Wednesday, January 06, 2010

Merlin Donald - The Definition of Human Nature

An older article from Dr. Merlin Donald, one of the more interesting people writing (see his online archives) about brain science and the philosophy of mind - see A Mind So Rare: The Evolution of Human Consciousness.

The following article (available as a PDF) is from The New Brain Sciences: Perils and Prospects (2004), edited by Dai Rees and Steven Rose (published as Chapter 2).
2 The definition of human nature
MERLIN W. DONALD

Our definition of human nature gives us a conceptual foundation for our ideas about human rights, individual responsibility, and personal freedom. These ideas were originally derived from the liberal humanities, and are ultimately the secular modern descendants of the concept of a `natural law' based on earlier religious and philosophical traditions. In this context, this is not a trivial exercise. It provides a conceptual foundation for our legal system, as well as our constitutional protections of human rights. Since the time of Charles Darwin there have been many attempts to define human nature in more scientific terms. In effect, this has amounted to an attempt to derive a new kind of natural law, based largely on scientific evidence, and especially on the theory of evolution. Here I am not speaking of Social Darwinism, an earlier intellectual movement that naively tried to extrapolate the laws of natural selection to human society, but of more recent empirical attempts to construct a culturally universal description of the human condition, and to explain it in terms of evolution and genetics.

In such attempts, human nature is usually defined as having been fixed when our species evolved in the Upper Palaeolithic, and this suggests that we have been genetically engineered to survive under the special conditions of late Stone Age civilisation. This raises the disconcerting possibility that human nature might prove to be maladaptive in today's high-tech, fast-moving, urbanised world. On the other hand, the logic leading to this conclusion is not compelling. It is based on two assumptions. The first is that human beings have a fairly rigid set of constraints on their mental and social life, imposed by an inflexible genetic inheritance. The second is that human mental and social life is determined largely by organismic variables, and that the human mind can be treated like that of any other species.

But there is an alternative view of human nature, based on scientific evidence and evolutionary theory that comes to a different conclusion, and fits existing scientific evidence better. It is based on a different set of assumptions. The first is that human nature has been characterised by its flexibility, not its rigidity. This is due largely to the overdevelopment of conscious processing, and those parts of the brain that support it. The second is that human beings, as a species, have evolved a completely novel way of carrying out cognitive activity: distributed cognitive-cultural networks. The human mind has evolved a symbiosis that links brain development to cognitive networks whose properties can change radically. Critical mental capabilities, such as language and symbol-based thinking (as in mathematics) are made possible only by evolving distributed systems. Culture itself has net-work properties not found in individual brains. The individual mind is thus a hybrid product, partly organismic in origin, and partly ecological, shaped by a distributed network whose properties are changing. Our scientific definition of human nature must reflect this fact, and free itself, not only of pre-scientific notions about human origins, but also of restrictive and antiquated notions about organismic evolution.

One consequence of this idea is that 'human nature', viewed in the context of evolution, is marked especially by its flexibility, malleability and capacity for change. The fate of the human mind, and thus human nature itself, is interlinked with its changing cultures and technologies. We have evolved into the cognitive chameleons of the universe. We have plastic, highly conscious nervous systems, whose capacities allow us to adapt rapidly to the intricate cognitive challenges of our changing cognitive ecology. As we have moved from oral cultures, to primitive writing systems, to high-speed computers, the human brain itself has remained unchanged in its basic properties, but has been affected deeply in the way it deploys its resources. It develops in a rapidly changing cultural environment that is largely of its own making. The result is a species whose nature is unlike any other on this planet, and whose destination is ultimately unpredictable.
The following is from later in the article.
The key question of human cognitive evolution might be rephrased in terms of this dichotomy: somewhere in human evolution the evolving mammalian nervous system must have acquired the mechanisms needed for symbol-based thought, while retaining its original knowledge base. To extend the metaphor, it is as if the evolving mammalian mind enriched its archaic neural net strategy by inventing various symbol-based devices for representing reality. This is presumably why the human brain does not suffer from the limitations of AI; it has kept the basic primate knowledge systems, while inventing more powerful ones to serve some non-symbolic representational agenda. But, how could the evolving primate nervous systems of early hominids have crossed the pre-symbolic gap? What are the necessary cognitive antecedents of symbolic invention? Cognition in humans is a collective product. The isolated brain does not come up with external symbols. Human brains collectively invent symbols in a creative and dynamic process. This raises another important question: how are symbols invented? I attribute this ability to executive skills that created a nervous system that invented representation out of necessity.

