Tuesday, July 11, 2006

Balfour and the Evolutionary Argument Against Naturalism

The Argument from Reason did not originate with Lewis. Something like it can be traced all the way back to Plato, and Augustine had an argument that said that our knowledge of eternal and necessary truths. Descartes maintained that the higher rational processes of human beings could not be accounted for in materialistic terms, and while Kant denied that these considerations did not provide adequate proof of the immortality of the soul, he did think they were sufficient to rule out any materialist account of the mind. However, naturalism or materialism as a force in Western thought did not become really viable until the 1859, when Charles Darwin published the Origin of Species.
The earliest post-Darwinian presentation of the Argument from Reason that I am familiar with, and one that bears a lot of similarities to Lewis’s argument, is found in Prime Minister Arthur Balfour’s The Foundations of Belief. Lewis never mentions The Foundations of Belief in his writings, but he does say in one place that Balfour’s subsequent book Theism and Humanism is “a book too little read.” According to Balfour the following claims follow from the “naturalistic creed.”
1) My beliefs, in so far as they are the result of reasoning at all, are founded on premises produced in the last resort by the ‘collision of atoms.”
2) Atoms, having no prejudices in favour of the truth, are as likely to turn out wrong premises as right ones; nay, more likely, inasmuch as truth is single and error manifold.
3) My premises, therefore, in the first place, and my conclusions in the second, are certainly untrustworthy, and probably false. Their falsity, moreover, is a kind which cannot be remedied; since any attempt to correct it must start from premises not suffering under the same defect. But no such premises exist.
4) Therefore, my opinion about the original causes which produced my premises, as it is an inference from them, partakes of their weakness; so that I cannot either securely doubt my own certainties or be certain about my own doubts.
Balfour then considers a “Darwinian rebuttal, which claims that natural selection acting as a “kind of cosmic Inquisition, will repress any lapses from the standard of naturalistic orthodoxy. The point was made years later by Antony Flew as follows:
[A]ll other things being equal and in the long run and with many dramatic exceptions, true beliefs about our environment tend to have some survival value. So it looks as if evolutionary biology and human history could provide some reasons for saying that it need no be a mere coincidence if a significant proportion of men’s beliefs about their environment are in face true. Simply because if that were not so they could not have survived long in that environment. As an analysis of the meaning of ‘truth’ the pragmatist idea that a true belief is one which is somehow advantageous to have will not do at all. Yet there is at least some contingent and non-coincidental connection between true beliefs, on the one hand, and the advantage, if it be an advantage, of survival, on the other.
However, Balfour offers this reply to the evolutionary argument:
But what an utterly inadequate basis for speculation we have here! We are to suppose that powers which were evolved in primitive man and his animal progenitors in order that they might kill with success and marry in security, are on that account fitted to explore the secrets of the universe. We are to suppose, that the fundamental beliefs on which these powers of reasoning are to be exercised reflect with sufficient precision remote aspects of reality, though they were produced in the main by physiological processes which date from a stage of development when the only curiosities which had to be satisfied were those of fear and those of hunger.
Interestingly, Balfour’s argument here finds surprising support from Darwin himself. In a letter to William Graham Down, Darwin wrote:
the horrid doubt always arises whether the convictions of man's mind, which has been developed from the mind of the lower animals, are of any value or at all trustworthy. Would any one trust in the convictions of a monkey's mind, if there are any convictions in such a mind?
As can be seen Balfour’s presentation of the argument, and his consideration of counter-arguments, anticipated much of the debate on this issue that is still going on a century after his book was written.

20 comments:

Blue Devil Knight said...

These discussions, posed in terms of beliefs, are sort of abstract. We could ground them in actual data from the literature on sensory coding, which shows that animal sensory systems are extremely accurate indicators of what is happening in the world. Bill Bialek, at Princeton, has done a lot to get people to think about this more quantitatively.

The leech, for instance, which predominantly uses its sense of touch (via mechanoreceptors) and chemical senses (to sense blood and fellow leeches to mate with), is a good model system. For instance, its mechanoreceptors are exquisitely accurate indicators of where it has been touched: for the leech this is of paramount importance for its survival.

