Monday Musing: Francis Crick's Beautiful Mistake

Many scientists don't know what they are doing. That is, they are so immersed in science, that they often do not step outside it for a wider philosophical perspective on what it is they do, while remaining convinced that science is somehow more correct than other ways of doing things. For example, a scientist might argue that she can treat malaria better than a witch doctor can. The witch doctor, of course, will say the opposite. If you ask the scientist why she thinks she is right, she will say that she can demonstrate her efficacy with an experiment: a large sample of cases of malaria which are treated by her method as well as with the witch doctor's method (and maybe even a control group), after which she will perform a sophisticated statistical analysis on the data that she collects on all these cases, thus showing that her method is better. Now, if you object that her reasoning is circular, after all, she has just used the scientific method to show that the scientific method is correct (thereby only really showing that the scientific method is self-consistent), and don't allow her to use science to prove science right (if the scientific method of proving something right were already acceptable to you, you wouldn't be questioning her in the first place), she will tend to start getting desperate and try to make appeals to common sense, or even question your sanity ("Are you crazy? It's obvious that witch doctor is a thieving fraud, taking people's money and pretending to help them with his wacky chants," etc.) And she will have a lingering suspicion that you have somehow tricked her with some sneaky rhetorical sophistry; she will continue to think that of course science is right, just look at what it can do!

So what's going on here? I am not claiming that witch doctors (or astrologists, or parapsychologists, or faith-healers, or Uri Gellar, or Deepak Chopra, or other charlatans) are just as good as scientists, or even that they are right about anything at all (they are not); what I am saying is that there is no neutral ground on which to stand, and from the outside, as it were, proclaim the supremacy of science as the best avenue to truth. One must learn to live without such an absolute grounding. Even as clear-headed and careful a thinker as Richard Dawkins can sometimes get confused about this. At the end of an otherwise fascinating and inventive essay entitled "Viruses of the Mind" (Dawkins's contribution to the volume Dennett and His Critics) in which he uses viruses as a metaphor for the various bad ideas (or memes) that "infect" brains in a culture (particularly the "virus" of religion), and also makes a parallel analogy with computer viruses, Dawkins asks if science itself might be a kind of virus in this sense. He then answers his own question:

No. Not unless all computer programs are viruses. Good, useful programs spread because people evaluate them, recommend them and pass them on. Computer viruses spread solely because they embody the coded instructions: 'Spread me.' Scientific ideas, like all memes, are subject to a kind of natural selection, and this might look superficially virus-like. But the selective forces that scrutinize scientific ideas are not arbitrary or capricious. They are exacting, well-honed rules, and . . . they favour the virtues laid out in textbooks of standard methodology: testability, evidential support, precision, . . . and so on.

Daniel Dennett spares me the need to respond to this very uncharacteristic bit of wishful silliness from Dawkins by doing so himself (and far better than I could):

When you examine the reasons for the spread of scientific memes, Dawkins assures us, "you find they are good ones." This, the standard, official position of science, is undeniable in its own terms, but question-begging to the mullah and the nun--and to [Richard] Rorty, who would quite appropriately ask Dawkins: "Where is your demonstration that these 'virtues' are good virtues? You note that people evaluate these memes and pass them on--but if Dennett is right, people (persons with fully-fledged selves) are themselves in large measure the creation of memes--something implied by the passage from Dennett you use as your epigram. How clever of some memes to team together to create meme-evaluators that favor them! Where, then, is the Archimedean point from which you can deliver your benediction on science?"

[The epigram Dawkins uses and Dennett mentions above is this:

The haven all memes depend on reaching is the human mind, but a human mind is itself an artifact created when memes restructure a human brain in order to make it a better habitat for memes. The avenues for entry and departure are modified to suit local conditions, and strengthened by various artificial devices that enhance fidelity and prolixity of replication: native Chinese minds differ dramatically from native French minds, and literate minds differ from illiterate minds. What memes provide in return to the organisms in which they reside is an incalculable store of advantages --- with some Trojan horses thrown in for good measure. . .

Daniel Dennett, Consciousness Explained

Below, Dennett continues his response to Dawkins...]

