Sunday, May 1, 2011

Knowing What We Know

The first book on physics I ever read was Stephen Hawking's A Brief History of Time. It's actually the book that catapulted me from a sort of weak theistic agnosticism into full-on atheism. It wasn't that I thought Hawking had disproved the existence of God or anything like that, but it was the thought process that intrigued me – he bravely took certain "big questions" out of the realm of mysticism and into the quantifiable world of science. This was also the book that introduced me to the famous double-slit experiment in quantum mechanics. In the experiment, a particle does not take one path from A to B, but rather all possible paths simultaneously. That's a profoundly counter-intuitive idea, one that's even more counter-intuitive than the weirdness of things like gravitational time dilation from Einstein's General Relativity.

Our minds play tricks

We tend to view the world from a rather insular kind of bubble. We're bombarded with a massive amount of sensory data which our brain constructs into a reasonably reliable model that we call "reality". We develop an intuitive understanding of the world, where we assume that things are going to work a certain way. We don't test every inch of ground before we step on it to make sure we won't fall through. We know from experience that if we let go of something, it will fall to the ground – and we don't bother making sure that applies to every object we encounter. In case you were wondering, cognitive psychologists have a name for these assumptions – they're called "intuitive physics".

Strange things like gravitational time dilation and the double-slit experiment are a big slap in the face to our intuitive understanding of reality. Like an optical illusion, we've learned that the model of reality that our brain gives us, while generally useful for our everyday frame of reference, isn't particularly reliable as a model of how things really work. Not only do some things completely defy our intuitions the way quantum mechanics does, but even on our everyday scale we find that our brains aren't always as reliable as we'd like to believe.

Here's a simple thought experiment: have you ever been driving down the road and slammed on your brakes as an animal scurried in front of your car... just before you realized that it wasn't an animal at all, but a leaf tumbling in the wind? Have you heard sounds that you believed were voices or footsteps, but turned out just to be your house creaking in the wind? Have you ever mistaken a shadow for a sulking prowler or a shy animal? Most of us have many such experiences. In every case, we are making a pattern-recognition error: we are subconsciously imposing a pattern of goal-oriented behavior where there is nothing but randomness.

Now, try to think of the opposite: have you ever mistaken an animal for a tumbling leaf? A voice for the howling of the wind? A prowler for a shadow? We do make such errors, but we make them far less frequently. To the gazelle, it's much safer to mistake the lazy teetering of the grass for a lion – and run away – than to mistake a hungry lion for the lazy teetering of the grass. Similarly, we're much better off if we mistake a prowler for a shadow than vice-versa. From a survival standpoint, it's far better to impose a pattern when it does not exist than it is to fail to see a pattern where one does exist.

A canvas of patterns

Our intuitive model of reality is a canvas of patterns. We see patterns in nature, patterns in behavior, patterns in events – even patterns in our minds. We can think of these patterns as the raw sensory data that bombards us every day. But what are these patterns? Why do they work the way they do? Which ones are real, and which ones are we mistakenly imposing upon the randomness of reality? How can we know the difference?

Science, at its most fundamental form, is a methodology by which we attempt to explain these patterns. We try to explain them by positing a mechanism. If we are correct in our understanding of the mechanism, it will do two things: First, it will explain the patterns we have already seen. Secondly, it will predict what specific patterns we will observe in the future.

Of course, we humans are quite imaginative, and we didn't have to wait around for science to start positing various mechanisms to explain the patterns we see. Primitive mechanisms have always tended to be rather anthropomorphic conceptualizations of ethereal beings, from gods to ancestral spirits. Animistic cultures, for example, do not believe that what they observe in nature is random, but the product of intent by conscious spirits of nature. It is certainly possible that gods or spirits are in control of the patterns we observe and experience. But that's not a particularly useful explanation; in order to move from possible to plausible we have to have a mechanism that is reliable – it doesn't just explain what we do see, but what we will see. This is what is meant when a mechanism (i.e., a theory) is falsifiable. If a we posit a mechanism that predicts a certain pattern but we subsequently observe a different one, the mechanism is falsified – it must be either reworked or discarded entirely. If a mechanism is unfalsifiable, that means it does not make any predictions about the patterns we should see, and thus cannot be tested for reliability.