Such accuracy is observed in sensory systems in all phyla that have them. It is not all that hard to imagine how such highly accurate sensory systems evolved. Also, since at least some of the content of our beliefs is fixed by the information carried by our sensory systems, since the latter are very good at tracking features of the world, wouldn't it be strange if our belief box (which in all likelihood depends on properly funcioning sensory systems) had a dramatic disconnect from this feature, suddenly became untethered from the reliable world mirror provided by our sensory systems?

I think the problem with the way Victor poses the debate is that it is too Cartesian: How can I justify my belief system? Since Quine and Sellars, and the naturalistic turn more generally, it has become clear that it is more fruitful to ask empirical questions about other systems, and the resulting methods/results typically extend in fairly straightforward ways to ourselves.

You could say, "How do you know the stimulus you gave to the leech? How can you trust the voltage signal you recorded in its sensory neurons?" I.e., you could take the skeptical turn and ask me to essentially justify everything, or even try to get me to justify the existence of the external world. This Cartesian epistemic stance is possible, and I can't prove that I am not presently dreaming. This doesn't particularly bother me. Here on Neurath's raft, where we are all situated, none of us can prove such things, so I assume you aren't claiming that naturalists need an answer to radical skepticism to be taken seriously.

JD: the theist, indeed, has the ultimate free parameter to explain anything. I'm not sure this is a good thing, or conducive to truth discovery. It certainly isn't conducive to truth discovery in science, where we try hard to kill our pet theories, and to take seriously theories that can be shown false.

Blue Devil Knight said...

Here is a link to a PDF of one of Bialek's classic papers. I actually wrote a review article comparing two approaches for quantifying how well a neural response tracks what is going on in stimulus space. It is here.

Steven Carr said...

'"For Descartes human reason was a faculty given to us
and guaranteed by God, no less, and that was why he
could rely on it to tell us about the essential nature
of mind and matter, and a good deal else besides'

How did Descartes know that his reasoning about this was true?

Steven Carr said...

'But will this improvement of responses result in the ability to achieve true insight into "the way things are"? Does the ability to see in infrared imply the ability to write down Maxwell's equations of electromagnetism?'

Is this why people do not rely on their innate abilities when it comes to answering exam questions on Maxwell, but spend years in a university studying?

The idea that , under naturalism, people's knowledge is limited to what is selected for is such utter junk that I wonder about people who are prepared to write such things in public.

Blue Devil Knight said...

I'm relieved we can all agree that sesory systems carry information (i.e., are accurate indicators) about what is happening in the world. I really didnt' want to waste time dealing with a general skeptic.

If our belief box depends on our sensory systems, that gets us a lot. It suggests the empiricist has something going for him. Beliefs evoked by our sensory systems are more reliable than those generated via complicated chains of inference which, though grounded in perception-based knowledge, are not tethered to it. This realization, which Kant really drove home, is the core of modern empiricism.

Whether you are an empiricist or not, facts about sensory systems are not philosophically inert. They may seem trivially true (to anyone who knows biology), but that doesn't mean philosophers have incorporated such facts into their worldview . Indeed, as recently as 1996, the philosopher Akins argued that we shouldn't look at sensory systems as providing information about the world: this article was very popular and published in J. Phil.. In other words, our knowledge that sensory systems really does provide accurate information about the world is fairly recent and has been recognized to have implications for epistemology and philosophy of mind (in addition to Akins incorrect piece see Dretske).

In reasoning about the structure of the world, the further we get from experience, the more we rely on complicated chains of inference in ordinary language, the more likely we are to get things wrong, to become a bit too enamored of their beautiful chains of logic. Philosophy is less successful than science partly because its practitioners don't struggle like we do to force their theories to make contact with experiment.

All of this success suggests the empiricists, and their naturalist cousins, were on to something.

