In thinking about early modern materialism, I’ve repeatedly come across the view that materialism implies panpsychism. This claim has some current resonance, in that Galen Strawson has been arguing for a version of it. And it has several early modern sources. Thomas Hobbes worried that his materialist account of perception would lead him to a sort of panpsychism. Henry More argued that the changes Hobbes made to his view to avoid this did not solve the problem. Margaret Cavendish was a panpsychist materialist, and thought that non-panpsychist materialists, such as Hobbes aimed to be, could not adequately explain the workings of the world. There’s also, I believe, a version of the claim that materialism implies panpsychism in John Locke’s Essay (in 4.10.10). And there’s another version — the one I describe below — in Pierre Bayle’s Historical and Critical Dictionary. This being, at least, a curiously persistent theme, it seems to be worth some investigation.
In note C to the article “Dicaearchus”, Bayle argues against the view (Dicaearchus’s view, as he has it) that body can think. Bayle’s argument works in something like the following way.
- If thought belongs to body, then either (a) it does so as a modification, or (b) it does so essentially.
- But not (a), because a modification lost is replaced by another of the same kind (a colour by a colour, a degree of motion by another degree of motion) but when a body loses thought it is not replaced like this.
- And not (b), because then all bodies would have thought. So
- Thought does not belong to body.
If we accept (a), we are effectively pushed towards the view that the only possibly acceptable materialism is a panpsychist materialism. Of course, Bayle thought that was wrong too. But that’s another story. Why should we accept (a)? Bayle says the following.
Someone will tell me that feeling could be a modification of body. From which it would follow that matter, without losing anything essential to it, could cease to feel as soon as it was no longer enclosed in the organs of a living machine. I answer that this theory is absurd, for all the modalities of which we have any knowledge are of such a nature that they cease only to give way to another of the same kind. There is not figure destroyed but by another figure, nor is any color driven out but by another color (Pierre Bayle, Historical and Critical Dictionary, translated by Richard Popkin (Indianapolis: Hackett, 1991) 66).
He follows this with the discussion of some possible counterexamples, such as the notion that the replacement of heat by cold in a subject is the replacement of a thing of one kind by a thing of another. He finds these claimed counterexamples ineffective, and thus concludes that “the conversion of sensation into the absence of all sensation is impossible, for it would constitute a conversion of something real and positive into nothing” (Bayle 67).
After publishing the Dictionary, Bayle received a letter from John Toland, objecting to the argument of note C. Bayle then responded to Toland’s objections in note L, in the second edition of the Dictionary. Toland’s letter does not survive, but we can reconstruct Toland’s argument.
Bayle had effectively said that materialists about the human mind must adopt a sort of panpsychism, if they are to hold on to their materialism: if some body can think, all body can think. Toland argued that you can hold that some corporeal things think, without holding that all corporeal things think. The key to this, he argued, is to see that the relevant view is that appropriately organized corporeal systems think. These can think when the organization is in place, and cease to think when the organization is broken, without it being the case that every part of them can think.
Bayle responded in note L to Toland’s objection.
I maintain that something is being supposed that has hitherto been inconceivable to all mankind, if one supposes that the arrangement of the organs of the human body alone make a substance that had never thought to become a thinking one. All that the arrangement of the organs can accomplish is reducible, as in the case of a clock, to various different kinds of local motion. The difference can only consist in the greater or lesser degree of motion. But just as the arrangement of the several wheels that make up a clock would be of no use in producing the effects of this machine if each wheel, before being placed in a certain way, did not actually possess an impenetrable extension, a necessary cause of motion as soon as it is pushed with a certain degree of force; so I also say that the arrangement of the organs of the human body would be of no use to produce thought, if each organ before being put in its place was not actually endowed with the ability to think (Bayle 70).
