The Evolutionary Argument against Rationalism

In the comments on my post on Armstrong on Naturalism and Empiricism Brandon and JS raised a concern about the evolutionary argument I sketched against rationalism. I have hinted at this kind of argument before but I guess I have never spelled it out. I have sort of thought it was just too obviously an extension of Hume’s kind of argument…but maybe it is worth spelling out.

Suppose that you have a stable environment E. In E there are events (e1, e2,…en) and it so happens that there is a regularity in E such that e2 regularly and reliably follows e1. Now suppose that there are creatures, C,  that live in E. Then evolutionary theory says (roughly)  that C will adapt to E via natural selection (I do not mean to be saying that C will become optimally adapted, that is a hotly contested claim). One obvious source of reproductive advantage would be being able to track the regularities in E. If e2 is such that it is (harmful or) beneficial for C then it would, through the course of natural selection, come to be the case that C  tracked the regularity, R, expressed by ‘e1 then e2’; that is to say C would come to associate e2 with the occurrence of e1. Now assume that R is some basic and (hitherto) extremely reliable regularity (like that when you have one object and you place another object next to it you then have two objects).

Now suppose that, for whatever reason, R continues to be regular and tokens of e1 are immediately followed by tokens of e2. Decedents of C will also track R, but will also have to track other regularities that they discover for themselves. So, decedents of C that have R ‘hardwired’ in will free up cognitive resources. So now suppose that C is a species from which a species evolved from which a species evolved…from which we evolved and that R has continued to be regular and reliable up till now. At this point R would presumably be something for which we are equipped to find intuitively obvious. This is exactly what developmental psychology (like work from Spelke and many others) has shown.

So, when the rationalist appeals to their experience of ‘finding something obvious’ as evidence that we know something in a special way that is revealed to be necessary and universal, it could just be the product of that regular and reliable conjunction of events has, around here and so far been that way. But that, of course, does not guarantee that it must be that way. Now how could the rationalist know which of the two theories better accounted for his experience? Our intuitions are shaped by the world that we happen to find ourselves in. And so if evolution is true rationalism is most likely false.

JS asks,

I’m also not sure evolution precludes intuition as evidence for rationalism. I suppose I just don’t think your reasoning pans out in real life. I don’t see how I can’t both agree with this: “we would expect that there would be certain truths which seem to be self-evident but are really just the product of a long process of natural selection in response to a stable environment”, and not believe in rationalism. Can’t I tell an equally nice just-so story like this: I agree with RB’s quote, but at a certain stage of complexity it’s been found that animals begin to intuit other actually self-evident beliefs, which then survive natural selection in virtue of their fitness. The end. I’m sure you disagree, however.

Well it would depend on what you mean when you say ‘animals begin to intuit self-evident beliefs’. If you don’t mean ‘discover solely by the use of reason necessary and universal facts about reality’ then we do not really disagree on anything. But if you do mean that, then the question is how could those kind of things be selected for or have any fitness at all if they did not causally interact with C or E? Sure there are spandrels and such, but in order for a property P to be reproductively beneficial it would have to play some role for C or in E. Notice that the mathematical truths play a hugely beneficial role for us. It is in fact their usefulness and indispensability to science that Quine and others have taken to show that they are empirical. David Rosenthal has given (in class) a similar argument for logic being empirical. He argued that one reason to think that it is is that it explains inferences that humans find valid and to that extent is hostage to the empirial–and so not analytic.

But maybe we did evolve to track necessary and universal facts about reality discovered by the use of reason. The point is that we couldn’t appeal to our experience of certainly as evidence for that view since we can’t rule out that the experience isn’t due to a very long association between certain events (that are not necessary). We would need some other kind of evidence. But what other evidence is there?

Advertisements

12 thoughts on “The Evolutionary Argument against Rationalism

  1. Then evolutionary theory says (roughly) that C will adapt to E via natural selection.

    Well, I take it that evolutionary theory says that either (1) the population of creatures C will as time goes on will become extinct; or (2) attributes found in the population that are favorable to survival and reproduction in E will, if hereditary, spread through the population, thus creating the general phenomenon of the adaptedness of C to E. That’s still rough, but even here I’m not sure how your argument would work. Suppose there is a regularity in E. What then? It doesn’t follow from it’s being a regularity of E that it will affect survival and reproduction of the problem; thus if we take a rationalist intuition to involve the discernment of that regularity, it doesn’t follow that there is even any evolutionary question at all unless we have good reason to think that it is a regularity having a significant effect on survival and reproduction.

