Since William Dembski’s Being as Communion is released in the USA today, I’ll make this post the last in the series. To be honest, I’m a bit hesitant to comment on the culmination of the book, since it deals with his Law of Conservation of Information, on which he has published stuff before and received the usual criticism (that is, usually, ideologically-motivated dismissal).
Chapter 17 is, at root, an explanation of how the concept of “search” is a useful one in dealing with questions of probability – one should not see it as a restricted human activity, but as something built into the very fabric of nature. This is not really impacted by the efforts made by skeptical biologists (in particular) to deny that the genetic code and its translation is “really” information, but only an “analogy” for it.
In my view, such criticisms are tendentious and the semiotic nature of genetics is beyond doubt. But if you remember, the book started with the concept of the “probability matrix”, a search landscape defined by the interests of the observer. You select a feature, or group of features, from the mass of reality, and look at the probabilities that the process you are interested in would bring the actual outcome about. “What are the chances that the king of Sweden will throw 13 with two dice?” Thus an outcome of nature, even if viewed apart from teleology, can be understood probabilistically in terms of search. In that sense, all science is search and the semiotic nature of DNA irrelevant.
Evolution, then, is legitimately viewed as an algorithmic search, which is agreed at least by those who produce evolutionary algorithms to simulate it. As is well known, Dembski utilized the then recently-proven “No Free Lunch” theorems to say that, when all factors are considered, no type of search is better than random search, including evolutionary searches.
More recently this led him to postulate a Law of Conservation of Information, or actually to consolidate the idea, first put forward by Nobel-prizewinner Peter Medawar in the 1980s. Medawar had shown, as others before him, that in mathematical and computational operations, no new information can be created, but new findings are always implicit in the original starting points – laws and axioms. Although Dembski doesn’t mention it, I’m interested how congruent Medawar’s Law is with what was argued by mathematicians at the infamous 1966 Wistar Conference (which he chaired), and which I mentioned here. He seems to have been persuaded. In that blog I quoted Kurt Gödel’s logical objection to Darwinian evolution:
The formation in geological time of the human body by the laws of physics (or any other laws of similar nature), starting from a random distribution of elementary particles and the field is as unlikely as the separation of the atmosphere into its components. The complexity of the living things has to be present within the material [from which they are derived] or in the laws [governing their formation].
As quoted in H. Wang. “On `computabilism’ and physicalism: Some Problems.” in Nature’s Imagination, J. Cornwall, Ed, pp.161-189, Oxford University Press (1995).
Gödel’s argument is that if evolution is unfolding from an initial state by mathematical laws of physics, it cannot generate any information not inherent from the start – and in his view, neither the primaeval environment nor the laws are information-rich enough. In other words, either information must be added later, or some currently invisible front-loading would be necessary. The one mathematical impossibility, he says, is the spontaneous generation of the (specified) complexity of life simply by random variation and selection from nothing.
As developed and applied to search by Dembski and Robert Marks, Medawar’s Law shows:
…that searches must employ existing information to successfully locate targets, and that locating targets through search never outputs more information than was inputted into the search initially. Simply put, searches, in finding targets, output information. CoI, as we have developed it, shows that the output cannot exceed the input.
Indeed, he shows by examples that “easier” searches may, in fact, need much more input of information than they output – there is a kind of entropic principle in facilitating searches. Dembski’s team has shown where, in all individual existing evolutionary algorithms, hidden information has been “smuggled in” by the programmer.
More generally, he shows how a generalisation by Ken Miller about what is necessary to generate information in evolution is demonstrably false: “Just three things: selection, replication and mutation … Where the information ‘comes from’ is, in fact, from the selective process itself.”
D. quotes J S Mill’s logical dictum that no set of circumstances that produces different outcomes can be regarded as the sufficient cause of one outcome. He points out that many computer simulations involving selection, replication and mutation in fact go nowhere – those algorithms that prove useful in optimization problems are those carefully and specifically designed by engineers, like one known to D. who styled himself a “penalty function artist.” Knowing the problem he has to solve, he “carefully adapts the penalty function to the problem and thereby raises the probability of successfully finding a solution.”
Commenting on my blog, linked above, materialist Lou Jost in effect conceded the legitimacy of Conservation of Information by saying that it is the environment that supplies all the necessary new information – the same point made to Dembski in 1999, by Ernan McMullin. But by the same Law, that simply shunts the question about the origin of that information into another complex and apparently random product of the laws and initial conditions of the universe, whilst doing nothing to explain how the system as a whole achieved a workable “penalty function.”
If teleogical processes were admitted to evolution, new information would not be a problem. But if, as currently, it is insisted to be non-teleological:
Conservation of Information entails that as we regress biological information back in time, the amount of information to be accounted for never diminishes and may actually increase.
Accordingly, ateleological views of evolution (and come to that, of the whole history of the universe, of which life is just the most troubling example) require that all the information we see has to be present “in embryonic form, at the Big Bang and at every moment thereafter.” That sounds congruent with some TE and ID “frontloaded” views of evolution, but where in the Big Bang could such information be found?
So where is it? How is it represented? How does it unfold? The environment is sure to figure into any answer to these questions. Yet, merely invoking the environmenst as evolution’s information source is, without further elaboration, empty talk, on the order of invoking the interstate highway system as the reason for Walmart’s business success.
As an example he pictures a robotic machine that lifts Scrabble pieces out of the box to spell out Richard Dawkins’ celebrated phrase, “METHINKSITISLIKEAWEASEL”. That machine is part of the Scrabble set’s environment, but that does not make “environment” in any sense a scientific explanation.
Dembski does not end his book on an argumentative note: his chapter on The Creation of Information examines the suitability of classical information theory being applied to creation, and even, specifically, Trinitarian creation. The final chapter, summarising the book, also reminds us that he is proposing a metaphysical system, at the heart of which is the title Being as Communion – that to exist is to be in informational communion with all else that exists, and most importantly with God. Information is not an epiphenomenon of matter, and intelligence is not an accidental outcome of matter’s interactions, but the primary reality of the world.
A viewpoint, in my opinion, well worth serious consideration.
