Genetic Background of Language

Information about the human and the chimp genome (The Chimpanzee Sequencing and Analysis Consortium, 2005) is now 'complete', and one can ask how far previous optimism seems justified in the light of comparative studies using this information. It is clear that a lot of work lies ahead: knowing all the genes of chimps and humans is not the whole thing: one should know how genotype is mapped to phe-notype, and this is a formidable problem. Genes are expressed in specific ways, under the influence of other genes and the environment. Interaction between genes is not the exception but the rule. One gene can affect several traits (pleiotropy) and actions of different genes do not affect traits (including fitness) independently (epistasis). It is the network of interactions that one should know, and one must not forget that there are networks at different levels, from genetic regulatory networks through protein interaction networks and signal transduction pathways to the immune system or neuronal networks. The question is how the effect of genes percolates upwards. Genes act on expressed molecules (proteins and RNA) that do their job in their context. There is something amazing about the fact that hereditary action on such primitive molecules percolates upwards resulting in heritability of complex cognitive processes, including language.

The chimp and human genomes are indeed similar, but one should understand clearly what this means (Fisher and Marcus, 2005). Substitutions make-up for 1.23% of difference between the two genomes, which translates into 35 million altered sites in the single-copy regions of the genomes! Insertions and deletions yield a further 3% genomic difference. It is convenient to distinguish between altered structural and regulatory genes. The first code for altered enzymes or structural proteins, the latter code for altered transcription factors, for example. Both kinds of changes did happen since the humans diverged from chimps, and both can affect language in critical ways.

There seems to have been acceleration in the changes of neural gene expression patterns in human evolution, although this should be evaluated against the background that liver and heart expression patterns have diverged a lot more between chimps and humans. The usual interpretation is that neural tissue is under stronger stabilising selection. Another observed tendency is the upregulation of human neural gene expression relative to the chimp, but the functional significance of this finding is unclear (it may be a more or less direct consequence of recent genomic region duplications).

It is not yet clear what is the gene expression difference behind the cytoarchitec-tonic differences among the Brodman areas: most known gene expression differences between chimps and humans are common to all cortical regions. This view has been refined very recently. Oldham et al. (2006) analyzed gene coexpression patterns in humans and chimps. They could identify network modules that correspond to gross anatomical structures including the cerebellum, caudate nucleus, anterior cingulated cortex, and cortex. The similarity of network connectivity between the respective human and chimp areas decreased in that order; consistent with the radical evolutionary expansion of the cortex in humans. It is intriguing that in the cortical module there is a strong coexpressive link between genes of energy metabolism, cytoskeletal remodelling, and synaptic plasticity.

There are genetic changes that probably did boost language evolution but in a general, aspecific way. Genes influencing brain size are likely to have been important in this sense. A note is in order, however. Genes involved in primary microcephaly seem to have been under positive selection in the past, but children with this syndrome have rather normal neuroanatomical structures despite the fact that their overall brain size can be reduced to a mere one third of the normal. They show mild to moderate mental retardation and pass several developmental stages. Fisher and Marcus (2005, p. 13) conclude: 'In our view the honing of traits such as language probably depended not just on increased "raw materials"

in the form of a more ample cortex, but also on more specific modification of particular neural pathways.'

Perhaps the most revealing recent finding concerning genetic brain evolution is the identification of an RNA gene that underwent rapid change in the human lineage (Pollard et al., 2006). It is expressed in the Cajal-Retzius cells of the developing cortex from 7 to 19 gestational weeks. It is coexpressed with reelin, a product of the same cells, that is important in specifying the six-layer structure of the human cortex.

Even if some of our linguistic endowment is innate, there may not be genetic variation for the trait in normal people, just as normal people have ten fingers. In contrast, our linguistic capacity may be like height: whereas all people have height, there are quantitative differences in normal people. To be sure, children as well as adults differ in their linguistic skills; the question is what part of this variation genes account for.

Surveying many studies Stromswold (2001) concluded that twin concordance rates have been significantly higher for monozygotic twins than for dizygotic twins. Twins are concordant for a trait if both express the trait or neither expresses it. Twins are discordant for a trait if one exhibits the trait and the other does not. If the concordance rate for language disorders is significantly greater for monozygotic than dizygotic twins, this suggests that genetic factors play a role in language disorders such as dyslexia and specific language impairment (SLI). The concordance rates for written and spoken language disorders are similar. For both written-language and spoken-language disorders, mean and overall concordance rates were approximately 30% higher for monozygotic twins than for dizygotic twins, with genetic factors accounting for between one half and two thirds of the written- and spoken-language abilities of language-impaired people. In studies of normal twins, depending on the aspect of language being tested, between one quarter and one half of the variance in linguistic performance was attributable to genetic factors. People have been tested on phonological short-term memory, articulation, vocabulary, and morpho-syntactic tasks. It seems that different genes may be responsible for the variance in different components of language and that some genetic effects may be language specific.

The sum of all genetic effects is usually not much greater than 50% for various aspects of cognition (Stromswold, 2001). Most individual genes are expected to have small effects. Candidate genes affect functions including the cholinergic receptor, episodic memory, dopamine degradation, forebrain development, axonal growth cone guidance, and the serotonin receptor. It is a great problem that cognitive skills are likely to have been at least in part inadequately parsed, thus so-called intermediate phenotypes with a clearer genetic background should be sought. By this token schizophrenia as such does not exist; rather, different genes may go wrong and the symptoms such as hallucinations are emergent outcomes (Goldberg and Weinberger, 2004). The situation may be similar to that of geotaxis in Drosophila, where the individual involvement of different genes that collectively determine this capacity is counter-intuitive (Toma et al., 2002).

