Society For American Archaeology

Society for American Archaeology
Boundary Conditions for Paleolithic Social Systems: A Simulation Approach

Author(s): H. Martin Wobst
Source: American Antiquity, Vol. 39, No. 2 (Apr., 1974), pp. 147-178
Published by: Society for American Archaeology
Stable URL: http://www.jstor.org/stable/279579 .
Accessed: 10/01/2014 11:02
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
Society for American Archaeology is collaborating with JSTOR to digitize, preserve and extend access to
American Antiquity.
http://www.jstor.org
This content downloaded from 200.54.94.206 on Fri, 10 Jan 2014 11:02:23 AM

All use subject to JSTOR Terms and Conditions
BOUNDARY CONDITIONS FOR PALEOLITHIC SOCIAL SYSTEMS:
A SIMULATION APPROACH
H. MARTINWOBST
ABSTRACT
Pleistocene societies are viewed in this paper as manifestations of band society from which Paleolithic
archaeologists can generate hypotheses about this cultural system and against which they can evaluate the
validity of general cultural explanations. A general model is posited toward the isolation of Paleolithic societies
in the archaeological record. Their numerical size is predicted from Monte Carlo simulations of model
populations, and ways are discussed in which distinguishable equilibrium states of such societies can be predicted
in time and space. Under the assumption of maximal constraints, simulations are used to predict the size and
longevity of social units. The frequency of settlements produced during an equilibrium state of a society is
simulated and the utility of survey for activity areas outside the settlements is demonstrated. The models
developed in this paper are intended to stimulate deductive research in Paleolithic archaeology.
Department of Anthropology
University of Massachusetts
July, 1973
I WAS STIMULATEDinto writing this paper by 3 observationson the state of Paleolithic

archaeologyin 1973:
(1) Paleolithic archaeologistsoften commit errorsin scale, if not in logic, regardlessof whether
they move from the generalto the specific, or extrapolatefrom a particularisticaspectto a general
model.
(2) Particularisticresearchhas reachedits limit in Paleolithicarchaeologysince, no matterhow
well an individualaspect is investigated,analyzed, and explicated, we lack a body of hypotheses
allowingus to integratethe particularresultinto a systemic whole.
(3) Paleolithicresearchcannot proceed to the systemic level without generalmodelswhich have
test implicationsfor the archaeologicalrecord.
Paleolithic archaeology will be able to generate its own coherent and logical frameworkof
hypotheses, predictions,and tests only if it can overcomeits handicaps.It is my contention that a
logical scale for researchat the systemic level can be defined, that previousparticularisticresearch
results can fruitfully be integratedon such a scale, and that it will be practicableand advantageous
to employ the models to be developedhere in future Paleolithic research.My paper receivedits
impetus from work with the EuropeanPaleolithicsequence,althoughthe methodology employed
can be appliedelsewhere.
The ProblemWithina CulturalFramework
The remains of Paleolithic hunters and gatherersare only a part of the total range of human
culture which anthropologistsconsider as their subject matter. It is the aim of the anthropologist
to discover the processeswhich have structuredthe anthropologicaluniverse,and to formulatethe
laws which have governed these processes. If culture is viewed as man's extrasomaticmeans of
adaptation (White 1959), then it is the ultimate aim of anthropology to formulate a body of
generaltheory which explains the behaviorof this adaptiveagent. As in any other field of science,
such a body of theory should be sufficiently generalto be applicableto all temporal and spatial
occurrencesof the processesto be explained,and should lend itself to empiricalverificationwithin
a logical frameworkof hypotheses, predictionsand tests.
147

148 ameRicanantiquity [Vol. 39, No. 2, 1974
In the investigation of cultural systems (for example, band society, ancient civilization), the
researcherneeds a scale of attack that is both analyticallymeaningfuland operationallypractical.
For the purpose of this paper, "cultural system" is dealt with in one of its most inclusive
manifestations:societies (for instance, Walbiri,Birhor). The systemic processes,mechanisms,and
relationshipsby which particularsocieties integrate themselveswithin, and adapt themselvesto,
their naturaland social environmentare viewed as instances of the cultural system in which they
participate. This approach was pioneered by White (1949, 1959) in a general anthropological
frame of reference, while Binford (1962, 1965, 1972) and Wilmsen (1970, 1973) addressed
themselvesspecificallyto the archaeologicaluniverse.
Paleolithic researchderivesits relevancewithin anthropologyfrom the aim of the disciplineto
formulategeneraland verifiableexplanationsfor the system of culture. Any culturaltheory has to
be valid for extinct as well as living culturalsystems in orderto satisfy the condition of generality.
Since the Paleolithic record (at least chronologically) comprises more than 99% of the cultural
universe,this segment has to be includedin any verificationprocedureof culturalexplanations.In
addition, the extinct societies of the Paleolithic form as relevanta basis for generatinginferences
about general cultural processes as do any other living or extinct societies. The Paleolithic
archaeologistshould investigatePleistocene hunters and gatherersin order to generateinferences
about general cultural processes and in order to contribute toward the verification of general
explanations inferred on the basis of post-Pleistocene and modern cultural systems. However,
Paleolithic archaeology can fulfill this role only if it can partition the Pleistocene record into
analytical units appropriate to the development of hypotheses, predictions, and tests about
cultural processes. Thus, Paleolithic archaeology, in developing anthropological goals, must
demonstrateits ability to isolate specific societies as instances of Pleistocene band society. These
societies constitute the sample from which generalcultural theory can be developed and against
which theories of culturecan be tested.
The following section will analyze some previous attempts of central and eastern European
archaeologiststo isolate Pleistocenesocieties and their culturalbehavior.It will document the need
to develop new ways in which Paleolithicarchaeologycan deal with societies as manifestationsof a
culturalsystem.
Some HistoricalConsiderations
Paleolithic research has only gradually reoriented itself from an antiquarianpreoccupation

