Lots of heat surrounding this question.
My take is that a code must be a system for conveying meaning.
In my view, an essential feature of a code is that it must be abstract and and able to convey novel messages.
DNA fails at he level of abstraction. Whatever “meaning” it conveys cannot be translated into any medium other than chemistry. And not just any abstract chemistry, but the chemistry of this universe.
Without implementing in chemistry, it is impossible to read a DNA message. One cannot predict what a novel DNA string will do.
DNA is a template, not a code.
Go to it.
You are dealing with an unknown designer, it may have done this millions of times.
And Harry Potter may be a real person in some alternate universe.
If you want to do actual science, propose a solution for the problem of emergence.
Otherwise, you are appealing to magic.
No, I am pointing out the possibility that life on Earth might not be a theoretical designer’s first time at the rodeo.
I understood that, but I don’t see how it changes anything.
It doesn’t affect the problem of design.
Consider Dembski’s numbers for a moment. Or, if you don’t like Dembski, consider Wagner’s estimate of the number of sequences that could code for proteins. (Never mind, for the moment the problem of regulatory sequences.)
The number is hyper-astronomical. It is not possible with the number of particles in the universe, to catalog the number of possibly useful coding sequences.
So that brings us back to the question of emergence, and the question of grammar and syntax. If there is no regularity to phase space, no way to read sequences for meaning without cut and try, then design becomes rather difficult.
petrushka,
That suggests that Morse Code would be more properly defined as simply an alphabet that is isomorphic to the English alphabet.
hotshoe_,
Ninja’d by hotshoe_! I should read ahead before responding when catching up.
walto,
Perry Marshall laid it out explicitly on another page on the site with the “3.1 million dollar prize”:
I don’t see an explicit definition of a code there, but my eyes were glazing over from the sheer hucksterism.
Yes. I would put all languages in the same definitional bucket.
I’m not smart enough to figure out whether any computer languages have the equivalent of connotation, figures of speech, double entendre. There is this:
http://www.ioccc.org/winners.html
But emergence does not appear to be obfuscation.
That approach just packs design into the concept of code, doesn’t it. Then you have to deny that DNA sequences are codes or you have a designer. But the first step is what I don’t see.
Petrushka says, if it’s a language and one understands the “words” one must be able to understand every sequence of the “words.” But I don’t see why that must be true. I’m not saying that it isn’t, mind you.
But your approach here, simply assumes that, I think.
What is supposed to follow from that? We can’t catalog the number of possibly useful sentences in English either, can we?
Here, I don’t know what you mean by distinguishing grammar and syntax. I take it those are both rules for well-formed formulas. Are there no rules at all with respect to DNA sequences?
And so what if design is “rather difficult”? What should we take from that in your view?
Patrick,
Like Braille! Still, I think both qualify as codes-of-sorts. To anyone who cannot access the mapping, it’s cryptography. There’s no particular rule that 1:1 mapping is an alphabet but not a code. An alphabet is after all roughly symbolic of sounds in the language, with caveats.
I bend like a willow on this one; I can see arguments for and against for most things to which the label may be applied.
Can you explain what you mean by that and what you take its significance to be, Alan? I’m not following petrushka at all here.
That’s my sense too. But petrushka is so adamant, that I feel like I’m missing something.
Allan Miller,
I don’t think it actually matters what we call it as long as we share the same definitions. It is what it is. It’s the lack of a clear definition that results in some of the disagreements with IDCists. And, as others have pointed out, that is also the root of some of the equivocation observed.
I’m not interested in a dictionary war. I don’t care about classifying sequences as codes or languages or whatever. I’m not a bucket person.
At the risk of driving off the road a bit, I will say that Chomsky (I believe) ran into the small problem of generating meaningful statements by following grammatical rules.
I think he encountered that problem because Chomsky is fundamentally incapable of taking evolution seriously and incapable of understanding that verbal utterances evolve.
But back to my roadway. the problem I’m addressing is not the definition of code. I couldn’t care less.
What I care about is whether the genetic code enables “talking” in genetic code. Whether it is possible to parse genetic-like codes and determine their meaning without implementing them in chemistry. Do they have an abstract layer that is independent of chemistry?
Joe/whatshisname made an honorable stab at this with his reference to Venter and the synthesis of a bacterial genome. But I assert that what Venter did was take known genes and instantiate them. I don’t think he invented any de novo genes.
And following a challenge made elsewhere by one of our members, I don’t think you can take a working genomic sequence, twiddle a couple of bits, put the original and mutated sequences side by side, and tell which is which, unless you know them. You can’t tell by grammatical analysis, which is working and which is garbage.