When considering the origins of a radical change in human cognitive skill, we must look at the sequence of cultural changes, including the cultures of apes and hominids. The cognitive culture of apes can be epitomised by the term 'episodic'. Their lives are lived entirely in the present as a series of concrete episodes, and the highest element of memory representation is at the level of event representation. Animals cannot gain voluntary access to their own memory banks, because, like neural nets, they depend on the environment for memory access. They are creatures of conditioning, and think only in terms of reacting to the present or immediately past environment this includes even their use of trainer-provided symbols, which is very concrete). Humans alone have self-initiated access to memory. This may be called autocueing, or the ability to voluntarily recall specific memory items independently of the environment. Consider an animal moving through a forest; its behaviour is defined by the external environment, and it can be very clever in dealing with that environment. But humans can move through the same forest thinking about something totally unrelated to the immediate environment - for instance the recent election, a movie or an item in the newspaper. In thinking about some topic, the thinker pulls an item out of memory, reflects on it, accesses another memory item, connects this to the previous idea, and so on, in recurrent loops. This reflective skill depends on voluntary autocueing; each memory item is sought out, precisely located and retrieved, preferably without retrieving a batch of other unwanted items, and without relying on the environment to come up with the relevant cues to help find the item. Our ability to transcend the immediate environment could not have developed without autocueing skill. Note that I am not saying we can introspect on the process by which we voluntarily access memory. We do not have to be aware of the retrieval process to have voluntary control over it. Language is `voluntary' cognition, but we have no awareness of where the words are coming from when we speak. The first symbolic memory representations had to gain explicit access to the implicit knowledge latent in neural nets. The initial adaptive value of the representational inventions of early humans would have been their ability to provide retrieval paths to a knowledge base that was already present, but not voluntarily accessible, in the primate brain. But, given the functional arrangement of the primate brain, where would such paths have been built?

THE FIRST STEP TO LANGUAGE: MIMESIS
The first cognitive transition occurred between 2.2 and 1.5 million years ago, when major changes in the human genome culminated in the appearance of Homo erectus, whose achievements indicate some form of improved memory capacity. This species produced (and used) sophisticated stone tools, devised long-distance hunting strategies, including the construction of seasonal base camps, and migrated out of Africa over much of the Eurasian land mass, adapting to a wide variety of environments.

Many evolutionary theorists are fixed on the idea that there was only one great cognitive breakthrough for humans: language, that this breakthrough came early, with Homo erectus, and that all higher human mental abilities followed from it. Bickerton (1990) argued that some form of proto-language must have existed at the time of Homo erectus, which might explain early hominid cultural achievements with a single adaptation - a sort of grammarless language - that later evolved into modern speech capacity. Pinker (1994) has suggested that grammar itself started its evolution early, and that some parts of a language module must have already been in place in Homo erectus.

I find this unconvincing. First, archaeological evidence doesn't place speech so early in evolution; neither of the principal mark¬ers for human language - the descended larynx and rapid cultural change - appears in the archaeological record until Homo sapiens, who evolved more than a million years later. Second, early hominids had no existing linguistic environment, and even proto-language would have required a capacity for lexical invention. This issue is crucial, because it raises the question of the autocueing of memory: lexical inventions must be self-retrievable, that is, autocueable. True linguistic symbols, even the simplest, could not suddenly pop up in evolution before there was some principle of voluntary memory retrieval in the hominid brain; to be useful, lexical inventions had to be voluntarily retrievable and modifiable, as well as truly representational acts, intentionally modelling some aspect of reality.

Before lexical invention became a realistic possibility, it was necessary to establish voluntary retrieval, or autocueing, in the pre-linguistic brain. The same adaptation would also have provided the cognitive prerequisite for a 'number of non-verbal representational skills. After all, language is not the only uniquely human cognitive advantage that has to be explained in evolution (Premack, 1986). If all our higher thought-skills were based on our linguistic capacity, how could we account for the virtual autonomy of some non-verbal forms of human intelligence? A good evolutionary theory of pre-linguistic adaptation should try to account for as many of these skills as possible, while providing the cognitive grounds for language.
My key proposal is that the first breakthrough in our cognitive evolution was a radical improvement in voluntary motor control that provided a new means of representing reality. Homo erectus's gift to humanity was mimetic skill, a revolutionary improvement in voluntary motor control, leading to our uniquely human talent for using the whole body as a subtle communication device. This body skill was mimesis, or a talent for action-metaphor. This talent, without language, could have supported a culture that, in terms of its tool-making abilities, was much more powerful refinements of skill, and flexible social organisation, than any known ape culture.

Mimetic skill logically precedes language, and remains independent of truly linguistic modes of representation. It is the basic human thought-skill; without which there would not have been the evolutionary opportunity to evolve language. Mimesis is an intermediate layer of knowledge and culture, and the first evolutionary link between the pre-symbolic knowledge systems of animals and the symbolic systems of modern humans. It is based in a memory system that can rehearse and refine movement voluntarily and systematically, in terms of a coherent perceptual model of the body in the surrounding environment, and is based on an abstract model of models that allows any action of the body to be stopped, replayed and edited, under conscious control. This is inherently an autocueing route, since the product of the model is an implementable self-image. Although the precise physiological mechanism of this system is not known, its functional retrieval path employs kinematic imagery. The principle of retrievability was thus first established at the top end of the motor system; and retrievable body-memories were the first true representations.

Mimesis is a supramodal skill.
Read the whole article.


No comments: