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.
Yes, that’s my take. And it’s the abstractness of codes that I see as important. Or, if you don’t like “abstract”, go with “arbitrarily representable”.
But it can be translated into just about any medium we like, because we can now read it.
I think that’s part of the problem in discussing whether or not it’s “code,” that we often view it as the “code of life,” and we wish to “decode” the meaning that is there. DNA knows nothing of any of this, of course, yet for us there is meaning, there is/was something to be deciphered.
Crick likely had some familiarity with codes in WWII, perhaps why he called the “DNA template” the “DNA code.”
For myself, it doesn’t mean much arguing either way as to whether it is a code. It acts as a template, fine, but it was “deciphered” and is read as a “code.” Humans see it as code, and I’m not going to argue with that. Clearly RNA polymerase doesn’t care either way.
Notably, “code” can have quite a broad meaning, etymologically meaning laws set out systematically in a book or some such thing, a meaning that is frequent today. Not much like DNA, yet not a system for conveying meaning, either. Hence I don’t see that one should be very picky on using such a term.
All that matters is how the “DNA code” works, the evidence that it evolved, and the fact that DNA happens to be far more forgiving of mistakes than are “computer codes.” I think that for our understanding, “code” seems to fit reasonably well (since to us a sequence means something), while it’s just a means of storing hereditary information for producing proteins and regulatory sequences. If we think it’s a code, it’s a code, but it existed well before the concept of “code” ever appeared.
Glen Davidson
But you can’t read it for meaning. It is not a blueprint. You cannot tell what a sequence will do or produce. you cannot tell what effect a change will have.
It is a template. It results in a sequence of chemistry.
The words information, code, read, and translate are metaphors. It’s chemistry. There is no meaning outside chemistry.
Without getting too philosophical, one could mechanically define it as a relation somewhat the notion of function f(x) in mathematics. Ideally we want f(x,y,z….) to map to one object to prevent ambiguity.
There are numerous codes which interpret the same strand of DNA differently.
Off the top of my head, there are alternative reading frames for coding, bidirection promoters, etc. etc.
The machines of life implement a model of a functional relation.
Example:
f ( ATG) = methionine amino acid
g(ATG) = start codon depending on context and for transcribing mRNA transcript
so we have two possible interpretations of the same DNA string.
Even more pointedly:
g ( AGGAGG) = Shine Dalgarno if not a coding sequence in a prokaryote likely as some sort of addressing and navigation road sign
whereas
f(AGGAGG) = Argenine Argenine if the DNA is a coding sequence
The same DNA string “AGGAGG” is interpreted two different ways in living organisms.
Even coding sequences there are certain string like palindrome endonuclease restriction sequences.
https://en.wikipedia.org/wiki/Palindromic_sequence#Restriction_enzyme_sites
5’GAATTC
has three meanings:
1. Glutamic Acid and Phenyl Alanine
but in the complementary reverse direction read it means
2. Phenyl Alanine and Glutaic Acid
and as navigation marker
3. Endonuclease Restriction site
So here we have at least three meanings for the same DNA strand depending on how the machinery reads it and according to context.
Let’s try another analogy. Take the cams in a sewing machine, or the holes in a Jacquard loom card , or the pins on a music box cylinder. A person familiar with the associated machinery can “read” these codes and predict the behavior of the machine.
You cannot do this with DNA. You cannot look at a novel DNA sequence and tell if codes for a protein or if it has a regulatory function. If you know the function you do not know what effect a change to the code will have. If you know the change, you do not know the consequences it will have, to the organism and to the ecosystem. You cannot know where a neutral change will lead.
In short, you cannot design DNA. It is not a code, not a language, not a blueprint. It has no grammar or syntax. it cannot convey novel meaning.
Those are just ways of following a template. You cannot predict the meaning of a sequence. You cannot tell if a novel sequence is a “design” or if it’s gibberish.
I think DNA can be arbitrarily representable in one sense. One could have an arbitrary sequence of templates that eventually produce a genome.
What is not arbitrary or abstract is the chemistry of life itself. That’s the part that is emergent, at least in the soft sense of emergent. One cannot predict the properties of novel DNA sequence.
There are probably some sequences that are “obvious” nonsense, but among those sequences that are superficially stochastic, you cannot tell if they are meaningful or gibberish.
DNA is just part of the genetic code, which is a real code complete with symbolization and actualization. The genetic code is arbitrary in that it is not reducible to any laws of physics and chemistry.
DNA is the template by which mRNA is transcribed. There isn’t any such template between the mRNA and the amino acids. The amino acids require special carriers and those carriers match up with the mRNA codons. That is part of the translation process. Nucleotides in, polypeptide out. Two different molecules that otherwise have nothing to do with each other.
