Can Matter Store Information?
Natural Selection
Evolutionary Theory
A theory of changes in organic design through 
controlled random mutations and contingent selection
(Francis Steen, revised 25 March 2001)

The Neo-Darwinian Synthesis

In broad terms, contemporary evolutionary theory builds on the synthesis of Darwin's ideas of natural variation and selection and Mendel's model of genetic inheritance accomplished by R.A. Fisher, J.B.S. Haldane, and Sewall Wright in 1930-32. For an overview, see George Williams, Evolution and Natural Selection (1966).

The Unit of Selection. The central realization of the neo-Darwinian synthesis is that natural selection is best understood to be acting on variation among elements that persist. Thus, natural selection does not act on me as an individual, because when I die, the structures that constitute my body and my genome - my full complement of genes - break down. Nor does it act on collections of individuals, since each of these eventually suffer the same fate (though see below).

However, if I have children, some of my genes persist, and it is here we must look for the appropriate unit of selection. The genome is organized into chromosomes that come in pairs. Each pair contains functionally similar but not identical alleles at corresponding loci. In the process of meiosis, sperm cells or eggs (gametes) are generated that contain single chromosomes instead of pairs. Each single chromosome is made up of alleles from both members of the pair in what appears to be a random shuffle.

Thus, the chromosome does not persist, but is scrambled with elements of its paired partner in the process of meiosis. This is repeated for each generation. What persists is the allele, which can be defined simply as a stretch of genetic material that is not broken up in the shuffle of meiosis, but is transmitted intact into the gamete and thus into the next generation.

Natural Selection. Once in the fertilized egg, the alleles regulate the construction of the organism. If this project is not successful, or if the organism lives but does not reproduce, all the alleles perish together. If there is even the slightest causal link between the failure to reproduce and a particular allele - that is, if a variant allele at the same locus in another organism enabled that organism to reproduce - then natural selection can be said to be operating on that allele. Such weak causal links can be detected by examining the reproductive outcomes of populations with slightly different phenotypic features; for an illuminating example, see Jonathan Weiner's The Beak of the Finch: The Story of Evolution in Our Time (1994). Evolution, accordingly, can be defined as a change in allele frequencies in a population.

The Strength of Selection. The selective survival of individual organisms due to functional differences, and thus the differential transmission into the next generation of alleles, is a pervasive force shaping the trajectories of evolutionary lineages. Natural selection takes several forms--directional, stabilizing, disruptive, indirect--and its strength varies in acting on different organismal traits. In the last several decades, researchers have attempted to measure the strength of the various types of natural selection, studying its effect on phenotypes and individual traits of organisms in the wild and in the laboratory. Kingsolver et al. examine this literature and find some unexpected patterns:

Kingsolver, JG; Hoekstra, HE; Hoekstra, JM; Berrigan, D; and others (2001). The strength of phenotypic selection in natural populations. American Naturalist 157. 3 (Mar 2001): 245-261.

Variation. For natural selection to operate, alleles that can occupy the same locus in the genome must differ somewhat between individuals. Such variation can appear because of replication or transcription errors, because of damage by radiation, or from other causes. Since the problem space of an organism is theoretically infinite, a random system of variation would be useful, though it need not be the only mechanism. For instance, regulatory genes that code for the introduction of random errors in specific alleles in response to certain environmentally generated cues may survive and spread.

Some evolutionary biologists argue that mutation rates are too slow to account for the observed variation. Since the functional effect of an allele depends on its position - the same allele at a new locus can subtly or dramatically alter the action of a whole suite of genes - it is possible that the primary cause of variation is chromosomal recombination rather than mutation (see McClintock 1987). Mutations in regulator genes is another candidate for rapid change (see King & Wilson 1975).

It is now also looking increasingly likely that the genome has evolved strategies of mutation (see Pennisi 1998 for an up-to-date discussion). If this turns out to be the case, mutations may still tap into the random, but the random may be tapped at controlled points so that the overall outcome is far from random. Localized novelties can be incorporated into the genome in non-random ways. Since random changes in a complexly ordered system may have moderately predictable outcomes, there is a vast field of possible genetic strategies that may evolve in response to patterns in the results of random mutations at determinate sites in the highly ordered genome. Life in this sense may be said to actively feed on the random in a controlled manner.