There is none. About this, I agree wholeheartedly with Rorty. But that does not mean (nor should Rorty be held to imply) that we may not judge the virtue of memes. We certainly may. And who are we? The people created by the memes of Western rationalism. It does mean, as Dawkins would insist, that certain memes go together well in families. The family of memes that compose Western rationalism (including natural science) is incompatible with the memes of all but the most pastel versions of religious faith. This is commonly denied, but Dawkins has the courage to insist upon it, and I stand beside him. It is seldom pointed out that the homilies of religious tolerance are tacitly but firmly limited: we are under no moral obligation to tolerate faiths that permit slavery or infanticide or that advocate the killing of the unfaithful, for instance. Such faiths are out of bounds. Out of whose bounds? Out of the bounds of Western rationalism that are presupposed, I am sure, by every author in this volume. But Rorty wants to move beyond such parochial platforms of judgment, and urges me to follow. I won't, not because there isn't good work for a philosopher in that rarefied atmosphere, but because there is still so much good philosophical work to be done closer to the ground.

Now I happen to agree more with Rorty on this, but that is not the point. What is important is that Rorty, Dennett and I, all agree that there is no neutral place (for Archimedes to stand with his lever) from where we can make absolute judgments about science (the way Dawkins is doing), or anything else. We must jump into the nitty gritty of things and be pragmatists, and give up the hope of knowing with logical certainty that we are right.

So how do scientists go about their business then? How do they know when they are onto something? These are questions that many sociologists, anthropologists, psychologists, philosophers of science, and scientists themselves have tried to answer, and the answers have filled many books. One thing comes up again and again, however, and especially when scientists themselves talk about what they do and how they do it: the importance of beauty. Scientists don't just sit there dreaming up random hypotheses and then testing them to see if they are true. There are too many possible hypotheses to work this way. Instead, they try to think of beautiful things. This intrusion of the aesthetic into the hard, cold, austere realm of science is unexpected to many people, but it is surprisingly consistent. When Albert Einstein was asked what he would do if the measurements of bending starlight at the 1919 eclipse contradicted his general theory of relativity, he famously replied, "Then I would feel sorry for the good Lord. The theory is correct." What he meant was that the theory is far too beautiful to be wrong. How do you tell when something is beautiful? That, I'm afraid, is a question too big for me. (Though if that kind of thing interests you, you may wish to have a look at this recent Monday Musing essay by Morgan Meis and the ensuing discussion in the comments area.) For now, we'll have to make do with some you-know-it-when-you-see-it notion of beauty. (Kurt Vonnegut once said that to know if a painting is good, all you have to do is look at a million paintings. I can only mimic him and say that if you want to know what is beautiful in science, all you have to do is look at a lot of science.)

Yes, yes, I am slowly coming to my subject. (Hey, it's my Monday Musing and I'm allowed to ramble on a bit!) We are now approaching the first anniversary of 3 Quarks Daily. The very first day that 3QD went online, July 31, 2004, I posted the sad news of Francis Crick's death. Crick, of course, along with James Watson (and Rosalind Franklin, and Maurice Wilkins), was the co-discoverer of the molecular structure of DNA. (In possibly the most coy understatement ever published in the history of science, at the end of the momentous paper in which Watson and Crick detailed their discovery of the double helix--which can be unwound, each strand then re-pairing with other bases to form a new double helix identical to the original--thereby solving the problem of DNA replication, they wrote: "It has not escaped our notice that the specific pairing we have postulated immediately suggests a possible copying mechanism for the genetic material.") Crick won a Nobel for this work, but this is not all he did. He spent the latter part of his life as a distinguished neuroscientist, publishing much in this new field, including the book The Astonishing Hypothesis.

The years following the discovery of the structure of DNA were busy ones, not just for molecular biologists, but also for physicists and mathematicians (Crick himself had come to biology after obtaining a degree in physics), and specialists in codes, because the code instantiated in the double helix took some time to understand. George Gamow made significant contributions, and other physicists also took a crack at the problem, including a young Richard Feynman, and even Edward Teller proposed a wacky scheme.

