For your fitness

There are various artificial life or synthetic life initiatives aimed at creating from non-living components, quasi natural life forms, chemical systems that are self contained and infinitely self-replicating. If such an effort were successful it would provide a great deal of knowledge about the nature of living systems. It would also provide methodology for engineering new life forms on a new level beyond the current genetic engineering methods.

Many scientists will assume that a living system is simply a molecular mechanism no different from in its fundamental nature from an artificial mechanism. However, there are other scientists who have suggested that perhaps this isn’t true. An early example of such a comment was made by the physicist, Erwin Schroendinger. In a thin volume entitled, “What is Life?”, he suggested that living systems are perhaps not fully covered by known laws of chemistry and physics. However, he didn’t really develop the idea in his book.

Perhaps the first formal development of such an idea is contained in an essay by another physicist, Eugene Wigner, entitled “The probability of the existence of a self-replicating unit”. His conclusion is that the probability is zero. In this model, a living system is represented as a state vector: v. Its environment would also have at least one state which permits the organism to multiply: w. The total state vector of the system, the organism and its environment would be represented by the direct Kronecker product of these two vectors: v X w. After replication, the state vector would be represented as v X v X r, that is two vectors representing a pair of organisms in the altered environment. This interaction was assumed to be random, more specifically, to be governed by a random symmetric Hamiltonian matrix. This assumption might be questioned, however, it was the same assumption that enabled John Von Neumann to complete a proof that the second law of thermodynamics is a consequence of quantum mechanics.

This is an article for a popular audience so these comments probably go over the heads of most readers. Suffice to say that Eugene Wigner and John Von Neumann most definitely were not cranks.

In 1964, P. T. Landsburg, a professor at University College in Cardiff, England published an article in Nature that reiterated Wigner’s results and developed them further. More recently, Prashant Chakrabarty has compiled various arguments along these lines in a paper called, “Non existence of quantum mechanical self replicating machine” that can be found on-line. So there has been some suggestion over the years that a self-contained, infinitely self-replicating system is paradoxical from the standpoint of quantum mechanics.

In this article, we try to develop this idea and suggest some implications and possible experiments.

Perhaps the most pressing and immediate implication is that any effort to create an artificial quasi-natural self-replicating system based on the the assumption that such a system can be purely mechanistic may encounter problems. The products of such an effort may not be infinitely self-replicating. They might divide a few times, but there may be a cumulative degradation of the species with each replication that causes each lineage to eventually terminate as a result of non-viability.

One of the problems may be that a quasi natural system must necessarily exist in an aqueous environment at room temperature. Under these conditions, there is a large amount of molecular motion within the system that will cause disruption to any structured activity. The result is likely to be a chaotic system. As such, it exhibits characteristics for which chaotic systems are known such as:

• Sensitive dependence on initial conditions.
• Bifurcations that exist at intervals of a reference variable that are dictated by Feigenbaum’s number.
• While definite statements might be made about average state of a population of such systems, it is impossible to predict the outcome in any specific instance because of the sensitive dependency and the effects of Brownian motion which are essentially random.

Therefore, with each successive generation, there may be a cumulative divergence from a vector state that is the functional equivalent of a known good initial state. The inevitable average outcome may be a divergence that is so great that it results in non-viability.

What sort of stabilization would keep the system “on-track”. It would have to be plausible yet obscure. An idea that occurred to me a number of years ago is derived from what is known as Landauer’s principle, first argued in 1961 by Rolf Landauer, a researcher with IBM. Roughly