GlenDavidson,
There isn’t any evidence that DNA evolved nor is there any evidence that the genetic code evolved. And there isn’t any way to test the claim that undirected evolution didit.
Agreed!
The machines however that instantiate such arbitrary relations are another story.
Those machines may not be the result of natural expectation since natural expectation is that there be no machines of any such complexity at all.
A Rube Goldberg machine is the pinnacle of arbitrary relations, but that doesn’t mean Rube Goldberg machines are the natural expectation of ordinary mindless processes with no pre-meditated goal.
A sufficient but not necessary circumstantial clue that there was a pre-meditated goal was that a system was put in a state far from natural expectation. It is not a formal proof of design, but it would be good enough for most creationists.
A machine that implements arbitrary relations in a repeated cybernetic manner like a Rube Goldberg machine or computer can be suggestive of design. The DNA string itself is not that persuasive to me of design except for the fact DNA itself only has a half-life of about 521 years. On physical and chemical grounds DNAs persistence is not the natural expectation of ordinary non-exceptional processes, but the strings in and of themselves may not be indicative of design, or at least not persuasively so.
What you are calling a symbol is better called a template.
It is not symbolic in the sense that language is symbolic.
With language, I can convey a new thought, or pass on new information. You can read a completely novel sequence and derive meaning.
You cannot know the meaning of a DNA sequence without implementing the chemistry. the meaning is not syntactical or grammatical. You cannot “understand” a novel sequence or predict what it will do.
In short, you cannot design.
My point is that you cannot design DNA sequences. The properties of sequences are emergent. You cannot predict their properties. You cannot “intend” an outcome because you cannot predict the outcome.
mRNA codons are the symbol. There isn’t a template for mRNA to amino acid conversion.
All your other spewage is meaningless. Strange that Venter designed DNA for an organism.
The properties of DNA are emergent? Evidence please.
Venter designed DNA for a bacteria. petrushka is refuted.
When you write down a sequence of letters to represent the bases, that’s a code. But the DNA itself is not a code (IMHO).
I’ll be happy to admit defeat if you can demonstrate that the genes were designed de novo. What I read is that they were synthesized, which is quite a different thing.
Show me any reference in the entire literature of biology to anyone being able to predict the properties of novel DNA sequences.
Show me any reference in the entire literature of biology to anyone saying that DNA’s properties are emergent- and what properties are those? DNA is basically inert. It doesn’t do anything without a lot of help.
Synthesized means they were designed.
I agree. DNA itself is not a code
Isn’t the real issue here the question of how DNA could have arisen by natural processes (not all of which might be classed as evolution).
Nick Lane summarizes some current ideas in the first two chapters of his his 2009 book. The second chapter is about how DNA could have arisen naturally.
It is called: “DNA: The Code of Life”.
Lane does use evolution as part of the explanation of how the coding scheme arose, and provides some details, but admits it is still very much an active and open research program.
His latest book expands on the original of life and how eukaryotes might have then arisen.
That’s not the issue I’m addressing.
I’m arguing that DNA lacks syntax and grammar and is not language like. You can’t read it in the abstract and you can’t design sequences except by trial and error.
You can’t speak DNA. You can’t utter new thoughts in DNA. You can’t really distinguish functional DNA from gibberish.
All it is is a template, a cookie cutter.
Synthesized means they were assembled. It says nothing about how the sequences were determined. I assumed that existing genes were copied from nature. I’ll eat my hat if Venter created entirely novel genes without reference to naturally occurring ones.
Language has syntax and semantics. Semantics involves representation. It does not have to do the full job of a natural language to represent.
Any formal language can have a syntax. DNA too has a syntax, I believe. Three letters per word, stop symbols, and so on. (I don’t have any significant knowledge of DNA structure so I may be wrong to say that is has a syntax like a formal language.)
Semantics involves representation. As I said in my other post in Noyau, representation in the language sense must have rules to determine correct versus incorrect, in cannot just be a causal correlation.
Where could these rules come from? If we are just using the concept of language to help explain the physical actions of DNA to another person, then we impose them.
If you want to argue that DNA has semantics like a language without involving us as the interpreter, then you can try to use evolution. The rules for correctness are whatever DNA was selected to do by evolution.
But I don’t think biology needs to justify DNA as language with semantics. We can say it represents but we also say tree rings represent tree age, or spots represent measles, or the position of the earth in its orbit represents the season. But that’s just us using representation to help explain. All that it really happening in all of those cases is just the operation of natural law and causation. No semantic representation beyond what we impose to explain.
Same for DNA.
How it came to function the way it does is the question I addressed in my first post.