toward a strongeralliance with the other social sciences. Collection and descriptionexposed the
range of variationin Paleolithic form (Hoernes 1903; Obermaier1912). But the archaeologistwas
still empty-handed:while sociology, geographyand economics were testing and refininga body of
social theory, the Paleolithicuniverseremaineda disjointedcollection of human artifacts.In order
to deal effectively with questions of cultural process, evolution, and ecology posed by its sister
disciplines,Paleolithic archaeologyhad to redefine its subdisciplinewithin the generalframework
of logic and science.
This trend had its beginningand has been most effective in the explication and reconstruction
of the history of individualsettlements. Its groundworkwas laid by centraland easternEuropean
archaeologists in the period between the 2 world wars. In 1919 and 1920 the Austrian
archaeologistJoseph Bayer excavated with meticulous care the first open air settlement structure
in central Europe (Langmannersdorf,see Angeli 1952). At the same time, historicalmaterialism
stimulated Soviet archaeologists to investigate the evolution of the modes of production, the
developmentof the division of labor, and the formationof class societies. The resultinglarge-scale
excavationsof open air settlements in the Ukraine(Cernys 1953, 1961), in Russia (Efimenko
1938, 1953, 1958), and in Siberia(Gerasimov1931; Okladnikov1940, 1941), exposed Paleolithic

Wobst ] PALEOLITHIC SOCIAL SYSTEMS 149
archaeology to classes of archaeologicalevidence such as living structures, portable art, and

ground-workingequipmentwhich speleologicaland francocentricarchaeologyhad neglected.More
importantly, it expanded the archaeologicaluniversefrom artifact classes to human settlements,
and providedPaleolithicarchaeologywith a new level of integrationwithin which the operationof
culturalprocessescould be demonstrated.
In the thirties, Alfred Rust began his well-designed problem-directedexcavations of late
Paleolithic camps in northern Germany (Rust 1937, 1943, 1958). Besides an exemplary
reconstructionof the way of life on these sites, Rust's investigationsdemonstratedthe utility of
systematicsurveysfor Paleolithicsettlementswithin a given period and environmentalniche.
Since then, the pioneers have been succeeded by a second generationof archaeologistswho
have carriedfurther the explication of culturalprocessesat the level of Paleolithicsettlements(for
instance, Bosinski 1969; Klfma 1963; Maniaand Toepfer 1973; Vertes 1964). This progresshas
been achieved through refinementsin the use of inductivelogic, throughadvancesin the auxiliary
sciences, and throughmore inclusive,resourcefuland rigidproblem-directedresearch.
Yet, when the investigationof individualPaleolithicsettlementsis evaluatedagainstthe general
aims of anthropology, our conclusion has to be more critical. There can be no doubt that the
achievement of a higher level of integration for the cultural data has led to the formulationof
more realistic hypotheses about cultural processes and about the culturalsystems within which
they were operating during the Pleistocene. At the same time, however, the concern with
explication at the level of individual settlements has prevented the isolation of analytically
meaningfulunits within which such hypotheses could be tested.
Therefore, the analytical universe of Paleolithic archaeologists is still only a disjointed
collection of sites. For any society, if such could be isolated, the sample is biased toward sites
which are late, close to the surface, and easily accessible to the Paleolithic archaeologist.The
sample also overrepresentssites with good preservationand easily preservedmaterials, that is,
those sites which show a high degreeof culturaluniqueness.
These individualsites do not properly reflect the cultural system of Pleistocene band society.
Instead, they represent only slices of its temporal, formal, and spatial structure.The sites only
reflect selected aspects of the operation of the whole system and can only be correctlyevaluated
once the operation of the system through time and space is understood. The sites were not
excavated to solve problems in the systemic behavior of culture, but ratherbecause they were
available."Explanation"was then providedas a by-productof site excavation,validand verifiable
only within the closed system defined by the aspects of form, time, and space of the individual
site.
The laws of logic do not prevent the generation of hypotheses about a higher order of
integration (society and its cultural behavior, cultural system, culture process) from a biased
sample of a lower order of integration(site universe).Such hypotheses, however,will be distorted
by the bias in the samplefrom which they were induced, and will receiveadditionalnoise through
the expansion in scale of the explanandum.In addition, the logical problemof positinglarge-scale
open (cultural) systems on the basis of the sites which archaeologistsare accustomed to treat as
small-scaleclosed systems has to be overcome. Thus, even though more realistichypotheses about
cultural process can be inferred from a biased sample of Paleolithic settlements than from an
unorderedcollection of Paleolithicartifacts,the a priorivalidity of such hypotheses is bound to be
small. Furthermore, since Paleolithic archaeology concerned with individual sites is not
appropriateto the isolation of the cultural systems of which these sites once formed a part, the
hypotheses still lack an appropriateanalyticalunit within which their test implicationscould be
verified.
Archaeologists have been more successful in discrediting some of the overly simplistic
traditional stereotypes of Paleolithic society. By formulating alternative systemic hypotheses

