Because of my long-standing interest in arguments for and against the existence of a nonphysical soul, mind, or spirit (I’m decidely in the ‘against’ camp), I’m happy whenever I run across a new argument to consider. One such argument comes from Joshua Rasmussen, professor of philosophy at Baylor University.
He describes it in the video above. The basic idea is that there are more logically possible thoughts than there are physical states available to represent them, which means that the mind cannot be physical, or at least not solely physical.
I think the argument fails, but I’ll save my explanation for the comment section.
I am underwhelmed by that argument.
The biggest problem with his argument is that he’s asking the wrong question. He’s basically asking
The answer is yes, and he takes that as proof of his thesis. But the question he should be asking is
There are many logically possible thoughts that are not humanly possible. For instance, consider a set consisting of a trillion numbers, selected randomly. One logically possible thought about that set is
…with each of the numbers named explicitly. No human can have that thought. It’s logically possible, but not humanly possible. There are many other thoughts like it.
What Rasmussen needs to do is to somehow quantify the set of humanly possible thoughts and show that it exceeds the set of available physical states. He hasn’t done so.
keiths,
Seems fairly straightforward. For his thesis to be correct, the human mind must be inconceivably vaster than it has been known to be, capable of grasping a near infinity of statements. I guess I would agree with his ideas if the human mind weren’t so limited. As our political threads demonstrate, some human minds seem incapable of understanding very simple things. To paraphrase Upton Sinclair, it is very difficult to get a man to understand something when his religious faith depends on his not understanding it. Again with Richard Dawkins, there is no sensible limit to what the human mind is capable of believing, against any amount of contrary evidence.
Flint:
In fairness to Rasmussen, it doesn’t matter whether all of the thoughts are correct. Confused thoughts are still thoughts, and they get counted regardless. “Mother Teresa was a turnip” is a logically possible thought, albeit a false one, and so it counts as part of the set that Rasmussen is trying to quantify.
The problem for him is that there are possible thoughts, both true and false, sensible and nonsensical, that are beyond humans’ ability to even have.
Thoughts aren’t sitting around in our minds ready to be thought: they only come into being when we think them. The potential to have an inconceivable number of thoughts exists in principle, but in actuality we only have the set of thoughts that are part of our stream of thoughts that we actually have. The mind makes thoughts, it doesn’t pull them out of a pre-existing set to choose from.
These are just a few quick thoughts of mine, which illustrate my point, I think. I’ve never thought these thoughts before, but I doubt they’ve been sitting around in my mind as a pre-existing neuronal state that iI’ve just never “seen” before. 🙂
aleta:
All true, but that doesn’t affect the argument, which is about possible thoughts, not actual thoughts. My “Mother Teresa was a turnip” example only became an actual thought when I wrote that sentence, but it was always a possible thought.
An analogy: A particular 64-bit register in my computer may never have contained the hexadecimal number 30904BFF11736191 in its entire existence until today, when it suddenly does. That only became the actual value just now, but it was always a possible value.
The idea isn’t that the thought “Mother Teresa was a turnip” was a pre-existing neuronal state somewhere in our brains that we had never “looked” at until now — just that it was a potential neuronal state (or states) that our brains had never taken on until now, just as the aforementioned computer register hadn’t taken on the stated value until now.
I think your analogy undermines your position. Human minds are not digital like a cpu register; they can have indeterminate states that might more-or-less have implications of many concrete ideas. So does the human mind have the capacity to articulate every idea, clearly enough to distinguish it from a whole class of related ideas? CPU registers hold static values, but our minds don’t click from one static state to another, they are not state machines. I think we need a very clear concept of an “idea” before we can count the number of ideas that can dance on the head of a neuron.
Flint:
The analogy isn’t “brains are digital just like CPU registers”, it’s “there are many possible brain states that aren’t actualized, just as there are many register states that aren’t actualized.” Rasmussen’s argument is about all possible states, not just actualized ones.
I should probably also emphasize that all physical systems have states, regardless of whether or not they are digital systems. Icicles and petunias have physical states despite not being digital. This is relevant because there may in fact be infinitely many possible brain states, and we have to consider that possibility when evaluating Rasmussen’s argument.
