| Human Nature Review ISSN 1476-1084 | Table of Contents | What's New | Search | Feedback | Daily News | Submit A Manuscript |
PDF of this article
Download Adobe Acrobat Reader
Email the reviewer
Reviewer's web site
Publisher's web site
Send a response to this article
Search the web for related items
Contact the Editors

The Human Nature Review Human Nature Review  2002 Volume 2: 367-374 ( 17 September )
URL of this document http://human-nature.com/nibbs/02/lewontin.html

Essay Review

Lewontin’s Living Legacy: Levels of Selection and Organismic Construction of the Environment

by Val Dusek

A review of Thinking About Evolution: Historical, Philosophical and Political Perspectives 
edited by Rama S. Singh, Kostas Krimbas, Diane B. Paul, and John Beatty
Cambridge University Press 2001

During a single week I was chastised by librarians of two different libraries at my college for making excessive noise while reading Richard Lewontin. In one I was loudly and enthusiastically exclaiming my appreciation of one of Lewontin’s articles on organism and environment, “Holy __! He really is doing dialectical biology!” While in the biology library I was hushed for laughing loudly at a satire by Lewontin and Levins written under the name of the sage “Isador Nabi” whose spurious biographies appearing in American Men and Women of Science (showing how little that publication checked its submission) and elsewhere variously claim that he was graduate of Cochabamba University, research scientist, executive of an oil company, and revolutionary editor of the journal El Fomento.) Lewontin’s work both in biology and politics has inspired equally strong reactions, both positive and negative, in many others.

Lewontin’s eloquence in speaking is legendary. Neural selectionist G. M. Edelman likes to tell the story of how Lewontin performed in a play at the French language high school in New York that Lewontin and Edelman both attended. For the occasion the school had invited a noted Franco-American Hollywood actor to evaluate the performance. The actor claimed that Lewontin was the most naturally talented actor he had ever seen and wanted to arrange a Hollywood contract for him, which, luckily for genetics, Lewontin refused.

This is the second volume of a festschrift for Lewontin, the leading evolutionary geneticist, Marxist, and critic of genetic explanations of human behavioral characteristics. This volume contains twenty-eight articles, including the work of some ten leading philosophers of biology, several of whom worked in Lewontin’s laboratory while on leave or exchange (Sober, Lloyd), took courses with (Brandon), was a colleague of (Wimsatt), co-authored with (Sober, Godfrey Smith) or were influenced heavily by Lewontin. There also are works of a general nature by several of Lewontin’s Marxist or radical biologist colleagues or comrades, including edited and abridged chapters from books by Steve Gould and Steve Rose, an article on identity politics by Ruth Hubbard, and a critique of chaos theory by Richard Levins. I shall concentrate on those articles that I think most clearly develop two theoretical themes originally adumbrated by Lewontin, the critique of genic selectionism and the defense of the claim that organisms construct their environments. For brief, clear abstracts of all twenty-eight articles see the review by Michael Bradie (2002) in this journal. An adequate essay-review of the score of topics covered in various articles would be at least as long as the book itself.

Festschrifts in philosophy and history sometimes serve as repositories for articles too eccentric in approach or too sloppy in execution to be publishable in journals. This is emphatically not the case with this volume. I was impressed by the uniformly high quality of the articles. A number summarize whole books with replies to subsequent criticism or epitomize decades of work of an author in two-dozen pages. Thus the volume is a valuable summary of the “state of the art” in the philosophy of evolutionary biology as well as containing a number of valuable articles critical of behavior genetics, sociobiology, and, by implication, parts of evolutionary psychology. One reason for the high quality and density of information in the articles is, I suspect, that many an author hoped (or feared) that Lewontin would actually read their piece, and is well aware of the dedicatee’s notoriously sharply critical tongue and pen.