As for naturalism itself, I don't think that is a claim like "I expect this neuron to fire more action potentials than that neuron." There is no such "crucial experiment" for naturalism. However, the alternative has simply failed as a hypothesis. Antinaturalism has all the hallmarks of a degenerative research program, naturalism of a flourishing productive research program. Anything else that had so consistently been shown to be useless in so many fields would evoke only scorn among scientists. And it does, among most. The Christian materials scientist would balk at you if you said that the shear in a beam can only be explained by god. Gods didn't used to be scorned in science: a telling fact. It fell out of favor because of its complete uselessness. Naturalism in science is not a dogma: it has been verified historically and people have just forgotten that it used to be one alternative among many. THis is unfortunate, as it gives license for people like Philip Johnson to say that it is simply an unjustified dogma, and people don't know he is utterly full of shit.

Incidentally, perhaps you could say utility isn't a guide to truth. I think it is, especially when arguing about our beliefs about the world. But that is another topic.

As a final thought, if someone's belief in god is based on gaps in our present scientific knowledge (e.g., our inability to explain why dogs turn around in circles when excited), I am worried about his faith. There are much better reasons to believe in god, and not a whole lot of worse reasons.

Again, working on this grant will probably prevent me from making any more comments here: I always think I will only take a minute to compose these posts but it ends up chewing an hour out of my day. But I will be sure to read any comments.

JD: I mentioned Bialek because I knew you were there. He is a fun and interesting guy.

Blue Devil Knight said...

OK, that rant against supernaturalism was a bit over the top. Riding into work I realized it. I think I am basically right, but I don't want to give the impression that only fools believe in gods. Naturalism/nonnaturalism are both worldviews that smart, reasonable people hold (though I do think the evidence, inductive and historical in nature, against nonnaturalism is overwhelming.) I do think 'god of the gaps' theists tend to be naive theologically and on thin ice, though.

OK, I will really try to focus on my grant. I rarely suceed in my attempts to steer clear of fun and explosive arguments on blogs.

Anonymous said...

BDK, really, I can't understand why you bother.

You gave a citation to the primary literature and JD has said -- in as many words -- that evidence about how neurological systems form beliefs is irrelevant to his view of how neurological systems form beliefs.

If you wanted a declaration of willfull, invincible ignorance, you wouldn't find it more plainly stated than that. Evidence is as irrelevant to the supernaturalist as steakhouse menus to the vegetarian. You're not dealing with someone who upholds the same conversational morality as you.

Blue Devil Knight said...

Good to see some other people posting. Before I call it a night, I thought I'd mention that I think one key to our scientific abilities, in things abstract and unobservable, is the power of public symbol systems like natural language (spoken and especially written) and mathematics.

While our inbuilt reasoning systems may well be evolutionary kludges, these public systems allow us to disengage from the cognitive systems to some degree, and to groom argumentative practices so that we make a lot fewer mistakes. E.g., people are really bad at understanding conditionals (believe me, I TAed logic a few times, and some people just couldn't grasp the conditional). By creating this public space of reasons and rules their use, rules that try to weed out error-propagating patterns of symbolic chains and preserve truth-propagating chains (e.g., 'if it is raining, then there are clouds', 'there are clouds', does not license you to assert 'therefore, it is raining'). With such public systems, we have made amazing strides. It is only via a ridiculous amount of study and practice that people get good at manipulating such symbols 'in their heads.' Just look at all the scratch paper people use in doing proofs.

I think that philosophers tend to overestimate the importance of private thinking, and have only just begun to appreciate the cognitive virtues of public cognitive prosthetics like language.

For a pretty good discussion of this stuff, Andy Clark's book 'Being There'. In Cognitive Science, Ed Hutchins' 'Cognition in the Wild' is an underappreciated empirical study of the cognitive processes at work in the navigation in naval vessels.

Blue Devil Knight said...

Think about it this way: if a group of people with poor eyesight work together, will that enable them to see any better?

If there are any scientists, they will make glasses.

Blue Devil Knight said...

I wouldn't put it in terms of 'transcending' evolution, though I understand what they are saying. For one, through public symbol systems, we have an extremely efficient means for transmitting knowledge via nongenetic means. An epistemic watershed! Even the honeybees use a primitive symbol system to good advantage.