Unintelligent design

I discussed an example of this in an old post about Intelligent Design back in the early days of this blog. Like evolution by natural selection, ID also posits a mechanism to explain the diversity and complexity of life. ID advocates insist that they are looking at the same data as evolutionary biologists, and simply making a different inference – that it was a conscious being instead of natural processes. But this mechanism has a fatal flaw: it is unable to predict the patterns we ought to see. For example, evolution predicts that species evolve according to the specific pattern known as the phylogenetic tree of life. Species on one "branch" will share more of their DNA than species on separate branches.

That post was called "Chicken Teeth", which was used to make a point: chickens have latent genes for producing teeth. Humans have latent genes for producing functional tails. Whales and dolphins have latent genes for producing functional limbs. We call these genes "junk DNA" – they're present in the genetic code, but they don't appear to do anything. Jellyfish and bacteria, despite both being fully modern animals, do not have latent genes for producing teeth, tails or limbs because they evolved on completely separate branches of the tree of life. So we can see that evolution by natural selection not only fully explains the presence of specific kinds of junk DNA, but also tells us what kind of junk DNA we should see – animals should always retain the DNA specific to their branch of the tree of life. ID is left to speculate about the motives and characteristics of the designer to retroactively account for this information.

This concept is of pivotal importance to the philosophy to which most atheists hold, and which is the backbone of all scientific inquiry: methodological naturalism. People often claim to possess knowledge about the fundamental nature of reality, and often attribute this knowledge to intuition, emotion, experience, or authority. But how can we really know which of our experiences are reliably true? How do we know whether the patterns we observe or the experiences we have correspond to a metaphysical reality rather than being mere projections and pattern-recognition errors?  As the weirdness of the quantum double-slit experiment showed, we can't always trust our intuitions.


So far, I've talked about how the raw sensory data that we take in is categorized by our brains into patterns,  how sometimes we make the mistake of imposing patterns on the randomness of nature, and how methodological naturalism is a means for explaining the patterns we see by producing falsifiable mechanisms with predictive utility. This gives us a reliable understanding of the world – we can predict, for example, that objects will fall when we let go of them. We can predict how fast they'll accelerate and how wind resistance will affect the rate of descent. All knowledge attained through methodological naturalism is provisional, because we don't know everything. There may be some unknown law of physics that, starting tomorrow, will cause the force of gravity to work in reverse. But based on empirical observation with predictive utility, we can make a valid provisional assumption that such a change is highly implausible.


Jerry Coyne writes a lot about "accommodationism", which is a popular new buzzword for the old "NOMA" argument put forward by the late Stephen Jay Gould. This is the idea that science and religion are "non-overlapping magisteria" – that both are valid, but different, means of understanding the world. In both cases, I'm treating these terms very broadly: I'm taking science to mean "empirical observation and rational inquiry" (basically, methodological naturalism), and I'm taking "religion" to mean "spiritual experiences and theology". I don't want my use of the word "science" to be confused with a bunch of guys in lab coats shooting lasers and mixing vials of smokey green liquid, and I don't want my use of the word "religion" to be confused with guys in goofy robes or people waving their arms while they sing cornball hymns. Humanities, such as historical inquiry, are still subject to the rules of empirical investigation; and spiritual experiences are by no means confined to religious dogmas.

With science, we have a sound methodology for attaining knowledge not only by validating claims with predictive utility, but by identifying and discarding erroneous or superfluous information. This creates a consensus. Einstein's theories of relativity were met with skepticism, as they radically re-wrote many of the assumptions underpinning Newtonian physics, which had been robustly substantiated through empirical evidence. But Einstein's theories improved on Newton's, and as more and more empirical confirmation was obtained, poorly supported alternative ideas such as the "aether" were discarded in favor of a consensus favoring relativity.

This is important for several reasons. Most often, when we hear the arguments of accommodationists, they sound something like this:
  • "We have to recognize the limitations of science"
  • "There are other ways of knowing besides science"
  • "There are many great mysteries that seem to hint at spiritual realities"
  • "Science can tell us how, but it cannot tell us why"
If indeed there are "other ways of knowing", it must be demonstrated that such "other ways" produce reliable and valid results. It is folly to assume that because science has not offered an explanation, that no scientific explanation exists (to do so is an argument from ignorance); it's similarly foolish to assume that religion can answer questions simply because science can't.