Somatic hypermutation. In an interesting special case, medical researchers have found that the immune system engages in a process called somatic hypermutation to increase antigen receptor diversity. In the body's immune response, B-cells are responsible for manufacturing and secreting antibodies, the protein molecules that bind to antigens. The ability of the immune system to recognize and respond to the enormous number of antigens encountered by an individual in a lifetime is due in large part to the diversity of antibodies (immunoglobulins) produced by B-cells. Each B-cell produces only a single species of antibody, and during the systemic immune response, the presence of a specific antigen results in the proliferation of B-cells producing antibody specific for that antigen (clonal selection). Antibody diversity depends on variability in the constituent amino acid sequences, made possibly by somatic recombination a scrambling and reassembly of genes in the DNA of the B-cell. During the past 15 years, it has become evident that in immune system B-cells, the part of the genome coding for the variable parts of antibodies is involved in a process of "hypermutation", a substantial increase in mutation rate, the effect of which is to provide the immune system with a rapidly changing enormous library of possible antibodies. This hypermutation process is highly specific to the immune system, and it occurs only within a DNA segment of approximately 1000 to 2000 DNA bases, the segment that encodes the bulk of the variable regions of the antibody polypeptides. The mechanism of the hypermutation process remains unknown. Since B-cells pass directly to the offspring from the mother, somatic hypermutation in the immune system provides a special case of a Lamarckian mechanism of inheritance. For an early overview, see French et al (1989).

Epistemological considerations. There is no reason to think we have a complete understanding of the operation of the genes. There are central areas of genetics, such as the question of how genes build bodies, that are still very poorly understood. It is also implausible that we have a full understanding even of those genetic processes for which we have highly specified models and vast amounts of experimental data supporting these models; this limitation is simply due to the inherent limitations in the activity of knowledge (see No Final Theory: Science as Perception). Such chronic incompleteness is a reason to be open to the discovery of new levels of order in the operation of the genes in ontogeny and under natural selection, to improved definitions of the scope of current models, and to a clarification of the mechanisms at work.

Current debate. There are various grounds on which to question aspects of the current evolutionary model, and a lively debate persists today. Evolution is in principle hard to model precisely, since the changes it describes usually takes place over time periods that are inaccessible to human beings. Consider the related situation in astronomy. Changes in the movement of the stars are slow, and until very recently were too slow to be detected within the lifetime of an individual. However, with the help of a continuous series of observations dating back to the fifth century BC, Copernicus was able to formulate a detailed model that fit two thousand years of data. Unfortunately, in the case of biology, two thousand years of continuous observation would in most cases reveal very little. We must thus rely in indirect evidence, such as fossil remains and systematic structural similarities and differences in living forms. This evidence leaves room for a variety of possible interpretations of past events, and thus of the mechanisms of change that underlie them. I can examine only a few focal points of contention.

Gradualism. All the way back to Darwin, the notion that changes accrue gradually over long periods of time has been a central proposition of evolutionary theory. As Ernst Mayr put it in Animal Species and Evolution (1963), "all evolution is due to the accumulation of small genetic changes" (p. 586).

In contrast, the fossil record suggests long periods of stasis followed by brief periods of rapid change - what Niles Eldredge and Stephen J. Gould dubbed punctuated equilibrium. This data has sometimes been taken as evidence against the neo-Darwinian model by people who believe the order of nature is due to the intentional act or acts of a supernatural being. Within the scientific tradition, the relative lack of continuous change in the fossil record is interpreted as evidence that speciation events have typically taken place in small populations over relatively short periods of time.

In addition, gradualism should not be discounted. For instance, in the period from 300,000 to 100,000 year ago, fossil remains of the genus Homo show a wide range of forms. It is not unlikely that we have inherited alleles from individual mutations that took place over a wide geographical area during this period. As the best mutations spread throughout the existing populations, the range of functionally meaningful variation drops towards zero. Archeologists of the future may well see only our remains, appearing as if by the hand of God, while the gradual accumulation of alleles that made us possible leaves little or no trace.

Group selection. There are situations where it is useful to think of natural selection acting on collections of alleles rather than on individual alleles, and thus on individuals, groups, species, and taxa. In most cases, however, group selection is thought to be so weak that the effect can be ignored; for a sustained argument, see Williams (1966). In contrast, David Sloan Wilson (2000) points out that "Group-level adaptations will evolve whenever group-level selection is stronger than individual-level selection" and argues this is often the case. Darwin, he points out, held that "the driving force behind the evolution of morality was the process of more moral groups replacing less moral groups, not the process of more moral individuals replacing less moral individuals within groups." He concludes that "group selection is a significant evolutionary force in nature and especially strong in the case of our own species." This article, a review of Wright's Nonzero, provides a useful overview of the debate.

Chris Boehm (1999) has proposed that group selection may have acted to create psychological adaptations favoring egalitarianism; see extracts.

Litterature on group selection

Boehm, Christopher (1999). Hierarchy in the forest: the evolution of egalitarian behavior. Cambridge, MA: Harvard University Press. Extracts (local), publisher's presentation (external) and editorial reviews ( Reviewed by Vincent Kiernan (external).

Keller, Laurent (ed.) (1999). Levels of Selection in Evolution. Princeton, NJ: Princeton University Press. Reviewed by Herbert Gintis (Amazon, external).

Wilson, David Sloan and Elliott Sober (1994). Reintroducing group selection to the human behavioral sciences. Behavioral and Brain Sciences 17. 4 (December): 585-609.

Wilson, David Sloan and Elliott Sober (1998). Unto Others: The Evolution and Psychology of Unselfish Behavior. Cambridge, MA: Harvard University Press

Wilson, David Sloan. A review of Wright, Nonzero. Full text (external).