about cultural processes where simple cause-and-effectexplanations had formerly been judged
sufficient, they stimulated the development of more complex, more realistic, and more explicit
researchdesigns. For example, the central EuropeanarchaeologistsL. Vertes (1955, 1956, 1958)
and F. Prosek (1953) brought argumentsof regional cultural ecology to bear on the perennial
Szeletian question. The malacologistLoZek(1964) also providedmany stimulatinghypotheses in
his discussion of the cultural ecology of the CzechoslovakPaleolithic. Vertes (1964, 1968) later
suggestedthat factors of adaptability,degreeof adaptation(comparealso Freeman 1964), and the
speed of cultural evolution may also be reflectedin the formalcharacterof Paleolithicassemblages.
Klein (1969a) attempted to integrate the material from the Kostienki sites into a functioning
whole, and dealt with the culturalecology of the CrimeanMousterian(1969b).
The destruction of the stereotypes and the rejection of the simpler cause-and-effect
assumptionswere necessary steps toward a higher level of integration. But rather than leading
Paleolithic archaeology to a new analytical universe within which these assumptions could be
tested, the stimulus was absorbedin the explication of behavioron individualsites (Banesz 1968;
Behm-Blancke1956; Brandtner1954; V. Gabori1968; Klima 1963).
A series of programmaticstatements for future researchin given regions of the Old Worldis
more explicitly directed toward the explanation of Pleistocene society and to the presentationof
testable hypotheses about cultural process. Such schemes have been proposed, for example, for
Czechoslovak(Klfma 1967) and Hungarian(M. Gabori 1970) Paleolithic archaeology.They tend
to suffer, however, from a logical shortcoming.They usually assume that the formulationof new
theory and the achievementof higherlevels of integrationare dependenton, and subsequentto, the
discovery of new and "better"data. Thereis no doubt that previousdata collection and previously
tested hypotheses constitute variableswithin the scientific feedback system of theory building.
Yet they are in no way essential to the formulation of any particularhypothesis as long as such a
hypothesis can be substantiatedwithin its own logical frame of referencethroughpredictionand
test. On the other hand, the discovery of new and better data is directly dependent on, and
subsequent to, the formulation of new hypotheses. This can be illustrated by developmentsin
Paleolithicarchaeologyduringthe last 20 yr.
The Bordiansystem of classificationand summarydescriptionof whole Paleolithicassemblages
(Bordes 1953, 1961; de Sonneville-Bordesand Perrot 1953-56) resulted in the discovery and
collection of new and better data in central and eastern EuropeanPaleolithic archaeology.Thus,
all artifacts including debitage had to be collected and analyzed where formerly a few type
fossils had been considered sufficient. Even so, Paleolithic archaeologistswould still be content
with cumulative graphs if a series of new hypotheses had not been formulatedfor the existing
data. These hypotheses led to the discovery of several new classes of "data", such as spatial
distributionsand multivariatecorrelations.New insightsare not generatedby new data, but by the
logic of archaeologists.
It appears from the preceding discussion that isolation of Paleolithic societies is not a
by-product of other researchendeavors,such as the explication of individualsettlements, or the
formulationof hypotheses about culturalprocesses,or the mere acquisitionof better data. Rather,
it constitutes a researchproblem by itself and forms a precondition to many other avenues of
investigation.This problemcan be approachedwithin a frameworkof deductivelogic by positing a
general model of a Paleolithic society, by generating predictions from such a model for the
archaeologicaluniverse, and by testing the implicationsof the model againstappropriatesets of
data within the Paleolithicculturalrecord.
In the following sections, I will outline the first 2 stagesin such a researchdesignas a first step
towardthe isolation of Paleolithicsocial systems.

Wobst] PA LEOLITHIC SOCIAL SYSTEMS 151
BASICASSUMPTIONS
A model appropriate to the isolation of Pleistocene societies should satisfy the following
constraints:(1) it should be generalenough to accommodatethe expected rangeof variation;(2) it
should be sufficiently simple so as to reduce the complexity of the archaeologicaland cultural
reality to intelligible dimensions; (3) it should have practicable test implications for future
research.
Initially, the 3 constraintsappearto be too restrictivefor this task. How are we to know which
rangeof variationto accommodateand what degree of complexity to expect without knowledge
of these variablesin the data?
Actually, the 3 conditions facilitate our task. The imposition of generalityrequiresthe model
to be as culturallynonspecificas possiblein orderto accommodatea wide rangeof culture-specific
data. The condition of simplicity forces us to include only cultureprocesseswhich are absolutely
essential to our model so that the less essentialprocessesdo not obscurethe operationof the more
pertinent variables.The need for test implications,f in ds us to keep our model focused
on the problem and to exclude all those aspects which have no bearing on the isolation of
Pleistocene societies within the archaeologicalrecord. All 3 constraintswork in the directionof a
minimal model which is built upon the least number of assumptionsand which incorporatesonly
variablesof relevanceto the problem.
Beyond the assumptionsabout cultural systems and cultureoutlined above, my model requires
the following basic assumptions:
(1) The maximumband is theculturocial correlateof the
he cula system of huntersand gatherers.
It thus constitutes the naturalanalyticalunit in the investigationof Pleistocene culturalprocesses.
(2) Paleolithic social groups are territorial."Territorial"implies that the membersof a given
social group moved within an area which was more or less delineatedby social factors, by the
proximity of other such groups, by considerations of distance, by familiarity with the
environment,and by naturalobstacles.
(3) A network of regularhexagons forms the best geometric approximationof the social and
geographicspace organizationwithin and between Paleolithicsocial systems.
(4) Giventhe generallylow Pleistocenepopulationdensities,and the relativelysmallnumberof
people requiredfor Paleolithic culturalunits, the size of the maximumband was maintainedat an
equilibriumlevel due to a balancebetween the procurementstrategies,the matingsystem, and the
populationsex ratio, mortality, and fertility rates.The spatialextent of a given society can then be
estimated from the equilibrium size of the maximum band and the carryingcapacity of the
regionalenvironment.
(5) Societies articulate the components of their cultural system in response to stimuli from
their naturaland social environment.A majorchangein stimuli will requirea differentarticulation
of the systemic components. Thus, we can estimate the temporalextent of specific constellations
of systemic components in particularPleistocene societies from the duration of specific steady
states in the set of stimuli to which they formed a response. This does not imply a changein the
culturalsystem, nor a knowledgeof the magnitudeor directionof the rearticulation.
Each of these assumptionswill be justified and discussedin more detail below.
Assumptionsabout Pleistocene Social Organization
Steward (1969) has introduced 2 terms useful to a discussion of the social organizationof
Pleistocene hunters and gatherers: the minimum and the maximum social aggregateor band.