Thanks for that.
I read this lengthy advert and note that the author says over and over that DNA is a code (not like a snowflake, goddamit, but like a language!). And he offers a bunch of money to anybody who can give an example of a code that was not created by an intelligent being. I think we can glean from this that he takes being a code to imply meaning/intentionality. OK, fine–he can define “code” any way he likes. I have no problem with that, anyhow.
But then he nowhere deigns to tell us why he’s so sure DNA *IS* a code. He just repeats that it is a whole bunch of times. Maybe you have to buy his book?
Hence the current interest in deep learning and the genome. If it was simple or even just hard, it’d already be done.
Ok, thanks, this is helpful.
As I understand it, your claims are that (i) nothing ought to strictly be called “meaning bearing” (or anything along those lines) if it would be impossible for anybody to “understand it” in anything like its entirety; and (ii) it would be impossible for anybody to ever “understand” DNA sequences in anything like its entirety (whatever that might be).
From those we can derive (iii) DNA sequences are not “meaning bearing” (or anything along those lines).
Is that basically it?
It’s the opposite of determinism. Of the 90 odd elements found naturally, only a few are needed for life; carbon, hydrogen, oxygen, nitrogen, phosphorus, sulphur and twenty or so others at much lower levels. These combine in various ways to form molecules such as the ubiquitous amino acids. Simply stringing amino acids together in particular sequences produces a panoply of biomolecules that are essential to life’s diversity. At each step up from atom to simple molecule to larger molecule to subunit to, say, ATP synthase to cell to multicellular organism there is an emergence of properties
not obviousunfathomable prior to the assembly.ETA more drama
I am not adamant about definitions and buckets. I am a bit firm on what I think is an important difference between genomes and language.
The ID assertion is that DNA is a language, and languages are characteristic of intelligent beings. I will punt, for the moment, on what constitutes an intelligent being.
I’m saying that genomes are not like languages. You cannot read them. You cannot look at sequences and say this one is valid sequence and this other one is invalid.
Now, there are languages we could not read without a Rosetta Stone, and there are language-like manuscripts (Vonich) that we cannot distinguish from gibberish.
But to a speaker of a language, utterances are meaningful, well structured, or not.
I don’t think it is possible to have a speaker of DNA.
Alan Fox,
And I take it the use of emergence here is the idea that design requires what you’re calling determinism above–that it’s inconsistent with emergent properties?
I agree that perhaps quantum computers might be able to do it. Simulating quantum events is definitely one thing quantum computers can do faster than classical computers. But a problem is that we have not reduced chemical theories to quantum mechanical theories!
In god of the gaps, I was not referring to gaps in species, although I suppose that could be one. I was referring to any gap which is still an open research program. Such as OOL from non-living matter. Or origin of sex. Or why universe appears fine-tuned.
petrushka,
I think I’ve got what you’ve been saying now (but take a look at my post above to make sure). Thanks.
Is it a priori vs a posteriori? It is that (I think Petrushska is saying this too) what can emerge is impossible to predict or design. What genetic changes may achieve in the phenotype can only be seen as a result of how the phenotype performs in the particular environment.
BTW, Alan, I don’t think that argument against design works I.e., I don’t think design is inconsistent with the existence of emergent properties. Also, I take emergence to be an issue involving reducibility, not determination.
I’ll let petrushka speak for himself here. I sense that your approach is a bit different from his, though.
Of course I don’t make good arguments against “Design” because I don’t take “Design” seriously. There are those that make better arguments for evolutionary explanations.
I don’t follow this at all.
DNA obviously has meaning. It is the basis of biochemistry.
The feature of language of that DNA lacks is syntax and grammar. It is not possible to look at a sequence of DNA and tell whether it is well formed or not. It is not possible to read it for meaning without implementing the chemistry. There is no level of abstraction that can derive its meaning in the absence of chemistry.
That is why I called it a template rather than a code. Perhaps my vocabulary is not felicitous. But it appears a few people are climbing mountains to avoid my intended meaning.
Design can be consistent with emergence. We call it evolution.
BruceS,
Hmmm. My perspective is more informed by Richard Dawkins’s: it serves a purpose for the DNA!
I don’t quite get the relevance of this. A physical, causal process need not be completely deterministic in order to be a physical causal process.
I don’t think we ‘must’ be involved in it to call it a code. But there are regularities we can observe which aren’t really ‘lookup’ rules, but can easily be made into lookup rules by writing the correspondences down on a piece of paper. But down in the system, it’s just binding energies, not any kind of lookup.