Baldwin effect. A longstanding evolutionary principle called the Baldwin Effect says that an advantageous behavior, once it has appeared in a population, will gradually reshape the genes of the species which has adopted it.

"At the end of the 19th century, biologist J. M. Baldwin enunciated the Baldwin Effect, which observed that when a species learns a useful new skill, the addition to its behavioral repertoire will reshape its biology.  Over time, says Baldwin, natural selection will bless the members of ensuing generations whose limbs and brains are suited to the maneuver, and cull out those whose anatomy is ill-suited to the innovative gambit." (Steven Levy, Artificial Life. New York: Vintage, 1993, p. 265.)

Literature on the Baldwin effect

Baldwin, J.M. (1896). "A new factor in evolution." American Naturalist 30: 441-451. Full text (external).

Behera, N. and V. Nanjundiah (1995).  "An investigation into the role of phenotypic plasticity in evolution." Journal of Theoretical Biology 172: 225-234

Belew, Richard K. and Melanie Mitchell, eds. (1996). Adaptive Individuals in Evolving Populations: Models and Algorithms. Reading, MA: Addison-Wesley. Series title: Proceedings volume in the Santa Fe Institute studies in the sciences of complexity, v. 26.

French, R. and A. Messinger (1994). "Genes, phenes and the Baldwin effect." Artificial Life IV, ed. Rodney Brooks and Patricia Maes. Cambridge, MA: MIT Press.

Hirst, Tony. "Search Space Neighbourhoods as an Explanatory Device: General Observations and a Reconsideration of the Baldwin Effect." Abstract.

Smith, John Maynard (1987). "When learning guides evolution." Nature 329: 761-762.

Turney, Peter, Darrell Whitley, and Russell Anderson (eds) (1997). "Evolution, Learning, and Instinct: 100 Years of the Baldwin Effect."  Evolutionary Computation (Special Issue on the Baldwin Effect), vol. 4, no. 3. Abstracts (external).

Weber, Bruce and Soren Brier (eds.) (2000). The Embodied Mind and the Baldwin Effect. Special Issue of Cybernetics and Human Knowing 7. 1. Publisher's presentation (external).

Non-linear dynamics. The realization that even mechanical systems are subject to non-linear dynamic effects may have important consequences for biology. For an overview, see Stuart A. Kauffman, The Origins of Order: Self-organization and Selection in Evolution (1993); here is a brief excerpt:

Consciousness and ethics. The basic explanatory framework of the neo-Darwinian synthesis is materialist (see for instance excerpts from Williams (1966). If this framework is adhered to, the phenomena of sentience and value must remain epiphenomenal. It is not clear this is a fully satisfactory situation, nor do we have any a priori reason to believe that the materialist framework is adequate for all phenomena. However, it is important to realize that the incompleteness of evolutionary theory is not in itself an endorsement of other modes of thinking we find easier to believe in. Rather, it may in general be inappropriate to use a theory of any kind as an ultimate framework for life as a whole. See for instance Bohm and Wiley on knowledge as perception.

Some basic principles of evolutionary theory

Some entailments Definitions

Mechanisms of Selection

Historical Notes

Before Darwin: Invertebrate Paleontology as Geology

"Little work of importance was done in paleontology until the 1700's, at which time both vertebrate and invertebrate fields began to assume importance. Intensive work in the invertebrate area arose from recognition of the fact, first clearly seen by William Smith, an English civil engineer and amateur geologist of the period, that a given set of beds tended to contain the same species of shells over vast and widely separated areas. Accurate determination of fossils could thus be of great practical use to the stratigrapher; as a result, invertebrate paleontology tended to develop not as an independent science, but as a handmaiden to the geologist -- a working tool for the stratigrapher looking for oil or ores or coal. The fossil shells were rarely thought of as the remains of once-living organisms, but merely as convenient markers for the identification of successive formations, and would have been as useful had they been identifiable mineral inclusions or distinctive assortments of nuts and bolts...

With this background, the invertebrate workers of Darwin's day not merely lacked interest in evolutionary ideas, but were inclined to view them with suspicion as detrimental to their work. For clear-cut stratigraphic work, the species in a given formation should be stable entities, clearly distinguishable from those in the strata above and below. The idea of gradual change and of transitional forms was abhorrent...

With this to contend with, it is apparent why Darwin was thrown on the defensive in his treatment of the fossil record. He could not call on the paleontologists for support; the most he could do was to attempt appeasement, to show that it was at least possible to interpret the geological story in evolutionary terms, and that there was no insuperable objection."

-- A.S. Romer: "Darwin and the Fossil Record" (1958). In S.A. Barnett (ed.). A Century of Darwin. From chapter 6.


Can Matter Store Information?
Previous: Can Matter Store Information?
Natural Selection
Back to EvPsych index
Next: Mechanisms of Selection
© 2001 Francis F. Steen, Communication Studies, University of California, Los Angeles