Sociality: Psycoloquy Call for Commentary (1038 lines)

Stevan Harnad (harnad@ECS.SOTON.AC.UK)
Wed, 22 Feb 1995 09:30:59 GMT

The target article on Sociality that follows below has just been
published in PSYCOLOQUY, refereed electronic journal of articles,
commentaries and responses sponsored by the American Psychological

Qualified professional biobehavioral, neural or cognitive scientists
are hereby invited to submit Open Peer Commentary on this article.
Please write for Instructions if you are not familiar with
PSYCOLOQUY format and acceptance criteria (all submissions are
refereed). The address for submitting articles and requesting
information is

The ABSTRACT includes the author's Rationale for Soliciting Commentary.

All past PSYCOLOQUY articles, commentaries and responses are publicly
accessible by ftp, gopher and World-Wide-Web at the URLs below.

Stevan Harnad

Department of Psychology
University of Southampton
Highfield, Southampton
phone: +44 703 594-583
fax: +44 703 593-281

Universal Resource Locators (URLs) for the Psycoloquy Archives:

----------------------------------------------------------------------- Monday 20 February 1995
ISSN 1055-0143 (51 pars, 1 table, 1 note, 44 refs, 999 lines)
PSYCOLOQUY is sponsored by the American Psychological Association (APA)
Copyright 1995 Linnda R. Caporael


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

ABSTRACT: Human interaction, as opposed to aggregation, occurs in
face-to-face groups. "Sociality theory" proposes that such groups
have a nested, hierarchical structure, consisting of a few basic
variations, or "core configurations." These function in the
coordination of human behavior, and are repeatedly assembled,
generation to generation, in human ontogeny, and in daily life. If
face-to-face groups are "the mind's natural environment," then we
should expect human mental systems to correlate with core
configurations. Features of groups that recur across generations
could provide a descriptive paradigm for testable and non-intuitive
evolutionary hypotheses about social and cognitive processes. This
target article sketches three major topics in sociality theory,
roughly corresponding to the interests of biologists,
psychologists, and social scientists. These are (1) a multiple
levels-of-selection view of Darwinism, part group selectionism,
part developmental systems theory; (2) structural and psychological
features of repeatedly assembled, concretely situated face-to-face
coordination; and (3) superordinate, "unsituated" coordination at
the level of large-scale societies. Sociality theory predicts a
tension, perhaps unresolvable, between the social construction of
knowledge, which facilitates coordination within groups, and the
negotiation of the habitat, which requires some correspondence with
contingencies in specific situations. This tension is relevant to
ongoing debates about scientific realism, constructivism, and
relativism in the philosophy and sociology of knowledge.

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


1. Most behavioral and social sciences assume human sociality is a
by-product of individualism. Briefly put, individuals are fundamentally
self-interested; "social" refers to the exchange of costs and benefits
in the pursuit of outcomes of purely personal value, and "society" is
the aggregate of individuals in pursuit of their respective
self-interests. To this view of "economic man," which long pre-dated
Darwin, sociobiology added the idea that individual advantage could be
measured in the currency of genes. In theories stressing the importance
of group living, conspecifics are viewed largely as a class of objects,
more unpredictable than others, but requiring substantial intelligence
on the part of the actor to use these "social objects" to achieve
genetic ends through alliances, manipulation, or exploitation (Byrne &
Whiten, 1988). In contrast, by "social," I refer to a structural
continuum of obligate interdependence, without which individual
prospects for reproduction and survival to reproductive age are

2. Humans are obligately interdependent, not only for acquiring their
daily bread, but also for the daily operation of their minds. When Adam
Smith proposed the "invisible hand" of self-interest, he took for
granted that the butcher, the brewer, the baker -- and their families
who worked in the business -- were organized in workshops, which were
in guilds, which were in villages, which were in districts, which had
seasonal fairs and religious celebrations. In the 18th century, the
skills for butchering, brewing, and baking were accumulated through
generations, passed from adult to child, and repeated in daily, weekly
and seasonal cycles of activity. Butchering, brewing and baking
demanded finely tuned sensory and motor coordination; familiarity with
variable materials, tools and methods; a marketplace, of course; and
coordination among these physical, mental and social components.
Today, telecommunications and transportation technologies expose the
significance of this coordination. People still organize themselves in
groups, but some of these no longer need to be face-to-face groups,
constrained by space and time in a nested hierarchy of guild, village,
district, etc.

3. Hull (1988) described a nested hierarchical organization in science
similar to the one in village life; a "demic structure" composed of
small research groups, "conceptual demes," and seasonal society
meetings. This description accords very well with psychological
research on the nested, hierarchical structure of social identity
(Turner, 1987). I propose that the demic structure of science described
by Hull is more general, and it is paralleled by a "demic structure" of
mind.[1] The parallelism suggests that the "mind's natural
environment" can be more adequately specified in terms of functional
organizational structure than it can be by invoking inclusive fitness
theory and "life in the Pleistocene" (Tooby & Cosmides, 1992). Given a
descriptive paradigm of the human evolutionary environment, it should
be possible to derive testable, nonobvious hypotheses about human
mental systems.

4. Before we can consider the "invisible hand" of demic structure,
however, we need an evolutionary framework that relieves us of the
individualistic assumptions built into the gene-centered view. The
traditional single-level, gene-centered evolutionary analysis, based on
genetic self-sacrifice, inclusive fitness, or number of offspring, does
not lend itself to discussions of hierarchical organization for social
structure or mental systems.