Leaving the nuclear family aside, the former is defined as the most permanent and strongly
integrated social unit in a hunting and gatheringsociety. Its size is large enough so that it will
survive prolonged periods of isolation through the cultural practices of cooperation among its
members,division of labor accordingto age and sex, and mutual food-sharing.On the other hand,
it is sufficiently small to not place an undue strainon the local food resources.
Such minimumbands tend to consist of severalfamiliesof consanguineand/or affinal relatives
who, at least duringpart of the year, share a common settlement and participatein a given range
of cultural activities. The size of these units allows the unimpairedtransmissionof the cultural
system from generationto generation.The collective experience accumulatedduring the lifetime
of its membersprovidesthe individualwith a sufficient range of choices to cope with day-to-day
crisis situations. The exact composition, size, and stability of the minimum band will, of course,
vary with the social and naturalenvironmentof Paleolithicsocieties.
Whileat least potentially self-sufficient, a given minimum band tends to participatein a larger
social network in order to enhance its chance of biological and cultural survival. Steward
(1969:290) defines this largersocial network (the maximum band) as "frequently. . .little more
than a group with which its members somewhat vaguely identify." It essentially constitutes a
memberso
marriagenetwork which guaranteesthe biological survivalof its members,sin the s of
minimumband have to rely on a largernumberof personsthan their own membershipin orderto
providea memberwith a mate upon reachingmaturity.
Mate recruitment is made possible by, and itself stimulates, integrativeprocesses between the
different minimum bands of the social network. The integrativeprocesses, in turn, enhance the
chance of survival of the minimum bands and their members. Thus, food sharing and visiting
between adjacent bands create an atmosphereconducive to the exchange of mates. At the same
time, and at least as importantly, they help to counteract variationsin food supply at the local
level and dynamically adjust the local population size to a level which can be supportedby the
resourcesat a given time. Bartermeetings and work parties between members of different bands
broadcastthe availabilityof mates within the communicationnetwork of the maximumband. At
the same time, the formerprocessprovidesa givenband with exotic rawmaterials,while the latter
increasesthe exploitative efficiency of local groups.
The maximum band is, thus, a loosely interlocking network of minimum bands maintained
through ritual communication and exchange. The communication density within this network
tends to level the social and stylistic idiosyncraciesof the participatingminimumbands and their
members, and integrate them into a more or less coherent social unit. This unit, loose as it is,
constitutes the highest level of social integrationamong hunters and gatherers.For this reason,it
usually serves as the descriptiveuniverseof ethnographers(compareMurdock1968). It thus forms
the naturalanalyticalunit for the investigationof Pleistocene culturalprocesses.
The Territorialityof Pleistocene Huntersand Gatheres
"It cannot be mere chance which accounts for close similaritiesin the equipmentof sites as far
separatedas Le Placard(Charente)and Mammoutova(Cracow)(160 km), and one should not rule
out the probabilityof great treks, over very long distances"(Breuiland Lantier1965:84).
This traditional view of Paleolithic man as the eternal nomad of prehistory cannot be
maintainedfor a variety of demographicand culturalreasons.The absenceof skis, sleds, the wheel,
boats, draft animals, and beasts of burden in the Paleolithic cultural equipment implies that
Paleolithic man had to walk and carry his necessities on his back wheneverhe wanted to move.
The purposeful acquisition of surplus for some future necessity appears to be foreign to those
hunter-gathererswho have been studied by ethnographers(Lee and DeVore 1968:12). These
factors place a restrictionon the distance to be covered and on the ease and speed of movement.

Band society is characterizedby low labor input and a highly flexible set of procurement
strategies(see Lee 1972; Sahlins 1972; Woodburn1971). Nevertheless,if a local groupleaves the
niches it is accustomed to exploit, its chancesof successin hunting and gatheringdecline while its
chances of failure due to unpredictableevents increase. Thus, long-distancemoves would tend to
lower the population density and introduce an element of instability into an interregionalsocial
network.
The movement of entire maximum bands, or their components, beyond the area which their
cultural systems permitted them to exploit, and with which they were familiaris also effectively
blocked by social boundaries. A given society was not located in a vacuum but in a social
environment, that is, in a network of neighboringmaximum bands. Although a few individuals
certainly move from 1 maximum band to another, this process cannot have been large-scaleor
directionalin the sense that a band migrationfrom Le Placardto Mammoutovawould imply, but
would have occurredin a random fashion between the immediately adjacentsections (minimum
bands)of 2 societies.
Even those instances when an empty niche "suddenly" became available,such as the New
Worldor Australia,would not lead to large-scaleor long-distancemigrations.Rather,colonization
would take place graduallyand randomly by small breakawaygroups along the fringes of the
settled area.
For the sake of argument, let us assume that the New World constituted some 41.5
million km2 of empty niche. The mean equilibriumdensity in the Old Worldprior to the time
when the New Worldbecame accessible might have been .008 persons/km2. Pleistocene hunters
and gathererswere probably able to surviveat lower population densities, such as .004, in areasto
which they were not specifically adapted. Initially, the New Worldwould then have supported
160,000 people. While Birdsell's(1968:230) observationthat human colonizing populations can
triple in size every generationin a previouslyunoccupiedniche appearshighly unrealisticunderthe
demographicconditions prevailingduring the Pleistocene, there is no doubt that an empty land
mass such as the Americancontinent would have facilitated rapid population growth among the
colonizing population. Assuming, for simplicity's sake, a uniform environment and an initial
colonizing group of 25 people (an unrealisticallylow number to assure survival), after 100
generationsa mean population density of .004 would have been reachedover the entire continent
with an intrinsicrate of populationgrowth of only 1.091. The resulting3000 yr would still appear
as a mass migration in the archaeologicalrecord. In reality the "migration"derives only from
population increase along the marginsof the colonizingpopulation and its gradualexpansioninto
previouslyunoccupied territory. At the maximum, the "migrationspeed" would be only 5 km/yr
and could hardly be conceptualizedas a conscious processof migrationand massmovement.
The territoriality of hunters and gatherers is determined at the organizationallevel of the
minimum band. The "territory" of these groups is usually not maintainedthrough an exclusive
claim but through habitual use. It is delineated by the proximity of other minimum bands, by
distance,by familiaritywith the environment,and by naturalobstacles.
The HexagonalAbstractionof Pleistocene SpatialUnits
The regularhexagon appears to be the best geometric abstraction for much human spatial
organization. Among the regular polygons, only the hexagon combines optimum packing
efficiency with minimal movement and boundary costs (Haggett 1966:49). While circles provide
better accessibility from the center and have the shortest peripheryin relationto their area,they
cannot be tightly packed without leaving unaccounted voids betweenneighbors. Trianglesand
squarescan be packedtightly but have pooreraccessibilitycharacteristicsandlongerboundariesper
given area. Because of these optimal features, networksof regularhexagonsplay a prominentrole