JD suggests we might be in a situation analagous to a group of nearsighted people that are simply myopic, that appeals to public linguistic systems can't guarantee our objectivity. Unfortunately, no subculture can guarantee they are not being myopic: this includes religious groups, scientists, etc.. We are fallible people making fallible claims about how the world works, using the cognitive systems we were given (whether evolved or designed by Zeus), and there is no way to step outside of these cognitive systems, peek at raw reality, and determine if the two match up. We are all in this boat. Public symbolic systems are an indispensible part of this process, but they are also fallible. There is no argument against naturalism, specifically, to be found here, but a general skeptical worry. When we start to argue about whether we can really know we are not all just deluded, we are being pulled into a Cartesian philosophical morass that is as unproductive as it is boring. The arguments will mimick teapot tempests about whether we can really know there is an external world. Have fun with that.

JD, in other words, you did not address my response to your claim. Through public symbol systems, especially those comandeered by people obsessed with throwing theories against the experimental wood chipper, we have developed rich and elaborate theories that bear little resemblance to those which our inbuilt cognitive machinery could have imagined. Infinite dimensional vector spaces, nonEuclidian geometries, Gods, etc.. Without these systems, neither science nor theology are possible for us. They give us access to worlds screened off from creatures without such public utilities.

Stepping back, I think these Balfour arguments are cute, and trade on our present ignorance (like all god of the gaps arguments), but I don't understand their pull as serious ideas. Good scientists take their own ignorance as a sign of their lack of imagination and creativity, of the need to think more deeply about how to integrate some phenomenon into the natural order. For the god-gap fetishists, our present ignorance is an invitation to laugh at science and proclaim (loudly) that God (and only God) has the answer. One example of this anachronistic style of thinking is found in Dembski, who in his assinine 'Design Inference' formalizes the argument and actually endorses it [Namely, if we can't (presently) explain X naturalistically, then X is the product of design]! Perhaps he was the tutor to the cavemen in their meteorological theories!

Just pointing out our ignorance implies nothing but a psychological fact about ourselves. It does not have metaphysical implications beyond that. Because of the quite justified resistance scientists (qua scientist) have to supernatural explanations, we require more than these subtle arguments about things we are all ignorant about. Much more will be needed to let the supernatural back into science after its long and fruitful absence.

Blue Devil Knight said...

Hmm. I can't get these links to work correctly, and the html code is fine, and they show up fine in preview mode...Oh well. If the links are still busted, I've provided them again below...

Hi Mike. I agree that mathematical systems are crucial (as I mentioned: without them, we'd be in a primitive scientific state), and these are a good example of a public symbolic system that gives us a great deal of power to think ourselves into new theoretical directions, directions we couldn't conceive of if limited to our home-grown cognitive machinery.

Your bringing up Bayes' nets is very interesting. If we discover, empirically, that our brains/minds implement Bayes' nets in our reasoning about the world, as many psychologists , and the mathematicians/statisticians tell us that such procedures are reliable for making inferences about the world [e.g., causal Bayes' nets are good at reliably inferring infer the presence of unobserved variables] that would be one example of a good start on the naturalization of cognition, a naturalization that dovetails with the epistemically normative claims of mathematics.

As for whether the accuracy of sensory systems is irrelevant for this debate, I think you are wrong. If we empirically discovered that they were systematically not accurate, for instance, then that would leave quite a philosophical puzzle for the naturalist. The fact that they are accurate makes the story about bootstrapping to more abstract cases, with the help of public systems of symbols, much easier to imagine. Plus, it suggests more concrete ideas: e.g., test the hypothesis that abstract thought piggybacks on the computations performed in our sensory systems.


Link 2

Blue Devil Knight said...

JD, I adumbrated the sense in which symbolic systems acquire epistemic normativity above. We discover that certain inference patterns are better than others for preserving truth.

For instance:

A-->B
~A
Therefore, ~B

is not reliable. (A=it is raining, B=there are clouds in the sky). That is, we should not be convinced that the conclusion is true by the above argument.