Toland suggests the view that the thoughts of humans depend on the more obviously physical properties of their bodies. However, Bayle argues, motion and impenetrable extension (mechanical qualities, we might say) are not a suitable basis from which thought could emerge. Body would have to possess some further part of its nature, a suitable basis from which thought could emerge. And the only possible such basis is the ability to think. There is no other, fundamentally different feature (such as a fundamental ability to be part of a system that will think if appropriately organized), from which thought could emerge. The only way in which something corporeal could have the ability to think, Bayle believes, is if it fundamentally and essentially possessed the power of thinking. Neither a mechanical explanation of thought nor a pan-proto-psychism are acceptable. Contrary to Toland’s objection, the only consistent materialism is a sort of panpsychism. (Given the substance-attribute-mode framework of the discussion, it’s tempting to gloss this as ‘the only way to be a materialist is to be Spinoza’, but that’s maybe a step too far.)
What a great post, Stewart! The parenthetical conclusion is precisely the one Bayle is aiming at!
Thanks Eric. I’m more hesitant about the parenthetical conclusion than you. In part that’s because of general hesitancy about interpreting Bayle. And that line seems to fit note L better than note C. Plus, ‘the only way to be a materialist is to be Spinoza’ seems likely to be confusing, because believing in substances with two attributes isn’t believing that there is only one substance.
It seems like Bayle’s objection can only properly construed as an objection against the view that thought emerges from mechanical complexity.
It is unclear why Bayle’s objection would cause trouble for the view that there are two types of matter: thinking matter and unthinking matter. Humans are composed of a combination of the two types of matter while rocks are composed simply of unthinking matter. Maybe you only get a full-fledged mind if you get the right quantity and arrangement of thinking matter.
So, to be clear the view is that there are fundamental bits of matter that think (or proto-think) and fundamental bits of matter that don’t. This is sufficient to avoid pan-psychism (and pan-proto-psychism).
I guess the question I’m asking is: why couldn’t someone endorse the view that some substances are dual-attributed and others are merely mechanical? I recognize that this is simply not the view that Toland is arguing for, but it needs to be ruled out to get the stronger conclusion that all materialists should be panpsychists, right?
Cheap objection: that’s not really materialism, because the so-called thinking matter isn’t really what people usually call ‘matter’.
More seriously, I agree that the position is available, for all that Bayle says here. And I suppose you could think of Henry More as having that position: there are spirits (thinking things) and bodies (unthinking things), and both spirits and bodies are extended. More didn’t think of that as a sort of materialism though (but see John Henry’s “A Cambridge Platonist’s Materialism: Henry More and the Concept of Soul”, http://www.jstor.org/stable/751295).
If panpsychism counts as materialism, I cannot fathom why having both thinking and unthinking matter would not count as such.
I don’t think it really matters whether we call it materialism or not, so long as we’re clear what we mean. “Cannot fathom” seems a bit strong though. It seems to depend on how we want to draw the distinction. If materialist views about the mind are just views that involve the thinking stuff being extended, then the view you describe is materialist, but so is More’s, which seems a little awkward. If the criterion is whether the thinking stuff is divisible (a more Morean criterion) then More’s view isn’t materialist, and the one you describe presumably is. If someone wants to insist that materialism is a sort of monism (about kinds of stuff) then neither the view you describe nor More’s view will count as materialism, because they’re both dualistic.
My thought was in part a response to the point you made when you said, “Given the substance-attribute-mode framework of the discussion, it’s tempting to gloss this as ‘the only way to be a materialist is to be Spinoza’, but that’s maybe a step too far”
If being Spinoza is a way to be a materialist, then it seems that there is at least one other way to be a materialist, namely to have some matter with dual attributes and some matter with only extension/mobility attributes. This avoids panpsychism, but has at least as good materialist credentials as a Spinozist ontology (after all the portion of the ontology that one might be tempted to think of as incompatible with materialism would be had in common with the spinozist ontology, and it is only with regards to the unthinking matter in which they differ).
Though I should note that I agree that arguing over which views count as “materialism” is not the most interesting thing to do, given that we have a common understanding of the view in question.