Let me now, finally, attempt to deliver on the promise of my title. At some point in time, this much was clear: the molecular code consisted of four bases, A, T, C, and G. These form the alphabet of the code. Somehow, they encode the sequences of amino acids which specify each protein. There are twenty amino acids, but only four bases, so you need more than one base to specify each amino acid. Two bases will still not be enough, because there are only 4², or 16 possible combinations. A sequence of three bases, however, has 4³, or 64 possible combinations, enough to encode the twenty amino acids and still have 44 combinations left over. Such a triplet of three bases which specify an amino acid is known as a codon. So how exactly is it done? What combinations stand for which amino acid? Nature is seldom wasteful, so people wondered why a combinatorial scheme which allows 64 possibilities would be used to specify a set of only 20 amino acids. Francis Crick had a beautiful answer. As we will see, it was also wrong.

What Crick thought was something like this: suppose you have a sequence of 15 bases (or 5 codons) which specifies some protein (remember, each codon specifies an amino acid), like GAATCGAACTAGAGT. This means the codon GAA (or physically, whatever amino acid that stands for), followed by the codon TCG, followed by AAC, and so on. But there are no commas or spaces to mark the boundaries of codons, so if you started reading this sequence after the first letter, you might think that it is the codon AAT, followed by CGA, followed by ACT, and so on. It is as if in English, if we had no spaces and only three letter words, you might read the first word in the string PATENT as PAT, or if by mistake (this would be easy to do if you had whole books filled with 3 letter words without spaces in between) you started reading at the second letter, as ATE, or starting at the third letter, as TEN, etc. Do you see the difficulty? This is known as the frame-shift problem. Now Crick thought, what if only a subset of the 64 possible codons is valid, and the rest are non-sense. Then, it would be possible that the code works in such a way that if you shift the reading frame in the sequence over by one or two places, what results are nonsense codons, which are not translated into protein or anything else. Again, let me try to explain by example: in the earlier English case, suppose you banned the words ATE and TEN (but allowed ENT to mean something), then PATENT could be deciphered easily because if you start reading at the wrong place you just end up with meaningless words, and you can just adjust your frame to the right or left. In other words, it would work like this: if ATG and GCA are meaningful codons, then TGG and GGC cannot be valid codons because we could frame shift ATGGCA and get those. Similarly if we combine the two valid codons above in the other order, we get GCAATG, which if shifted gives CAA and AAT, which also must be eliminated as nonsense. This kind of scheme is known as a comma-free code, as it allows sense to be made of strings without the use of delimiters such as commas.

Now, Crick worked out the combinatorial math (I won't bore you with the details, Josh) and found that with triplets of 4 possible bases, one has to eliminate 44 of the 64 possiblilities as nonsense codons, to make a comma-free code. Voila! That leaves 20 valid codes for the 20 amino acids, saving parsimonious Nature from any sinful profligacy! This is what beauty in science is all about. Now, Crick had no evidence that this is indeed how the genetic code works, but the beauty of the idea convinced him that it must be true. In fact, the exact elegance of this scheme was such that all attempts at actually figuring out the genetic coding scheme for the next many years attempted to be compatible with the idea. Alas, it turned out to be wrong.

In the 60s, when the actual genetic coding schemes were finally figured out in labs where people managed to perform protein synthesis outside the cell using strings of RNA, it turned out that there are real codons which the comma-free code theory would have eliminated, and this nailed the coffin of Crick's lovely idea shut forever. In fact, more than one codon sometimes codes for the same amino acid, while other codons are start and stop markers, acting as punctuation in the sentences of genetic sequences. It is now understood that nature is not prodigal, and uses this redundancy as an error correction measure. Computer simulations show that the actual code is nearly optimal when this error correction is taken into account. So it is quite beautiful, after all. Still, why did so many scientists think for so long that Crick must be right? Because in science, as in life, beauty is hard to resist.

Have a good week!