Validating claims about reality

Believers often have very powerful (and personal) "spiritual experiences", and they want to believe that these experiences correspond to a metaphysical reality – perhaps one that eludes understanding with the tools of science. But in every case, these mysterious "other ways of knowing" turn out to be rather familiar: subjective experiences, intuition, emotional appeals, and personal revelation. It's important to point out, for the sake of avoiding the genetic fallacy, that simply because knowledge comes from these sources does not make that knowledge invalid. I mentioned intuitive physics in part 1, and that's a great example of something that we learn intuitively that does indeed correspond with an objective understanding of reality. But things like intuition and revelation are, in themselves, insufficient to establish a claim as objectively true.

This can be illustrated in the distinct lack of consensus among the world's religious faiths. Christianity alone has some 16 broad branches of theology, with over 30,000 denominations. There is not, and has never been, any world-wide consensus on what God is, what God does, or what God wants from us. Two geographically isolated cultures will form conceptualizations of spirituality that are so dissimilar as to be absurd, ranging from the worship of ancestral spirits to pantheistic paganism. But that's not the only problem: there is no methodology by which we can objectively discern the validity of such information. The faithful often talk about mankind's seeming propensity for spirituality; but if this really corresponded to a metaphysical reality rather than simply being a by-product of evolution (namely, the categorical and pattern-recognition errors I discussed in part 1), we ought to expect not only a great deal more consensus on spiritual matters, but a means by which to objectively evaluate, confirm and falsify spiritual claims to knowledge. Otherwise, there is simply no way to determine whether someone is just making it all up.

We intuitively strive for the greatest possible congruence between the raw sensory information we are bombarded with, and a reliable understanding of the patterns we observe. Sometimes, we form assumptions that are incongruent with reality – for example, we might be convinced that we have experienced the presence of a spirit or deity. When these kinds of sensory-pattern interpretations cannot be validated through any objective mechanism, we create rationalizations to preserve congruence between our beliefs and our experiences. There is perhaps no greater example than that of prayer:
  • Joe has stage 4 leukemia, so we are praying for him
  • Joe did not get better, and eventually died
  • (Rationalization) It was not God's will, we didn't have enough faith, etc.
Prayer has never been demonstrated to have any effect that transcends mere coincidence, and has failed empirical investigation repeatedly and miserably. But that doesn't stop believers, who count the "hits" – no matter how trivial – and simply rationalize away "the misses".  If prayer really worked – if appeals to deities had a real effect on the natural world – we would be able to validate such a fact through the tools of science. Believers, though, want to have it both ways: God does intervene in the physical world, but not in any way that can be empirically observed. It's an absurd paradox.

I can't throw a rock at a conversation with a believer without hitting claims about how concepts such as the self, love or morality are things we necessarily accept as true, but elude empirical inquiry. This is simply false. What we call the "self", or "I", is simply a category of sensory information – those who attach mysticism to individuality would do well to bone up on some basic cognitive psychology. Abstractions such as love and morality, if they are to be useful to us, must correspond to objective facts about the human condition. (There's more to be said about those topics specifically, but I've covered them extensively in past posts.)

The limits of knowledge

Scientific inquiry is the only means we humans have of attaining reliable, valid knowledge about our experiences. We do not possess some sort of spiritual "sixth sense" that allows us to objective discern the truth or falsity of supernatural claims, and this is evident in the growing discordance, rather than consensus, that emerges as human spirituality "evolves". Methodological naturalism gives us the greatest possible congruence between our experiences and our beliefs, since it is the only means by which we can objectively validate or falsify our assumptions. It might not give an answer to something like, "Why do we exist" that alludes to some grand purpose that transcends human interest; it can, however, demonstrate that even if such a grand purpose does exist, it eludes our capacity for reliable, valid knowledge – in other words, that it's a pointless question.

We can't disprove subjective claims about personal spiritual experiences. What we can do, however, is demonstrate that we have no reason to assume that such claims are in any way representative of an objective reality. Time after time, I encounter believers who misguidedly attach mystery to abstractions like love, creativity and morality, and claim that such mystery is evidence of spiritual realities. I encounter many who claim to have experienced this god or that, and who don't seem to mind when the assumptions they derive from their experiences conflict with other people's spiritual experiences – but of course they pay great attention to even trivial similarities. In the intellectually honest spirit of methodological naturalism, I can't discount the possibility that spiritual realities exist. But if they do, the burden to prove it falls squarely on the ones making the claim: how do you know what you think you know?

No comments:

Post a Comment