The questions I’m addressing is whether new meanings can be uttered and understood in the DNA lexicon. Is it possible to understand a DNA sequence?
I think not.
That’s not how anyone else uses the word “synthesized”.
We synthesize ethanol. Does that mean ethanol was designed?
Of course not.
Not unless you’re the kind of fanatic who thinks gravity is really angels pushing things towards a center of mass.
In which case, every single quark and atom and molecule of the universe is specially designed. (God has its eye on the sparrow.) They’re all made to the order of god’s will and held in their proper places by god’s host.
Of course, that means when Albert Hofmann first synthesized LSD, which he himself did not predict, then he was just manufacturing one of god’s pre-existing designs. Thank you, angels of heaven!
I think that understanding what protein or regulatory action DNA represents is an example of what is meant by understanding the language of DNA.
That understanding is in us, not in the DNA transcription process.
We can use DNA to say new things by splicing sequences. Again, it is us that add the meaning in order to that explain what happens by splicing as creating new meaning/representation. It is not something inherent in the DNA. (Unless you subscribe to biosemiotics, I suppose)
What we can’t do is design sequences from scratch and predict what they will do. We have something like words in protein domains, but most eukaryotic evolution has been in regulation, and that’s hit or miss.
I can be wrong, but I believe design is impossible without trial and error. The notion that some super advanced being knows the outcome of changes, or front loads changes, is rubbish.
The problem with the design hypothesis is that design is impossible without evolution.
I’ll second the importance of abstraction. All human designed codes use arbitrary symbols to abstractly represent other quantities, meanings, or ideas. The medium doesn’t matter. You can sent Morse code with electronic pulses, or by tapping your fingers, or by smoke signal. That’s because of the abstract nature of the symbols involved.
DNA on the other hand has no arbitrary symbols or abstract representations anywhere. Genes are complex biochemical molecules that follow the laws of chemistry and physics in chemical reactions to produce other complex biological molecules. There is no message or meaning passed as abstractions. You can’t take a Lego codon and have it produce a Lego amino acid.
The closest DNA comes to being a code is the mathematical definition where code is a function that maps source set members to target set members. DNA may be a real code, it’s just not an intelligently designed abstract one.
I was considering commenting until Adapa chimed in.
Perhaps another time then.
I agree. Codes are abstractions.
Morse Code is a code because it exists only in the human mind. But the genetic ‘code’ exists in the physical interaction between aa-tRNA synthetases and their substrates. Codes exist in the natural world in the same sense that beauty exists in the natural world.
Back in the 80s a professor in my core biochemistry class told us about a comment made by a famous movie director ( I think it might have been Kubrick)
The director said that humans wont need to travel to the stars in spaceships. We’ll just beam the genetic code into space and aliens will recreate humans using the code. I think its a productive exercise to consider all the reasons why that wouldn’t work.
As a bit of an aside, it seems to me this discussion is touching on a rhetorical trick that theists use to support the existence of God and IDers use as evidence for the intelligent designer.
The trick involves subtly defining some aspect of the natural world in terms of an abstraction. Then somehow suggesting this abstraction is some concrete property of the natural world. Then guiding the audience into making the connection that if an abstraction exists there must be a mind ( ie. God) to conceive of it.
Mistakng the map for the territory. Also, reifying a metaphor.
Calling a mark on the map a mountain does make the mountain made of ink.
I absolutely agree. What typifies human design is protyping and refinement.
This is also true of the process proposed by Darwin.
It also seems to be how living things changed over time.
To me, the reason living things “look designed” is that they look as though they emerged from a process of prototyping and refinement, just as human designs do, but which is also the process of adaptive evolution.
They do NOT look as though they emerged from the mind of a designer who knew the best solution beforehand. Indeed, we know of no intelligent agent who designs this way.
petrushka,
Your diet can remain hat-free. Venter is reverse engineering at best. He synthesised an ‘artificial’ DNA strand, but its sequence was that of an existing one. Then he placed it in a cell already furnished with everything else. But of course it’s statements like this that get Creationists all excited: “Dr Venter likened the advance to making new software for the cell.”.
There are of course senses in which the analogy works. But like all analogies, there must be some point at which they diverge from the analogised system, otherwise they would be the same thing.
There are (as I said in Noyau) senses in which I am happy to use the terminology of codes, languages, etc. But my offhand comment to Mung that triggered this discussion was ‘You know I think the genetic code is not “really” a code too, don’t you?’ It was a tease following on from some discussion about what was ‘really’ controlling what in the emergent cell, which has some relationship to Hofstadter’s “who pushes whom around inside the Careenium”.