5. If evolution is conceived merely as changes in population gene
frequencies, then the only factors of evolutionary interest will be
those passed on in the zygotic DNA. Accordingly, environmental factors
are, at best, secondary to evolutionary change, and development is a
side issue (Oyama, 1992). For the study of most non-human organisms,
this view has been useful, at least to the extent that a high
correspondence between genotype and phenotype could be assumed. In the
human case, that assumption has been much more troublesome. For genes
to be of any interest in the study of mind and behavior, they must be
conceived as much more than a step in protein production. They are
"for" a useful attribute, as in "genes for" kin altruism or "genes for"
rape, mathematical ability, mate preferences, or cognitive mechanisms
specialized for producing the behavior. Although the environment is
secondary, it may fill in gaps in "open programs," "shape" innate
proclivities, or "shunt" behavior from one option to another. This is a
vague gene-environment interactionism that privileges the gene and
invites the useless imagery of the habitat as the environment for DNA.
There are several problems with this version of evolution, not least of
which is the difficulty of demonstrating genetic specificity for particular
mechanisms, or reproductive variation corresponding to particular
genotypic variation. If the appeal to genes is interpreted
metaphorically, it presupposes aspects of behavior that are matters of
empirical investigation: the universality of behavior, the difficulty
of changing behavior, and the fixity of its form (Oyama, 1985).

6. The restructuring I propose draws on three strands of thought:
hierarchical views of evolutionary phenomena, which locate genes in
their appropriate environment -- the cell; developmental systems
theory, which emphasizes the contextual and contingent events in the
flux living phenomena, and teleofunctionalism or "proper functions,"
which offers a historical definition of function and a taxonomy for the
classification of function. The result is a notion of "repeated
assembly," where nature-nurture and biology culture distinctions are
eliminated. They are replaced with questions about the reliability and
durability of replication rather than conflicts about "how much" can be
attributed to genetic or environmental determinism.


7. A hierarchical, or expanded, evolutionary theory (Brandon, 1990;
Buss, 1987; Campbell, 1974; Wilson and Sober, in press) recasts
Darwinism in a form that can accommodate the human case without
resorting to nature-nurture dualism. The nested, hierarchical
organization of life allows for selection on multiple levels of
organization -- macromolecules, genes, cell lineages, individuals, and
on occasion, groups. Each of these entities is situated in its specific
selective environment. Entities at one level can function as
"environment" or context for entities at another level. Thus, the
"gene-environment" interaction occurs between DNA and the local
cellular machinery. Genetic variants are sorted in differential births
and deaths, which are a function of organism-environment interactions;
they are selected in the context of cellular machinery. The
persistence, or continued replication, of lower level units is
crucially dependent on the maintenance of the organized unit
interfacing with the habitat. Hence, evolutionary analysis, distinct
from other kinds of biological analysis, focuses on the interfaces
between levels -- the coordination or lack thereof among levels of

8. The hierarchical perspective re-orients us in two important
respects. First, the focus is on functional relations between entity
and environment, rather than just traits of the organism.
Distinguishing between traits and functions is significant for adaptive
explanation generally (Griffiths, 1992). In the human case
specifically, the distinction allows that a functional relationship may
be achieved through others or even artificially. For example, humans
can have disabilities that in other species would be terminal, but
humans may use "artificial traits" (e.g., wheelchairs, canes,
eyeglasses) to achieve basic functional relationships such as
locomotion or vision. The results of natural selection are not "traits"
in the usual sense of static features possessed by an organism, but
relational linkages between organism and environment. Second, these
relations are concretely situated in specific contexts, available to
empirical investigation. For example, there is no basis for assuming,
as the trait view does, that an individual who may act socially
dominant in a dyad will also be socially dominant in a group of five
people working on a shared problem.


9. Social scientists are uncomfortable with the determinism suggested
by genecentric evolutionary thought. For example, Crawford (1989), a
committed evolutionist, reassured his colleagues in psychology that
they need not believe that natural selection is currently occurring to
use an evolutionary approach; they need only assume that evolution
occurred in the past. Evolution, this view suggests, has stopped.
Ideally, we would like to restate human evolutionary theory in a form
that satisfies the general constraints of Darwinism (Endler, 1986;
Brandon, 1990), facilitates "matches" between evolutionary theory and
human psychological and social science data, and does not lend itself
to mindlessly legitimating existing social and political inequities
(Caporael & Brewer, 1991).

10. Contra Crawford (1989), organisms are the continuation of
evolutionary processes rather than the result of them. This view
allows relations between entity and contexts, over the lifespan and
over generations, to function as both cause and effect over time (cf.
Brandon, 1990; Gray, 1992; Levins & Lewontin, 1985; Oyama, 1985). We
are no longer talking in the billiard ball universe of linear causes
where we can imagine that genes cause or predispose traits, which are
then modified by the environment. The analogy is closer to a shared
language (relation) where (entity) speakers' utterances are both caused
by and cause other (environment) speakers' utterances. Just as
speakers are more interested in the "fit" between their utterances
(rather than, say, the specific language they are using), so is
our interest in the "fit" between entities and environments that
feature in evolutionary processes (i.e., some relations persist, others

11. "Repeated assemblies" are recurrent entity-environment relations
composed of hierarchically organized, heterogeneous components having
differing frequencies and scales of replication. This mouthful
definition needs to be unpacked in pieces. To start, a mammalian
zygote, for example, is the repeated assembly of two sources of DNA,
centrosomes from the sperm, and other components, all of which, if in
the right place, at the right time, automatically result in a zygote.
There is no "genetic program" that directs or controls the assembly of
the zygote. Instead, the analogy is to a chemical reaction, where,
given a set of constituents, against a background of enabling
conditions, an explosion, a precipitate or a vapor results -- or
repeatedly assembles.

