Of Minds and Language (17 page)

In our current state of ignorance on the ontogeny and phylogeny of mind and all of its component parts, including the device for language, it is safer to move to simpler biological systems in our efforts to distinguish between biology based on universal principles of physics/chemistry, and biology based on local, modular, interactive promiscuity. For this I turn to the reaction-diffusion models first proposed by Alan Turing and which still form an active focus of theoretical biology. The case in hand concerns the appearance of seven stripes of activity of genes involved in segmentation of the larva of
Drosophila melanogaster
along its proximal-distal axis. Turing developed equations (taking on board differences in rates of diffusion of two interactive molecules and subject to random perturbations of Brownian motion), which showed an initial homogeneous solution settling down into a series of standing waves of concentration. The inference here being that something similar occurs during segmentation. Ingenious but wrong. In essence, as with everything else in biology, each stripe is the result of very local networks of interactions between a variety of modular units in which a particular permutation of interactants is specific for each stripe. Stripes do not arise as a consequence of gene-independent chemical and physical processes operating in a “field.”

D'Arcy Thompson similarly proposed in his once influential book
On Growth and Form
¹
that the laws of growth are independent of genes in that diverse animal body plans can be circumscribed by Cartesian coordinates, with a little appropriate bending here and there.
²

The well-known early nineteenth-century debate between Geoffroy Saint Hilaire and Cuvier has been introduced by Noam (see page 23) as another example of early antecedents in the argument for what he has called “rational morphology,” a position he claims is supported by recent results derived from comparisons between species of the molecular genetics of ontogenetic processes. Geoffroy argued that there is one animal body plan embracing both vertebrates and arthropods, as any sharp morphologist could deduce by examining a lobster on its back. Over the last decade many of the networks of genes responsible for body plans have been elucidated, and many, if not all, of such genes are shared by lobsters and humans. Notwithstanding some fashionable return to Geoffroy by some biologists, does such widespread sharing support the concept of an ur-body plan? Are the tens of major body plans (phyla) in the animal kingdom, and individual biological variation in general, an illusion, as Marc Hauser has advanced? In the background of what I have introduced above the answer has to be no.

Biological variation arises from differences in combinatorial interactions between shared modular units, from genes to neurons. Such sharing does not specify a “rational morphology” of an ur-body plan, rather it indicates, as Darwin taught us 150 years ago, that life is a process of continually evolving differences and that, so far as we know, there is one tree of life on earth occupying a minuscule fraction of the totality of phenotypic space. Hence, it is not at all surprising that genetic modules are shared by all subsequent life forms once such modules were established, long ago in the ancestry of animals. As with all historical processes, subsequent steps are contingent and constrained by earlier steps. Furthermore, we are not in a position to consider the ur-modules or the ur-plan “rational” or “optimal” for we do not have an alternative tree for comparison, any more than we can say that the genetic code is “rational” or “optimal.” Given what we know about the large amount of stochasticity in evolutionary processes (see section 6.1), we are on safer grounds viewing all such features, in the words of Francis Crick, as successful “frozen accidents.” Noam might suggest that a biology of language as “one damn thing after another” is a “worst possible solution,”
³
but there seems no alternative in the current state of our understanding of biology in general. It is nothing but one novel permutation after another of a relatively small handful of gene/protein modules (possibly as few as 1,200) whose chemistry makes them highly susceptible to such promiscuity of interaction and co-evolution, thus leading to the generation of novel functions.

To answer this we need to explore sex – which is an odd phenomenon. From the point of view of the stresses I am making above, it is indeed odd that as a consequence of sex, all of the ontogenetic networks have to be reconstructed from scratch. Newly fertilized eggs contain randomized sets of parental genes that have never before co-existed, and that need to renegotiate, step by step, the patterns of contact required by the history of a given species. In this respect, the making of each individual is unique, in addition to the unique influences of locally expressed epigenetic and environmental factors. I have argued elsewhere that sex inevitably leads to the construction of a completely novel individual; that is to say, individual ontogeny is a highly personalized process of
total
nurturing from the moment of fertilization onwards. Importantly, it needs to be emphasized that the genes are as much part of such widespread nurturing as the more traditionally recognized environmental inputs. If, say, a given gene is participating in a network of 100 other genes, then, from the point of view of that gene, the other 100 genes are part of its nurturing environment. There is no false dichotomy between nature and nurture in this scheme of things – all is nurture in a world of modular biology; an ongoing process throughout an individual's lifetime. Furthermore, there is a sense in which the zygote (the first diploid cell) is a blank slate (give or take some epigenetic influences) in that a process of reconstruction starts at this point (Dover 2006).

It is because of sex and the constant generation of new, unique phenotypes that I emphasize the central role of individuals as units of selection or drift in evolution, and as a potential explanation for the subjectivity of consciousness and free will. Individuals produced by sex, whether uni-, bi-, or multi-cellular, are the only real units of biological operations. Their constituent genes and proteins are not: they have no functions, no meanings, in isolation. Neither have populations nor species. They are all abstractions as we willfully ignore the variation within each category.

I do not think that individuals are just my choice of “abstraction.”
⁴
For example, there is no one “human nature” – only millions upon millions of different takes on human nature as each individual emerges, alive and kicking, from its highly personalized process of nurturing (Dover 2006). “Average” has no heuristic meaning in such a situation. Men are taller than women, on average – true – but this does not help either in the prediction of height of a given man or woman, or in the prediction of sex of a given height. Nor can we measure the height of an abstraction. We objectively measure the height of an individual at a given moment in that phenotype's real lifetime.