We can call it a code if we like, just as we can call anything a species if we like. We may or may not get the whole of humanity to agree with us. But this game the Creationists play is silly. This dogged insistence that it must be a code in every respect is one premise of a shaky syllogism. It inherits all the linguistic, cryptographic, communication, programming and protocol baggage of the poor, overworked little word? When one argues for clarity, by trying to avoid the equivocation inherent in using one word with multiple meanings, one is accused (in typical topsy-turvy fashion) of employing equivocation. There seems, as I have noted before, a private Creationist meaning for ‘equivocation’ which is the opposite of the convention. And they get annoyed, out of all proportion to the crime, with people who don’t accept their definitional approach.
All that really matters, for evolution and general understanding, is what the system does, not what we decide it is. There is plenty of evidence that it evolved, and no evidence, beyond bad analogy, for design.
OK, how is that related to the issue of calling the mapping of genetic sequence to protein a code?
Plus, I suppose one could argue that just because we cannot do it by design does not mean some other designer could via processes we cannot understand. I don’t think it is it enough just to call such a hypothesis “rubbish”.
Better to say like the issue with super-designers then is whether such a hypothesis is scientific. It seems to me to be just another god-of-the-gaps argument.
But as you point out later, the code is the abstract relation between the sequence of letters and the protein.
DNA just happens to be the implementation substrate for the code. But we could build a different machine to implement the same code; eg one where typing in the three letter code produces a physical sample of the corresponding protein (something like a soda dispenser).
Now one can argue that we impose the code on a physical process; it is not something inherent in the process.
But consider bee dances. Their movements can fairly be called an abstract representation of distance and direction to a pollen source. That seems to be a code with meaning to other bees. No need for us to impose it.
You could say that bees have neurons which allow them to execute and interpret codes, but neurons are just electrochemical implementations. Is there a principled difference between them and DNA transcription as implementation vehicles for abstract code mappings?
Right–and beautifully put, I think.
Why is it a requirement that we have to design DNA? We are not the designers of life.
But nice to see that petrushka is a strawman maker and equivocator. Undirected evolution, ie drift and natural selection, is impotent and couldn’t produce DNA, ever.
BruceS,
Bee dances, and other phenomena such as predator-specific alarm calls, are communication methods. They really do use symbols. A DNA codon, however, is not a symbol. It’s physical stuff, with a profile of charge density and variable affinity to its complement. It happens to form part of a physicochemical cascade which has reasonably tight coupling between that codon in similar contexts and the amino acid that sits on the other end of its cognate tRNA, mediated by the set of 20 AARSs which happen to have high specificity for the thing they stick on one end of a tRNA and somewhat looser at the other.
But mapping – that is done by us. The 3D charge densities can be written down as abstract symbols, and the consistent relationship with other 3D charge densities called a mapping, which can be transferred from lab to lab via ASCII or even Morse code. But this does not make the system being mapped a code. Or, if it is, where does the boundary lie? Many molecular configurations produce a consistent result.
A primitive system could involve a monotonous polypeptide. There is only one tRNA and one amino acid stuck on its end, with specificity at the ‘anticodon’ end permitting docking with between 1 and 63 different codons (it does not matter how many). We wouldn’t call that a code (IMO), except with hindsight.
Duplicate the AARS and tune specificity at either end to partition the set, and I guess now you have a code, because now there are two possible consistent outcomes. But I don’t think it has thereby become a symbolic or representational relationship. In the hypothetical 1-acid system, an indeterminate number of codons function as non-STOP positions. They don’t represent the acid any more than codons not corresponding to a charged tRNA ‘represent’ STOP. They cause a physical effect. This doesn’t change with AARS duplication and codon set partition.
Allan all you have to do is step up and demonstrate that unguided evolution can produce what we are calling the genetic code. Continually whining about semantics isn’t going to help you.
mRNA codons are the symbols. They represent amino acids.
A traffic light is a symbol and it is physical stuff.
Duplicate the AARS? You can’t even account for any AARS. You can’t account for a more primitive system, either.
Allan,
It was predictable that your opponent would ignore all the other points you made. It proves that your opponent is not interested in an honest and open discussion.
mRNA codons represent amino acids. They do not become amino acids via some physiochemical process. That means mRNA codons are symbols in the translation process. The genetic code is arbitrary in that it is not reducible to any law.
The genetic code is a code in the exact same sense as Morse code is a code.
You could try Lane’s stuff in the books I linked early as an entrance into the scientific research program which is trying to address the issues you raise.
That research far from complete, of course. But I am aware of no reason to pre-judge it as something which must fail.
Thanks- I will get the books and have a look. I am very willing to read what the opposition has.
Thanks again, Bruce- both books are ordered and I should have them by Friday- Prime is good! 🙂
Allan Miller,
Another good inscription for the clue-by-four.