My other recent Monday Musings:
The Man With Qualities
Special Relativity Turns 100
Vladimir Nabokov, Lepidopterist
Stevinus, Galileo, and Thought Experiments
Cake Theory and Sri Lanka's President

Comments

[sacundim]

If you ask the scientist why she thinks she is right, she will say that she can demonstrate her efficacy with an experiment: a large sample of cases of malaria which are treated by her method as well as with the witch doctor's method (and maybe even a control group), after which she will perform a sophisticated statistical analysis on the data that she collects on all these cases, thus showing that her method is better. Now, if you object that her reasoning is circular, after all, she has just used the scientific method to show that the scientific method is correct (thereby only really showing that the scientific method is self-consistent) [...]

Um, I'm sorry, but there's just no circularity here. The scientist hasn't used the scientific method to defend the scientific method—she's used the scientific method to defend a claim about the efficacy of different treatments in treating malaria. The only thing she might be faulted with is taking your question to be a scientific one about treatments for malaria, when you intended it as a philosophical one about the scientific method. And if we do fault her with that, then we should remind you that she is your very own hypothetical scientist, and her words are those that you yourself put in her mouth. (Though I agree that scientists by and large are very, very bad at philosophy...)

I think you do equivocate much earlier, though, with the tacit assumption that the witch doctor and the scientist are both trying to "treat malaria." The witch doctor may well object to that description of what he's doing, on many counts.

[Tasnim]

Science and your essay are beautiful things.

[J. M. Tyree]

Never boring, Abbas! Maybe by the end of next year I will actually know something about science...Very enjoyable.

[ymatt]

I posted this comment on lies.com, but thought I'd mention it here with the original article as well. With respect, this just reads like post-modernism without the obfuscating vocabulary:

Many scientists don’t know what they are doing. That is, they are so immersed in science, that they often do not step outside it for a wider philosophical perspective on what it is they do, while remaining convinced that science is somehow more correct than other ways of doing things.

No, I believe science really is more correct and philosophy has no place in it. You seem to suggest that science relies on the sensibilities of scientists — that their western way of thinking or personal notions of beauty will inevitably twist the science. This is nonsense. The only hard principle on which science relies is consistency. When a theory is not consistent with measured reality, it must be modified or discarded. And if we’re going to call consistency itself merely a construct of western thought or somesuch, I’m not sure there is anything left in the world that can be considered absolute.

Notions of beauty do however enter into extrapolative theory. A scientist may see an imcomplete set of data and intuit a wider theory based on its simplicity or beauty. Certainly personal biases can be introduced in this way. But it is critical to remember that a theory only endures if it is consistent with the world around us! Even the basic “laws” of science would have to be discarded if we found evidence to contradict them.

Extrapolative theory is still immensely important to science however. There is an infinte amount of data that can be taken and Occam’s Razor very often holds true: the simplest explanation is usually the correct one. The simple, elegant, beautiful theories are frequently borne out by supporting data. Others, such as Crick’s, are discarded. But as you note, more data eventually showed that the correct theory was just as elegant, but more complete.

[Abbas Raza]

Dear ymatt,

I might be guilty, but not of what you accuse me. I am possibly guilty of having given a poorly chosen example in my very first paragraph, and not making it clear enough that I actually agree that the scientist is right and the witch doctor is wrong. I was trying (obviously not well enough for some people) to make a subtle point: that while science is right, there is no neutral place outside of it from which we can judge it. I happen to believe that science is correct, but as Dan Dennett says, there is no way to convince the mullah or the nun of this, because they are not willing to examine things scientifically. I agree with you that there is no place IN science for philosophy, but there ARE interesting philosophical questions ABOUT science. If you don't believe that, then I have nothing more to say to you, because we just have too little common ground from which to continue a discussion, and I have neither the time, nor the space here to try to convince you of it.

I am no pomo thinker, and you would be hard-pressed to find someone with a greater aversion to obscurantist jargon: I am absolutely NOT suggesting that scientists' western way of thinking or personal notions of beauty will inevitably twist the science! Not at all, and I don't see why you choose to interpret me in this uncharitable way. Just as you say, and I show in the essay, in developing theories, scientists take aesthetic factors into account, but if this results in a wrong answer, it is eventually corrected, just as it was with Crick's elegant comma-free theory. This is the whole point I am making!

