Of Minds and Language (19 page)

That epigenetics results in real changes in how genes function is a fact. A clear example of how this happens is provided by the case of the black mice. These mice are all genetically identical, in DNA sequence, but it does not take a geneticist to see that they are quite different phenotypically, in terms of the color of their coats. What has happened is that the mothers of these mice are given diets containing different amounts of substances that provide methyl groups. As we discussed, DNA methylation is a major epigenetic regulator of gene expression. After the mothers are fed different amounts of methyl donors and the pups are born, their coat color is checked. Depending on the amount of methyl donors the mothers received, and depending on the different colors of the coats, different levels of methylation are found in the DNA locus that regulates this trait, the color of the coat, with a nice linear relationship between methylation and coat color (Morgan et al. 1999).

This may be true not only of mice; there are interesting data in humans as well, for instance the famous case of the Dutch hunger winter, the famine in the Netherlands during World War II, when mothers who were pregnant at that time had very small children. The children of those children (the grandchildren of the mothers pregnant during the famine) remained small despite receiving a perfectly normal diet.
²
It is possible that this feature, this trait, was transmitted across generations.

What we propose is that this
kind
of mechanism may account for some of the features of
L
at least (those in red in
Fig. 7.1
). Here are some cases in support of our proposal.

Plasticity
is certainly a paramount feature of biological trait L. A relevant well-known case is that of the Achillea, a plant. Plants are masters at using epigenetics because they are exposed to weather and heavy environmental insults and they need to react to light and temperature. This they do epigenetically. For Achilleas, the same plant at low altitude is very tall, at medium elevation is very short, and at high elevation it becomes again very tall. Nothing changes in the genome of this plant, but the phenotype changes heavily in response to environmental cues, in this case climate and altitude.
³
This is the concept of
norm of reaction
that Richard Lewontin, in the wake of the Russian geneticist and evolutionist Ivan Ivanovich Schmalhausen (1884–1963),
⁴
has so
clearly formulated: what the genotype specifies is not a unique outcome of development, it is a
norm of reaction
. A norm of reaction is constrained by genotype, but specifies a pattern of different developmental outcomes depending on the environment.

The concept of
windows of opportunity
is quite familiar to immunologists. In the stables of a Bavarian farm, the mothers work while their children sit in a cradle. As a result of that, we now know, these children are incredibly well-protected from allergic disease, but only if they sit in the stables up to the age of one year, or even better if the mother goes and works in the stables when she's pregnant. Prenatal exposure to stables and barns has the strongest effect. If exposure occurs when the child is 5 years old, it matters much less or not at all.

For
multiple discrete final states
, we already discussed how functionally and morphologically distinct cells (in our case, red and white blood cells) can derive from a single precursor. This process stresses two points. One is about plasticity, as we said, but the other is
partial irreversibility
. Once a cell becomes highly differentiated and its epigenetic differentiation program is fully implemented, this cell cannot go back. In fact, only stem cells retain plasticity all the time. For most other cells, the features acquired through epigenetic modifications are fixed and irreversibly preserved throughout life.

Now do we need to say the
L
we have been talking about is language? We think the genetic components of
L
are species-specificity and the common core (Universal Grammar) with room for large but highly constrained parametric variation (variation is going to become important to some extent, but it requires of course a robust common core). These components may correspond to FLN (the faculty of language in the narrow sense, in the terminology of Hauser et al. 2002). All the other plastic, dynamic components of
L
, we propose, are mechanistically implemented through epigenetic mechanisms – these could be the broader language faculty (FLB). We may have to go beyond this “division of labor” for another feature – the fact that
L
is or seems to be extremely robust, resistant to degradation, and also extremely stable, at least over a lifetime. From a strictly biological point of view, this feature suggests simplicity of design, because simplicity of design gives very high effectiveness. However, a simple design is also vulnerable to stress, unless it is balanced with some redundancy. The stability of a very small system is difficult to understand without postulating that somewhere, somehow, there is some compensatory repair pathway that allows a very compact core to repair. But this is even more speculative than our previous speculations.

Our last point, and this is entirely Massimo's doing, depicts two potential (alternative) scenarios: (1)
All parameters are innately specified
. This would put a very high burden on genetic encoding, something that we immunologists
are acutely aware of. And the problem of how you encode an enormous amount of diversity in a limited genome would of course come back here. This possibility would put very little or no burden on learnability. At the other end, (2)
unconstrained variability
, would however put an excessive burden on learn-ability. So I guess that what we are trying to say is that perhaps having principles and parameters might represent an optimal compromise.

DOVER: Epigenetics is a very active and important research field at the moment and it is highly appropriate that you should attempt to link it to the supposed difference between FLB and FLN as I understand it. But I need to add one important caveat, which is that epigenetics is fast becoming a catch-all phenomenon covering anything that moves in the workings of biology. The turning on or off of any gene, whatever it's doing, requires the prior engagement of tens upon tens of proteins which are the products of other genes of course. Now, some of these other proteins are opening and closing the chromatin near to our gene of interest in preparation for transcription; others are involved with nearby DNA methylation; others with the initiation and termination of transcription of the gene, and so on, so you can go on forever. If that is the case, then everything is both epigenetic and genetic at one and the same time, that is, no gene exists in a vacuum, its expression is carefully regulated and depends on the state of its local chromatin, which in turn depends on the comings and goings of many other gene-encoded proteins. In such a situation we might well ask what is the real operational distinction between genetic and epigenetic? Can this really be the basis to distinguish between
core
processes, which are supposedly ancient and go way back, and the more recent
peripheral
processes?