    Now suppose, as you do, that the regularity is harmful or beneficial to C. It still doesn’t follow that C will actually come to track the regularity, only that those who do come to track it will (usually) have an advantage over those that don’t. If the life of a mollusk depends crucially on the cycles of the moon (e.g. by their influence on tides), and the mollusk population thrives, it doesn’t follow that the mollusk has some way of tracking the cycles of the moon, but only that any harmful effects are swamped out due to other factors or that the beneficial effects are such as to be beneficial without being tracked. Indeed, it doesn’t follow that there is any likely evolutionary path under which the regularity is capable of being ‘hardwired’ so as to become heritable.

    But suppose that there is such a path, and it is essential to track it, and that as it happens the tracking ability comes about by mutation in the population. Then it spreads. It still doesn’t follow that the regularity will come to be seen as self-evident. There are other possibilities: e.g., it might be ‘hardwired’ in such a way that all the assumption is just automatic and unconscious rather than conscious, like a sort of mysterious ability to predict what’s going to happen next; or it might be taken to be not self-evident but just very, very plausible, etc. The sun’s setting and rising is reliable, but no one thinks it is a self-evident truth that the sun will rise and set. Thus there is a gap in the explanation. The spread of an ability to track a regularity and the spread of a tendency to regard it as self-evident and universal are not the same.

    Moreover, I don’t think rationalists simply appeal to ‘finding something obvious’ ; it’s finding something obvious where that something obvious meets certain rational criteria. It’s not a mere appeal to certainty, since every rationalist knows that there are feelings of certainty that are deceptive. (Indeed, I don’t think rationalists are committed to appealing to the experience of certainty at all. But I take it that you are just arguing that when they do there is an evolutionary problem.)

  2. If we consider the survival of specific genes/organisms evolution/natural selection appears to designs things with a fair degree of precision. this is because it does actually matter over thousands of generations if you have a 1% better chance of survival. So natural selection, at least, has something to say about pretty much anything you would find significant enough to talk about (even though determining the direction of the effect might be very hard).

    Now considering evolutionary paths – it is possible that there is not a path to do something like detect the moons cycle. But I suggest that is very unlikely. the moon’s cycle (and beliefs about logic for that matter) effect a huge number of things and there is a huge number of ways they might be hard wired. For example your mollusk might tell where the moon is by detecting the presence of a plant that grows better at full moon, or average water flows or any of a million other things.

    In light of your final paragraph, and in defense of the rationalist, they may just be arguing “if I find it obvious that is some evidence for it (i.e. I generally find true things obvious) – and this becomes good evidence if it also fits with the rational criteria. That reasonably modest claim survives even if we provide a description of how the belief came to exist. (although I think that that opens up a potential avenue for evidence for or against the hypothesis).

  3. But I suggest that is very unlikely. the moon’s cycle (and beliefs about logic for that matter) effect a huge number of things and there is a huge number of ways they might be hard wired. For example your mollusk might tell where the moon is by detecting the presence of a plant that grows better at full moon, or average water flows or any of a million other things.

    I quite agree; and if the mollusk simply detects the presence of a plant, it may well know nothing, and never know anything, and never be capable of knowing anything, of the moon. There are many kinds of ability to track for purposes of survival that are not linked to cognitive apprehension of any sort. Natural selection is simply not that fine-grained: it serves such a wide explanatory purpose precisely because just about anything that has more or less the right effect will do. At the level of natural selection, it’s not even possible to pick out cognitive abilities from noncognitive ones, and therefore none can be evaluated as to truth-aptness.

    It’s worth noting that while the rationalist can’t simply mirror Richard’s argument, substituting the word ’empiricist’ for ‘rationalist’, the rationalist can nonetheless turn the tables: for however much evolutionary explanations might be taken to explain rationalist explanations, they could equally explain any bias or tendency to recede from a rationalist position. After all, from a rationalist viewpoint, rationalism is required by reason, but empiricism depends on our bias in favor the senses, which will even more clearly and straightforwardly fall under evolutionary explanation than the rationalist’s preferred form of cognition. (In fact, this is what the Cartesians actually did: if you read, say, Malebranche, he explicitly explains the survival of empiricism in the face of rationalist arguments on the basis of the fact that our sentiments our adapted not to truth-finding, which requires rational insight, but to survival and reproduction, which depends chiefly on the senses. Thus, as far as our feelings and sentiments go, we are inclined to put more trust in what we can see and touch than what we can rigorously demonstrate; and this, he argues, is absurd. It’s much like Richard’s argument, but in reverse. In fact, Malebranche makes Richard’s point in the paragraph about ‘freeing up cognitive resources’ — against empiricists.)

