in tons of predictors and interaction terms, even if you have a large enough sample size to calculate your parameters accurately, overfitting meant that your predictive powers didn’t actually improve, even as your were wrecking your ability to draw causal conclusions from your model. So for … I dunno, half a century? … when linear modelling was the only game in town, there was no tradeoff, in practice at least, between causal explanations and predictions.
The next important type of model is the family of Principal Components Analysis (PCA) and relatives. By the time I was in graduate school in the 2010’s, PCA was already well-established. With PCA, you throw away the linear modelling best practices. Instead of curating your predictor variables, your gather every possible predictor you can think of. You don’t worry about predictors being correlated with each other, you don’t worry about them being useful, you just throw in everything you got. Then, you make a bunch of made-up variables that are combinations of your real variables, and identify which of these predict your data the best. One of these synthetic variables might be something like 53% x1 + 4% x2 + 14% x3, etc, etc. Oftentimes, systems that are resistant to linear models will cluster nicely in a PCA analysis. For instance, you might have two species that are very difficult to distinguish. You measure a whole bunch of traits (head width, leg length, body mass, wing area, spine number, etc), none of which distinguish the species. But some synthetic combination of those traits will: anything above 1.6 on that mystery madeup synthetic variable is species 1, anything below is species 2, or whatever. Once you have your PCA model calculated, making predictions is going to be more successful than if you had tried to do a linear model.
The beautiful thing about PCA is that it is basically immune to overfitting, so you don’t need to curate your variables at all, and you can include any number and quality of them, which is especially nice in a world of Big Data. One drawback to it is that it is kinda just linear-modelling-with-extra-steps, so it also has the same difficulties that linear modelling has when interactions between variables are important. Another drawback is that it is much harder to pull out causation. It’s not totally impossible—you can go back and look at your mystery made-up synthetic variable that you are actually using to make predictions, and you can decompose them into the real variables that make them up. That will at least give you some clues about what distinguishes your two species (“well, body length and leg length are both big contributors here, so maybe it’s something to do with the ratio of the two?”). But it’s quite difficult, and the only situations where you are going to get clean answers out of it are simple scenarios where linear modelling would have worked anyway.
By the late 2010s, more state-of-the-art machine learning models like neural nets were beginning to make their way into biology. As a CS guy, you probably know more about neural nets than I do (which is almost nothing). For anybody reading along who is interested, I recommend looking up the Random Forest Algorithm. This is a competitor of neural nets, and I don’t think it’s very popular these days, but it has the virtue that it is relatively easy to understand what is going on under the hood, and it is also easy to understand why nowhere in the model is there any number that would even approximate causation. These models were often big improvements on PCA approaches, in terms of prediction, because they were well suited for dealing with interactions. Alas (and maybe, "Because of this..."), they are even more inscrutable than PCA in terms of causation.
All three approaches are very much used to this day. So if you are modelling in biology, you are probably going to pick one of these three approaches to make sense of your data, and which one you use will depend on what you want—do you want to find causal factors? Then you’ll do some linear modelling and pray the whole time that the real world is simple enough to capture that way. Do you want to make predictions and damn the rest? Out come the neural nets.