154 ameRican antiquity [Vol. 39, No. 2, 1974
in theoretical works on locational analysis (Losch 1954; Christaller1933) and recently have been
successfully applied to the archaeologicalrecord(see, for instance,Wilmsen1973; Johnson 1971).
The degree of regularity which this abstraction imposes on modern spatial distributions and
human spatial networks is often not warranteddue to the forces of agglomerationand regional
specialization inherent in the modern market economy (Isard 1956). The hexagonal model,
however, is less likely to introduce gross distortions of reality if appliedto hunting and gathering
societies. This is due to the egalitariannature of Paleolithic societies, the lack of intraband
specializationexcept by sex and age, and the ecological restraintson population agglomerationsof
hunters and gatherers.
The relevance of hexagons to the spatial organization of hunters and gatherers can be
documented from the ethnographicliterature.We can express the degree to which human spatial
units conform to the hexagonal standard by means of the following parameter: the mean number
of boundaries with neighboring territoriess (per territory) excluding
perionly natural obstacles and the
periphery of a spatial network (Haggett 1966:51). This parameter has a value of 5.67 (median 6)
for a group of 31 minimum band territories among the eastern subarctic hunters (determined from
a map in Rogers 1969:23). Birdsell (1958) obtained values of 5.5 for the Australian tribes, and
Wilmsen (1973) calculated a mean of 5.6 (N = 155) for the North American tribes based on
Kroeber(1939).
The shape of the actual territories utilized by hunters and gatherers may not at all look like
whe
equilateral hexagons n plotted on a map. But the fact that minimum or maximum bands of
hunter-gathererson the averagefind themselves surroundedwith close to 6 neighboringbands
recommendsthis spatialpatterningas an ideal approximationof Pleistocene conditions.
The SpatialExtent of Pleistocene Societies
Paleolithic social units should display a range of size variationat least as largeas that observed
among modern hunters and gatherers.This variability is not random, but forms a predictable
response to stimuli in the naturaland social environment,and to the integrativeand centrifugal
forces of band society (see Woodburn1971). We are, at present,unable to determinethe size and
range of Pleistocene societies because the stimuli are of such infinite variety, because the
archaeologicalrecord is too biased, and because hypotheses with test implicationshave not yet
been advanced. However, we can approximatethe spatial dimensionby means of the maximum
band as the social correlateof the culturalsystem of band society.
This approximationis guided by 2 considerations.The minimalequilibriumsize of any society
may be defined as the number of people which will consistently guarantee the presence of a
suitable mate for a group member upon reachingmaturity. The maximal equilibriumsize of the
society may then be defined as the maximal number of people which can be consistently
integratedby the culturalmechanismsof a given culturalsystem and which is consistently required
for the successfuloperationof such a culturalsystem.
These definitions approximate the size of Pleistocene social groups ratheropen-endedly.For,
while the lower size limit of the social groupmight representa constant, dependingon the mating
system, on population mortality and fertility rates, and on the sex ratio at birth, the upperlimit
seems to leave much room for variation.It is doubtful, however,whether the maximalequilibrium
size has any bearingon the culturalsystem of Pleistocenehuntersand gatherers.
Let us assume, for the sake of simplicity, that the units of a Pleistocene society (the minimum
bands) are evenly spaced over a territory,that a definite intensity of communicationis requiredto
maintain the maximum band as a social unit, and that the social units can spend only a finite
amount of their time in communication as opposed to other tasks. As population density
decreases, a point will be reached at which it becomes impossible to maintain the necessary

intensity of communication without impairingthe probability of survivalor without the social

groupbreakingapartfor lack of communication.
In analogy to modernband societies, Pleistocenehunter-gathererscan be expected to have such
low population densities that the minimal equilibriumsize will hover only slightly above this
point, or will fluctuate around it in a random fashion, thus not permittingany largerculturally
maintainedgroupsthan those consistingof a few minimumbands.
This is also evident from the comparativelysmall size of hunting and gatheringwork partiesor
exploitative task groups (Lee and DeVore 1968:12; Steward 1969:265). These tend to stay
conveniently below the biologicallyrequiredminimumsize, only exceptionally involvingmembers
from more than 1 minimumband. One expects that this is not only due to the vagariesof food
supply, but also to distance, communication,and population density, that is, to the same factors
which might keep the Pleistocene social group at the minimalequilibriumlevel.
Once the minimalequilibriumsize of Pleistocenemaximumbandsis determined,we can predict
the spatial extent of Pleistocene cultural systems by matching the caloric requirements of such a
group with the carryingcapacity of specific environments(see Birdsell1953). The isolation of the
minimal equilibrium size of Pleistocene social groups is thus a necessary prerequisitefor the
isolation of social systems.
Whilethis sounds unrealisticgiven the present state of Paleolithicresearch,populationdensity
predictions and hypotheses about the spatial extent of Paleolithic societies have test implications
for archaeologicalfieldwork. A cursory investigationof the anthropologicalliterature on band
society, however, suggestssome importantmethodologicalcaveats.Carryingcapacity is a function
of stochastic variables.Each of these variableshas its own probabilisticdistributionfunction and
its own chance to generatea limiting condition within the lifetime of a population. If a society
of hunters and gatherersisis in essential equilibrium, it will be adapted to the worst possible
conditions occurringwithin a generat
thin reasonablememory of its members.Thus, it is
useless to employ mean values of environmentalvariablesto determine the carryingcapacity of
various environments, as for example has been recently attempted by Casteel (1972). Secondly,
band society puts a premiumon minimizingwork effort and tuning demandsdown to the level of
relativelyeasily obtainable supplies (Sahlins 1972). Thus, work effort has to figureimportantlyin
any prediction of band society carryingcapacity, and diminishingreturnfunctions will depressthe
population density that could conceivably be supported by a region's resources.As population
increases toward a given carrying capacity, the tighter packing will increase social labor, the
increased population density will increase productive work effort, and the increased person to
person contact, coupled with increasedpopulation clustering,will increasepopulation morbidity
and mortality (requiringan increasein female fertility to maintainthe populationsize). All factors
combined significantlydepressthe potential carryingcapacity of an areafor band society.
The Time Depth of PleistoceneCulturalSystems
Archaeologists frequently generate a model of a given society on the basis of a specific

