Two superordinate ideas unite and drive the biological sciences at each of these space-time segments. The first is that all living phenomena are ultimately obedient to the laws of physics and chemistry, with higher levels of organization arising by aggregate behavior at lower levels. The second is that all biological phenomena are products of evolution, and principally evolution by natural selection. The two ideas are expressions of consilience in the following way: Cells and thence organisms, being organized ensembles of molecules, are physicochemical entities, which were assembled not at random but by natural selection. Looked at this way, consilience in biology is the full sweep through the space-time scale, from near-instantaneous molecular process to the transgenerational shifts of gene frequency that compose evolution.

To many critics, especially in the social sciences and humanities, such an extreme expression of reductionism will seem fundamentally wrong-headed. Surely, they will say, we cannot explain something as complex as a brain or an ecosystem by molecular biology. To which most biologists are likely to respond, yes, we can, or we will be able to do so within a few years. The critics in turn call that impossible; such a complex systems are distinguished by holistic, emergent properties not explicable by molecular biology, let alone atomic physics. The only fair response to this is yes, put that way, you are right.

Thus arises the paradox of emergence: Complex biological phenomena are reducible but cannot be predicted from a knowledge of molecular biology, at least not contemporary molecular biology. Each higher level of organization requires its own principles, including precisely definable entities, processes, spatial relationships, interactive forces, and sensitivity to external influences, which permit an accurate characterization and perhaps a stab at prediction from knowledge of its elements. Still, the principles, if sound, can be reduced from the top down and stepwise to those formulated at lower levels of organization. An ecosystem, to take the most complicated of all levels, can be broken into the species composing its biota.

The species in turn can be analyzed according to the demography of the organisms composing them (population size and growth, birth and death schedules, age structure), along with their interactions with other kinds of organisms and with the physical environment. As part of this study, the organisms can be divided into organ systems, the organ systems into tissues and cells, and so on. The ecosystem, like other biological systems, is not truly hierarchical but heterarchical. It is constrained by the nature of its elements, and the behavior of the elements is determined at least in part by the sequences and proportions in which they are combined. By and large, however, the entities of each level can be reduced; and the principles used to describe the level, if apposite and correct, can be telescoped into those lower levels and, especially, the next level down. That in essence is the process of reduction, or top-down consilience, which has been intellectually responsible for the enormous success of the natural sciences.

To proceed in the opposite direction, bottom-up, by the synthesis–simple to more complex, general to more specific–is far more difficult. Physical scientists have succeeded splendidly at the task. They have interwoven principles of quantum theory, statistical mechanics, and reagent chemistry into stepwise syntheses from subatomic particles to atoms to chemical compounds. Advances in biology, if we measure their success by predictive power, have been much slower. Scanning the space-time scale along which biological complexity increases, we can see progress decelerate to a near stall at the level of protein synthesis. This is a critical juncture in the life sciences. About one hundred thousand kinds of protein molecules are found in the body of a vertebrate animal. Along with the nucleic acids that encode them, they are the essential materials of life. In particular, proteins form most of the basic structure of the body while running its machinery through catalysis of organic chemical reactions.

Thanks to advances in technology, biochemists find it relatively easy to sequence the amino acids composing at least the smaller protein molecules, and to map the three-dimensional configuration in which these units are arrayed. It is another matter entirely, however, to predict how amino acids will fold together to create the configuration. Three-dimensional form is all-important in the case of enzymes, which are the protein catalysts, because it determines which substrate molecules the enzyme molecule captures and which reaction it then catalyzes. When procedures are worked out to predict the exact shapes that arise from particular amino acid sequences, the result is likely to be a revolution in biology and medicine. It will permit the design of artificial enzymes and other proteins with desirable properties in biochemical reactions–perhaps superior to those occurring naturally. The difficulty is technical rather than conceptual: Prediction requires the integration of binding forces among all the amino acids simultaneously, an enormous computational problem; and in order to proceed that far it must also measure the forces with a precision beyond the capability of present-day biochemistry.

Even greater challenges are presented by the conceptual reconstruction of cells and tissues from a knowledge of the constituent molecules and chemical processes obtained through reductive analysis. In 1994 the editors of Science asked a hundred cellular and developmental biologists to identify the most important unsolved problems in their field of research. Their responses focused prominently on the mechanisms of synthesis.

In rank order, the problems most often cited were the following: 1) the molecular mechanisms of tissue and organ development; 2) the connection between development and genetic evolution; 3) the steps by which cells become committed to a particular fate during development; 4) the role of cell-to-cell signaling in tissue development; 5) the self-assembly of tissue patterns during development of the early embryo; and 6) the manner in which nerve cells establish their specific connections to create the nerve cord and brain. Although these problems are formidably difficult, the researchers reported that considerable progress has already been achieved and that the solution of several may be reached within a few years.

To summarize to this point, the consilience of material cause-and-effect explanations is approaching continuity throughout the natural sciences, binding them together across the full span of space and time. Of the two complementary processes of consilience, reduction and synthesis, the more successful has been reduction, because it is both conceptually and technically easier to master. Synthesis good enough to be quantitatively predictive has progressed much more slowly, but it is now inching its way within biology to the level of cell and tissue.