Normative questions exist at two levels. First, what analogs of symbolic processes are occuring in our minds? Second, what are the best inferential strategies in our public symbol systems. It is an empirical question whether the two will match up.

Once we have discovered the strategies used by our minds, we can use them to generate predictions about errors people will tend to make, etc, using the much more powerful resources of mathematics.

Blue Devil Knight said...

Mike. I need to be clear about some distinctions I am making. Against JD, I was arguing that even if our cognitive mechanisms are not trustworthy, that is not all we have to go on. In fact, science is fundamentally social and relies on the kinds of symbol systems I was talking about. I leave open the possibility that our brains actually implement one of these symbol systems, which is the point I was making with the Bayes' nets stuff.

As for confidence that we can get things right, even when we go beyond simple perceptual reports ('There is a book on the table'), this may be true, but what are we to make of such confidence? I would imagine that an Uluguru tribesman is as confident in his animistic explanations of the weather, as we are in our modern-day explanations. It seems to me that our minds are pretty bad at judging how confident we should be when we extend our theories way past the evidence. We accept what our parents and other authorities say without question for many years (Santa Claus), we get persuaded by charismatic charlatans about quack remedies, etc.. The evidence and mathematically obsessed culture of science provides a pretty useful brake to our free-wheeling flights of imagination. It often leads to things quite at odds with our theoretical intuitions about the structure of the world (e.g., constancy of c in all reference frames, time dilation in general relativity, even F=ma).

Blue Devil Knight said...

JD, at this point I'm gonna refrain from repeating myself, as you are bringing up issues I've discussed in my already too numerous posts. Thanks for the discussion.

Edwardtbabinski said...

From "collisions of atoms" to a brain-mind filled with conscious ideas in constant interactive movement. That's a question worth the scientific investigation of everything from colliding atoms to brain-minds and all things inbetween, i.e., every cell and animal in the loop so to speak.

But as a starting philosophical premise (i.e., "from colliding atoms to brain-minds") it is no wilder than any other premise. Supernaturalism presumes "from 'God stuff' to brain-minds," and we know less about "God stuff" than we do about colliding atoms. In both cases we are dealing with our present ignorance, not with "proofs" or "disproofs" of anything.

Blue Devil Knight said...

Naturalists typically aren't in the business of providing 'foundations' for human knowledge. Since Quine and Sellars, we have embraced the fruitful rejection of this Cartesian project. There is no first philosophy. There is no privileged set of core beliefs that can provide a foundation for the rest of what we know.

The Cartesian, axiomatic, formal model of human knowledge is an anachronism (not too old, though: the Logical Positivists were subject to its call, as was Husserl). I can understand its pull: I think we all start our initial musings in things epistemological wishing to find epistemic bedrock upon which to rest our inference chains, as otherwise an infinite regress threatens. Fortunately, while there are circles in our web of belief, not all of them are vicious, and some of them are tethered to the world via data and evidence (though even data and evidence are subject to revision, for instance if my voltmeter needs calibration; also, the nature of the tethering is itself a subject of scientific study in studies of sensory systems).

It is a much less secure epistemic situation, to realize we start (and end) en media res, and that there is no unassailable epistemic fulcrum that will capture the human condition in its net. Where Cartesianism gets you is stuck in the "I think", trying to work your way out.

Rather than focus on what core beliefs we should 'start out with', the naturalist has focused more on the best methods to revise her beliefs in a way that maximizes some norms such as error-avoidance. Nowadays, this is done most fruitfully in probability and statistics, fields in which, ironically, most philosophers are not even trained.

Blue Devil Knight said...

You are right to recognize that naturalism provides no foundation for any beliefs.

I think your misunderstanding, JD, is that you conflate having foundations with having justifications for beliefs. There is an illicit jump from 'There are no bedrock beliefs' to 'There are no justified beliefs.' That this is a mistake was a key insight provided by Quine/Sellars. They realized that the pretensions of philosophers to ground knowledge in some core of claims, that can be known independently of science, and from which science somehow derives, is a Quixotic endeavor based on the phantom menace of an infinite regress.

