Elliott D. Ross’s paper on the brain (abstract here) discussed in my last post has plenty more food for thought. In particular, Ross discusses the subject of brain efficiency which, I confess, was a new idea for me but one that immediately shows its utility in suggesting topics for human and speech evolution.
This detour process is very different from what goes on in a modern, commercial computer such as the high-powered laptop I was given just this week. If you fired a small-caliber bullet at random into the laptop you might do only a little bit of damage, but the damage you did would be permanent. If, for example, you broke the connection between the video card and the processors, that connection would remain severed and no other processor would take up the slack. That’s because computer functions are localized properties of the machine, not emergent ones. Thus, it is a mistake to go directly from the familiar computer model to thinking about brain evolution because the functions that evolved emerged.
An example of something that emerges in the body is the fist. Is a fist stored in my open hand? Yes, in a manner of speaking, but not in the normal way we think of storage. The muscles and nerves are present and follow a practiced routine when I make a fist, but the fist emerges from the interaction of muscle and nerve; it has not been stored whole waiting to make its appearance. Emergent properties and functions are stored in a very different way from localized ones.
The result of damage to emergent processes is loss of efficiency rather than loss of function. Ross tells the story of a physicist he once treated for a stroke. After some time of therapy, Ross examined the patient and told him he seemed fully recovered. The patient denied it. He used to be able to work—in his case, think, lecture and write—for six hours at a stretch, but in the “recovered” state he became exhausted after an hour, and began to make many mistakes in grammar and spelling. He had to limit himself to one hour of lecturing or writing per day. Ross concludes from this and other data:
Broca’s and Wernicke’s areas are essential for modulating in the most cognitively efficient manner possible the articulatory-phonological and comprehension-phonetic aspects of language…however, they are not absolutely critical for these cognitive operations to occur because each of the functions is an emergent property… [pp. 225-6]
The notion of efficient brain use seems quite valuable. It immediately explains, for example, what is going on in basic learning or practice. “Overall,” Ross reports, “practice dramatically reduces the spatial distribution of activated regions” . Repetition of an action, which underlies both practice and stimulus-response learning, is a way of getting the brain to organize its synapses so that an action can emerge in the most efficient manner possible. At some point the system may become as efficient as it can be and further practice can only maintain a skill level, not advance it.
For this blog, the idea of efficient emergence suggests that we did not evolve new, localized functions. We evolved more efficient, emergent ones.
If true, the argument over innate versus learned abilities has less relevance than supposed. Innate actions are not stored somewhere in the brain. The inborn wiring for an emergent function like crying simply needs to be already fully efficient or nearly fully. Meanwhile, a learned ability such as accounting does not have to build on an innate counting, module. It can be patiently taught over many years as more and more arithmetical skills become efficient.
This idea is very different from the traditional assumption of a universal grammar based on a dictionary and parser localized somewhere in the brain. Children are not born with some kind of algorithm, but in a few short years they acquire efficient neural pathways for organizing words into sentences and sentences into something more elaborate. Presumably the children have pathways that give them a good start, but things like noun phrases are not lying in wait for activation. If their target language uses noun phrases, eventually they will emerge; if not, not. And the noun phrases are never in the brain; they emerge on the tongue, or through the hand.
This analysis also indicates that we shouldn’t be looking for the evolution of functions, but of efficient emergence. And it suggests a way to understand the growth of the brain in the Homo lineage. How was the line able to support the continuous growth of such a demanding organ? Group-level support provided one breakthrough, but that probably happened 2.5 million years ago, right at the start of Homo. Another breakthrough was probably the introduction of cooked food, which provides many more calories than raw food. Richard Wrangham puts that start at 1.8 million years ago, so there is a second leap at the start of Homo erectus. But the brain kept growing. Neural efficiency suggests a third, continuous change. As the brain became more efficient, it could support a larger brain at the same level as the smaller, less efficient brain. Thus, brain B1 supports an N sized inefficient brain, but as the brain becomes more efficient an individual without much greater calorie intake can support an N+n sized brain B2. And what is behind this greater efficiency? At least part of it has to be the more efficient ability to speak.