17 February 2006

The Rain In Spain Falls Only In The Human Brain

By Rusty Rockets

New studies showing that both grammar usage and symbol-to-object association are innate and universal to humans may put to rest some long standing controversies in cognitive research. Research suggesting that humans have an innate and universal faculty to form sentences supports the idea that humans may have been born with a ready-made language "module." Claims such as these have been coming out of a relatively new but burgeoning field called evolutionary psychology (EP), with its proponents making the claim that the brain evolves through natural selection, as does any other body part.

It seems that Psycho-Darwinists - or evolutionary psychologists - such as Noam Chomsky and Stephen Pinker now officially have the right to gloat. Findings from a number of different studies are all starting to point to one conclusion: that language structure is an innate ability that is probably the result of humans evolving a sophisticated language organ through natural selection. Evolutionary psychology, established and popularized by Leda Cosmides and John Tooby, advocates the existence of many types of brain modules that EPs call "evolved psychological mechanisms" (EPMs). EPs contend that these modules handle specific tasks throughout the brain, such as vision, hearing and motor control, for example.

EP theory is in stark contrast to the competing theory of connectionism, which, as you'd imagine, suggests that the brain carries out specific tasks, such as language structure, by "thoughts" bouncing between simple, discrete neural units to form neural networks throughout the brain. The connectionism model claims that a word, say, is formed by any number of interconnecting neurons throughout the brain and how this word network relates to other networks, and so on. The neural network theory has so far been the pre-eminent theory for explaining how the brain operates. But in the face of new evidence, it now seems that connectionism theory is an endangered species.

A recent University of Rochester study, published in the Proceedings of the National Academy of Sciences, may have answered the long-standing question of why certain fundamental characteristics of grammar are present in all languages. Researchers at the University of Rochester claim they have discovered evidence that a universal ability for grammar among humans is "hardwired" into our brains. The team of researchers arrived at this conclusion by studying the gestures of deaf people who had been isolated from conventional sign, spoken, and written language their entire lives, yet still managed to develop a distinct form of gesture-based communication.

"Our findings suggest that certain fundamental characteristics of human language systems appear in gestural communication, even when the user has never been exposed to linguistic input and has not descended from previous generations of skilled communicative partners," says Elissa L. Newport, Professor of Brain and Cognitive Sciences and Linguistics at the University of Rochester.

The research involved Newport and co-researcher Marie Coppola spending eight years studying the behaviors of three deaf Nicaraguan boys. The researchers claim that the boys had no knowledge of or exposure to any formal Nicaraguan sign language, and that their deafness from birth ensured that they had no exposure to spoken language. The researchers also added that the boys had little or no formal education, so they had no knowledge of written Spanish. The boys' communicative isolation led each of them to develop their own gestural language (referred to as "home sign systems"), which presented the Rochester team with a unique opportunity to study how the brain creates language. "We examined a particular hallmark of known grammatical systems and found that these signers also used this same hallmark in their gestured sentences. They designed their own language and wound up with some of the same rules of grammar every other language uses," explained Newport.

The team studied the boys closely as they watched videos of various images and described what they saw using their own sign language. What the researchers found was that the boys repeatedly used the grammatical construction of "subject" in a manner consistent with how it is used in other languages. Subject, usually a noun, is a mandatory component in any sentence construction, but is a complex and difficult concept to explain. A simple explanation would be that the subject is a person, place or thing that performs the action in a sentence. But this is not always the case, and how the subject is placed grammatically in a sentence is not always obvious. Despite this potential pitfall, the three boys managed to create a language on a par with any other language grammatically sufficient on which to hang the concept of subject.

"The notion of subject does not appear to require either linguistic input or a lengthy history within a language to develop," says Newport. "We're starting to see that the grammatical concept of subject is part of the bedrock on which languages form." Even though Newport's study suggests that language structure is an innate, hardwired ability in humans, it does not imply the concept of a language organ. However, we don't have to make speculations on Newport's findings alone.

The Max Planck Society has also released some interesting findings on grammatical structure that deal predominantly with the hierarchies of sentence structure. Their findings are fascinating in that they not only show a specific area of the brain that deals with language, but that humans have developed an extra language area that allows us to understand more complex, higher order sentence structures. This finding goes a long way in supporting the EP claim that humans evolve specific brain modules, or EPMs, through natural selection.

The intricate linguistic rules that allow us to understand long complex sentences are called a hierarchy. While non-human primates have been known to process simpler language structures, known as local probability based rules, they cannot process the more complex hierarchy rules. According to the Max Planck study, this seems to be because unlike non-human primates, humans possess a "phylogenetically younger" language area in the brain. Using functional magnetic resonance imaging (fMRI) and diffusion tensor imaging (DTI), a team of researchers from Leipzig set up an experiment that tested both the older and younger areas of the brain using artificially created grammars with meaningless but structured syllables. This principle was meant to reduce grammar into the simplest formal rules. The advantage of experimenting with artificial grammars - as opposed to naturally spoken grammars - lies in the fact that other elements of language (semantics, phonology and morphology) do not influence the neurological processing.

During the course of the experiments both the older and younger sections of the brain were easily distinguished. When an individual was tested for the simpler "local probability" processing, the older part of the brain (known as the frontal operculum) common to primates lit up. Conversely, when higher order processing was required, the person's younger area (known as Broca's Area) kicked in. The experiment seems to have highlighted what EPs have been saying for some time: that humans have evolved a specific language area, organ or module through natural selection.

If these modules have only evolved relatively recently in primate evolution, then Newport's discovery of an innate foundation concept of "subject" might be how humans distinguish self among other objects; and could also be what has given rise to self-awareness and consciousness. If this is the case, how will these new findings affect other areas of science and technology? A theory that uses modules to describe brain function is likely to have profound implications for artificial intelligence (AI) development. If intelligence relies heavily upon self-awareness and that self-awareness is reliant upon a language organ that can distinguish "subject" as opposed to self, then AI researchers may be barking up the wrong tree with connectionism models.

Sources: Max Planck Society, University of Rochester