How do you see Locke’s remarks about superaddition fitting into this story? On Lisa Downing’s reading (if I remember it right), superaddition is merely epistemic, so that would mean Locke is allowing the epistemic possibility that there are two different types of matter.
I think I want to say the same thing here: it depends how we want to draw the materialist/non-materialist distinction, and that doesn’t matter too much as long as we’re clear. (I suspect using 17th-century standards won’t really help, as the difficult cases had contested descriptions then too.)
I don’t think that superaddition is always merely epistemic; like Ayers, I want to allow for a kind of divine architect reading where God might be “configuring” (more than merely mechanical) matter. But I definitely think that Locke has to allow for the epistemic possibility that there are two different kinds of matter: that is to say, the real essence that corresponds to the nominal essence– extended, solid stuff– might be disjunctive. Locke is clear what is required to count as matter, but otherwise I agree with Stewart that there is a question there.
Something I’m not sure about is whether Locke could make sense of the “dual attribute” proposal. If thought isn’t derived from or integrated with the thing’s other qualities, what makes the difference between there being one substance with two attributes vs. two substances? I don’t think it fits with the way that he thinks about real essences. But those are Locke-specific observations.
There might not be a non-circular way to do this, but could one appeal to Locke’s principles of anti-coincidence to try and establish that it is a single substance with multiple attributes? That is, treat the two-substance view as a violation of the place-time-kind principle? The main worry would be whether there would be independent grounds for invoking the principle without having already settled the question of whether there is one substance or two.
Allowing for two kinds of matter also seems unprincipled to the New Philosophy (and, for those who care, the PSR), which uses homogeneity of matter theses as a trump card against the Scholastics.
I don’t mean to turn the conversation away from Bayle (does anyone know if he ever eased up on the unity of the mind in his later works?), but another important discussion of whether materialism implies panpsychism can be found in the Clarke-Collins correspondence (there is a new Broadview edition of it edited by William Uzgalis, and I have an online version here). As Clarke puts it towards the beginning of the debate:
“And if Cogitation in like manner could possibly be a Quality really inherent in a System of Matter, it must likewise necessarily be the Sum and Result of the Cogitations of the several Parts: And so there would be as many distinct Consciousnesses, as there are Particles of Matter, of which the System consists; Which I suppose will be granted to be very absurd.”
Collins is trying to defend a sort of materialist emergentism, but ultimately speaks the language of panprotopsychism:
“I allow Consciousness to consist of Powers of the same Kind; for that, according to me, consists of Parts which have a Tendency to Thinking or Consciousness, as Roundness does of Parts that have a Tendency to Roundness.”
It’s not entirely clear whether Collins’s view is robust panprotopsychism (where each material part contains something protopsychic in addition to solidity) or the sort of trivial panprotopsychism that follows from any emergentism (where each material part is merely fit to be part of a system from which consciousness emerges). In any case, Clarke recognizes no room for panprotopsychism as distinct from downright panpsychism: “nothing that is not Consciousness, can have any Tendency towards Consciousness.“.
[…] my previous post I mentioned Hobbes’s worry that his materialist account of perception would lead him to a […]
[…] an earlier post I talked about some arguments in Bayle’s Dictionary. In notes to the article […]
I think pan-psychism should be thought of as the non-emergent form of property dualism. The justification for pan-psychism is the logically inconceivable emergence of subjective experience from objective brain activity because they are two fundamentally different things. If what we call ‘subjective experience’ is ‘really’ objective brain activity then there is no radical emergence, just complicated physical activity brought about by preceding quantitative physical changes. If matter-energy is all that exists, as materialism claims, we don’t have to explain subjective experience because it doesn’t exist. If it did, materialism would be false (unless you want to stretch the word ‘physical’ to include something that is non-spatial, has no mass and can’t be converted into anything that does and can’t be observed through sensory perception. Clearly the external physical world is different from the internal ‘physical’ one of mental states and it’s still inconceivable that one could emerge from the other).