Elizabeth Lloyd’s “Units and Levels of Selection” (pp. 267-291) is an especially valuable piece that anatomizes the many various meaning of “group selection” as used by different advocates and critics. Her treatment of the various claims of Richard Dawkins makes one think that, although Steve Gould was the target of Maynard Smith’s canard that he was too confused to take seriously, Dawkins himself is at least equally open to that charge. (Obviously both, even if sometimes confused, are well worth taking very seriously.) Lloyd distinguishes four different questions that are often run together as the question of group selection. They are the question of a.) The interactor (David Hull's more neutral replacement for Dawkins "vehicle") b.) The replicator c.) The beneficiary of selection d.) The possessor of adaptation. She notes that Wynne Edwards, the original group selectionist whipping boy, combined the benefit question with the adaptation question. The early critics, such as G. C. Williams and Maynard Smith, as well as Dawkins, in his lucid popularization of the formers’ views, tend to do this also. This is understandable given that Wynne Edwards was the original target. Later group selectionists, such as D. S. Wilson, Michael Wade and others did not hold this position, but Dawkins criticizes them as if they held it. Sewell Wright correctly noted that group selection for group advantage (benefit and adaptation) is not the same as his shifting balance theory, in which groups were vehicles, but individuals reaped the advantage. Both G. C. Williams and Maynard Smith both recanted in the late eighties and recognized that D. S. Wilson’s "group selection" was not the sort of thing they were criticizing as group selection. Dawkins apparently did not notice this and held his original characterization of all group-selectionists in the second edition of The Selfish Gene. Meanwhile, Dawkins, oddly enough, like Monsieur Jordain speaking prose, seemingly unknowingly discovered higher-level selection in writing an article on artificial life. (Dawkins elsewhere has been unwittingly led to positions contrary to his official one out of proneness to being excessively impressed by techno-hype. In his Unweaving the Rainbow Dawkins ridicules the relativism and subjectivism of the postmodernists. Later in the book, however, in his enthusiasm for virtual reality, is led down the slippery slope to a subjectivism that would make Bishop Berkeley blush.)

Gould in his article (pp. 230-32) notes that the later Williams (1992) ends up accepting selection on interactors (as opposed to replicators) at the material level but places genic selection at a non-physical, informational, or "codical" level. Interestingly, this revised position of Williams converges with the position that Lewontin (1976) in an early, brief review suggested Dawkins would have held if he had been less confused. Lewontin noted that Dawkins claimed that an extra-terrestrial attempting to gauge the level of human civilization would ask, "Do you understand natural selection?" Lewontin counters that a better question would be, "Do you understand the difference between sets and their members?" (implying that the early Dawkins does not). Many have noted that the immortality that Dawkins attributes to genes applies not the physical DNA but to the whole set through time of the copies or to the form of the sequence as manifested in successive, physically different, individual DNA molecules. Williams new "codical" realm likewise resides in this quasi-Platonic region, separate from the material world.

Gould (p. 233) mentions that the first question Lewontin ever asked him (in an airport on the way to a conference) was concerning Aristotle's distinction between efficient and final causes. Gould (pp. 218-19) himself sketches Aristotle's four causes: material cause (constituent stuff), efficient cause (today identifiable with mechanical causality), formal cause (today structure, including notions such as symmetry) and final cause (purpose, goal or teleological cause). Gould then claims that the gene is material cause of selection while the interactor (organism, group or species) is the efficient cause of selection. Gould does not note that the later Williams of the "codical" realm or the corrected Dawkins (distinguishing sets from their members) would be treating the gene, not as either material or efficient cause, but as formal cause. Almost a century ago W. Bateson moved in this direction, but most biologists subsequently shied away from this Platonic rejection of materialism. Some others such as phage group leader and molecular biology founder Max Delbrück and scholastic philosopher Brian Cooney have characterized the gene as an Aristotelian substantial form (a kind of form of the individual) as opposed to a material or efficient cause, though this notion would be as alien as that of Platonic general forms to the strict materialist.

One line of criticism of Dawkins by Lewontin, Wimsatt, and others, is that because of context-dependence of genes’ operations the fitness of single genes cannot be evaluated (due to linkage disequilibrium, early studied by Lewontin, as well as complex interactions of genes) does not faze Dawkins, as Lloyd points out. Dawkins (1982), in The Extended Phenotype, grants that larger chunks of DNA, such as a whole chromosome in diploid organisms could operate as the object of selection. Thus, although arguments of Lewontin and his disciples Wimsatt and Sober are valid, they do not undermine the notion of the genome as object of selection, only that of particular genes as objects of selection. R. Brandon has presented a dual hierarchy of "units" and "levels" of selection against the single hierarchy of Wimsatt, Lloyd or Gould. Brandon restricts use of "units" of selection to issues of the replicator (for instance, genes. chromosomes, genomes, groups, species) and "levels" of selection to levels of interactors (including genes, organisms, groups, species.)

Lewontin wrote a classic paper on levels of selection in 1970 and most of the leading figures in the philosophical analysis of levels of selection (Wimsatt, Sober, Brandon, Lloyd, Godfrey Smith, Kitcher) are students of or sometime collaborators with Lewontin. All, with the exception of Kitcher, agree that universal genic selection is inadequate. However, the defenders of an hierarchical levels of selection view do not speak with one voice. Wimsatt, followed by E. Lloyd have used additive (or linear) variance of fitness as a criterion of a level of selection (where lower levels lack additive variance). Godfrey Smith (1992) and more briefly Lewontin (1991) have raised problematic cases or issues for this view. Robert Brandon and Elliott Sober have presented causal analyses of selection. Lloyd (1989) in turn has criticized Sober's causal account in claiming he trusts idealized models too literally. Brandon introduced causal analysis in terms of "screening off." (stemming from the logical empiricist Hans Reichenbach via Wesley Salmon). One cause screens off another when it alone can account for the probability of the effect, while the other (screened off) cause alone cannot do so, and probability of the effect, given the two causes together is the same as that yielded by the screening off cause alone. The second is screened off by the first. Brandon [1992, 83-8, 1996, 60-4, 128-30] claims that the phenotype screens off the genotype with respect to ordinary, individual selection. Similarly, in any case of genuine group selection, the group characteristics screen off the individual phenotypes.). Sober (1992) criticized the screening off analysis and offered a slightly different causal analysis borrowed from what Brian Skyrms called "Paretian" causality. (This refers to a structure of probabilistic causal conditions similar to Pareto optimality, not to Pareto's own account of sociological causality.) Brandon and a number of biologist and philosopher collaborators (1994) replied to these criticisms. Part of the disagreement hinges on whether event chains or genuine causes are appealed to and whether frequency theory or propensity theory of probability is used. Nevertheless, as Brandon points out in a recent review (2001), the technical differences between Brandon and Sober should not overshadow the fundamental agreement that almost all philosophers and biologists who have specialized in the units of selection problem have come to, that there is a hierarchy of levels and/or units of selection (and that universal genic selection is a non-starter).

Gould (pp. 229-30) notes that Dawkins since 1982 has taken up a conventionalist rather than a scientific realist account of genic selection. This suggests something about the nature and structure of the theory, for not only Dawkins, but the philosophers P. Kitcher, K. Sterelny, J. Cassidy and K. Waters, who defend universal genic selectionism as viable, all do so in terms of conventionalist philosophy of science. A conventionalist (such as physicist Henri Poincaré) claims that scientific laws or principles are not empirical descriptions of reality but arbitrary conventions or stipulations. In contrast, most of the defenders of hierarchical theories of selection such as Sober and Brandon are scientific realists (though Lloyd attempts to remain neutral on the issue of realism). Dawkins (1982) explicitly shifts to defend his view in terms of one point of view among several. According to Dawkins (1982) genic selection and organismic selection are like two interpretations of a Necker cube illusion. Kitcher and others all emphasize that genes can "represent" the outcome of selection. Wimsatt had introduced the notion of genes as "bookkeeping" devices and not causes of selection. (Wimsatt himself had a kind of implicit screening off view of causality here). What the critics of genic selectionism argue is that although genes can indeed represent results of selection at all levels, the representation misses the causal connections in selection. Lewontin and Godfrey Smith (1993) describe this in terms of genic selectionism models having too low a dimensionality to capture causal processes in many cases of selection, although they always can represent results of a selection process.

Russell Gray in "Selfish Genes or Developmental Systems?" (pp. 184-207) summarizes developmental systems theory and replies to some of the criticisms of it. Both Lewontin and developmental systems theorists reject genetic determinism and the primacy of the gene. Gray (p. 198) notes that Dawkins argues against the nucleotide being the unit of selection because it is so context dependent, but that Dawkins rejects this same argument when it is deployed to defend units of selection above the gene on the basis of the context dependence of individual genes.

Gray (p. 200-201) also replies to P. Kitcher's criticism of both developmental systems theory and Lewontin in the Kitcher's "Battling the Undead: How (and How not) to Resist Genetic Determinism" appearing in the festschrift. Like Brandon (see below) with respect to Lewontin's organism/environment account, Gray denies Kitcher's and Sterelny's claims that developmental systems theory or Lewontin's dialectical biology dissolve into intractable holisms that do not allow specific empirical research. Gray (p. 201) notes that the tradition of developmental biopsychology as in the work of Daniel Lehrman tease apart physiology, behavior and the environment.

Kitcher (pp. 398-399) claims that genetic determinism consists of the empirical claim that certain forms of norms of reaction (such as constant, constant slope consistently above that of another gene, or constant over most of its range or its non-pathological range) are most common. Kitcher's account (p.400), however, implies the odd claim that Lewontin (and developmental systems theory) reject the norm of reaction notion as incoherent. This would seem not at all to be the case. Lewontin makes much use of the norm of reaction notion in his criticisms of genetic determinism. Some critics of Lewontin attempt to counter him by claiming the norm of reaction notion is not used by most working biologists (hardly an effective strategy is Lewontin himself rejects it as incoherent.) Indeed, Kitcher grants in the Acknowledgements (p. 203) that correspondence with Lewontin has convinced him that Lewontin's own positions are not identical with those Kitcher criticizes.

Two of the most valuable articles in the book are attempts by Robert Brandon "Organism and Environment Revisited" and Peter Godfrey-Smith "Organism, Environment and Dialectics" to spell out Lewontin’s “dialectical” view of organism and environment in terms of analytical philosophy and biology. (They articulate better than my shouted obscenities noted above the “cash value” for non-Marxists of Lewontin’s ideas.) Many, I suspect, are needlessly put off from Lewontin’s fruitful ideas on this topic by the references to and terminology of Marx’s sidekick Friedrich Engels. (Because of political opposition to Marxism, some would, I fear, deny 2+2=4 if they found out Marx had held this equation to be true.)

An accusation made by many, including those agreeing with much of Lewontin’s work, such as Kitcher (p. 408) is that Lewontin criticizes mainstream selectionist approaches without offering an empirically testable alternative, and that Lewontin’s approach would lead those attempting to use it into an intractable holism. Robert N. Brandon reviews two articles, one by J. Antonovics, N. C. Ellstrand and himself (1988) and one by K. N. Laland, F. J. Odling-Smee, and M. W. Feldman (1996) that attempt to develop biological models involving niche construction by organisms. Both involve frequency dependent selection. As with most models in population genetics there are various extreme simplifying assumptions. The Laland model has a complex genetic situation, diploid two-loci but takes only global frequencies of organisms. The Brandon model is genetically simpler (haploid, one locus) but has a more complex spatial structuring of populations. The Laland model shows that niche construction can inter alia affect inertia and momentum, fix otherwise deleterious alleles, support stable polymorphisms where otherwise not expected and influence disequilibrium. Time lags between niche construction and selection pressure change resemble those in maternal inheritance of Lande and gene-culture co-evolution of Cavalli-Sforza (presumably both of which phenomena are varieties of niche construction.)

Although Lewontin’s organism/environment dialectic and niche construction have not otherwise been explicit inspirations for population genetics work, Brandon argues that works on within-population frequency dependent selection and works on co-evolution are in fact a cases of organism constructing environment even if not described explicitly as such.

One criticism by Brandon is that Lewontin uses his notion of “construction” of the environment by the organism too broadly. Just as the notion of construction used by the social constructionists in sociology of scientific knowledge is multiply ambiguous, so, it would seem, is Lewontin’s.

Social constructionists sometimes mean that scientists physically construct scientific instruments and technological devices (a totally uncontroversial claim), sometimes that scientists construct concepts (a fairly non-controversial claim), either from pre-existing elements or out of whole cloth, or sometimes that scientists literally construct facts or reality (a highly controversial claim, to say the least). Part of the apparent force or shock-value of social constructionism is the unacknowledged slide from the innocuous forms of construction to claims about the construction of reality -- as in "the social construction of facts and artifacts." Some social constructionists are Kantians in which the construction organizes a pre-existent stuff. Others are Fichteans in which construction is creation ex nihilo.

Lewontin uses construction to include cases in which organisms select or determine the relevance of features of their environments as well as cases of physical modifications of the environment by organisms. Brandon and Godfrey Smith wish to use “construction” only in cases in which organisms physically modify their environments, not in cases where organisms determine by their structure which parts of the physical environment are relevant.

To describe the matter this way assumes a distinction, such as Brandon makes, between the pre-existent physical environment, the ecological environment, and the selective environment, and a distinction Godfrey Smith (p. 263) makes between the absolute properties of entities and the relational properties of entities.

Need the “constructionist” evolutionary ecologist hold to these traditional distinctions? One way he/she might avoid them, and the criticisms and restrictions that flow from them, is to consider apparently absolute properties of objects as really relational ones. Idealism, such as that of G. Berkeley erased Galileo’s and Locke’s distinctions between secondary qualities (color, taste) that like sensations (tickles, pains) were relative to the knower and primary qualities (space, time and mass) that were absolute. Berkeley made the spatio-temporal qualities relative to the subjective mind. However, relational theories of space, time and mass need not be subjective. Arthur Murphy coined the term “objective relativism” for a view, like that of the relational interpretation of Einstein’s special relativity theory, that all qualities are relational, but that the relational structure is an objective one. The logician and process philosopher A. N. Whitehead and the pragmatist G. H. Mead are Murphy's examples of objective relativists. The early Bertrand Russell, though unsympathetic to his erstwhile collaborator Whitehead’s metaphysics, presented a logical construction of the world in terms of perspects and aspects (the subject-oriented and objective features of the world.) The early Rudolf Carnap developed Russell’s ideas with more formal logical rigor.

Clearly Lewontin’s Marxism would be critical of both idealism and positivism. Some writers, however, have noted the parallelisms between an objectivistic Marxism and Whitehead’s objective relativism. On an objective relativist view the physical environment would be taken as the system of perspective of all the selective environments of the objects of selection (primarily organisms, but possibly also genes, cells, groups, and species). I think it is significant that developmental systems theorists and others sometimes describe the environment in terms of "the point of view of the organism" without literally meaning the perception or cognition of the organism.

We paragons of nature may consider the perspectives of an earthworm or an eagle inferior to ours. We self-described Homo sapiens correctly consider science to be the best developed way of knowing, and thus consider our description of the environment the most accurate one, and those of earthworms more confused than ours. However, one may be a realist without claiming that our present science mirrors nature accurately in all respects. Our theories and models are approximations and idealizations of nature.

Lewontin’s dialectical account of organism and environment is inspired by Hegel via Marx. Dawkins, whose views on adaptation and selection appear to be the antithesis of Lewontin's, is led to the "extended phenotype" notion. Though Dawkins' purpose is eliminating the organism, there are some peculiar similarities between the dialectical view of organism and environment and the extended phenotype notion. Both emphasize that at the very least significant aspects of the environment are produced by organism for Lewontin and genes for Dawkins. If the Dawkinsian gene really is a quasi-Platonic, formal entity, rather than the particular, material thing it appears to be for Dawkins and most of his readers, then the contrast between Dawkins and Lewontin on constitution of the environment is between Neo-Platonic emanationism and Hegelian mutual constitution. For the Platonized Dawkins, like the ancient and renaissance Neo-Platonists Plotinus and Ficino, the genetic form emanates at least some features of the environment.

Finally, I mention that Lewontin (1974), in an article unlikely to be run across by many biologists (a multi-disciplinary Berkeley symposium volume dedicated to a celebration of Copernicus), characterizes the inseparability of contributions of heredity and environment in a passage that could have been written by Niels Bohr. Lewontin claims that the epigenetic process destroys the separability of the “factors” of heredity and environment in terms sounding like Bohr’s account of how the uncontrollable disturbance caused by the interaction of measuring instrument and subatomic entity leads to the inseparability of the effects of either. Thus development plays the role that measurement does for Bohr. Pushing this analogy between positions in subatomic physics and nature/nurture yet further, Gould’s position on genetic possibility parallels that of the middle or later W. Heisenberg. For Heisenberg quantum states are potentials that only collapse into actuality of particular events upon measurement. Appropriately, Dawkins’ genetic determinism resembles, contra- Gould and Lewontin, Bohm’s deterministic alternative to Bohr and Heisenberg. Like Bohm (and Newton for that matter), Dawkins, in order to save an atomistic, deterministic view, posits a kind of action at a distance (“genetic action at a distance”) in The Extended Phenotype.


Antonovics, J., Ellstrand, N. C. and Brandon, R. N. (1988) Genetic Variation and Environmental Variation: Expectations and Experiments. In Plant Evolutionary Biology, eds. L. D. Gottlieb and S. K. Jain, London: Chapman and Hall.

Bradie, M. (2002). Review of Thinking About Evolution: Historical, Philosophical, and Political Perspectives edited by Rama S. Singh, Costas B. Krimbas, Diane B. Paul, and John Beatty. Human Nature Review. 2: 257-263.

Brandon, R. N. (1990) Adaptation and Environment, Princeton University Press.

Brandon, R. N. (1996) Concepts and Methods in Evolutionary Biology, Cambridge University Press.

Brandon, R. N. (1994) Antonovics, J., Burian, R. et al, Discussion: Sober on Brandon on Screening Off and the Levels of Selection. Philosophy of Science 61: 475-86.

Brandon, R. N. (2001) Philosophy of Selection: Units and Levels.

Dawkins, R. (1982) The Extended Phenotype New York: Oxford University Press.

Dawkins, R. (1998) Unweaving the Rainbow Boston: Houghton Mifflin.

Godfrey Smith, P. (1992) Additivity and the Units of Selection PSA 1992, vol. 1, Philosophy of Science Association, East Lansing, Michigan, pp. 315-328.

Godfrey Smith, P., and R. Lewontin, (1993) The Dimensions of Selection Philosophy of Science 60: 373-395.

Laland, K. N., Odling-Smee, F. J., and Feldman, M. W. (1996) The Evolutionary Consequences of Niche Construction: A theoretical Investigation Using Two-Locus Theory, Journal of Evolutionary Biology 9: 293-316.

Lewontin, R. C. (1974) Darwin and Mendel-the Materialist Revolution, The Heritage of Copernicus, Theories More Pleasing to the Mind’ Jerzy Neyman, ed., Cambridge MA, The MIT Press pp. 166-183.

Lewontin, R. C. (1991) [review of] The Structure and Confirmation of Evolutionary Theory, Biology and Philosophy, 6: 461-66.

Lloyd, E. A. (1989) A Structural Approach to the Units of Selection, Philosophy of Science 56: 395-418.

Singh, R. S., Krimbas, K., Paul, D. B., and Beatty, J. (2001) eds., Thinking About Evolution: Historical, Philosophical and Political Perspectives Cambridge University Press.

Sober, E. (1992) Screening-off and the Units of Selection. Philosophy of Science 59: 142-52.

Williams, G. C. (1992) Natural Selection: Domains, Levels and Challenges Oxford: Oxford University Press.

Buy Thinking About Evolution: Historical, Philosophical, and Political Perspectives from Amazon United States of America Amazon.com  Amazon United Kingdom Amazon.co.uk  Amazon France Amazon.fr  Amazon Deutschland Amazon.de  Amazon Japan Amazon.co.jp Amazon Canada Amazon.ca

Computer-generated translation of this page French français German deutsch Spanish español Portuguese português Italian italiano Russian Russian JapaneseJapanese Chinese (Traditional) Chinese (Traditional)Arabic Arabic― also try this alternative fast translation service.

© Val Dusek.

Professor Dusek teaches Philosophy of Science; Science, Technology and Society; Eastern Philosophy; Logic; and the Humanities. His primary research interests are in the philosophy of biology (sociobiology, biotechnology); philosophy of the history of science; social epistemology of science; alternative approaches to science (Romantic, Hermetic, Chinese); and Marxism. He has published articles on Peirce, on the Sociobiology debate, and on Bukharin and has written a book, Holistic Inspirations of Physics: The Underground History of Electromagnetic Theory (Rutgers University Press, 1999). Recently he has written on the "Science Wars" dispute between physicists and the sociological and literary analysts of science.


Dusek, V. (2002). Lewontin’s Living Legacy: Levels of Selection and Organismic Construction of the Environment. Human Nature Review. 2: 367-374.

US -

Amazon.com logo

UK -

Amazon.co.uk logo

The Human Nature Review