One of the strengths of Rasmussen’s argument, which is unfortunately not enough to save it, is that it is agnostic about how thoughts are implemented in brains (or would be implemented, since he doesn’t accept physicalism). All he’s depending on is the fact that if physicalism is true, distinct thoughts must correspond to distinct brain states.
Rasmussen’s description of Cantor’s theorem and how it applies isn’t the best, so let me flesh it out a bit.
First the set theory terminology:
1. Sets — A set is just an unordered collection of things, usually denoted by putting curly braces around a list of the things contained within it, which are known as ‘elements’. So {Cuba, Venezuela, Canada, Iran} is a set, and {Venezuela, Iran, Canada, Cuba} is the same set, since the order of the elements doesn’t matter. There is a set with nothing in it known as ‘the empty set’ and denoted as {}.
2. Sets can be nested; that is, sets can contain other sets. {3} is a set, and so is {7,9}, and so is the set containing both of them: {{3}, {7,9}}. Note that this last set is not the same as {3,7,9}, which is not nested. The structure matters.
3. Subsets — Set B is a subset of set A if all of the elements of B are also elements of A. The set {cat, dog, quokka} has eight subsets:
{cat}
{dog}
{quokka}
{cat, dog}
{cat, quokka}
{dog, quokka}
{cat, dog, quokka} (because every set is a subset of itself)
{} (because the empty set is a subset, which sounds strange at first but actually makes sense since all zero of its elements are contained in {cat, dog, quokka}).
4. Power sets — The power set of A is the set containing all of the subsets of A. In the case of {cat, dog, quokka}, the power set is
{{}, {cat}, {dog}, {quokka}, {cat, dog}, {cat, quokka}, {dog, quokka}, {cat, dog, quokka}}.
5. Cardinality — The cardinality of a set is, informally speaking, its size — the number of elements in it. (There can be infinitely many, which is where things get tricky.) {head, toe} has a cardinality of 2, because there are two elements in it. {} has a cardinality of 0, since there are no elements in it. The set {{bison, buffalo, moose},{chocolate,vanilla}} has a cardinality of 2, because it contains 2 elements: the set {bison, buffalo, moose} and the set {chocolate, vanilla}.
6. Cantor’s Theorem — Cantor’s Theorem, which Rasmussen’s argument depends upon, simply says that the cardinality of a power set is always greater than the cardinality of the set itself. The set {cat, dog, quokka} has a cardinality of 3, but its power set (shown in #4 above) has a cardinality of 8. Cantor’s Theorem applies even to infinite sets, and it also applies to the empty set: The cardinality of {} is 0, because there are no elements in it, but its power set is {{}}, which contains 1 element: the empty set {}. 1 is greater than 0, so Cantor’s Theorem holds.
Now let’s weave Cantor’s Theorem into Rasmussen’s argument:
1. For every physical system, there is a set of possible physical states.
2. For every physical state, there is at least one possible thought about that state.
3. But we can also think about sets of states. Suppose we’re talking about a system with three possible states: S1, S2, and S3. Not only can we think about each state individually; we can also think about combinations of them.
4. Another way of expressing that is to say that we can think about {S1}, {S2}, and {S3}, but we can also think about {S1,S2}, {S1,S3}, {S2,S3}, {S1,S2,S3}, and {}. The pattern should be familiar — those are all of the subsets of {S1,S2,S3}. Together they form the power set of {S1,S2,S3}.
5. There is a possible thought about each of the subsets, and there are 8 of them, but there are only 3 physical states. Possible thoughts outnumber states, and by Cantor’s Theorem, this will always be true, no matter how many physical states our system has.
6. If thoughts correspond to brain states, then each possible thought must map to a distinct brain state (or a distinct set of brain states).
7. Because of #5, there aren’t enough brain states to go around. The possible thoughts outnumber the possible brain states.
8. Therefore thoughts can’t be purely based on brain states. The mind must be nonphysical, or at least not solely physical.
Rasmussen’s main error, as I pointed out above, is that he is implicitly assuming that all possible thoughts are humanly possible. That’s clearly not true, as my trillion-number example shows. The question isn’t whether the logically possible thoughts outnumber the physical states, it’s whether the humanly possible ones do, and he hasn’t shown that.