12. At the level of the organism, genes will always be among the
multiple, heterogeneous resources of a repeated assembly (the
organism), but they are inert without the epigenetic components at
their own and other levels. Genetic action is dependent on the reliable
recurrence of appropriate contexts, from cellular machinery to social
events to constancies of atmosphere (Griffiths & Gray, 1994). Neither
genes nor environment are privileged (although the microbiologist would
still focus on genes and the educational psychologist might focus on
classrooms). Some resources will themselves be repeated assemblies:
ideas, customs, artifacts, learned skills, languages, and group
configurations can all be repeatedly assembled.

13. Some repeated assemblies have, in Millikan's (1984) sense, "proper
functions." That is, a relation repeatedly assembles ("is selected")
because in the past (see below) the co-occurrence of organismic and
environmental components in some particular relation (e.g., timing,
frequency, contingency) contributed to the persistence of the assembly
contingent on particulars of setting. Repeated assemblies are all but
infinitely variable because they are concretely situated. No two
digestive tracts are identical; sometimes there is a difference that
makes a difference (e.g., lactose intolerance). No two breakfasts are
ever exactly alike, but breakfast is repeatedly assembled; not by
everyone, not every where in the world, but often enough to be
extracted and named as a category for human minds. Change in any
component, genetic or epigenetic, can alter an assembly, increasing or
decreasing the likelihood of its future assembly, or having no
particular effect at all. "In the past" may refer to time measured in
terms of ontogeny (in "habits of a lifetime"), to cultural-historical
time ("2,000 years of Western civilization"), or to geological-
evolutionary time. Thus, assemblies may persist in evolutionary time,
among Americans, in the Smith family, or in J. Smith.

14. The proper function of a repeated assembly is not necessarily the
normative outcome. Most sperm never fulfill their proper function;
most acorns become humus. Some repeated assemblies have no function;
they result from illusory contingencies, such as the superstitious
circling of pigeons in a Skinner box. Other repeated assemblies may
have had a function at one time, but become dissociated from it.
Langer (1989) tells the story of a friend who would always cut a slice
off the end of a roast before placing it in the oven. To make a long
story short, the colleague said she did this because her mother did it;
mother said she did it because her mother did it; grandmother shed
light on the practice: the pan she used for roasting was too small to
accommodate the whole piece of meat. Such dissociations may occur in
evolutionary time, giving rise to vestiges, and in historical-cultural
time, giving rise to customs that have no function or even incur costs
(Boyd & Richerson, 1985).

15. Sources of heritability (i.e., acquired or innate traits) have no
a priori significance in descriptions of repeated assemblies.
Paradoxically, however, the notion of heritability is expanded. As
humans, we inherit not only genes, but also attitudes, practices,
place, nationality, expectations for behavior, and so forth. There may
be change within lifetimes, but this simply indicates low reliability of
repetition. Heritability itself is thus no indicator of the
universality, fixity, or difficulty of changing organism-environment

16. The difference between the gene-centered view and the expanded
heritability of repeated assembly can be illustrated. Tooby and
Cosmides (1992) argue, as do developmental system theorists (e.g.,
Levins & Lewontin, 1985; Oyama, 1985), that phenotypic features are
fully codetermined by genes and environment in development throughout
the lifespan; that the phenotype cannot be analyzed into separate
genetically determined and environmentally determined components; that
events at one stage of development are contingent on the products of
prior developmental outcomes, and that both genes and developmentally
relevant environments are outcomes of evolutionary processes. Yet,
Tooby and Cosmides (1992) also assert that developmental programs in
the genes create the relation between environmental conditions and
developmental outcomes. Thus, a child raised in social isolation may
not acquire language, but nevertheless will have a species-typical
language acquisition device.

17. If we view language as a repeated assembly, however, we could allow
that such a child might have genes for a language acquisition device,
but lacking a critical component, a language environment, that device
was never assembled. Both the genes and the language environment must
be inherited. English speaking and Kikuyu speaking differ because the
various elements have different cycles of repetition. One set, which
includes genes, has a longer cycle of repetition relative to another
set, which includes the language environment. Both sets are repeatedly
assembled, but on different scales of time. The cycle of language
environment (English or Kikuyu), in cultural-historical time, is nested
within the cycle of other components, including genes, in evolutionary
time. Although the frequency and scale of the cycles may be roughly
distinguished, it makes little sense to separate language into an
innate and an acquired component; both are parts of inherited
resources. Nested within these two cycles of evolutionary time and
historical-cultural time, is a third one, the sine qua non of variation
and evolutionary process: the situated relation of "organism-in-
setting," (e.g., a child raised in social isolation) where the
conditions for an assembly to be repeated, be it later in the day or in
succeeding generations, take place or fail to take place.

18. In the next section, we consider the relations humans repeatedly
assemble in evolutionary time. These form the contexts for repeated
assemblies of shorter cycle lengths in historical and lifespan time and
the basis for superordinate coordination.


19. Human evolutionary theorists agree that group living was a critical
feature in human evolution. They disagree about the causes and
consequences. In my view, groups functioned and continue to function as
an interface between individual and habitat. Although this is a
substantial assumption, it does allow us to unburden ourselves from a
host of lesser assumptions about the past and to construct a
"minimalist scenario." We assume there would have been both minimum and
maximum constraints on group size: too small a group would have a
higher risk of perishing; too large a group strains the carrying
capacity of the environment. The "envelope" for selection for sociality
is thus a function of the physical parameters of the species morphology
and ecology. The combination of minimum and maximum group size
constraints has several important theoretical consequences.

