From the IEP article on Dualism:

Decartes argues that the mind is indivisible because it lacks extension. The body, as an object that takes up space, can always be divided (at least conceptually), whereas the mind is simple and non-spatial. Since the mind and body have different attributes, they must not be the same thing, their "unity" notwithstanding.


Still, he cannot doubt his own existence, since he must exist to doubt. Because he thinks, he is.

It seems to me that both of these arguments are almost trivially refuted by modern science.

A) thoughts and ideas do have extension and can be easily divided when viewed as configurations of neural patterns or digital memories, and B) Computers "think" all the time without having a corresponding ego that does the thinking, i.e. thinking is possible without existing.

These refutations are so obvious and simple to me, I fear that I might be overlooking something.

My questions:

  1. Are these indeed refutations of Descartes arguments for dualism or is there something more fundamental that I am missing here?
  2. Have any modern dualist presented arguments that are NOT mere refinements of Descartes original arguments for dualism?
  • 3
    You overlook that computers do not think. How do you get this impression? Further you can't map ideas to "neural patterns". Thinking is possible without existing? How is this?
    – John Am
    Oct 28, 2015 at 11:42
  • @JohnAm Computers a) were originally designed as simulators of the human thought process. Turing's Machine was intended to be a model of how a human performs calculations, not as a theoretical underpinning for computational hardware. b) You can map ideas (both images, impressions, etc,.. and propositional statements) onto neural patterns, this has been demonstrated repeatedly in the field of artificial neural networks and backed by results in neuroscience. b) Computers aside, thinking is possible without existing, see the bundle theory of mind and Hume, W. James and B. Russell. Oct 28, 2015 at 15:58
  • 3
    Main arguments for dualism are listed by SEP, not all of them go back to Descartes, the most popular argument from knowledge for example plato.stanford.edu/entries/dualism/#ArgForDua Cogito ergo sum is an argument for existence, not for dualism, Descartes gives modal argument for dualism, which goes back to Avicenna's Floating man, and was modernized by Kripke.
    – Conifold
    Oct 28, 2015 at 19:29
  • One could call Kant a dualist and a "more fundamental" argument. Oct 28, 2015 at 20:47

6 Answers 6


For 1. - No.

Your refutations beg the question.

If I divide the neuronal path, or the part of the brain activated, have I divided the thought? Only if you reject all forms of idealism on principle.

Do computers think? Again, only if you have no standard for separating thinking and computation, which is again only the case if you reject idealism completely on principle.

The contrasting problems are obvious (and so are better refutations):

If 'red' is a complex pattern of activation involving the whole brain, into what can I divide it that remains a thought? So I can dismember thoughts, but I cannot necessarily divide them into other thoughts, which is what Descartes meant. When I divide matter, it remains matter, when I divide certain thoughts, they seem most likely to become something else entirely. (Of course, bodies are not divisible either, at a certain level you divide matter and get only energy.)

If computers think, is it killing to turn one off? At what point, and on what basis do we decide what kind of thinking constitutes life? If thinking is something computers do, then can we happily 'turn off' people less intelligent than computers because we don't hesitate to turn off computers? Are they just close enough to brain-dead not to make any difference? (Of course, quantum mechanics implies Heraclitus: If you are never the same configuration of parts from one moment to the next, how well-defined is this 'you' thing that is doing the doubting? Some one thing might be doing the doubting in one instant and a wholly different thing in the next. So there may be no "thing that does the doubting".)

