Clues to how human language evolved

A study published online in Nature Neuroscience yesterday provides some clues to how human language evolved, and could make a valuable contribution to the on-going debate over language evolution.

The work, by groups at the National Institute on Deafness and Other Communication Disorders (NIDCD) and the National Institute of Mental Health (NIMH), shows that macaques (rhesus monkeys) use brain regions corresponding to the main language centres in the human brain when perceiving the vocalizations of other monkeys of the same species.  

Most species, from insects to primates, have some system of communication. It is generally agreed that the computations underlying these systems are shared among all species, but that human speech differs qualitatively from the modes of communication used by other species.  

There are several hypotheses suggesting how human language evolved. One of these holds that human language is homologous to animal communication and is produced by the same functional components which underly communication in other species. According to another hypothesis, human language is a complex adaptation that is unique to our species, and whose core components have been subjected to selective pressures. A third hypothesis, proposed recently by Chomsky and his colleagues, states that most aspects of human language are shared with the modes of communication of other species, but that certain aspects of it, especially recursion, evolved recently and are unique to Homo sapiens. 

One issue in the debate about language evolution therefore concerns the steps required for the evolution of human language from that of our ancestors. Were the components of the neural mechanisms underlying human language remodelled and extended from components  inherited from our ancestors, or did they evolve de novo to accommodate the linguistic capacities of our species?    

The current study sheds some light on this question. Using neuroimaging, the researchers found that the ventral premotor cortex (PMv), the temporoparietal area (Tpt) and posterior parietal cortex (PPC) are active in macaques when they perceive the vocalizations of other macaques. The authors therefore conclude that the common ancestor of macaques and humans, which lived 25-30 million years ago, possessed the neural mechanisms required for the evolution of human language.

The PMv corresponds to Broca’s area in the human brain, and the Tpt and PPC to Wernicke’s area. Both regions in the human brain are named after the physicians who first described their association with speech production. Broca’s area, located in the inferior gyrus of the frontal lobe, is often damaged in strokes, leading to aphasia (the inability to produce speech). Wernicke’s area is located in the superior temporal gyrus. It was previously thought that Broca’s and Wernickes’ areas were separately involved in speech production and comprehension, respectively; the general consensus now, however, is that both areas work together in both aspects of speech. 

Another contentious issue in the debate concerns exactly how human language evolved. Was it the direct result of selective pressure, and, if so, what selective pressures led to its emergence? If language is a direct result of Darwinian evolution, how can mutations in the DNA sequence give rise to it? Did those mutations result in the elaboration of existing structures which supported other cognitive abilites in our immediate ancestors, or did they give rise to new structures unique to our species?

The evolutionary advantages of human language are obvious. Language enabled our species to purposefully communicate vital information about the environment and to form cohesive social structures. Language also has a fundamental role in thought, allowing for complex mental manipulations.

Language must be a product of natural selection. It is, however, difficult to imagine how mutations could give rise to completely new neural structures. If the components of the human brain underlying language evolved de novo with the emergence of our species, we would expect these components to be solely adapted for the purposes of language.

This does not, in fact, seem to be the case – recent research has implicated Broca’s area in the executive control of cognitive functions. This involvement of Broca’s area in other functions suggests that it and other neural components involved in language are built upon pre-existing structures. 

Language is something that makes us different from other animals. Anthropologists and evolutionary biologists therefore sometimes claim that language was a major driving force in the expansion of the neocortex that occurred as Homo sapiens emerged as a distinct species, although alternative hypotheses have been suggested.

The late Stephen J. Gould argued that language is a spandrel, or a by-product of the evolutionary process. According to Gould, “Natural selection made the human brain big, but most of our mental properties and potentials may be spandrels – that is, nonadaptive side consequences of building a device with such structural complexity.”

I agree with Gould that the human brain expanded for reasons other than the accommodation of language. I believe that the transition to bipedalism had a large part to play in the expansion of the neocortex. Freeing the forelimbs from the constraints of their use for locomotion allowed for far greater versatility, and would have required an expansion of the motor cortices. (Broca’s area is actually part of the motor cortex, and is involved in the control of muscles necessary for speech production.)

This cortical expansion, in turn, would have enabled a rudimentary, non-verbal gesticulatory mode of communication to become more elaborate. By increments, and with the co-evolution of the larynx, each gesture may have had a particular sound associated with it. In this way, over the course of hundreds of thousands of years, human language may have evolved.

The authors of the Nature Neuro paper suggest that the macaque homologues of Broca’s and Wernicke’s area are involved in prelinguistic functions such as assigning meaning to sounds made by other macaques. The findings reported here provide one piece to a very complex puzzle, and we are still a long way off from any real understanding of how language evolved. 

Teaching grammar to songbirds

A new study carried out by Tim Gentner of the University of California, San Diego, suggests that songbirds can be taught simple grammar.

Gentner generated his own versions of birdsong, which were played when starlings pressed one of three buttons on a wall with their beaks. The birds were trained to press the same button a second time if the song they heard contained an explanatory clause, and were rewarded with food if they recognized the pattern.

Of the 11 starlings used in the study, 9 recognized birdsong with warbling inserted into it, and did so 90% of the time. Gentner was “dumbfounded that they could do as well as they did.”

The study shows that the cognitive skills of animals are far more complex than was previously thought. According to Marc Hauser, director of Harvard University’s Cognitive Evolution Laboratory, it shows that “some of the cognitive sources that we deploy may be shared with other animals.”

In experiments carried several years ago, Hauser tried unsuccessfully to make tamarin monkeys recognize recursive grammar. This seemed to confirm Chomsky’s assertion that recursive grammar was unique to humans. On the face of it, Gentner’s work seems to show that this is not the case. (An example of recursive grammar is changing the sentence “John walked to the shop” to “John, who doesn’t like to drive, walked to the shop”.)

According to Hauser, the study shows that although starlings may have an ability to learn simple grammar, their communication does not contain the semantics of human language. “[Gentner’s] article is based on an elementary mathematical error,” says Chomsky, “and has nothing remotely to do with language; probably just with short-term memory.”