constellation of the system's components at a fixed point in time. This approachhas certain
shortcomings,particularlyif applied to the Paleolithic record, as I will illustrateby means of a
model.
Let us suppose that it would be possible to isolate the year 30,000 B.C. in the archaeological
record. Let us further assume that at this time the equilibriumpopulation size of a particular
maximum band was 500 people and that the population density was fixed at .02 persons/km2.
Finally, let us assume that only habitation sites would yield sufficient cultural material to be
recognizableand datable, that each personproduced 10 m2 of habitationdebris(Cook and Heizer
1965), and that only 50%of this areahas escapederosion.

156 amenican antiquity [Vol. 39, No. 2, 1974
Our task then consists in finding the 2500 m2 of remaininghabitation space within the 2.5
billion m2 originally exploited by the given maximum band. The chance of hitting any of the
originally occupied square meters in a single try is depressinglylow (1:10 million), while any
samplingscheme designed to obtain a representativesample would requireso much excavationas
to discouragethis approach.
It is, of course, impossible to establish convincinglythe contemporaneityof any 2 Paleolithic
sites. When applied to the dating of Pleistocene cultural materials, chronometric and
paleoenvironmentalmethods cannot differentiatemore than gross chronologicaldifferences with
any degree of reliability. Thus, archaeologistshave to be content with chronologicalsections large
enough (several hundred years for the Upper Paleolithic, several thousand years for the Middle
Paleolithic and severaltens of thousands of years for the Lower Paleolithic) to be differentiated
with the availablemethods.
While important to the archaeologistinterested in exact historical reconstruction, a specific
constellation of systemic components is of small consequence to the operation of a cultural
system. Immediatelybefore or after the particularyear in our model, the same system will still be
operating but its components will be articulated slightly differently. Lacking time depth, the
archaeologistinterested in the precedingand succeedingyear would generatea differentmodel of
manifestationsof the same self-regulatingequilibriumsystem (see Maruyama1963).

To avoid such errors in scale, one should isolate equilibriaof sufficient duration so that the
processes which structure the interaction between the members of a society, and between the
society and its naturaland social environment,become intelligible.In the Pleistocene context, the
changes in fauna and flora associated with glacial events in the northernhemispheremay be the
least arbitrary boundary markers by which a societal continuum can be partitioned into
analytically distinct temporal units. Such changes will affect regional carrying capacity, and
population densityd and istribution, in a lasting way sot inth e predominanceand
distributionof tool kits will differ, the procurementstrategieswill changein their seasonalround
and spatial focus, and different cycles of seasonal dispersaland aggregationwill result. While the
archaeologistis faced with the same society before and after such a change, as well as with 2
instances of the same cultural system, the articulation of the systemic components will be
sufficiently distinct to make it analyticallyeasierto talk about the before and after separately.
The analytically distinct equilibriumstates of Paleolithic societies should closely follow the
duration of equilibrium states in Pleistocene regional environments. This conclusion has 3
beneficial consequences for archaeologicalanalysis: (1) it allows us to deal with analyticalunits
which can be isolated archaeologically,chronometricallyand paleoenvironmentally;(2) it allows
us to deal with the processeswhich operate cultural systems ratherthan with time-specificfossils
of systemic structure;and (3), as will be shown below, it allows us to verify hypotheses about
culturalprocessesduringthe Paleolithicby testing them againstrepresentativesamples.
THEMINIMUMEQUILIBRIUMSIZE OF PLEISTOCENE
POPULATIONS
Choice of Approachand Predictions
The basic assumptionsof the foregoing discussionsuggesta logical approachto the isolation of
Pleistocene societies. The size of social groups can be approximatedin terms of the minimal
equilibrium size (MES) of Paleolithic populations. Under the assumption of territoriality,the
spatial extent of Pleistocene societies becomes a function of the carryingcapacity of regional
ecological systems, and of the size of the maximum bands. The time depth of specific cultural
systems can then be predicted from the duration of equilibriumsystems in the naturalecology of