  1. I gave a summary of such an argument from a specific intermediate position, as part of an answer to a previous question https://philosophy.stackexchange.com/a/29048/9166 -- part 2
  • "When I divide matter, it remains matter, when I divide certain thoughts, they become something else entirely." Are we sure? One could use some form of logical atomism to divide thoughts, if not indefinitely, then at least make them as divisible as matter is. Oct 28, 2015 at 19:10
  • 1
    "If computers think, is it killing to turn one off?" - People opposed to the computational theory of mind tend to forget that computers were originally built as explicit simulators of the human thought processes. Imagine a very rudimentary robot that moves on a set of wheels à la R2D2. Many would deny that it is walking the way a human walks, and deny that it is therefore achieving "strong artificial walking", but at the end of the day, it is performing the exact same task that humans do when they use their legs. Oct 28, 2015 at 19:16
  • If don't forget that - they are a concretization of our map of ourselves ca. the 19th c. The map is not the territory, and the map itself has changed.
    – user9166
    Oct 28, 2015 at 19:29
  • And no, we are not sure. But we still don't know. What we have not yet established cannot be used to prove much or refute anything. We are pretty sure there is no part of the brain for distinct qualia we expected to find maps for. I will add a 'seem likely to'
    – user9166
    Oct 28, 2015 at 19:30
  • But to what extent can you imagine an idea that would be 'part of' red, rather than a variant upon it? If we can't segment it logically (which seems clear) or spatially (which seems quite likely), what other alternatives present themselves?
    – user9166
    Oct 28, 2015 at 19:36

Not necessarily. Seems there are two viable approaches to not conclude this:

If you are a non-materialist, the answer is no: modern neuroscience does not answer many questions about the identity of self, nor is it capable of accounting for free-will. In this view, there is still something to the mind other than its physicalness, yet "mind" as such is still indivisible.

If you are a materialist, there is still a great deal of complexity theory that makes the answer complicated. I don't know primary sources, but as a programmer the two books I've encountered that discuss this are Management 3.0 and Agile Thinking and Learning. They both discuss how the mind, being a complex system, cannot simply be broken down into separate pieces and understood separately - in some sense, it has to be understood as a whole. The mind also exhibits some "holographic" properties - parts of the mind duplicate other parts, meaning it's hard to point to a particular neuron as the "cause" of a thought.


a) suppose you say thoughts and ideas can be divided because you accept their identity with neural firing - why is that not begging the question?

b) why did you write that computers "think" - in parenthesis? is it possible that you and Descartes mean different things by that word?

  • My original background is in AI, and for a very long time I took it for granted that thinking and computing were the same thing. Now that I realize that not everyone agrees with that, I overcompensate. Oct 28, 2015 at 6:16

Ad 1: The concept of mind does not just cover perceptions, ideas, thoughts and concepts as the content of our thinking but also the mental activity itself like thinking.

I agree, one can structure the content into sub-content and divide it in parts. Also mental activity can be decomposed into sub-activities. I think this holds already on the level of everyday argumentation. One does not even need to argue on the scientific microlevel of neural patterns or digital memories.

Concerning the questions whether the excution of algorithms by a computer can be considered “thinking”, different opinions exists. Hence many philosophers would not accept the premiss of your counter argument. I consider this point an interesting separate question.

My main issue criticizes the general setting in which Descrates discusses the mind body problem. It is the setting of two distinct substances: mind (res cogitans) and body (res extensa).

Today and particular in the context of neurosicence we consider mind no longer a separate substance but a capability of the brain. It is not a separate object, but primarily the function of a certain organ. The concept starts as a function like “mental activity” not as an object like “mind”. It is risky to reify the function as a substance.

Hence I vote for discussing the whole mind-body problem in quite a different setting than Descartes did.

Ad 2: No, modern dualists like Eccles and probably Penrose employ a similar approach like Descartes to solve the interface-problem, which is inherent to a dualist approach.

But in the philosophical tradition after Descartes different proposals from dualists have been proposed to solve the mind-body problem: Leibniz (Parallelism), Geulinx and Malebranche (Occasionalism), Huxley (Epiphenomenalism).


No matter what one thinks of dualism, to suggest that consciousness or mind is simply extended, divisible, and measurable is fantastically wrong.

So I would argue that, yes, if this physicalist model appears glaringly obvious to you, you are missing something. Something called philosophy, for lack of a better word.