I just can’t picture the mind as having states as implied in this argument. Thoughts flow like rivers, in constant flux, perpetually incomplete, mostly undefined. If we could take a brain snapshot of the exact state of every neuron at some instant of time and consider that inconceivably complex picture a “brain state”, the very act of doing so is to deliberately miss the essence of what’s happening, the notion of process and movement and change. Counting the number of “states” is like counting the number of values between 0 and 1.
Flint:
Rivers have states, too, just like brains. All physical systems have states. At some point in time the swinging pendulum is at an angle of exactly 11.4° from the vertical, right? In other words, at that instant it is in that state.
The pendulum is in constant flux too, but that doesn’t mean that it isn’t progressing through states. There is an instant of time at which the angle is exactly 11.4°, after all. The system passes through that state (and many others).
“Mother Teresa was a turnip” seems like a complete thought to me. I’m not still working on that thought — I finished it a while ago.
Would you say that “4 + 1 = 5” is an undefined thought?
Capturing a snapshot of the pendulum at one instant of time, in one state, doesn’t imply that the pendulum isn’t swinging, so why should capturing a snapshot of a brain imply that it isn’t changing?
That’s why I commented earlier that
The question is whether nature is discrete or continuous at the relevant levels and in the relevant respects, and whether the brain therefore has finitely many or infinitely many states. I haven’t worried about it much here, because it doesn’t matter for Rasmussen’s purposes. Cantor’s Theorem applies even to infinite sets, so Rasmussen can press his argument regardless.
I think this is the original paper in which Rasmussen presents the counting argument:
Building Thoughts from Dust: a Cantorian Puzzle
I’ve only skimmed it so far, but it appears that he doesn’t address any objections of the kind I raised above.
Your notion of a “state” is so expansive and inclusive that there can’t be anything that is NOT a state. Even the process of transitioning from one state to another, in your definition, might encompass an indeterminate number of intermediate states, but somehow this infinity of transitional conditions are all “states”.
This is why I said earlier that we need a much more specific notion of what an “idea” is, to know if, say, incoherent dreams are streams of “ideas” or instead just random neuron firings. You can, of course, regard time as being composed of an infinite number of discrete instants. I doubt time actually works that way, but like Trump, you can say it does because you SAY it does. Your entire argument rests on assumptions I don’t share.
Flint:
Sure there can. Plenty of things aren’t states, including toasters, wetness, money, and movie plots.
Correct. If time and space turn out to be continuous and not discrete, then there can indeed be infinitely many states (uncountably infinite, in fact) in between any two states, just as there are infinitely many real numbers between any two real numbers.
I know that as a fellow digital guy, you’re used to thinking of a state as something that persists over time, as in the case of a circuit in which the state changes on clock edges but remains static in between. That isn’t the only kind of state. Instantaneous states are a thing too, and physicists and philosophers deal with them all the time.
Haven’t you ever solved a physics problem that read something like “At t=0, an artillery shell is headed upwards at an angle of 45°. It’s speed is 500 m/s…”? Those numbers are specifying its instantaneous state at time t=0.
Yes, they’re all states. Take my pendulum example. If the angle is 11.4° and the bob is moving at 5 meters per second at time t, then those are part of its instantaneous state at time t. Both of those quantities will vary continuously if it happens to be the case that time and space are continuous and not discrete.
That isn’t necessary for Rasmussen’s purposes. His argument only requires the assumption that every physical state can be matched with a possible thought, with no physical states left over.
I don’t say that it does. That’s just one possibility.
That isn’t an assumption of mine and my argument doesn’t rest on it. Nor does Rasmussen’s, as far as I can tell. Cantor’s Theorem applies to both finite and infinite sets, so that part of his argument works either way.
In the Rasmussen paper I linked to above, there is a section in which he deals with possible objections to his argument. The objection most related to the one I’ve been raising is this one:
The conclusion doesn’t follow. Nothing about identifying a physically unrealizable thought undermines the thesis that there are more possible thoughts than physical states. I don’t know where Rasmussen is getting this.
It doesn’t read as an objection at all. It reads as supporting Rasmussen’s position. Rasmussen wants to show that not all possible thoughts can be realized physically and that physicalism is therefore false. The counting argument is one way to do that, but identifying a specific possible thought that is unrealizable physically is another way. The example he gives above is of such a thought, which should make him happy, not strike him as an objection.