20. First, to the extent that exploiting a habitat is more successful
as a collective group process than as an individual process, not only
would more successful groups persist, but so also would individuals
better adapted to group-living. Because a group mediates individual
contact with the environment, and the number of "niches" within groups
is limited, fitness should have been correlated with the evolution of
perceptual, affective and cognitive processes that support the
development and maintenance of group membership (Caporael, Dawes,
Orbell & van de Kragt, 1989; Brewer & Caporael, 1990). One consequence
of these dynamics is a ratcheting, feed-forward evolutionary process
where other constraints, including cognitive ones, constrain group size
(cf. Dunbar, 1993; Wilson & Sober, 1994). We would expect humans
to be obligately interdependent; that is, their prospects for survival
and reproduction outside group contexts would be greatly diminished.

21. Second, "self-interest" at the individual level would be
necessarily, albeit incompletely, constrained by requirements for group
coordination and group fissioning. A corollary is that groups would
reproduce by fissioning, not by dispersing individuals or dyads. Third,
humans have not evolved to apprehend "true beliefs" in the sense
implied by prescriptive rationality, realism, or objective knowledge,
but rather to develop, maintain and negotiate face-to-face group
membership. (Note that I am not saying humans cannot do these things,
but rather that they have not evolved to do them.) Fourth, mental
systems specialized for face-to-face interdependency in evolutionary
time must be "reweavable" for the production of large-scale social
coordination, which has appeared in historical-cultural time; that is,
humans have no evolutionarily specialized adaptations for agricultural
or urban living.


22. Human face-to-face interaction, repeatedly assembled,
generation-to-generation, over evolutionary time, is a continuation of
the nested hierarchy of life's functional organization. I propose four
"core configurations" of situated, face-to-face activity: dyad,
family/workgroup or team, face-to face group or deme, and macroband or
macrodeme. These labels are not intended to represent social roles, or
even necessarily individuals: they represent kinds of interactions. A
dyad is an interaction between two entities, one of which can be non-
human. "Family/workgroup" does not specifically refer to families or
workgroups, but to "small-group, common task orientation" interactions.
Of course, such interactions are frequent in both families and
workgroups, but, as we shall see, are not limited to such social

23. Configurations are "core" because they are repeatedly assembled in
hunter-gatherer groups (Birdsell, 1972; Jarvenpa and Brumbach, 1988) --
presumably in evolutionary time, but also in human ontogenetic
sequences, and in day-to-day activity. Each configuration is associated
with a group size, which should be considered more like a center of
gravity than a fixed condition, and "modal tasks" for interacting with
the environment. The general idea is that size/task configurations are
stable and repeated generation-to-generation in evolutionary history as
a functional consequence of the physical interaction between species'
morphology and ecology. From a strictly evolutionary historical
perspective, core configurations are affordances for the evolution of
proper functions. That is, a size/task configuration is a specialized
affordance for the repeated assembly and evolution of various proper
functions. An affordance allows a proper function to occur, but there
is no necessity that the proper function will assemble given the

24. Table 1 sketches a model of this structure. (I have used the terms
"deme" and "macrodeme" to indicate a greater generality than the use of
"band" or "macroband" by anthropologists.) The combination of size and
modal task (not just size alone) constitute a specialized affordance
for the evolution and repeated assembly of proper functions.


Core Group Modal Proper
Configuration Size* Task Function
Dyad 2 Sex, infant
interaction with Microcoordination
adults & older

Work/Family 5 Foraging, hunting, Distributed
Group direct interaction cognition
with habitat

Deme (Band) 30 Movement from Shared construction
place to place, of reality
general processing (includes folk
and maintenance, psychology),
work group social identity

Macrodeme 300 Seasonal gathering, Stabilizing &
(Macroband) exchange of standardizing
individuals, language
resources and

* Except for dyads, these numbers should be considered as "centers
of gravity," modal estimates in a range roughly plus or minus a
third of the number.


25. There may be multiple proper functions, but for clarity and
simplicity, one per configuration will do for this discussion. The dyad
affords possibilities for micro-coordination (e.g., facial imitation in
a mother-infant dyad, the automatic adjustment of gait that occurs when
two people walk together); the workgroup affords possibilities for
distributed cognition; the face-to-face group affords a shared,
construction of reality ("common knowledge"), which also mediates
interaction in macrodemes and other superordinate group configurations.
Macrodemes afford the stabilization and standardization of language,
signs and symbols. Higher-level configurations cannot be reduced to
lower-level configurations, but once a proper function is available
from higher-level configurations, it may be usefully employed in
lower-level configurations. For example, there is no reason to propose
that language evolved "for" dyadic interaction. However, once language
has evolved, it can be used in dyadic interactions.

26. Human core configurations are also repeatedly assembled in an
ontogenetic sequence. Infants develop microcoordination in dyads; as
their coordination increases, they participate in workgroups
(families), and through them, face-to-face groups (extended networks of
kin, family friends, etc.). Rather than just independence, human
development is increasing interdependence. There is a broadening of the
range of social interaction accompanied by increasing requirements for
reciprocity, skills, memory, social judgment and so on.


27. Psychological research on humans also suggests a nested
hierarchical organization for social identity. Social identity plays
an important role in ingroup-outgroup relations, the distribution of
resources, self-categorization, and expectations for behavior. It is an
automatic redefinition of "self" in terms of shared group membership
(Brewer, 1991; Turner, 1987). To go back to Adam Smith's period, the
brewer works in a small group for brewing; he is also a member of a
guild, a village, and so on. Each of these may be distinct social
identities, posing different conflicts and synergisms. Social identity
is more complicated for many people in the modern world because they
potentially belong to multiple "lineages" of groups: the modern brewer
may also be a member of an environmental activist organization, which
may be a chapter in a larger organization, which may be part of a
federation of environmentalist groups. Is there a way of bringing
together, from the one side, a vision of evolution as hierarchical,
repeatedly assembled levels of organization, and from the other, a line
of empirical research on the hierarchical organization of human social

28. I propose that the topography for the evolution of human relations
and practice is simultaneously social-organizational and cognitive.
Dynamic psychological shifts in level of identity (which can result
from various conditions including group size, shared fate or outcomes,
and salient group boundaries; Turner, 1987) are hypothesized to
maintain core configurations even when the physical constraints of time
and space are reduced, as for groups of globally dispersed individuals
linked by telecommunications technology.

29. Another way of saying this is that configuration units are
group-level entities having proper functions. Their psychological
correlates, that is, how they are cognitively maintained and repeatedly
assembled, are redefinitions of "self" that minimize distinctions
between self and others (Brewer, 1991). Thus, from one perspective, the
"skin-bounded" organism consists of multiple selves; from another
perspective, the skin bounded organism is "dissolved" into multiple
individuals, that is, units of coordinated activity. Personal identity,
the locus of conscious awareness of goals, plans and beliefs, is the
product of multiple interacting and dynamic selves. These are two
perspectives of the same phenomena (similar to the way that light can
be described as waves or packets depending on one's perspective).

30. This unfamiliar way of thinking about social cognitive phenomena
can be intuitively illustrated: five people in an elevator are an
aggregation, not a group; that is, they are not a core configuration.
The same five people stuck in the elevator between floors are a group
(core configuration), the "glue" being shared fate. When group
identity is salient, personal identity is not. Although we can access
the experience of self at different levels of organization, it is, in
effect, partially encapsulated.

31. Some reviewers complain that by introducing a dual perspective, I
have not completely dashed dualism. That was not my objective: some
minimal dualism is essential because without it, we have no way of
conceptualizing (or perceiving) contrasts. I have relocated the
panoply of dualisms that we tacitly assume exist in nature
(nature-nurture, individual-society, etc.) and put them in the
researcher's project, where they properly belong; have been all the
time; and -- when placed into nature -- have caused no end of confusion
and dispute among realists and relativists.


32. Once a proper function, associated with a core configuration,
evolves, it can be extended to other domains of interaction; that is,
it can become part of another behavioral assembly. For example, dyadic
micro-coordination (involving perception of movement dynamics,
emotional expressiveness, etc., whose proper function probably includes
sex and infant-elder interactions) can be extended to interaction
between animal trainer and animal, and human and artifact (where it
becomes the basis for embodied knowledge, tinkering or highly skilled
performance). Language is one of the most salient examples of
extension, having both upward and downward causal effects, although
these do not seem to be equivalent. Language has considerable impact
on band-size humans, and largely allows modern humans to be members of
multiple macrodemes presumably an innovation in recent history.
Nevertheless, the proper function of macrodemes persists. Various fields
of work, from dog grooming to biochemistry, have specialized languages,
and one purpose of professional meetings is to standardize terms and
references that have no ordinary usage. Though language is still
important for workgroups, these smaller groups rely more than higher
level groups on gesture, visual aids, and "private language" for
telegraphing information. The rhythmic features of language, more than
content, are probably significant in microcoordination.

33. With an appropriate bridge, proper functions specialized for one
level of organization may be evoked at others. The rhythm of a
marching band, for example can generate microcoordination among
hundreds of marchers. We would not expect the evolution of a capacity,
"group-marching ability" to evolve for marching in large groups, as
there is no reason to suppose that either large groups or marching
bands (or an equivalent) were repeatedly assembled until recently in
human history. Technology, crucial for achieving agriculture and
settlement life, also bridges domains of application and the
psychological correlates of coordination, particularly where technology
eliminates the constraints of time and space that have traditionally
characterized human social organization (Caporael, 1987). We "reach
out and touch someone" thousands of miles away by using a telephone.
Radio and television can be used to extend group identity to millions
of people by invoking national category boundaries rather than ethnic,
religious, or professional ones.

34. Specialized affordances may be combined (e.g., a heart surgery
team combines affordances for microcoordination and distributed
cognition). There may be other recognizable configurations for human
activity in addition to core ones, but all human action, and all
selective action at the organism environment level, is situated in
group contexts plus the solitary or individuated state (e.g., being
alone in a crowd). For analytic purposes, affordances can be
decomposed. Piloting an airplane involves highly evolved
microcoordinative abilities, and, counterintuitively, would not be
evolutionarily novel. Getting into an airplane -- sharing a small
contained space with hundreds of strangers -- is evolutionarily novel.


35. David Hull (1988) provides an excellent description of scientific
organization that can be used to illustrate how repeated assembly, core
configurations, and their psychological correlates could be used. He
proposed that scientists competed for citations, priority, and
recognition, but they had to cooperate to achieve their ends. He also
saw that the social organization of science, its demic structure, was
very important. My interest here is to use his excellent observations
to illustrate an analogy between the demic structure of science and
that of traditional societies.

36. Hull (1988) identified a scientific demic structure consisting of
the individual scientist, research groups, "conceptual demes," and
seasonal society meetings. These correspond to the typical organization
among nomadic hunter gatherer groups: a nested hierarchy composed of
work groups, microbands, and seasonal macrobands (Jarvenpa & Brumbach,
1988). To these configurations, we must add the dyad. In both instances
-- in science and among hunter-gatherers -- group-size numbers at these
four levels are fairly constant; dyads, about 3-5 individuals, 30-50
individuals, and 100-500 individuals (Birdsell, 1972; Hull, 1988). In
both cases, isolates have reduced viability, and must be part of a
subgroup with sufficient "critical mass" to persist. As Hull noted,
isolated scientists are not very productive. Also, scientists do not
usually begin a new research group as social isolates; they must bring
some of the old group members with them or attract new recruits.

37. Learning to fashion specialized tools, in stone or other materials,
or to use specialized laboratory or food preparation equipment and
skills is a hands-on, situated activity (like golf or tennis) requiring
feedback from materials, equipment, and an experienced user to develop
finely-tuned sensory-motor microcoordination. The most intense
conceptual interaction with the material environment occurs in
workgroups, the activities of which are influenced by the microband
(conceptual deme) and feed back into it. Small research groups or
intensely social small hunter-gather groups working on the environment
seem to do two things. They distribute cognition; that is, they share
tasks such as perception, classification, inference, memory, and
contextually-cued responses in the conduct of practice and
interpretation of data, or in interaction with uncertain habitat
features. In the process of doing so, they draw on and contribute to a
socially constructed and shared "world-view," which permits groups
(hunter-gatherers or scientists) to interface with nature (cf. Jacobs
and Campbell, 1961; Amann & Knorr-Cetina, 1990).

38. The product of that interface feeds into higher levels in the group
system. Bands and conceptual demes are "staging communities" (Jarvenpa,
1993), which serve as "general processing and maintenance centers" for
information and resources retrieved from the smaller dispersed groups;
they are also loci of shared group identity. For both scientists and
hunter-gatherers, seasonal macroband meetings (or yearly conventions)
are important for the exchange of myths, gossip, and information about
more distant areas and groups. Macrobands are also arenas for
competitive games as well as the affirmation of common worldviews, the
maintenance of languages (hunter-gatherers) and idiolects (scientists),
and the exchange of people (mates, new PhDs, or disgruntled members).
For science, unlike hunter-gatherer groups, the production of articles
in journals may find their way into the popular press and eventually be
incorporated as "common knowledge" (e.g., almost everyone believes in
atoms, germs, and innate sex differences), which remains unsituated
(e.g., atoms) or is realized in interaction (e.g., sex differences).

39. The correspondence between these two demic structures, scientists
and hunter-gatherers, suggests some counterintuitive hypotheses:
divergence in scientific ideas may not lead to fissioning;
fissioning or rather too large a group size may lead to divergence in
scientific ideas (or, alternatively, bureaucratic suppression of
productivity in too large a group). Likewise, the common practice in
academic departments to hire faculty for coverage of a discipline may
come at the cost of having critical mass for effective research


40. The system of meanings and practice implied by terms like "culture"
or "society," must be realized in specific organism environment
interaction, which becomes increasingly elaborated during development.
Superordinate coordination partitions the "blooming, buzzing confusion"
of situated activity into conventionally shared, orderly descriptions.
These in turn influence situated activity. Such interactions result
from, and contribute to, specialized domains of repeatedly assembled
belief, knowledge, and practice, such as those distinguishing between
sexes, generational cohorts, technical specialties and cultures.

41. For example, a young child may be puttering around in the kitchen
"mixing stuff," as children typically do, but a boy's mixing is more
likely to be described by adults as evidence for an incipient interest
in "chemistry" and a girl's as incipient interest in "cooking." For
both children the situated activity is the same. Yet the adults'
description of activity, based on common cultural assumptions, is
different. One child is directed to ponder a future career, the other a
future meal.

42. Coordination occurs to the extent that knowledge and practice
domains overlap or are complementary. I suggest that values serve as a
medium. Humans live in a value-saturated environment; values are known
from interactions with people, natural objects, and artifacts, all of
which interact with social life and customs in complex ways (Caporael,
Panichkul & Harris, 1993; Merchant, 1980).

43. The word "values" tends to trouble psychologists. I am not
suggesting there exists a set of unalterable "shoulds," moral
provisions, or ethics that can be revealed by biology or the
neurosciences. By values, I am referring to the evaluative
(positive/negative) dimension of aggregate social preferences, usually
tacit and invisible, shared by members of a community, diagnostic of
shared social identity, and, to some interesting extent, channelling
behavior in some directions and away from others.

44. Values could not play a coordinative role unless they were both
accessible and automatic. Accessibility allows tacit values to be
exposed and used to negotiate new prescriptive values. Automaticity is
critical; values could not serve coordinating functions if we always
had to contemplate them before acting. We should, therefore, expect a
tradeoff or compromise between values (for coordination) and
experience (of situated activity), which in turn should result in a
discrepancy between actual experience and the description of it.

45. Obviously, humans can describe situated activity, but the
circumstances for doing so are rather rare (e.g., a radio sportscaster
delivering a "real time" description of a baseball game, a field
researcher gathering "real time" observations of animal or human
behavior). Instead, people summarize experience or concatenate it, in
widely shared, conventional ways. These descriptions are not just a
shorthand, expandable to a full description of behavior. People are
often unaware of their actual experience, and instead they describe it
in terms of widely-shared expectations (Nisbett & Ross, 1980).
Descriptions of other people's behavior is often no more precise. Even
a very short time delay can produce descriptions of behavior that
correspond more highly to measures of conventional linguistic meaning
than to actual observed behaviors (Shweder & D'Andrade, 1980).
Expectancy effects can result in self-fulfilling prophecies about
others' behavior (Eccles, Jacobs & Harold, 1990; Rosenthal & Jacobson,
1968), as well as structure access to physical spaces, knowledge and
objects (Caporael, Panichkul & Harris, 1993). Research on eyewitness
testimony, self-description, social judgment, causal attributions, as
well as other social psychological phenomena, suggest that
discrepancies between actual behavior and descriptions of behavior may
be quite common. Because cognition is considered to be fundamentally
"asocial," these phenomena are typically explained in terms of
side-effects of a tradeoff between accuracy and cognitive efficiency
rather than between accuracy and sociality.

46. The discrepancy between situated activities and their conventional
descriptions allows perception and talk to function in the production
of coordinated activity despite the novelty of every living moment. At
least two interacting psychological systems appear to be involved in
superordinate coordination. One is based on folk psychology as
"value-making" talk and the other, following Hodges & Baron (1992), is
based on perception (of objects, people, things) as "value-realizing"
(Hodges & Baron, 1992). Both systems would operate in face-to-face
groups, but easily extend for higher levels of coordination.

47. Psychologists and philosophers have extensively debated the role of
folk psychology as an adequate or heuristically useful description of
human cognition (Greenwood, 1991). However, the prescriptive content of
folk psychology (e.g., the evaluative dimensions of trait terms)
suggests another interpretation. Folk psychological talk about beliefs,
intentions, and desires -- including judgments concerning which are
"natural" and which are not -- develops a context for action that
limits and entrains which actions are conceivable, possible,
desirable, essential, tolerated and forbidden. Folk psychology can serve
for making heuristic predictions of behavior because, to some extent, it
produces the behavior it predicts. The coordination involved is not
simply directing behavior (as in the sense implied by norms), but in
its social cognitive construction.

48. Like folk psychology, perception itself can also tacitly limit the
conceivable, possible, desirable and essential. When presented with a
stimulus, humans are capable of judging whether or not they like it
even if they cannot identify it -- evidence of the primacy of value in
perception (Zajonc, 1980). Moreover, such evaluations correlate with
prior exposure to a stimulus: familiarity breeds liking. Hodges and
Baron (1992) go beyond the notion of values as preference. They argue
that values are criterial and embodied in public, objective
affordances. This view allows that perception can be coordinative on
a large scale. For example, a simple artifact such as a toddler's
spouted cup, is weighted on the bottom to discourage spilling and
encourage drinking (the dual affordances of a cup filled with liquid).
It also connotes a value on independence and self-sufficiency
consistent with a culture with early weaning and training for
independence (Hodges & Baron, 1992). The device, which is only one of
many possible designs for transporting liquids to the mouth,
contributes to the direction of parents' child-rearing activity,
influences the child's development, and reverberates through the
culture, constraining other activities. Most Americans would
disapprove of a woman breast-feeding a two-year old toddler, even in
private, although it is common in nonindustrial cultures and was
throughout human evolutionary history. In perception, "the social"
actively interacts with "the material" through the translation of value
in situated interaction (cf. Pacey, 1983).


49. The descriptive paradigm presented in this article is preliminary;
it is a target to shoot down, a thesis to test. Flaws that are
invisible to me shout for recognition from critics, and much more
empirical and conceptual work remains. My claim is that, in a
fundamental sense, the natural environment of humans has not
dramatically changed. It is maintained by social/cognitive mechanisms
and a structure of activity through which humans "parse" the stimulus
information of complex environments into a small-group "grammar"
throughout the lifespan (Caporael, 1987). The persistence of nested
demic structure, in industrial as well as hunter-gatherer societies,
the existence of novel, cross-cutting "bureaucratic" organization
(typically indifferent to issues of group size and task structure), and
now, the existence of groups unsituated in space and time, linked
through email, fax, phone, and "groupware" provide unprecedented
opportunities for both pure and applied investigations by
evolutionarily-oriented researchers.

50. Although I have stressed sociality, almost as a synonym for
coordination, I am not claiming that human cognition is always a
group-level process: individual cognition is an important source of
variability. Sociality theory most definitely is not a claim that
humans are basically altruistic, prosocial or selfless. I have not
denied that humans can be self-interested, competitive, altruistic,
individualistic or intentional, as the case may be. However, these
terms may be less useful as analytic categories for science than they
are for the critical function of negotiating, coordinating, and
constructing the human social environment (Caporael, 1994). If there is
a distinction, as I have argued, between the "real-time" description of
situated activity and the conveyance of expectation (value) through
unsituated descriptions, there may be a genuine failure to discriminate
between scientific activity as a description of human nature from
scientific activity as the construction of human nature. If this is
the case, then there may be no transcendental "human nature," but only
historical sequences of descriptive paradigms, of which self-interest
and sociality are two, comparable in their generativity more than in
their truth value.

51. Whether repeated assembly is of any value to biologists depends to
a large extent on how useful and appropriable it is for the study of
other species and the development of evolutionary theory. To the extent
that genotype and phenotype can be collapsed into a single unit of
analysis for some research purposes, repeated assembly will likely be
an unnecessarily complicated description of evolution. Still, this
rephrasing seems to pose a challenge -- to recent notions, at least. One
is the interactor/replicator distinction (Hull, 1988), which divides
entities (or their attributes) into those things that interact and
those that replicate. Repeated assembly, coordinated over levels of
organization, implies that repeated interaction is replication, or at
least the source of replication. Similarly, there has been a trend to
divide evolution into two domains of inquiry: the process of evolution
and the results of evolution (Reeve & Sherman, 1993). If organisms (and
other entities) are the continuation of evolutionary processes, than
process and product are indistinguishable; the durability of repeated
assemblies, or repeated assemblies nested in other repeated assemblies,
replaces the process-product distinction.


1. The term "deme" comes from the Greek word, demos, referring to
district, and it is used in biology to refer to a breeding population.
Interestingly, in ancient Athens, there was a law requiring a citizen
to use his deme name rather than his family name. The explicit reason
for this requirement was to foster ties with fellow citizens over kin.


I am indebted to Reuben Baron, Marilynn B. Brewer, P. Thomas Carroll,
Glen Culbertson, Elihu Gerson, Maryanne Garry, Paul Griffiths, Cecilia
M. Heyes, Susan Oyama, Henry Plotkin, Michael G. Shaftoe, and David
Sloan Wilson. This work was made possible by by the National Science
Foundation, Grant No. SBR-9321461. Any opinions, findings, and
conclusions or recommendations expressed in this material are those of
the author and do not necessarily reflect the views of the National
Science Foundation.