First, by what sort of evidence do you conclude that you are looking at thoughts when you are looking at neuronal patterns? There is a peculiar paradox here. If you look at a brain and neural system in a lab, you are apparently looking at the complete container of a person's consciousness. The entire world--every sensation, memory, and idea--was or is "in there."

And yet, strangely enough, that neural system and the world it generated are also "in here," in the world you are in. You would like to say that, by analogy, this world you are in is also "inside" a brain and neural system, your own. But which side of this mind-body ouroboros has priority? It is worth remembering that you can never be conscious and consciously observe your own brain and neural system. You can't get outside of it so that it is, in turn, inside the world it is supposedly generating.

Whatever it is that consciously observes brains and neural systems can never itself be directly observed. We tend to forget this perplexity, just as we forget that we can never see our own faces, which we so readily identify as ourselves. We can only see the reversed image and tip of the nose. You can observe neural patterns, but you can never observe whatever is observing this immediate world that those neural patterns are in.

Moreover, there is a problem with the divisibility of those neural patterns. It is not just the problem of the infinite divisibility of matter. It is also the problem of what "goes between" the partitioned parts. What "keeps the parts apart"? The mind, apparently. Or is it the mind that is "kept apart" somehow by those objects? The problems only deepen if you turn away from computers and dots on screens, shut your eyes, and try to simply "divide up" your thoughts. How do you separate them without merging them in another thought? Can they really be separated?

I'm sorry I don't have to hand some references or arguments to cite, outside of the whole tradition of German idealism, but I am just starting to think about it myself. This says nothing about dualism or monism, but I do not think the identification of mind with brain states is so glaringly obvious. Nor is Descartes simply "refuted" without some act of physicalist faith. A physics that explains physicists is a long ways off. And were it accomplished, it could still be the work of a deceptive demon.

  • "First, by what sort of evidence do you conclude that you are looking at thoughts when you are looking at neuronal patterns? " How thoughts can be encoded in neural patterns has been demonstrated amply, see artificial neural networks and computational neuroscience. Echo State NNets, Liquid State Machines and Deep Learning NNets have all shown impressive results in that topic. Oct 29, 2015 at 17:40
  • 1
    "But which side of this mind-body ouroboros has priority? It is worth remembering that you can never be conscious and consciously observe your own brain and neural system." I disagree, how this can be possible has been demonstrated by Hofstadter with his strange loop concept, and more generally in the self-representational approach to consciousness. Ironically Hofstadter turns DesCartes on his head: An "I" emerges because the human brain is capable of observing its own states. Oct 29, 2015 at 17:44
  • First, I will certainly have to look at some of that literature. And I not doing a good job of thinking through and expressing my problem with this. I am not at all opposed to AI, nor very humanistic or essentailist. But I cannot see that mapping and modeling (especially if they must correlate to self-reports) somehow solve the infinite regress problems of self-reference. I would also agree that some model of "self-reference" is crucial to consciousness. But what crucially remains to be defined is the excess-something that observes that "now dead" model. Oct 29, 2015 at 18:30
  • 1
    "somehow solve the infinite regress problems of self-reference" Why is this a problem? There's nothing paradoxal about a computer going into an infinite loop and not stopping until some shuts the power off (this happens all the time and not always by accident). Oct 29, 2015 at 18:44
  • 1
    Doesn't the fact that individuals can have partial mental impairments (e.g. expressive aphasia: understands speech, but cannot effectively speak or write it) belie the idea that the mind is indivisible?
    – Dave
    Sep 22, 2016 at 19:14

The body, as an object that takes up space, can always be divided (at least conceptually)

We need to consider what this means in practise and not just as words:

Extension is a specific property of matter: I can pick up a piece of bamboo that has a certain length (or extension); and then I can break it into two pieces, one piece which I let drop to the ground and the other i take with me.