In addressing the objection, he writes:
Demonstrating that there are thoughts that can’t be physically realized is the whole point, so identifying such a thought is good for him, not something he has to make excuses for. He should be saying “Not only does my counting argument show that possible thoughts outnumber physical states; I’ve also identified a particular possible thought that cannot be represented physically.”
Now, his overall argument still fails because he’s seeking support for the wrong thesis — namely, that logically possible thoughts outnumber possible physical states. He should be asking that question about humanly possible thoughts, not logically possible ones. But given the thesis that he’s (mistakenly, in my opinion) trying to validate, the objection doesn’t strike me as an objection at all. It’s perfectly compatible with the thesis.
The argument that Rasmussen presents in the video aims to show that logically possible thoughts outnumber physical states, thus undermining physicalism. (It doesn’t actually undermine physicalism, for reasons already given, but that’s what he’s aiming for.) However, showing that thoughts outnumber states only proves that at least some possible thoughts can’t be grounded in those states, not that none of them can.
In an article on Substack, Rasmussen presents an expanded version of his argument in which he attempts to go a step further and show that it isn’t just some possible thoughts that can’t be physically grounded, it’s all of them.
The basic idea is this: There is a principle he calls “Uniformity” which asserts that thoughts are all fundamentally the same, meaning they all must be grounded in the same way. If one thought is physically grounded, they all are. If one thought can’t be physically grounded, then none of them can.
Here’s his analogy:
If not all possible thoughts can be grounded in physical states, which is what the Cantorian part of his argument attempts to show, then by Uniformity, none of them can and physicalism is false.
Take a brain, with all of its possible physical states. Assuming physicalism is true, then for every physical state there is a corresponding thought. Multiple physical states may map to the same thought, but multiple thoughts can’t map to the same physical state. A physical state can only give rise to (to ground, in other words) a single specific thought from among the possibilities.
The counting argument proper shows that you run out of physical states before you run out of thoughts. The leftover thoughts can’t be physically grounded, because there are no physical states left to ground them. By Uniformity, that means that none of the thoughts can. Every thought must be grounded in something nonphysical. Physicalism is false.
The problem with this argument is that it doesn’t show that there is something fundamental about thoughts — something in their very essence — that prevents some of them from being physically grounded. Instead, the reason those particular thoughts can’t be grounded is that the brain you’re trying to ground them in has a limited capacity, and the capacity limitation is based on the size and nature of the brain. A thought that can’t be grounded in a chimpanzee brain might very well “fit” into a human brain. A thought that is out of reach for Donald Trump, and there are many, might very well fit into a smarter person’s brain. The thing that determines whether it fits or not isn’t the nature of the thought itself — it’s the capacity of the physical system you want to ground it in.
Unless Rasmussen can show that there are thoughts that can’t be grounded in the states of any physical system, his argument doesn’t work. Put differently, Uniformity may very well be true, but if it is, the common trait among possible thoughts isn’t “cannot be physically grounded in system X”, where system X is some specific system. Instead, the common trait is something else, and one candidate is “can be grounded in principle by a physical system of some kind.” He hasn’t ruled that out.
The above reasoning gives two cases in which Rasmussen’s argument fails to prove that there are no thoughts at all that can be physically grounded: 1) Uniformity is true, but the common trait isn’t the one Rasmussen needs it to be, as I’ve explained above; or 2) Uniformity isn’t true in the first place, in which case we don’t need a common trait. In that case there’s nothing inconsistent about some thoughts being physically groundable while others aren’t.
To summarize: the fundamental nature of a thought is what should determine whether it can be physically grounded, not the capacity of the particular physical system you’re trying to ground it in.
While looking into the counting argument, I signed up for Rasmussen’s mailing list. Today he presented a new argument under the title You Can Know There Is Life After Death — Without Dying.
The argument goes something like this:
Premise 1:
Premise 2:
Conclusion:
I’m quote mining for the sake of laughs, of course, but it’s actually not that much of an exaggeration to say to say that the argument boils down to the above. Those are the actual examples Rasmussen uses to make his points, and here is the image that accompanies the article:
Rather than discussing it further in this thread, I’ll start a new thread to keep the two arguments separate.
My firm belief is that matter is non-physical.
That was easy.
And glib.
And just as useful as any philosophy.
Where my personal view differs from the sim hypothesis is, I believe reality is fractal, and there is no smallest unit.