Hello Dan,
Thanks for your kind comments. Here are my answers to your interesting questions:
You say:
""Lossless compressibility" [p9] may apply perfectly to a pattern that is already defined (like files in your computer), but only imperfectly to natural patterns (data from observations). Your result that "most empirical data carries an algorithmic signal" seems simply to restate the fact that there are evident patterns in the world. I cannot take this to mean, however, that the world in itself is pure ordered pattern, fully accountable in a set of algorithms. That seems too great a leap. Am I missing something?"
You are not missing anything, you make a fair point. That, as you say, may be the case. Ii find, however, interesting the fact that this 'algorithmic signal' seems to be present everywhere and that it can actually be quantified. There is a researcher that takes your point to the extreme, that empirical data sets are algorithmically random, you may be interested (although I cannot agree with his conclusions):
James McAllister's 2003 article, "Algorithmic randomness in empirical data" Studies in History and Philosophy of Science Part A 34 (3):633-646.
See also a strong reply:
Charles Twardy, Steve Gardner & David Dowe (2005). Empirical Data Sets Are Algorithmically Compressible: Reply to McAllister. Studies in the History and Philosophy of Science, Part A 36 (2):391-402.
You say:
"[p5] You say that "Producing random bits in a deterministic universe...would actually be very expensive..." But isn't this an argument AGAINST determinism? Perhaps nature cannot be forced into a mold that defines it either as deterministic or as non-deterministic. This could be the case if 'determinism' is actually a purely logico-mathematical concept and not ascertainable as an ontologically physical reality. The meaning of 'determinism' may be nothing other than logical implication (computability)."
Yes, if a deterministic universe were able to produce true random bits that would be an argument against determinism if I were ready to accept that that actually happens (as Quantum Mechanics may imply). If I consider this possibility is only to explain that classical mechanics implies a deterministic universe yet quantum mechanics is supposed to be a source of free randomness which is one of the deepest contradictions between these two mainstream theories. I only point out that if the universe is deterministic as one may believe (as I do), then there is this fundamental incompatibility.
You say:
"We are free to project this upon nature, but isn't it really our own invention? Similarly, we are free to imagine the universe as driven by simple algorithms (after all this works to some extent!). The really interesting thing, to me, is our apparent human need to know what the universe is in itself, in ultimate terms, apart from simply understanding what knowledge is (or can be) for us. Perhaps it is human thought that is driven by simple algorithms!"
Yes, it might be a projection, or even a mirage. But when a mirage works well (i.e. seems to explain and predict something) we use to call it a scientific model. I think it is fair to think that the world is based in simple rules if it turns out that, as it seems to be the case, it is comprehensible in a large extent with simple models of the world (including current scientific theories governed by simple formulae). If these simple rules turn out to produce the complexity we see around I think one can safely assume that they may be the responsible for the organized complexity in the world. It could, of course, be the case, that nature is fooling us making us to believe that rules are simple but actually are very complicated, looking only simple at the surface.
"[p4] From an engineering point of view, "what makes a cup a cup" is information; but from a physics point of view, what makes a cup a cup is structure. They may seem to coincide in the cup, which is an artifact, more than in the case of the human body, which is not. Information is effectively a set of instructions to the engineer to build the cup (a program). No engineer, however, knows how to build a human body or any natural thing. (The cup too--as a physical thing rather than a conceptual thing--is made of natural materials.) Such knowledge presupposes a blueprint from which to construct the natural entity. But a natural thing does not come with a blueprint, and anything looking like its blueprint is actually a product of an analysis that can never be assumed complete or exhaustive. The natural thing is found, the simulation (and the information behind it) is made. I suspect this applies to atoms as well as cups."
But a blueprint is a description which tells someone (if not you then nature) how to build something. The claim that only information makes a cup a cup rather than a human being is because both human beings and cups are made exactly of the same elementary particles and it is nothing but the way they are arranged that make one or the other. But let me know how that could be wrong from a purely materialist point of view.
You say:
"[p5] You say "if information is even more fundamental than the matter of which it is made and the physical laws governing that matter, then the question of whether these effects violate physical laws may be irrelevant." That is a big 'if'!"
Of course that is a big 'if', I would only dare to say so in a foundational question essay. Notice, however, that several authors think of information as more fundamental than physics itself. And it is a common practice in science to find all the time more fundamental structures on which previous ones were lying on. I don't have may troubles seeing information as more fundamental than matter but of course some may disagree (as they do, such as e.g. David Deutsch).
"The very question to be decided! It would undeniably be convenient if natural reality were fundamentally "informational", but that does not make it so. In the medieval world, violation of physical law was similarly irrelevant, because an omnipotent God was ultimately the cause of everything, including natural law and miracles!"
Right, but I don't think we are going backwards, replacement of explanations that were once laws before is a common practice, if not the goal, of science, and the replacement seems to have a direction in the form of models that explain more and more phenomena and have greater prediction power. Of course saying that the world is something won't make it that something, but when you find that something smells, looks and behaves as something one can be persuaded that it is this something. In this case, it seems clear that information is, in the worst scenario, something as fundamental (even if it is a worldview and not necessarily an ontological truth) as other variables in the physical world (matter, energy). Whether this is the case or not we will see or we may never know. Notice, however, that many physicists have jumped to similar conclusions even if they are treated differently, by giving the concept of symmetry (that you may see as information or an abstract mathematical object) a foundational role for even predicting the existence of new particles, that so far has been quite successful.
"[p8 re: "DNA construction] I think it is unfair to dismiss the interaction with environment as not a "true random function operating on the DNA"."
It is a generalized agreement that the macro world is fully deterministic and does not produce indeterministic randomness, so mutation may be only truly a random process if based on quantum mechanics based on current physics.
"What is a 'true random function'? Even by a mathematical definition, one cannot prove randomness."
When I write true random function I mean a function capable of producing truly independent random bits just as predicted by the mainstream interpretation of QM.
Thanks Dan, very interesting questions.
- Hector