So just to get away from language, let me say something about legs, because it is easier to make my point. We all have two legs, yet we all walk very differently. Now it has long been thought that having two legs is one of those core, basic things that universally characterizes our human species – any healthy fertilized human egg will develop into an individual with two legs. But the shape and manner of usage of legs, peculiar to each individual, is considered to be something peripheral, something that might be “epigenetically influenced” during individual development. Now the whole point of Richard Lewontin's earlier concept of “norms of reaction” (he might not have said this in precisely the same way at the time, but it is certainly the way it's being interpreted now) is that the developmental emergence of two legs, and not just the ways we use the two legs, is as much “epigenetically” modifiable, and is as much a key part of that total
process of ongoing, ontogenetic nurturing that I spoke about earlier.
⁵
In other words, those complexes of genes that are involved in making two legs are no different in kind from the genes, or the very complex milieu of interactions of genes with genes, and genes with environment, that affect the individual shape and use of those legs. So it is very hard to distinguish between them, between “core” and “peripheral,” given that this is happening from the moment a specific sperm enters a specific egg and on through each individual's highly personalized route of development.

Each individual's personal history of cell differentiation, tissue patterning, organogenesis, emergence of consciousness, language acquisition, and all the rest of it involves many complex and fluctuating networks of gene (protein) interactions, also subject to much environmental input. There is variation and constraint, simultaneously, at all times. The only thing we can be sure about is that, as a consequence of the sexual process of making sperm and eggs, we essentially get back to a genetic blank slate from which all human developmental processes, “core” and “peripheral,” “genetic” and “epigenetic,” “variable” and “constrained,” need to re-emerge. Anything produced by evolution is bound to be a mess and even the original concepts of principles and parameters might be difficult to unravel when considering biological, ontogenetic processes and their inherently sensitive networks – but here I reach the edge of my understanding.

V
ERCELLI
: I think we need to tread lightly because we are on tricky ground. That the development of an organism involves, as you put it, “many complex and fluctuating networks of gene (protein) interactions, also subject to much environmental input” I certainly will not deny. Nor will I argue against the continuous interplay between (and the likely co-evolution of) genetic and epigenetic mechanisms and processes, which at times may blur the distinction between them. But a distinction
does
exist and emerges when one thinks about the
kind of mechanisms
that may account for certain essential features of language as a biological trait. Some of these features (species specificity and uniqueness to humans, first and foremost) appear to be rooted so deeply and constrained so strongly that one would expect them to be inscribed in the genetic blueprint of our species – that is, to be genetically encoded. But most of the other defining features of language reveal a degree of
plasticity
in development and final states that best fits under the epigenetic paradigm. In other words, not everything in language is nurture – but not everything is nature either.

P
IATTELLI
-P
ALMARINI
: Let me add to this the following: take the case you present of movement and the fact that we all have two legs and yet each walk differently. There is the famous two-thirds power law;
⁶
all biological movement obeys this two-thirds power law. All natural movement in humans and animals obeys the law that the two-thirds power of the ratio between linear speed and radius of gyration is always constant. It is universal and we immediately perceive it. Indeed, each one of us walks in a slightly different way. You can look at someone and say “Oh that's Jim, because see the way he walks.” But it's very interesting to see that there is a universal law for biological movement. So, what are we interested in? The big effort that has been going on in language – we use different words, different accents, different tones of voice – but the big effort has been to go beneath these and see at what level there may be something universal, something that is common, that is deep. And it is no mean feat. You have seen these days what is in the lexicon, what is in the syntax, what is in the morpho-lexicon, what is in semantics – very, very difficult questions, all subdivided in order to deal with them one at a time. And so the FLB/FLN distinction is complicated to make, but it is a good way of distinguishing things, seeing which components are innate and which components are not. You are a geneticist but I have been a molecular biologist and continue to follow the field, so we both know that there are certain things you can do to genes with very specific effects. Of course, the effect of a gene on a phenotype usually depends on the effect of many other genes, that is called epistasis, and sometimes subtle or not so subtle effects come from apparently unrelated genes. But there are also clear examples of the effects of only one gene. For example, there is the outstanding phenomenon of Hsp90, with its chaperone protein which, if knocked out, gives rise to all sorts of mutations, all over the body of, say, a fruitfly.
⁷
That is, there are very specific things you can do to specific genes with very specific effects. Moreover, the distinction between genetic core processes and peripheral (also called exploratory) processes is unquestioned these days. I find it all over the current literature, often under the label of developmental and evolutionary modularity.
⁸
The biochemical pathways and their enzymes, for instance, just to name one clear case, are evolutionarily strictly conserved, often all the way down to bacteria.

D
OVER
: I don't think I've argued against genetics, otherwise I'd be out of a job; nor have I argued against universality, in terms of human-specific features which are shared by all humans. That's not my point. The point is that the
ontogeny of a given individual is a highly personalized dynamic in which many factors are involved unavoidably nurturing each other. You cannot, with regard to the ontogeny of an individual, say that the “universal genes” and all their participatory networks for two legs are more of a “core” process than the genes and all their participatory networks for the manner in which we use those two legs. The two are ontogenetically unfolding together and there are many, many diverse and interactive influences at play in each unique individual – genes, proteins, environment, culture – the whole catastrophe!