regions. Our major aim should thus be the determinationof the MES of Pleistocene maximum
bands.
My approachis based on the assumptionof maximalconstraints.Therefore,the resultsreported
in the following chapter do not establish constants to be found in all hunting and gathering
societies, rather, they establish the lower limits of a range. However, the models are sufficiently
simple so that they can be modified for the habitat of specific societies, and thus can be adapted
to the requirementsof the individualarchaeologist.
As defined previously,the MES describesthe numberof people which can consistentlyprovide
group memberswith suitable mates upon reachingmaturity.It is operationalizedhere as the mean
and median number of persons that live in the interveningdistance between 2 marriagepartners.
These partnersare nearest neighborsamong the availablemates in the mating pool. This mean or
median value should adequately reflect the required size of the MES. It is frequently used in
historic demographyand population genetics (Harrisonand Boyce 1972; Spuhler 1961) and has
recently been applied to hunter-gatherersby Yellen and Harpending(1972). The MEScannot be a
cross-culturalconstant due to the large variety of factors which may be expected to influence it.
At the same time, the value of the MES should be predictablewithin narrow limits from the
constellationof the variablesthat exert an influence on it.
Thus, all other factors being equal, we would expect a negative correlation between life
expectancy at age 15 and the MES. As life expectancy increases, more persons per given
population size would be in the pool of potential mates. Juvenile mortality and MES should be
positively correlated with each other. Whenit increases,fewer persons per given population size
would reach reproductiveage. This would imply increaseddistancesbetween potential mates, and
thus cause an increasein MES.Strong deviationsfrom a sex ratio of 100%should reducethe MES,
if the closest neighborswithin the pool of mates are removed first, since, per memberof the less
frequentsex, there would be more potential matingpartnersper givenpopulationsize.
The MES should also be stronglycorrelatedwith the culturalrulesreflected in a matingsystem.
Here we can predict that the more restrictivethe rules on mate selection, the largerthe required
MES will be since fewer mates will be availableper given populationsize. Thereforeexogamy, age
restrictions,and preventionof remarriageafter widowhood would all tend to decreasethe number
of eligible mates and increasethe MES.
The effect of polygyny on the MESwill depend on the sex ratio at birth. If the sex ratio of the
membersof the pool of potential mates is close to 100%,polygyny should not changethe size of
the MES appreciablyas comparedto a monogamoussociety. If, however,the sex ratio of the same
matingpool stronglyand repeatedlydeviatesfrom 100%,the MESshouldincreaseperceptibly.For,
while in the monogamous society, one sex of mates can pick out of the excess of the other-sex
mates those that are closest and leave the remainderunmated,in a polygynous society those excess
mates (which on the averagewould be furtheraway from their potential partners)would be mated
too, and thus increasethe mean distancefigures.
Let us suppose that a group of Pleistocenehunter-gathererssuffersfrom a recurrentshortagein
marriagepartners.Under these conditions, 3 nonviolent alternativesare open to the group: (1) it
could relax the culturally restrictivemating rules and thus make more mates availableper given
population size; (2) it could include neighboringminimumbandsin the matingnetwork to achieve
the same effect; or (3) it could simply let the queue of unmatched mates accumulate in the
mating pool until other potential mates reach reproductive age. Alternatives 1 and 2 are normal and
expected responses of band society and have frequently been reported in ethnographic literature.
Strategy 3 can be expected to endanger the survival of the group. If a sizeable fraction of the
group membership is kept unmated, the chances for group survival should decrease. Conversely, if
the population were kept stable under these conditions, a strong increase in female fertility rates
would be required to make up for the lost reproductive potential of the unmated group members.

Only alternatives1 and 2 appearto be strategiesthat are adaptivein the long run in the context of
band society.
The value of the MES in Pleistocene societies cannot be induced from presently available
archaeologicaldata. It would also be dangerousto derive it from modern hunting and gathering
populations due to distortions introducedby their modern culturalenvironmentsuch as imported
epidemic diseases, means of transportationabsent in the Paleolithic archaeologicalrecord, and
negativemigrationbalances.Only the simulationof a hypotheticalPleistocenepopulation offersan
economicalmeans to arriveat a numericalapproximationof the MESduringthe Paleolithic.
In its most abstractform, simulationreducesa complex reality to a set of generalmathematical
equations which then can be manipulated for particularresearchaims. I am aware of only 1
attempt in which this approachhas been appliedto Pleistocenesocieties(Thomas1967). Deduction
from this model for the purposesof my researchdesignis made difficult by the logical framework
of higher mathematicsin which it was conceivedand by its researchfocus on primitiveeconomics
ratherthan primitiveculturalsystems.
I chose to simulate hypothetical Pleistocene populationsby means of Monte Carlotechniques
which permit a mathematicallyless demandingand culturallymore rewardingmodel. Monte Carlo
techniquesimply "that a society of 'robots' is created in which 'life' goes on accordingto certain
probability rules given from the start" (Hagerstrand 1968:372). A particular Monte Carlo
simulationcan best be conceptualizedas an educationalgame consisting of a gamingtable (area),
pieces (people), rules (biological or cultural rules of behavior),and a seriesof different outcomes
dependingon the specificationsof the components. Its educationalpurposeis either the isolation
of the specifications under which a given outcome was produced, or the approximationof the
outcome given a set of complex instructions. The latter purpose is ideally suited to the
determinationof the minimumequilibriumsize of Pleistocene populations.
The SimulationModel
A complex simulation cannot be carried out without the use of the computer. I wrote the
simulationprogramoriginallyin FORTRAN-IVfor use on the IBM 360 computer system of the
Universityof Michigan.Subsequently,I revisedthe program,included severaladditional options,
and adjusted it for 3600 FORTRAN to be used on the CDC 3800 system of the Research
ComputingCenter at the University of Massachusetts.A technical description, flow chart,listing,
and sample output of the programare contained in Wobst (1974). This section will deal with the
assumptionsand rulesof the simulationmodel in a nontechnicalway.
In my model, the gamingtable is considered to be a geographicregionof hexagonalshape. Its
size is immaterialexcept that it should be large enough to place an upperlimit on the amount of
intraregionalcommunication. The regionalenvironmentis uniform over space, constant through
time, and lacks geographical barriersto communication.The entire area is taken up by a tightly
packed network of smaller-sizedregularhexagons. Each of these subterritoriesis exploited by a
minimum band. The region as a whole is occupied by a regional population without any
assumptions about intraregionalsocial and stylistic differentiation beyond that given by the
minimumbands. The size of this populationhas to be largeenough (1200-2000 people) to include
all expected sizes of the Pleistocene MES. The number of minimum bands is dependent on their
size. In the simulationruns on which this paperis based, the minimumbands(61 in number)had a
mean of 25 members.Thus, the regioncan be visualizedas a network of hexagonalminimumband
territories,4 tiers in radius(see Fig. 1).
Each individualin this populationis identified by name, age, sex, place of residenceand marital
status, and by the namesof mate(s), parents,grandparents,and great-grandparents. Individualsare
born, grow up, mate, produce offspring, and die. Individuals,nuclearfamilies,or minimumbands