  4. Hey Brandon, thanks for the comments; as usual they are very helpful.

    1. I agree that there is no gauranteed path from C to us (just as I take it that there is no gauanteed path that gets us the eys) but given that there are creatures like us the question becomes what the most likely story is…in fact it is worse that. If the story is likely at all and it accounts for the subjective sense of certainty then we need an alternative argument for rationalism.

    2. I also agree that the appeal to the subjective sense of certainty is not something that the rationalist need appeal to. But I do think that it is very often appealed to (Descartes and Locke are notorius for this…) and I find it hard to see what other considerations or arguments have ever been given that do ultimately reduce to an ‘intuition’ backed by a subjective sense of certainty.

    3. I rather like the Malebranche argument. But I suspect that it is employing some very heavy-duty conception of truth. Aoccording to the empiricist the wordly fact that p makes the belief that p true, with certain qualifications obviously. So the truth of ones beliefs will be very important for reproduction and survival. What Malebranche must think (based on his Cartesianism…I don’t really know much specifically avout him) is that a truth is a necessary fact about reality, like 1+1=2 and we may not think that creatures would evolve to know this kind of thing. But the empiricist rejects the necessity of mathematics, and points out the usefulness of applied mathematics. So it is plausible that creatures who were ‘interested’ in maximizing their true representations of their environment, would, once they had acquired the necessary mental capacities and processing power, start to produces individuals who were disposed to wondering about truth, mathematics, and the limits of knowledge. Thinking of this sort has been enormously valuable to our species. So, again it comes down to the question of which of the two theories is more plausible. One theory appeals to decently understood phenomena (worldly facts) tghe appeals to…what? Supernatural entities that have no causal powers or cause things in a way that is fundamentally at odds with science. Given a choice between these two kinds of theories, and the fact that the empiricist can give an empirical account (in terms of naturalism/materialism) of everything the rationalist can, then rationalism is at a severe disadvantage.

  5. I’m not sure I get it yet Richard. I’ll grant that if evolutionary theory can offer a plausible explanation of the source of or intuitions then it is better of than the rationalist who seems to have at best a crazy explanation and at worst no explanation at all for the reliability of intuitions. I am just not convinced that your sketch of the evolutionary story makes sense to me. Consider one philosophical application of intuitions as evidence: for theorizing about modal epistemology. How do we know that, rather than being merely contingent, it is in fact a necessary truth that water is H20? We combine the empirically derived chemical knowledge of the constitution of water with our intuitions about the necessity of identity. Now, suppose for the sake of argument that all of this right… in this case the intuition is reliable and it does deliver the knowledge that it is a necessary truth that water is H20. What is the evolutionary story for the reliability of this intuition? It doesn’t look like we can track the modal facts by becoming responsive to regularities in our environment. So how come it works?

  6. I agree that there is no gauranteed path from C to us (just as I take it that there is no gauanteed path that gets us the eys) but given that there are creatures like us the question becomes what the most likely story is…in fact it is worse that. If the story is likely at all and it accounts for the subjective sense of certainty then we need an alternative argument for rationalism.

    Well, again, I’m not sure how this is supposed to work; to the extent that natural selection accounts for any cognitive ability or sense at all it accounts for all of them, and in ways that aren’t precise enough to distinguish them from any other form of tracking. (That’s related to Colin’s point, I think.) And natural selection doesn’t in itself put any limits on abilities: it merely governs how they spread through a population when by mutation they enter it. Thus the matter seems to remain what it always was: whether the rationalist (or the empiricist) is right is determined not by an etiological account of its spread through the population but by an examination of the mind’s actual abilities.

    On (2), surely rationalists make use of procedures for determining consistency and inconsistency. (That’s what the ‘clear and distinct’ part of Descartes’s clear and distinct rule is doing, for instance.) And unless we are to argue that there is no particular reason to think truth is consistent (in which case all bets are off in any case), that would seem to go beyond a mere feeling of certainty.

    So, again it comes down to the question of which of the two theories is more plausible. One theory appeals to decently understood phenomena (worldly facts) tghe appeals to…what? Supernatural entities that have no causal powers or cause things in a way that is fundamentally at odds with science.