It is worth calling attention to the fact that the genetics of human cognitive skills is a notoriously difficult problem. One common reason is that usually the clinical characterisations are not sufficient as descriptions of phenotypes (Flint, 1999). A consensus seems to emerge that the genes involved are 'liability genes' that, when present in the right allelic form, significantly enhance the probability of developing the respective cognitive skills.

Perhaps the most important neurodevelopmental syndrome for our topic is SLI, where there is significant difference between verbal and non-verbal skills. Several candidate chromosomal regions have been identified (SLI consortium, 2002); importantly, they involve the gene USP10 (an ubiquitin-specific protease), which encodes a protein involved in synaptic growth, and shows an increased copy umber in the human lineage (Fortna et al., 2004). In Drosophila, its overexpression of its homolog results in increased synaptic branching and altered synaptic function (DiAntonio et al., 2001).

A by now famous gene is FOXP2, that was first called attention to by Gopnik (1990). In a certain English-speaking family there is a dominant allele that causes the syndrome, formerly grouped under SLI, but recently termed developmental verbal dyspraxia (DVD). There is no disagreement that SLI is real. It is more contested how closely it is limited to, or rooted in, a specific grammatical impairment. The Gopnik (1990, 1999) case has been very stimulating because of its characterisation as 'feature-blind' dysphasia and its obvious genetic background (a single dominant allele). Whether other cognitive skills are also, or even primarily, affected, has been debated ever since (Vargha-Kadem et al., 1998). More evidence with other linguistic groups is accumulating (Dalalakis, 1999; Rose and Royle, 1999; Tomblin and Pandich, 1999). A study (Van der Lely et al., 1998), sadly without genetics, claims to demonstrate that grammatically limited SLI does exist in 'children' (although only one child is analysed in the paper!).

The FOXP2 protein is an old transcription factor present in vertebrates, and there is evidence that it has been under positive selection in the human lineage. It seems to affect development of distributed neural networks across the cortex, striatum, thalamus, and cerebellum. The DVD condition makes the phenomenon different from SLI, but it is important that speech and language deficits are always present, even in otherwise normal children. In other affected individual general intelligence is also impaired. It is also important that the grammar deficits (difficulty with morphological features such as the suffix -s for plural or -(e)d for past tense) occur in written language as well. The selective sweep that affected this gene in the human lineage occurred within the last 200,000 years (Enard et al., 2002; Zhang et al., 2002).

Analysis of the expression patterns of FOXP2 in other species suggests that this gene has been involved in the development of neural circuitry processing sensory-motor integration and coordinated movements, lending support to the notion that language has its roots in motor control (e.g. Lieberman, 2007), which makes the involvement of basal ganglia in speech and language less than surprising.

Recent studies (reported by White et al., 2006) demonstrate that FOXP2, although without accelerated evolution, plays a crucial role in the development and seasonal activation of relevant brain areas in songbirds. Interestingly, although the avian FOXP2 is very similar to the human version, neither of the human-specific mutations is found in the former. Also of interest in the fact that the ganglia involved in bird song learning seem to be analogous to the basal ganglia involved in human vocal learning [Scharff and Haesler, 2005].

Researchers have called attention to the fact that not only FOXP2, but also FOXP1 is expressed in functionally similar brain regions in songbirds and humans that are involved in sensorimotor integration and skilled motor control (Teramitsu et al., 2004). Moreover, differential expression of FOXP2 in avian vocal learners is correlated with vocal plasticity (Haesler et al., 2004). Mice, like man, have also two copies of the FOXP2 gene. If only one of them is affected in mice, the pups are severely affected in the ultrasonic vocalisation upon separation from their mother. This suggests a role of this gene is social communication across different species. The Purkinje cells in the cerebellum are affected in the pups (Shu et al., 2005). Determination of the expression pattern in the developing mouse and human brain is consonant with these investigations: regions include the cortical plate, basal ganglia, thalamus, inferior olives, and cerebellum. Impairments in sequencing of movement and procedural learning thus may be behind the linguistic symptoms in humans (Lai et al., 2003).

We think a key issue is the biologically motivated dissection of the language faculty. Put differently, what are the intermediate phenotypes composing language? This question cannot be answered, we believe, without an appropriate formulation of aspects of language. Thus linguistic theories must ultimately be biologically constrained. A good start in this direction may be Luc Steels' fluid construction grammar (Steels and De Beule, 2006). There is no coupling yet between details of linguistic theories and those of brain mechanisms.

Was this article helpful?

0 0
How To Bolster Your Immune System

How To Bolster Your Immune System

All Natural Immune Boosters Proven To Fight Infection, Disease And More. Discover A Natural, Safe Effective Way To Boost Your Immune System Using Ingredients From Your Kitchen Cupboard. The only common sense, no holds barred guide to hit the market today no gimmicks, no pills, just old fashioned common sense remedies to cure colds, influenza, viral infections and more.

Get My Free Audio Book


Post a comment