I really don't think we are very far from each other in our views, but for some reason, despite my repeatedly saying that I myself believe that science IS correct, and clearly denigrating other forms of chicanery, including religion, in the essay, somehow you have managed to conclude that I am anti-science.

By the way, I hold degrees in Electrical Engineering and Computer Science as well as philosophy, and worked as a scientist as well as an engineer for a decade. Perhaps you could try to be a bit more charitable in trying to understand what I am saying, instead of imagining slights to science where there are none, and then being outraged by them.

In any case, I thank you for taking the trouble of posting your comment here, as well as at lies.com.

[ymatt]

My apologies if I misinterpreted your essay. This is the first thing I've read of your site, so I didn't have any background to frame it with. I think what I missed is that your argument is that while science may be based on absolutes, it is difficult to convince others of this. I certainly didn't mean to accuse you of being anti-science. I may just be a little sensitive to arguments I've heard before that science is just another form of faith, which I find idiotic or at least misguided.

I think what I still disagree with, however, is that there is no neutral ground to use to make the argument. The consistency of reality is the only absolute that we can rely on. I would think that the witch doctor would agree that if he cures the patient, that any other observer should be able to witness that the patient was cured. This kind of independent observation and verification is the only foundation that science needs.

The trouble comes when the supporting data for a theory, no matter how concrete, is too complex for a layperson to understand (or its conclusions contradict closely-held beliefs). I think this leads many (certainly not including yourself) to conclude that some scientific theories are just atheistic activism (as weird as I find that notion) or self-delusion.

But I agree, the practice of hard science can raise a number of interesting philosophical questions. I've always enjoyed the fascinating symmetry and beauty that emerges from the relatively simple laws governing the universe as we know them. Life itself is a testament to the depth of complexity that can arise from such an elegant system, smoothly combining in a natural way what it takes most of our scientific and engineering knowledge to explain and mimic.

[Abbas Raza]

Dear ymatt,

Thanks again for your comment. There are two statements in your comment about which I might urge you to think a little more:

I would think that the witch doctor would agree that if he cures the patient, that any other observer should be able to witness that the patient was cured.

The reason many people think that the witch doctor actually cures people is that they DO witness him doing it. Most diseases are self-limited in time, and people will become better after a certain amount of time has passed. Now, if the witch doctor chanted over them while they were sick, when they get better (and are clearly seen to be so by others), people believe the witch doctor somehow caused them to get better. This is common even closer to home: someone gets a cold (which is usually a 24-48 hour virus), someone else tells them a crazy new remedy to try; they try it; they get better! Now they believe that, for example, drinking clove tea (or whatever), is great for curing colds and go around spreading this idea to others.

The way someone like you (or me) takes care of this problem is by careful controls, and statistical analysis of the data. This is the part that is hard to get the witch doctor to agree to, because, for example, he may not KNOW ANY math and can't even understand what you are saying, and this brings me to the second statement from you:

The trouble comes when the supporting data for a theory, no matter how concrete, is too complex for a layperson to understand (or its conclusions contradict closely-held beliefs).

Exactly. Just as the witch doctor of my example has not been prepared by years of training (through high school, college, etc.) to understand the math (or other science) necessary to grasp what you mean by a controlled study and your statistical analysis of data (which can be so sophisticated that most scientific studies hire a PhD in statistics to do it), he can claim that YOU have not had the years and years of experience with witch doctoring that he has (or a religious person might say that you need years and years of meditation and prayer and detailed study of divine texts and abstruse religious arguments before you can understand why prayer works, or develop the faith that he or she has). They could say: we are willing to spend ten years to learn your method, if you are willing to spend ten years in a religious monastery (or seminary, or witch doctoring apprenticeship, or whatever) learning why we are right. Now what do you say to them? If you are like me, you will decline their offer, but then they will say, you are not willing to undergo the strenuous preparation that allows you to see that we are correct, which is exactly what you had accused them of!

This is what I meant when I said that there is no neutral place from which to automatically show that science is correct. Science is correct to the people who know science, like you and me, and we must be content to try to explain it, and convince others of its correctness, from within. And this I try to do all the time.