And what I can do with bamboo, I can do with rocks, with houses, planets and galaxies.

Or I can do it to the body of a dead man - a cadaver or corpse; or to a Barbary duck hunted in the long grass on some autumnal day creased with clouds.

Or to a molecule, and then the atom, and then a proton and then till we reach a true simple atom, and in itself nothing composite being pure and simple extension.

But then time - is time extension? In one way, yes: each day is divided from from another by the night, and every hour into minutes; and in another way, no: for I cannot break a day in half, except by metaphor - I cannot even hold it; what I see by outer sense is motion, is change - energia - and time is measured by motion, and by motion that moves in regular cycles - the pendulum.

And also space; for space surely has extension but unlike a piece of bamboo cannot be broken; though one can and does mark off one part of space from another - and thus the invention of territory and territorialisation - except of course this is not space itself; but what is in space: ground, earth and air; like time it seems too, we cannot grasp it; but that it is but a stage, in the sense of ground.

And in this sense, too - time; but notably Descarte took time to be an inner sense (and then Kant took space too to be an inner sense).

Now we note, that marking is prior to breaking; for to break a piece of bamboo I must break it at a specific place, though I may care not at which place I break it.

And this is true for atoms - in their true sense; for an atom as a pure extensionless point is arguably nothing; and merely marks a position and a point in space; whereas as a string - it has parts; so that it can curve, twist or close up into a loop; but merely to have extension in one direction and not another is to replay this first error; and therefore and thus branes.

Decartes argues that the mind is indivisible because it lacks extension ... the mind is simple and non-spatial

But all these simplest attributes of extension, can I do this with a mind? Is it even conceptually possible to concieve this?

Two different computers - not just in location but also in terms of hardware can run the same program; but then again the same thought can be expressed in writing and written in two different scripts in two different languages, as in:

cogito, ergo sum

I think, therefore I am

and written in many books, or as graffiti on some walls.

But we do not think of this as thought itself being split.

Perhaps, genetically identical twins are splittings of the same person? But no, they are from conception two different persons sharing exactly the same genetic material; their consciousness of being here, is seperate in each - and was never unitary as in a single consciousness or mind.

Thus your A does not hold: a thought expressed is not the thought itself; and merely places or translates the thought from the world of thought into the world of matter - and matter is essentially described by extension, and one of the properties of extension is that it can be split - so being able to split such a translated thought doesn't demonstrate the splitting of thought, but the splitting of extension.

Thus also your B does not hold; for no mind has been demonstrated in a computer; but simply that the expression of thinking can be automated; and this automation - this motion of matter - or extension in motion - is in essence no different from a thought expressed in writing in a book; and we do not think a book thinks, or that it has mind; even though we might ourselves think new things or recollect old feelings when our mind passes

Since the mind and body have different attributes, they must not be the same thing ...

Liebniz principle of indiscernability says if two things have exactly the same properties, then the two things are indiscernible (but not neccesarily the same); here we have two things which differ in some essential properties (and not accidental ones).

Therefore they cannot be indiscernible, but discernable; and generally we do exactly that - almost all thought has distinguished the animate from the inanimate .

... their "unity" notwithstanding.

Spinoza, have a conception of Unity, still did not take the world of thought and extension to be the same, but differing modes of something other.

And adopting too a principle of plenitude, argued where there is two which cannot be reduced to one; there must be more and in fact, an infinite number not discernible to us - ever and in principle.

  • "and merely places or translates the thought from the world of thought into the world of matter" Thanks - you just made me realize that multiple realizability has its roots in Platonism :-) . But then does every random neural pattern, say me thinking about "how attractive my colleague appears on Facebook compared to how she looks in real life while there's a taste of peanut butter and harissa in my mouth and Metallica is playing in the background while my boss shouts at the new guy" create a new element in the world of thought, or do you subscribe to some form of logical atomism? Oct 28, 2015 at 16:16

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .