Linnda R. Caporael (1996) Repeated Assembly. Psycoloquy: 7(04) Group Selection (6)

Volume: 7 (next, prev) Issue: 04 (next, prev) Article: 6 (next prev first) Alternate versions: ASCII Summary
PSYCOLOQUY (ISSN 1055-0143) is sponsored by the American Psychological Association (APA).
Psycoloquy 7(04): Repeated Assembly

Reply to Harris on Group-Selection

Linnda R. Caporael
Department of Science and Technology Studies
Rensselaer Polytechnic Institute
Troy, NY 12180


Internal, essential nature contrasted with external, molding forces has had an irresistible grip on the imagination, one long preceding modern biology. Add to that the impoverished language for generalizing about the contingent interactions of genetic and epigenetic factors in the formation of phenotype, and difficulties are bound to arise. Repeated assembly is a vocabulary for replacing the Scylla and Charybdis of nature-nurture dualism.


developmental systems theory, group coordination, group selection, hierarchy, human evolution, social cognition, social identity, teleofunctionalism


1. Harris's (1995b) commentary provides an opportunity to clarify the rationale for introducing repeated assembly. Repeated assembly is a description. It was needed because the standard description of evolution, changes in allele frequencies in populations, provides virtually no toehold for establishing an evolutionary framework in the human sciences. While it is quite true, as Harris notes, that sophisticated audiences know that natural selection does not imply an agent, the lack of an adequate language forces equally sophisticated writers into borrowing awkward metaphors: highly respected scientists "search for the soul," (Crick, 1993); reveal God's utility functions (Dawkins, 1995, November), and argue that that people can rebel against their genes (Williams, 1989). This language simply will not do. It implies that genes are like souls, the essence of individuals, somehow more "real" than the ephemeral body or phenotype (vessel of the soul and vehicle of the genes, respectively). Such metaphors cannot be translated into concepts useful in the human sciences.

2. Genes and environment interact in a developmental system. Developmental systems are hierarchically-organized; their outcomes are contingent on prior states of the system, and they emerge from constructive interactions among multiple inherited resources including genes, cellular machinery, social resources, and the reliable presence of the atmosphere (Griffiths & Gray, 1994; Oyama, 1989). To encompass organized, supra-individual entities (e.g., groups), a broader term was necessary; hence, I used the term "repeated assembly." Thus, (human) phenotypic nature is repeatedly assembled, generation-to-generation, and interacts with the repeated assembly of different group structures.

3. Harris wants repeated assembly to provide convincing explanations of existing data and make specific predictions. However, repeated assembly is not a theory. It is more a vocabulary; it enables theoretical construction. Before turning to that topic, we need to consider Harris's other main objection: that repeated assembly "throws away useful information." This charge must be evaluated with respect to nature-nurture dualism. If genes are the "essence" of behavior, modified by the environment, then a vocabulary of repeated assembly couldn't be particularly helpful.


4. Harris is concerned that I will throw out the baby of nature-nurture dualism (i.e., genetically determined traits modified by context) with the bathwater, which she does acknowledge is by now very dirty. She believes "the baby is salvageable." I don't believe there's a baby in the water.

5. Harris correctly reports my assertion that the phenotype cannot be analyzed into separate components. She responds that this claim "is inaccurate. Identical twins have exactly the same genes; therefore any differences between them must be due entirely to environmental influences" (para. 12). I disagree. Consider the hypothetical case of identical twins who have a gene that makes them sensitive to a rare chemical: exposure to the chemical has large phenotypic effects later in development. One twin is exposed; the other is not. Later the phenotypic effect shows up in the exposed twin, and the unexposed twin remains unaffected. The gene had to be present for the chemical to have an effect; similarly, the chemical had to be present for the gene to have an effect. Is it true that 100% of the phenotypic variance is genetic? Yes. And it is equally true that 100% of the phenotypic variance is environmental. Given that relevant epigenetic factors enable the activity of genes, it makes little sense to declare genes or environments as "determinants" of behavior. The "determinants" are the developmental process itself.

6. Harris fails to distinguish between phenotypic causes in development and the completely different matter of phenotypic variance. Phenotypic variance, unlike phenotypes, can be partitioned. The partition, however, is not between genes and environments, but rather between shared and unshared resources in particular local populations. The analysis of variance does not tell us "how much" of a trait is "genetic" for an individual, nor does it indicate resistance to intervention, presence at birth, or context-independence as Harris suggests. The finding that growing up in the same home "does not make siblings more alike" cannot be interpreted as a causal statement or as a description of human universal. We would not expect an infant adopted from China into an American family to grow into a child more alike in personality to her siblings remaining in China than to the siblings with whom she grows up in the United States.


7. Harris objects that the concept of repeated assemblies is too elastic. The elasticity is intentional. Repeated assembly is a vocabulary. It is supposed to describe what we see and say what we mean. The vocabulary can be used at different levels of analysis, facilitating the investigation of fit between levels and the multiple possibilities for conflict and synergism between levels, including upward and downward causal processes (Buss, 1987). The vocabulary allows us to make predictions. Separating levels allows us to attend to their integration, particularly in the way the products of development at one stage and context can be integrated and used in other stages and contexts. For example, I proposed that microcoordination developed through interactions in infancy is also adapted for sex and plays a role in tool-use. Harris regarded it far-fetched that the "mechanism that enable people to use a hammer is the same one that enables them to please a lover." Yet that appears to be the case: extended observational research in singles bars indicates that movement synchrony is a critical feature in the mutual assessment of potential sexual partners (Perper, 1985). There is, of course, more to using a hammer and pleasing a lover than just microcoordination, but the ability is critical for both activities.

8. A major advantage of the vocabulary of repeated assembly is the recognition that heritable changes in either genetic or environmental interactants can be a source of phenotypic variability, and hence can result in selection. Harris writes that I consider variability simply as noise and that I regard differences among people as "not real because they are context-specific." Quite the contrary-what could be more real than context-specific behavior? I wrote that the coordinated relations between people and their environment are concretely situated and available to empirical investigation. The almost infinite variability of people's situated activity is the sine qua non of evolutionary processes.

9. The values, beliefs and attitudes that seem to give a society its apparent integrity, I called unsituated coordination. It is a property of culture, not, as Harris writes, something humans use to facilitate communication. What gives culture its apparent integrity? I hypothesized that evaluative talk and artifacts realize cultural and subcultural values. They (talk and artifacts) entrain and constrain behavior, providing links between unsituated coordination and the situated activity of face-to-face interaction (Caporael & Baron, in press). Harris chides me for being "to quick to conclude" that the use of artifacts in the home have important effects on children, but in my example, the use of artifacts are not "parental effects"; they are a link in the concrete realization of cultural patterns. Without such concrete links, the relationships between individual and culture are at best ephemeral.

10. At its simplest, repeated assembly allows researchers in the human social, behavioral and cognitive sciences to refer to a common observation -- some things involving biological systems reliably recur -- without having to make a priori commitments as to whether such repeated occurrences should be deemed adaptations, traditions, habits or some combination of all three. Repeated assemblies are, by definition, cyclic; that is, they have their own generation times, which may be the same as, shorter, or longer than human generational times (see below). Repeated assemblies can be characterized in terms of their durability through time -- evolutionary, historical or individual lifetimes, for example. They may also be characterized in terms of their reliability of replication: my argument that some group configurations should be considered "core" is based on considerations of both durability (reasons to believe such configurations are repeatedly assembled throughout evolutionary history) and reliability (they recur pretty much the same way and in an orderly developmental sequence).


11. Oyama (1985) analyzes the work of several researchers who appeal to nature-nurture dualism; yet, at the same time, the details of their research contradict their general claims. Harris's (1995a) own work falls into this mode: despite her defense of nature-nurture dualism, she provides an excellent illustration of many concepts and advantages of repeated assembly. Her basic proposition is that socialization and personality formation occur in a series of peer-group cultures outside the home. Social identity processes, which have recurred throughout evolutionary time, interact with shorter duration historical trends (the privatization of the home about 400 years ago), cultural variation, and local ecology. Thus, the social category, "our family" is salient in Asian cultures, but much less so in Western ones, which tend to emphasize individuals more than groups. As a result, within-family differentiation predominates over group identity in Western families, and peer-groups become the major arena for processes involved in group identity and individual differentiation. (Cultural values, however, are still situationally contingent, as when the social category "our family" becomes salient for the American family traveling in unfamiliar territory.)

12. Harris's theory invokes a series of repeatedly assembled children's cultures. They are durable, with some elements traceable to Roman times; they are cyclic and reliable, "capable of remaining relatively unchanged while cohort after cohort of children pass through [them]" (1995a, p. 470). Details of the assembly of children's cultures are also contingent upon local conditions. In high-density populations, where there are enough children to form separate groups, salient category distinctions pivot on gender. In low density populations, such as among hunter-gatherers, where there are not enough children to form separate groups, the salient group category distinctions are children versus adults. In her work, Harris also refers to artifacts (but not as a class of relevant phenomena) and their role as a link in the concrete realization of cultural patterns (parents' peer cultures). Her example is fork and spoon rather than a weighted drinking cup, which I used.

13. Although Harris (1995a) says her theory is about the environmental modification of an innate substratum, in fact, she delivers a finely textured description of phenotypic development, where nature is not innate, but constructed; where phenotypic nature -- its form and function -- is not prefigured by the genes and superficially influenced by environment, but rather is a transient and shifting "product of the processes of the developmental interactions we call nurture" (Oyama, 1985, p. 5). These engage multiple resources that vary over temporal scale, group-size, ecology and culture. The language of repeated assembly would have served Harris better than nature-nurture dualism.


Buss, L. W. (1987). The evolution of individuality. Princeton: Princeton University Press.

Caporael, L. R. (1995). Sociality: Coordinating Bodies, Minds and Groups. PSYCOLOQUY 6(1)

Caporael, L. R. & Baron, R. M. (in press). Groups as the mind's natural environment. In J. Simpson & D. Kenrick (Eds.), Evolutionary social psychology . Hillsdale: NJ: Lawrence Erlbaum.

Crick, F. (1993). The astonishing hypothesis: The scientific search for the soul. New York: Scribner.

Dawkins, R. (1995, November). God's utility function. Scientific American, 273, 80-85.

Griffiths, P. E. & Gray, R. D. (1994). Developmental systems and evolutionary explanation. Journal of Philosophy, 91, 277-304.

Harris, J. R. (1995a). Where is the child's environment? A group socialization theory of development. Psychological Review, 102, 458-489.

Harris, J. R. (1995b). Individual Differences Within Human Groups. PSYCOLOQUY 6(40)

Oyama, S. (1985). The ontogeny of information. New York: Cambridge University Press.

Oyama, S. (1989). Ontogeny and the central dogma: do we need the concept of genetic programming in order to have an evolutionary perspective? In M. R. Gunnar & E. Thelen (Eds.), Systems and development. The Minnesota Symposia on child psychology. Vol. 22 (pp. 1-34). Hillsdale, New Jersey: Lawrence Erlbaum.

Perper, T. (1985). Sex signals: The biology of love. Philadelphia: ISI Press.

Williams, G. C. (1989). A sociobiological expansion of Evolution and Ethics. In J. Paradis & G. C. Williams (Eds.), Evolution and ethics (pp. 179-214). Princeton: Princeton University Press.

Volume: 7 (next, prev) Issue: 04 (next, prev) Article: 6 (next prev first) Alternate versions: ASCII Summary