This doesn't mean that certain claims are not more reliable than others (e.g., there is a real distinction between theory and evidence!). As we've all agreed, the information provided by sensory systems is a wonderful constraint, and science has turned this realization into the most productive (in terms of providing true and useful theories) set of methods for understanding the structure of the world, i.e., metaphysics.

The best introduction to these issues is in the book
Knowledge, Mind, and the Given
, an extended introduction to Sellars' seminal 'Empiricism and the Philosophy of Mind', an essay which can also be found online here.

Blue Devil Knight said...

JD, you argue like you are trying out for a job at the Discovery Institute, not earnestly looking for truth.

You haven't paid attention, or perhaps understood, my claims above. Public symbol systems are a key to getting past our innate sensory/cognitive take on the world, but you keep pointing out that such innate systems have some problems. I have addressed that. My above posts make this point multiple times, from multiple angles.

van Frassen is in the fringe of empiricism. Most modern empiricists are realists.

Some relevant classes to consider: Semantic Theory, Psychology of Rationality, Epistemology (coherentism vs foundationalism, hopefully with Quine and Sellars).

I would go for the BSE rather than the AB though, to be better prepared to engage with things as a scientist. I don't know how much flexibility there is.

Blue Devil Knight said...

For every case in which we have come to understand a phenomenon thoroughly, it has turned out that deities weren't needed (e.g., meteorology, astronomy, muscle contraction, geology, etc).

You don't have to make a pessimistic induction over the history of supernatural explanations, but I think it isn't unreasonable at all to do so.

Of course, there are still lots of things science can't explain, for which scientific explanations are hard to imagine. This interesting psychological fact might turn out to turn on a metaphysical inadequacy of scientism. Or, it may turn out to be a boring result of our impoverished imaginations.

Rather than get all dogmatic, certain, and inflexible (the cornerstone of the Discovery Institute's strategy), we need to do more science. I think everyone should agree on that: to proclaim that our ignorance about something shows that naturalism is false is pretty weak. While I know you wouldn't agree that this is what you are doing, it is precisely what you are doing. You are predicting what you will think about what science has to say about some topic before the science is even close to done, hell, it's barely started.

It is quite an interesting coincidence that the phenomena on which the theists base their arguments are things which even naturalists will admit we don't understand very thoroughly (e.g., morality, consciousness, reason, free will).

This seems to be a foolish strategy. Why not pick something which the naturalist thinks we have a really good naturalistic explanation for, and show that they are wrong? E.g., muscle contraction, AIDS, planetary motions? At least in those cases, you'd have lots of concrete data upon which to make a more sound judgment, and perhaps even some predictions. As it is, we are left with people being really confident about what science cannot tell us, but paradoxically basing this on things which science is just starting to study.

Blue Devil Knight said...

Sigh.

Shrug.

More stupid web-isms.

Grin.

LOL

:P

JK

My grandmother was a Christian Scientist. Now there is one philosophy that goes full tilt against the naturalist canon, straight on. No shifting or hemming/hawing: flat out, medical science is wrong. That's what I'm looking for.

I think, if the antinaturalists are ever going to be taken seriously, they need to take on similarly established canons, and stop firing at the fringes of science. I'd like to see the nonnaturalist explaination of AIDS.

Show me the gods in your physics textbooks. I'm not talking about the psychology of theory generation. Kekule dreamt the solution to the structure of hexane in a couple of snakes sucking each other's tails. That's fine. But nobody thinks that hexene is anything but a physical phenomenon. Anything else is explanitorily invisible in science.

I could be convinced by evidence that I am wrong. Flat-out crazy big-ass undeniable miracles. Dead people coming to life, telling me that I better become a Christian because they have seen the afterlife. That would make me a Christian. Pretty fast. And if such things happened regularly, science would undergo a radical transformation.

But such things don't happen, and science merely reflects this fact. It would only be a dogma if science stopped modifying its doctrines in the light of good evidence that it is on the wrong track with naturalism.