    Perhaps, but that’s not the only factor relevant to plausibility. Malebranche has long arguments arguing that empiricists cannot account for certain features of human intellectual life at all. (Thomas Nagel’s The Last Word is a sort of modernized version of this type of argument.) So from a Malebranchean rationalist perspective the real options are an appeal to sensible facts (for which we have a survival-and-reproduction argument that we are for biological reasons over-inclined to give too much weight to) that cannot explain everything it is supposed to explain, and an appeal to intelligible facts that (like serious scientific facts) take an immense amount of intellectual work to get clear about (for the same reason we tend to give too much weight to sensible facts) but that explains everything. That is, from a Malebranchean perspective the question would really be: Which is more plausible, a form of explanation based largely on bias that doesn’t explain everything it is supposed to, or a form of explanation, however difficult it may be, that explains everything and makes it possible to go beyond empiricism to do advanced science and mathematics? And in that sense I’m rather worried that what really is going on is that both sides are simply building refutations of the other side that work on their own principles but not on the other side’s.

  7. hey Colin, thanks for the question.

    We know that identities are necessary (if we do) because we can prove that they are. here is a simple proof (given by kripke and many others)

    1 (∀x)(∀y)(x = y → (Fx → Fy) )

    2 (∀x)□(x = x)

    3. (∀x)(∀y)(x = y →( □(x=x) →□(x=y) ) ) [from1 and 2].

    4 (from 3) (∀x)(∀y)( x = y→□(x=y) )

    premise one is Leibenz’s law that says if two objects are identical then they have all the same properties, and premise two is the closed axiom of identity that says that every object is such that it is necessarily self-identical. Given these two premise it follows that identites are necessary. But according to the empiricist (1) and (2) are themselves empirical hypotheses and so could ultimately be rejected or reformed or whatever. So, the empiricists rejects any strong sense of neccessity. So I wasn’t arguing that an evolutionary account could explain how we could actually come to know necessary facts about reality, but rather it would explain why it would seem to us that we know neccessary facts about reality that aren’t really necessary.

    Brandon, you say

    To start with you last point first. I think you know that I am very sympathetic to that kind of claim. But there are well worked empircal accounts of these things that that people claim can’t be captured by empiricism. They, obviously, do not account for all of the intuitions that rationalist have about neccessity and such, but we have a nice explanation for why it would seem to us that these things are necessary when they really aren’t. even so empiricism has the advantage since it gives us all the benefits of rationalism without all of the baggage. Whether mathematical truths are neccessary or not surely doesn’t impact the fact that they are enormously useful to us. Who cares if they are contingent? So we CAn explain everything from an empirical point of view. the burden is then on the rationalist to motivate his program over ours. The only argument I have ever seen is an appeal to intuition, but that’s exactly what I am trying to target.

    So, in response to (2) you say

    “On (2), surely rationalists make use of procedures for determining consistency and inconsistency. (That’s what the ‘clear and distinct’ part of Descartes’s clear and distinct rule is doing, for instance.) And unless we are to argue that there is no particular reason to think truth is consistent (in which case all bets are off in any case), that would seem to go beyond a mere feeling of certainty.”

    But I don’t think that is quite right. Descartes says that something is clear if it is ‘manifest to an attentive mind’. He defines clearness in terms of what I have been calling the subjective experience of certainty. Distinctness is the property if an idea that it has when it contains ‘nothing but that which is clear’. So, I don’t see how you think that Descartes is doing anything other than what I am targeting. the stuff about consistency is important, but the familiar problem with that is that all kinds of systems are consistent. In particular, and a propos this discussion, both rationalist and empirical epistemologies are consistent. So we need something else besides consistency.

    As for the first point about how this is supposed to work in detail. I have to admit that all of this is a bit hand wavy, as are all evolutionary explanations. But isn’t it reasonable to assume that a creature that can represent propositions (that is represent worldly facts as worldly facts) is better off than one that cannot? This, for instance, will allow the creature to notice that, sometimes, the way that the world is is not the way that the world appears to the creature. Creatures that are better able to accurately represent the external world will do better than creatures that have less accurate representations of the world. This will provide some evolutionary pressure towards creatures like us who have developed complex representational systems. Of course once one starts to talk this way one runs head first into the problem that Jerry Fodor has been talking about for the last few years.

  8. Even so empiricism has the advantage since it gives us all the benefits of rationalism without all of the baggage. Whether mathematical truths are neccessary or not surely doesn’t impact the fact that they are enormously useful to us. Who cares if they are contingent? So we CAn explain everything from an empirical point of view. the burden is then on the rationalist to motivate his program over ours.

    But the rationalist has no reason to concede any of this: whether mathematical truths are enormously useful to us doesn’t affect whether rationalists are right (and empiricists wrong) about necessity, infinity, and the like. Thus from the rationalist point of view it doesn’t seem that you can explain everything from an empirical point of view, unless you deliberately ignore facts (and being able to explain everything if we ignore facts is not really an advantage for an explanation!). Thus from the rationalist point of view the burden is on the empiricist to motivate his rejection of the facts to which the rationalist points. (You seem to be assuming that we all start from the empiricist side, and then the question is just: why should we go farther, into rationalist territory? But, of course, most rationalists won’t concede that empiricism is, or even can be, the default position.)

    Clarity, as you note, is for Descartes manifestness to an attentive mind; but manifestness to the mind is not a feeling of something’s being obvious. The analogy Descartes always gives is of taking something in hand, in good light, and examining it closely. Thus a clear and distinct idea is one which you are examining closely in such a way that you are not (a) overlooking anything; and (b) confusing anything with anything else. And it is clear, I think, from some things that Descartes says that he regards this as important in part because he regards it as a procedure for rooting out inconsistency. I don’t deny that Descartes puts a hefty emphasis on subjective experience of obviousness, but I think there is more to the clear and distinct criterion than that. But I’m not really trying to argue that Cartesians or other rationalists are not guilty of arguing on the basis of subjective experience; it was just an example of the point that they needn’t be seen as always arguing solely on this basis.

    I agree that creatures that can form propositions and remember them will in general be more flexible in their responses to their environments than those who can’t. But all this shows is that if a proposition-forming creature shows up in sufficient numbers with the ability to reproduce, it will, all other things being equal, increase rather than go extinct, and, in cases where that flexibility becomes very important to survival, creatures of that kind will tend to dominate the population. Natural selection is an explanation about how things become widespread; it doesn’t tell us about how they started in the first place or what their nature is. (These are studied not by studying natural history but by studying things like physiology, genetics and mutations, etc.) Very different things — e.g., the sort of empiricist mind you are proposing and the sort of rationalist mind the rationalist proposes — can perform the same evolutionary function; and because of this we cannot distinguish the two by way of their function in evolutionary history.

  9. Hi Brandon, I posted some thoughts inspired by your remarks in a new post.

    I should say that I agree that the rationalist will see things in the way you describe. They will think that the empiricist hasn’t explained certain facts. But the question is; why think that they are facts in the first place? The empiricist has an account for why it would seem to us that they were but actually weren’t. That this is conceivable shows by the rationalist’s own critereon that there are no necessary facts. It is a priori conceivable that empiricism is be true and so that naturalism be true. Of course some rationalist will argue that one of the two options must be only epistemically possible (i.e. either it is necessary that there are numbers or it is necessary that there are not numbers). But I have never seen an argument to the effect that I cannot coherently conceive that numbers are contingent existents. This is of course because no one has a well worked out theory of a priori knowledge.

    It is hard for me to see what else Descartes could have in mind when he compares clearness with taking something in hand besides pointing out that those criteria for consistency work because they are experienced as being reliable. How else could I know that I am not missing anything confusing anything if I weren’t relying on my subjective sense of certainty? So maybe they don’t rely on it explicily…but everything they do rely on is supported by that, and as far as I can tell, only that.

    Finally, I must say I completely disagree with your last statement. An evolutionary theory must posit that whatever speads throughout a population does so because it is causally efficacious. Creatures who evolved in order to know things about their environment would have to causally interact with the things which were known. So only empiricist minds could have evolved.

  10. There are real life examples of regularities similar to e1, e2 such as the existence of the sun that call into question your assertion that evolutionary-based truths are incompatible with the concept of rationalism since these truths are not exactly expected to remain ‘truthful”.

    Human beings across all cultures and epochs have now evolved to regard the sun and its effects as intuitively obvious, however that doesn’t negate the rational aspects of the cosmic forces that make the sun behave the way it does, nor does it preclude our evolutionary-based belief in the sun from being a rational one.

    The events or patterns associated with e1, e2 as with the sun are only part of a much larger scheme of rationality that encompasses infinite manifestations of micro instances of rationality.

    Your assertion here is not unlike Emmanuel kant’s assertion that our experiences are not to be trusted to reveal to us the true nature of reality since we have no other way of corroborating them, besides our own senses. My answer to kant was even simpler; What does kant rely on to even offer up this skepticism? Obviously, he must rely on his own senses. Thus, his criticism, even to him, would be effectively null and void.

    In your post, there is a similar skepticism of the rational nature of our experiences, but unlike kant who attacked their limitations, you simply attack the imperfect nature of the process of evolution (through which our experiences are formed) that equips us to associate intrinsic rationality and truth to events that may prove to be otherwise.

    Evolution’s main function is survival, it is not rationality. Logic and rationality are only some of the tools which species have made use of to gain an edge during their evolutionary trek, the same way some species like humans have made use of such tools as empathy to gain an edge over others.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s