Every one of the 64 possible triplets has a unique shape and charge density. One can divide the triplet set in half, such that each of the 32 in one subset has its exact complement in the other. These complements fit together snugly. They are physically attracted. A whole bunch of charged tRNAs are floating around, but that complementary to the triplet currently at the growing peptide end binds most strongly, and the others are excluded. Likewise for the AARSs that charge the tRNA, and are the closest thing to a ‘mapping’ mechanism.
For bees, there is actually symbology. The bee’s waggle is not a molecule, falling into place impelled by forces that we struggle to conceive of at our level. It’s a macro-level object, moving in a way that the brains of other bees can translate into a foraging flight. I guess nervous systems seem to be the key distinction in my mind!
It doesn’t solve the issue – we are the issue. Codes are a human conception. we can look at some things and say they are, and others and say they are not, and quibble endlessly about others. (eta – but these extensions are still adjudicated by reference to our own systems).
I guess the issue might boil down to one of scale as well. ‘just biochemistry’ does not work in the same way at every scale. The electrostatic force has limited range, but in the range it operates, it is very powerful. The translation system is driven by such thermodynamic interactions, which are more ‘law-like’ (though stochastically so) than the more detached, still less ‘compulsory’ emergent interactions of whole-organism systems-with-brains, and their various comms methods.
I’m enjoying Nick Lane’s The Vital Question. He gets to grips with the first two issues.
Patrick,
Yep – among those others being me! It’s actually a textbook fallacy of equivocation. But one is derided for not going along with the equivocation, for arguing against false equivalence. Shrug.
At some point it has to be. Both the waggle and the response to the dance. Innate behaviour (or the information required to produce it) has to be in that zygote somewhere!
ETA clarity
petrushka,
You’re making perfect sense to me, Sir.
I do not pretend to be a great or deep thinker, or a careful writer. I also reserve the right to be completely wrong about emergence. Perhaps tomorrow someone will be doing protein folding on an iPhone.
I see no contradictions between what I’m thinking and what most others are saying. Perhaps some differences in emphasis and differences in levels of certainty. Some people think it’s just a matter of time before we will be able to design biology. But not many think it will be easy.
My own gut feeling is it belongs in the same bucket as faster than light travel, time travel, ESP and such. When I’m in a bad mood, I’d include strong AI.
petrushka,
As I said, I think I understand your argument now.
Alan Fox,
Sure, but it is not ‘a’ molecule. Nor is the information being transmitted encoded in DNA. After all, the bee is saying where to find nectar in the current environment. Innate triggers help it do so, but the info it is conveying is not instinctive, but locally contingent.
I don’t claim that you’ve contradicted Alan, only that the point you were making is somewhat different than the point he was making.
Hmm. This might be getting a bit off topic and it’s late for me so this will be my last comment probably till the weekend. I’m not expressing any form of skepticism here. The fact is bees, without any obvious learning curve, are imbued with the ability to convey information to fellow workers. Your pointing out that the information is contingent upon where the food source is, for me, adds to the mystery as to how this innate behaviour is inherited – as it must be. I find it fascinating. Another layer of emergence.
Yes, I read most of it. But I must admit I was intellectually exhausted by it and ended up just skimming the closing chapter or two, which include the origin of sex explanation, as I recall.
I’ll probably go back to finish it properly at some point.
His style is definitely more detailed and more technical than Dawkins’ style.
Its not determinism I was trying to refer to. The addition beyond simple causal process is the problem of how to understand misrepresentation or malfunction.
We an say transcription failed, even though that failure will still be due to some causal process. We can say the bee dance failed to communicate, even though there will still be a causal explanation of the events underlying the failure. So there is a judgement, a norm, a right/wrong condition involved in DNA transcription or in bee dances that is not there (for the organism) in tree rings or in spots and measles. Where does that come from? Is it just us imposing our judgement? Or is it inherent in the organism?
I should also say that I may be reading more into the argument than was intended. After all, one could argue the very existence of a mapping means there must be an intelligent entity who designed the code. No need to get into semantics versus causal relation and no need to worry about right and wrong.
But then it would just be another “it’s too complex to have evolved” argument, like irreducible complexity. So I decided that when Frankie used words like “representation”, I would take him to mean it in the philosophical sense, so the argument would be more than just another instance of complex therefore designed.
Got it, lots and lots sequences estimated and with what we know with a couple thousands years of accumulated knowledge an insurmountable hurdle. No argument.
I would say that is also true.
Would it be more or less difficult than creating the universe and all those particles?
ID includes such a designer in the set of possible designers.
Creating a universe would appear to be magic. I don’t see that the appeal to a creator has any explanatory power over I don’t know.
