Science is not a disinterested examination of the structures of reality. Objective truths about the world can be captured only in the thicket of cultural belief, refined experience and honed intuition. The nature Darwin met on his Beagle voyage in the 1830's did not flow through hollow eyes to imprint features mechanically on a pliant brain; it came tumbling through a mind whose romantic contours had been prepared by extensive study of the work of the German adventurer Alexander von Humboldt. In ''The Origin of Species'' (1859), that nature -- creative, value-laden and goal-directed --had not shed its original cultural guise.

Of course, scientific ideas, once formulated, must be tested against systematic experience, which itself might rest on more stable cultural belief. When the subject is the origin of human beings, where anchoring evidence is relatively light, cultural assumptions might carry it away on a tide of possibilities. Since Darwin wrote ''The Descent of Man'' in 1871, many evolutionary constructions have been conceived, usually bearing the features of their cultural and social contexts. With Ian Tattersall's smoothly developed argument about the evolution of distinctively human nature, cultural belief plays a decisive role.

In ''Becoming Human,'' he initially leads his readers into the dark recesses of the Combarelles cave in France. Moving some distance down into the suffocating lower end of the cave, with the aid of artificial illumination, a visitor will come upon an astonishing array of delicately engraved figures -- horses, lions, mammoths, wooly rhinoceroses and other creatures that inhabited the region some 13,000 years ago. Over these remarkable murals, the late ice-age artist layered geometric designs of undeciphered significance. This cave and many others in France and Spain give poignant evidence that our Cro-Magnon predecessors exhibited a distinctively human need for symbolic expression of values and beliefs.

The question Tattersall sets from this arresting beginning is this: What evolutionary history led to the appearance of a symbol-making creature?

He elaborates a scenario that traces hominid evolution from its origins some five million years ago in Africa, with the appearance of the Australopithecines, through the emergence of various species of the genus Homo -- Homo habilis, Homo ergaster, Homo erectus, Homo neanderthalensis, and the rest -- right up to us, Homo sapiens, when we first appeared displaying symbolic behavior, perhaps about 50,000 years ago.

Chimpanzees and other apes offer the best models for understanding the cognitive capacities of early hominids. Tattersall reviews the evidence for chimp language and toolmaking, and decides that chimps -- and by extension the early hominids -- display no real capacity for these behaviors, which depend on symbolic cognition. More orthodox evolutionary theorists would find his judgment too rigidly dichotomous.

Darwin and most contemporary neo-Darwinians have held that the anatomical and cognitive traits of modern humans gradually arose in tandem; faint beginnings would be ascribed to our remote ancestors and greater capacities to our more proximate relatives. Tattersall believes humanity was achieved in a quantum leap. After consideration of anatomical and artifactual evidence, he is reluctant even to admit that the Neanderthals had language and the kind of symbolic understanding that would mark them close cognitive cousins to their contemporaries, the cave-painting Homo sapiens. This quantum evolutionary theory seems more a conclusion derived from deep cultural belief than from strong evidence or convincing hypothesis. Tattersall does provide, however, an interesting, if unorthdox, theoretical support for his belief.

He endorses a punctuated equilibrium model of hominid evolution. That was originally suggested by Thomas Henry Huxley, but advanced in its modern form by Stephen Jay Gould and Niles Eldredge, Tattersall's friend and colleague at the American Museum of Natural History. It suggests that hominid evolution was punctuated by speciation events (usually caused by dramatic change in climate or geography) during which groups split from one another, creating new species that might remain morphologically and behaviorally stable for hundreds of thousands of years. As various species of the genus Homo separated from one another, new traits would erupt, keeping species distinct but not necessarily cognitively improved. This model permits -- and Tattersall supposes -- that speciation occurs not simply by natural selection operating on individuals but also on species themselves. He recognizes, however, that hypotheses about species selection are highly controverted in the scientific literature.

He deviates most from established evolutionary ideas in his assumption that function follows anatomy rather than the reverse. In the standard account of hominid evolution, functions of incipient language use and requisite mental representations would provide the pressure for consequent molding of the voice box and improved cerebral connections. Individuals who were slightly more efficient in communication and understanding would have advantages in hunting, child rearing and surviving. Such selective processes would gradually improve the cognitive capacities of those hominid lines whose previous adaptations allowed such traits, finally producing creatures that could cleverly and finely depict animals in caves.

Tattersall argues that gross anatomical changes in early species of Homo would generally occur prior to new functional acquisitions. He thinks the evidence that simple tool design remained static during the evolution of many hominid species argues for the disjunction of large brain size and cognitive function. Nonetheless, at speciation junctures, when one hominid species would arise out of another, certain ''exaptations'' -- traits that would later become adapted to new uses -- would have come together by sheer chance. A lengthened larynx, enlarged prefrontal lobes and cerebral wiring for speech would have arisen independently. They would not, according to Tattersall, have been mutually selected for the gradual improvement of symbolic consciousness. Why they came together to allow the emergence of such mentality he regards as a mystery. Yet the birth of modern man came to our late-Pleistocene ancestors some 30,000 to 50,000 years ago. Perhaps it was a small neural change, fed by a comparably small genetic alteration. But with all the other exaptations in place, that small change would have completely transformed our ancestors. Self-reflection and symbolic consciousness, Tattersall argues, would then have suddenly emerged as wholly new properties of the brain.

This scenario will not please many evolutionary biologists. A lot of hominid history will remain mysterious. And the general reader will at least be puzzled. For the apparent intent of this history -- to find out how and why we became human -- seems to dissolve in mist two-thirds of the way through the book. These consequences result from Tattersall's employment of the punctuated equilibrium model. In his hands, that model suggests that no analysis of early hominid development can finally illuminate the dark abyss separating us from even our recent ancestors. Chance accumulation of just the right traits to allow emergence of symbolic consciousness and an improbable gene mutation that instantaneously activates the whole -- such happy accidents would have to remain inscrutable. Moreover, the behavior of ancestors left on the other side of the divide would have little relevance for understanding our own actions and inclinations. Though it seems to undermine his historical intent, Tattersall himself draws this paradoxical conclusion: ''Whether these ancestors were hunters or gatherers, or both; whether they lived in large or small groups; whether early human males were promiscuous, or faithful providers -- all this has little bearing on what we are (or can be) in the late 20th century.'' Tattersall exhibits a reverence and awe for human accomplishments, from French cave paintings to our inadequate efforts to form an idea of God. This cultural attitude inclines him away from gradualistic and reductive explanations. He will not, for instance, attempt to lift all human behavior by the lever most favored by evolutionary scientists -- reproductive success. Indeed, even human sexual activity, he rather disconcertingly suggests, does not have its ultimate explanatory source in that biological principle. Esthetic and moral insight may have led him to advance his particular evolutionary construction, but without deeper argument and extensive evidence it only drifts on a sea of possibilities.

Robert J. Richards, a professor of history and the philosophy of science at the University of Chicago, is the author of ''Darwin and the Emergence of Evolutionary Theories of Mind and Behavior.''