As for the "consistency of reality", it is the same story: as I have pointed out, the witch doctor and the religious person have their own ideas about what reality is. The religious person may claim God as an absolutely obvious part of reality, more real than quarks, he might say, claiming that you have never seem quarks directly, but have just inferred them from a theory based on the observations of other things, that it took you a decade or more to learn, just as it took him a decade of meditation (or whatever) on the things around him to infer the existence of his God from them. The problem is in how we construct our picture of reality, which is not just given to us by nature (like the notion of quarks, say), but deduced from years of study; and there, once again, we can't easily show that our picture is correct from OUTSIDE of science. Does this make what I was trying to say any more clear? Let me know what you think.

Thanks again for the thoughtful comments.

[ymatt]

Yes, there is the problem of convincing someone who doesn't have the background to understand the argument, but the real test of consistency is predictability, which I think anybody should be able to understand. If the illness is severe enough and of long enough duration, the scientific cure should predictably cure patients whereas the witch doctor's methods will be less certain. The world could no longer be flat when the round earth theory predicted how to navigate large distances correctly.

Of course theories which apply to subjects where prediction is difficult, like evolution today or ether 50 years ago or heliocentricity several hundred years ago, tend to draw more critics. I'm not saying that any given person can be convinced of any given scientific theory. But to say "there is no neutral ground on which to stand, and from the outside, as it were, proclaim the supremacy of science" seems unfounded.

Good scientific theory does nothing more than predict the workings of the world -- in fact it must by defninition, since a theory that does not work ceases to be valid. Science is not faith, it is practice. It is means, not end. And those means are based on nothing more than observation of the world and finding any explanation that has demonstrable predictive power.

I know you agree with all this. I think I just worry when science is made to sound like just another way of looking at the world. It gives ammunition to those, unlike yourself, who would discredit science.

In any case, thanks for a stimulating conversation.

[Basil Fernie]

"Science is not faith, it is practice. It is means, not end."

Scuse me guys, you've just stated - perhaps unintentionally - what it took me a while as a practising mathematical statistician, helping scientists to make better theories about the world around us, to conclude: what most people call science is actually technology. Understand that, you're free to use and appreciate science for what it can do via its predictive power, even for its beauty, without exalting it to some kind of absolute status and bowing down to it.

There's only One Who is absolute, and His Name says it all. Everything else, including "science", is derivative and dependent. To be respected, enjoyed, used maybe; but not to be idolized.

[Martin Packer]

NO comment, but see:
http://www.packer34.freeserve.co.uk/rememberingfranciscrickacelebration.htm

[Apoorva Bhandari]

Hi,

I wish to point out an inaccurate allusion made in your blog regarding Francis Crick's idea on coding.

It is correct that Francis Crick had proposed such an idea. At the same time, he was always very clear that this is a theory for which no evidence exists and, if you examine his notes to the RNA Tie Club, you will find that he did not take the idea seriously himself simply because was evidence that suggested that the code was degenerate. Crick has himself pointed out that he was rather embarassed that people took the idea seriously because he did not. You also say that "In fact, the exact elegance of this scheme was such that all attempts at actually figuring out the genetic coding scheme for the next many years attempted to be compatible with the idea." This is completley inaccurate because it was very quickly established that the code was degenerate.

You use this example to say that, therefore, the methods in science are flawed because they are attracted to 'beauty'. This is silly argument, because the history actually shows that the methods in science are very robust, because they do not let even remarkably powerful ideas that are not correct to persist.

[jjensenii]

I know this is a few years on, but I want to reply to ymatt, if only for later readers:

The point is that whatever justification you give for science (or any other belief), someone could always demand justification for that justification, ad infinitum. Most people won't, but that doesn't invalidate the point.

So any view of the world has to start with certain beliefs that are accepted for no better reason than that they seem obviously true. But since what's obvious to one person need not be obvious to another (to claim that certain things are so obvious that they must be obvious to everyone is an argument from lack of imagination), it's quite correct to say that there is no Archimedean point of reference, or at least that we have no way of knowing which point of reference it is.