Amann, K. & Knorr Cetina, K.(1990) The fixation of visual evidence. In
Lynch, M. (Ed.), Representation in scientific practice, 85-121, MIT

Birdsell, J.B. (1972) An introduction to the new physical anthropology.
Rand McNally.

Boyd, R. & Richerson, P.J. (1985) Culture and the evolutionary process.
University of Chicago Press.

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

Brewer, M. (1991) The social self: on being the same and different at
the same time. Personality and Social Psychology Bulletin 17: 475-482.

Brewer, M.B. & Caporael, L.R. (1990) Selfish genes versus selfish
people: Sociobiology as origin myth. Motivation and Emotion 14:

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

Byrne, R.W. & Whiten, A. (1988) Machiavellian intelligence. Clarendon

Campbell, D.T. (1974) "Downward causation" in hierarchically organized
biological systems. In Ayala, F.D.,& Dobzhansky, T. (Eds.), Studies in
the philosophy of biology, 179-186, Macmillan.

Caporael, L.R. (1987) Homo sapiens, Homo faber, Homo socians:
Technology and the social animal. In Callebaut, W. & Pinxten, R.
(Eds.), Evolutionary epistemology: A multiparadigm program, 233 244,

Caporael, L.R. (1994) Of myth and science: Origin stories and
evolutionary scenarios, Social Science Information 33: 9-23.

Caporael, L.R. & Brewer, M.B. (1991) The quest for human nature,
Journal of Social Issues 47: 1-9.

Caporael, L.R., Dawes, R.M., Orbell, J.M. & van de Kragt, A.J.C.
(1989) Selfishness examined: cooperation in the absence of egoistic
incentives. Behavioral and Brain Sciences 12: 683-739.

Caporael, L.R., Panichkul, E.G. & Harris, D.R. (1993) Tinkering with
gender. In Rothschild, J. (Ed.), Research in Philosophy and Technology.
Vol. 13. Technology and feminism, 73-99, JAI Press.

Crawford, C.B. (1989) The theory of evolution: Of what value to
psychology? Journal of Comparative Psychology 103: 4-22.

Dunbar, R.I.M. (1993) Coevolution of neocortical size, group size and
language in humans. Behavioral and Brain Sciences 16: 681-735.

Eccles, J.S., Jacobs, J.E. & Harold, R.D. (1990) Gender-role
stereotypes, expectancy effects, and parents socialization of gender
differences. Journal of Social Issues 46: 183-201.

Endler, J. (1986) Natural selection in the wild. Princeton: Princeton
University Press.

Gray, R.D. (1992) Death of the gene: Developmental systems strikes back.
In Griffiths, P.E. (Ed.), Trees of Life, 165-209, Kluwer Academic.

Griffiths, P.E. (1992) Adaptive explanation and the concept of a
vestige. In Griffiths, P.E. (Ed.), Trees of Life, 111-131, Kluwer

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

Greenwood, J.D., Ed. (1991) The future of folk psychology. New York,
Cambridge University Press.

Haraway, D. (1989) Primate visions. Routledge.

Hodges, B.H. & Baron, R.M. (1992) Values as constraints on affordances:
Perceiving and acting properly. Journal for the Theory of Social
Behaviour 22: 263-294.

Hull, D.L. (1988) Science as a process. University of Chicago Press.

Jacobs, R.C. & Campbell, D.T. (1961) The perpetuation of an arbitrary
tradition through several generations of a laboratory microculture.
Journal of Abnormal and Social Psychology 62:649 658.

Jarvenpa, R. (1993) Hunter-gatherer sociospatial organization and group
size, Behavioral and Brain Sciences 16: 712.

Jarvenpa, R. & Brumbach, H. (1988) Socio-spatial organization and
decision-making processes: Observations from the Chipewyan. American
Anthropologist 90: 598-618.

Kitcher, P. (1985) Vaulting ambition: sociobiology and the quest for
human nature. MIT Press.

Langer, E.J. (1989) Mindfulness. Addison-Wesley.

Levins, R. & Lewontin, R. (1985) The dialectical biologist. Harvard
University Press.

Merchant, C. (1980) The death of nature. Harper & Row.

Millikan, R.G. (1984) Language, thought, and other biological
categories. MIT Press.

Nisbett, R. & Ross, L. (1980) Human inference: strategies and
shortcomings of social judgment. Prentice-Hall.

Oyama, S. (1985) The ontogeny of information. Cambridge University

Oyama, S. (1992) Ontogeny and phylogeny: A case of metarecapitulation?
In Griffiths, P.E (Ed.), Trees of life, 211-239, Kluwer Academic.

Pacey, A. (1983) The culture of technology. MIT Press.

Reeve, H.K. & Sherman, P.W. (1993) Adaptations and the goals of
evolutionary research. Quarterly Review of Biology 68:1-31.

Rosenthal, R. & Jacobson, L. (1968) Pygmalion in the classroom. Holt,
Rinehart and Winston.

Shweder, R.A. & D'Andrade, R.G. (1980) The systematic distortion
hypothesis. In (ed), Fallible judgment in behavioral research. New
Directions for Methodology of Social and Behavioral Science, No. 4,
37-58, Jossey-Bass.

Tooby, J. & Cosmides, L. (1992) The psychological foundations of
culture. In Barkow, J. H., Cosmides, L. & Tooby, J. (Eds.), The
adapted mind, 19-136, Oxford.

Turner, J.C. (1987) Rediscovering the social group: a self
categorization theory. Basil Blackwell.

Wilson, D.S. & Sober, E. (1994) Reintroducing group selection to
the human behavioral sciences. Behavioral and Brain Sciences, 17:

Zajonc, R.B. (1980) Feeling and thinking: preferences need no
inferences. American Psychologist 35: 151-175.