Alan Fox,
By the same token, much of our communication has an innate basis too. We are language sponges: Genes give us the capacity to transmit visual and verbal information, and an apparent eagerness to do so. There is less hard-wiring than in the bee case, but still.
My basic point, though, was that molecules are ‘obeying’ physical rules at the molecular scale, which is the level of the genetic code. To say that certain instinctive communication methods are molecularly encoded in some way is … equivocation! 😉 The electrostatic rules of the genetic code are not applicable in higher-level communication, even though ‘molecular’ genes have a role.
Or for equivocating!
Elizabeth,
Yep, that’s what they seem to call ‘trying to avoid ambiguity’ in some circles …
petrushka,
Strictly, that’s not possible with (most) sentences either. One frequently needs extratextual, contextual information.
In the case of proteins, it’s a rather complex onion of secondary, tertiary and quaternary structure, promotion and inhibition, ultimately feeding back through net effect on replication frequency which is environmentally conditioned. None of this can be determined from primary structure of course. But the complexities of emergence aren’t so much the issue as that peptide and xNA strings simply don’t have a syntax – or certainly not at the codon level, unless you count Start and STOP, both of which any random RNA string of sufficient length will possess.
BruceS,
I think the ‘adjudication’ is at the level of the DNA itself. A strand of DNA is either successful or unsuccessful as a result of the phenotypic ‘tricks’ it performs. Success is measured by persistence, which includes successful replication. If a supposed error is made in replication, it’s not an error at all if it enhances the persistence of DNA strands copied from it. The generality of misreads, nonetheless, tend towards optimising fidelity. Likewise if transcription errors occur, they ‘really are’ errors if the result is detrimental to persistence. The persistence of better and worse ‘bee dance genes’ can be understood in the same way.
Well, I do dispute that it is, internally, a ‘mapping’. It can be represented as such in a matrix. But its simply one AARS having specificity for a set of tRNAs, another for another etc, which creates a pool of charged tRNAs covering most of the codon set. Those tRNAs have physical affinity for their complementary codon. One could analogise to a set of paint brushes with screwdriver ends – say Flat, Philips and Hexagon. 3 charging systems grope for their specific screwdriver ends and put paint on the other – each only has one kind of paint, and can only accommodate one bottom end.
Then these charged bushes dock into recesses in a pattern of screws and daub the paint on the other end onto a sheet of paper. Because of the physical limitation on docking, a given set of screws will always produce the same pattern on the sheet. But I don’t see the shape, either that on the brush end or its counterpart, as ‘representing’ the colour it ends up associated with, any more than the paint colour represents the shape.
We can easily create a mapping from this. We can draw the shapes – we can even represent them symbolically in 2 dimensions – but that does not make the represented association a mapping.
This is, I realise, simply a semantic exercise.
I think he does mean more – it is ‘a code therefore designed’. But because of the physical constraints on the linkage between parts, I don’t think the language of symbolism is appropriate. It’s not wrong, but no hard conclusions on evolvability can be drawn from it.
As an aside, the energy of tRNA activation and mRNA docking actually provide some of the energy for peptide bond formation. It’s not just Scrabble tiles.
In linguistics, the theory goes that you always need extratextual, contextual information. Always, without exception. There’s no word or sentence that can be understood (i.e. that can convey any meaning) without knowledge of the entire framework of grammar.
This theory is easy to verify. Do the terms “word” or “sentence” have any meaning themselves without a theory about what they are? How about the term “meaning”?
Yes, that is the teleosemantic theory of meaning/representation; it has to do with what representation was associated with reproductive success in the history of the organism.
So a taxonomy of ID arguments on finding a complex structure:
1. It is irreducibly complex: it could not evolve by known mechanisms.
2. It has a function, and a function requires a designer.
3. It has a meaning like a language, and a meaning requires an intelligent agent. Further, the meaning is there, it was not added by us
One can interpret Frankie’s argument as #3. Evolution can explain 1, can explain 2 by saying that it is as-if function. We can also use evolution through teleosemantics to try to explain 3, but there are some philosophical arguments which challenge that. But enough philosophy for this thread.
We may not think that meaning is actually there for DNA, but rather that we provide it simply to explain the biochemistry. But that point of view is harder to maintain for entities with neurons, like bees or us, even though neurons are still “just” biochemistry.
Of course, one could deny that meaning actually exists in any case, but that is the extreme reductionist that (correctly, I think) upsets most commenters at this site.
On the hand hand, many of those commenters will effectively accept that sort of reductionism for moral values, ie that they don’t really exist. But that is a topic already well covered at TSZ.