Individual biological variation is not an illusion, it is at the heart of all that happens in evolution and ontogeny. And the same can be said of all sexual species – including Noam's ants
⁵
– for these too we can dismiss the old irrelevant nature-versus-nurture debate in terms of the individualized processes of nurturing involving all of the networking genes. There seems little need to say it, but ants too have “blank slates” at the single-cell stage of a fertilized ant egg.

Are “principles” and “parameters” to be found in the forms and functions of networks? Networks are evolved structures and their topology (the pattern of connections between interacting units) reflects the history of successfully functioning contacts. Some network nodes are highly connected, perhaps indicative of their early origin. Other nodes form into tightly connected sub-networks which have been shown to be conserved as sub-networks across widely separated taxa. The quality of the contact between units at the nodes reflects the differences in their chemistry, as explained earlier, in addition to a large number of local influences of temperature, pH concentrations, and so on. Are topologies (or at least the widely conserved sub-networks) equivalent to “principles,” and are the local influences equivalent to “parameters” of language acquisition?

In the discussion on optimization properties with reference to Massimo's and Donata's “minimax” concept, Noam suggests that “if you take a parameter, and you genetically fix the value, it becomes a principle.”
⁶
There seems to be a clear operational distinction here, allowing us to ask the question whether network topology is the genetically fixed “core” component responsible for network functional stability, with the local parameters at the node imparting functional flexibility. Or could it be the other way round? Computer simulations, based on real networks, reveal in some cases that topology is the key to stable network function, and in other cases stability is a consequence of buffering in contact parameters. Hence, there is no clear distinction between what might be considered “core” and “peripheral” components. Both operate simultaneously during network formation and their influence on network function depends on the types and number of modular units that go into the making of each node, which are of course genetically encoded. So far, there is no obvious distinction between “principles” and “parameters” in network biology, nor with respect to “core” and “peripheral” operations.

My emphasis on individual personalization during ontogeny is perhaps no more relevant than in the dissection of the biological basis of consciousness, and from that the phenomenon of free will. In so far as I am less a philosopher of mind than I am a linguist (!), I have a sort of amateur freedom to join the dots where the professionals might say no lines exist. Nevertheless, I have the sense that there is general agreement that human consciousness is a first person subjective phenomenon of experiences (
qualia
) that cannot be described in their totality to another conscious mind. Whatever the correct wording might be, there is no doubt that it is a real, not illusory, biological process that can be expected to be unique and subjective in its precise operations to each individual phenotype. So, is there anything about evolved biological networks and their ontogenetic reconstruction, post-sex, which figures in the existence of consciousness?

To answer this question I need to introduce one other confounding feature of biological systems, which is the phenomenon of “degeneracy.” This is the capacity for different routes to be taken through a network, with each route yielding the same or similar functional outputs. Degeneracy was spotted early on in the history of molecular biology with regard to codon–anticodon patterns of recognition in which some amino acids have more than one designated codon. Degeneracy is invariably found wherever it is looked for, and one relevant new study by Ralph Greenspan (2001; and Van Swinderen and Greenspan 2005) has found degeneracy operative in a network of genes regulating neuronal behavior in
Drosophila
. He was able to show that the topology of connections in the relevant network could differ widely, depending on the mutant state of different participating units, yet with only subtle alterations of the behavioral phenotype under investigation.

Coupling widespread degeneracy with random background noise is one of the strong arguments in favor of my advocacy that development is a highly personalized set of operations from the early inception of the networks regulating gene expression through to the ever-changing neuronal connections in the brain. From beginning to end there is a subjective process of individualization that is perhaps no different in kind from that mode of first person subjectivity that is considered to be the basis of each individual's mind.
Subjectivity is the name of the game at all levels, even though we are only mindful of it in the brain.

Could it be then that there is some biological basis to free will residing in such personalized degeneracy? I consider free will to be the feeling that, although we make decisions based on a long series of cause-and-effect steps, there is nevertheless a gap in the chain of causality at the very last step. Acceptance of this “gap” means abandoning for a moment the basis of Western science. How can we overcome this dilemma?

According to the philosopher Ted Honderich (2002) there is a sense in which, when we look back on our lives, we have an inescapable conviction that we were always “our own man” (or woman); that “things could have been otherwise.” Our subjective feeling that this is so is no illusion, any more than our subjective experiences of qualia are an illusion. The latter might be a first person phenomenon emanating from the highly personalized structures of degenerate networks, as is everything else in the totality of living processes in an individual, but this does not mean that qualia cannot be dissected, as emphasized by John Searle, using the third person, objective methods of Western science bounded by its acceptance of cause and effect. There is a real phenomenon of personalized free will that is open to scientific investigation starting with the genes, continuing with the processes of total nurturing as individualized degenerate networks are configured, and ending with the subjective reality of mind.

With all of this in mind it might not be totally off the beaten track to see free will, not as an abandonment of cause-and-effect determinism, but as a situation of rapidly and subtly changing outcomes as degenerate neuronal networks switch from one quasi-stable state of topology to others. Our sense of what is going on is that we, each and every lonely individual, feel that a freely willed, subjective decision has been made. At the level of biology (all that chemistry and physics if you will), there is an unbroken route of cause and effect passing through each and every personalized degenerate state, but at the level of our sense of what has happened, we feel that at the threshold of the final step (the gap to the one remaining degenerate state with its final functional output) is one for us alone to decide.