Wobst] PALEOLITHIC SOCIAL SYSTEMS 159
61
60 38.
59 37 39
58 36 20 40
57 35 19 21 41
34 18 8 22
56 17 7 9 42
33 6 2 23
55 16 1 10 43
32 5 3 24
54 15 4 11 44
31 14 12 25
53 30 13 26 45
52 29 27 46
51 28 47
50 48
49
Fig. 1. Arrangement of minimum bands in the simulation runs. The bands are
numbered in terms of their closeness to Band 1 within the hexagonal matrix.
can move from subterritoryto subterritoryor outside the regionif culturalrulesprescribeit. Time
proceedsyear-by-year.
The basic stochastic variables controlling the operation of the system are the age-specific
mortality rates (the chance of an individualdying within a givenyear of life), age-specificfertility
rates (the chance of a mated female givingbirth to live offspringin a givenyear), and the sex ratio
(the chance of a live birth being male or female).
It would have been difficult to generate these variablesfrom known Paleolithicpopulationsas
the largest preserved series contain less than 50 members (Krapina,Priedmosti,Ofnet Hohle).
Sampling error would have distorted any life tables generated for these populations to such a
degree that this approach had to be discarded. Instead, model life tables for preindustrial
populations were utilized (Weiss 1973) which can be closely matched against the demographic
characteristicsof living hunters and gatherers.Most of the simulation runs were carriedout with
Weiss'sMT:25-40 (1973:144) which approximatesthe demographyof severalEskimo and Alaskan
Indian groups. For comparison,a few simulationswere run with MT:27.5-45 (1973:154), best fit
by other Eskimo groups. The relevantparametersof these models are given in Table 1. For lack of
data, I was not able to use sex-specific mortality rates. Since the age composition and mortality

Table 1. Model life tables utilized in the simulations.
1) MT:25-40 (Weiss 1973:144)
Fertility Rates
Age Mortality Rates (annual) Age Composition
0-1 .3330 4.4%

1-5 .2130 13.0%
5-10 .1350 13.9%
10-15 .1191 12.1%
15-20 .1476 10.5%
20-25 .1516 .070 8.9%
25-30 .1556 .191 7.5%
30-35 .1598 .191 6.4%
3540 .1640 .155 5.3%
4045 .1684 .108 4.4%
45-50 .1728 .045 3.7%
50-55 .1774 .009 3.0%
55-60 .2134 2.5%
60-65 .2823 1.9%
65-70 .3619 1.3%
70-75 .4704 0.8%
75-80 .6125 0.4%
81 1.0 0.2%
2) MT:27.5-45 (Weiss 1973:154)
Fertility Rates
Age Mortality Rates (annual) Age Composition
0-1 .3000 3.9%

1-5 .1850 11.9%
5-10 .1220 12.9%
10-15 .1016 11.5%
15-20 .1305 10.2%
20-25 .1336 .059 8.8%
25-30 .1367 .161 7.6%
30-35 .1400 .161 6.6%
3540 .1432 .131 5.6%
40-45 .1466 .091 4.8%
45-50 .1500 .038 4.1%
50-55 .1536 .008 3.5%
55-60 .1879 2.9%
60-65 .2513 2.3%
65-70 .3270 1.6%
70-75 .4315 1.0%
75-80 .5698 0.5%
81 1.0 0.2%

Wobst ] PALEOLITHIC SOCIAL S YSTEMS 161
rates in the model tables are listed by age group, the programcalculatesvaluesspecific to the year
of age from this input. Fertility is alreadyin the form of annualratesin Weissso it only has to be
multipliedby 2 to reflect births of both sexes.
Weiss's model life tables assume a stationary population. This implies that the simulated
population would rapidly die out if no adjustmentwere possible in the population mortality or
fertility rates. Since the stationary population model assumes an infinite population, any
significantdeviation from infinity will work to decreasethe size of the populationsteadilyuntil it
dies out (produced by randomly fluctuating births, deaths, and sex ratios). The age-specific
fertility rates were self-adjustingto preventthis. Since "the relativeamount of fertility at each age
in the child-bearingperiod follows generallysimilarpatterns in most human populations"(Weiss
1973:31), the age-specific fertility rates were multiplied by a constant wheneverthe population
had increasedor decreasedto a specified level. The sex ratio in most runswas set at 100%,while a
few simulations were run with differing sex ratios for comparison. The programallows for the
simulationof a largenumberof differentculturalassumptions.Thus, the simulatedsociety (among
others) can be monogamousor polygynous, endogamousor exogamous, incestualor with various
differentincest taboos, and patrilocalor matrilocal.
The operation of the estochastic variables is governed by the use of uniformly distributed
random numbers. Thus, given an individual'sage and the chance of death at this age, a random
number between 0 and 1 is drawn. If this number is smaller than the probabilityof death, the
personwill die in the givenyear; if it is larger,the individualsurvivesthe year in question.
Each simulation run consists of the following sequence of events. As a first step, the initial
population is generated on the basis of the given age and sex distribution.The annual cycle of
events then begins. First, all individualsare aged by 1 yr and their status is adjustedas required.
Next, those individualsnew to the marriagepool, due to widowhood or due to havingreached
maturity, enter the mating loop. The males-to-be-marriedcheck through the pool of
females-to-be-marrieduntil a compatible partner is found. This is for the sake of programming
efficacy only and does not imply any cultural assumption. If there should be more than 1
potential partnerper male, the one closest to the male's place of residenceis preferred.The status
of the marriagepartnersis adjustedand moves requiredby marriageresidencerules are carriedout.
Following the mating loop, all females of reproductiveage enter the procreationloop. Here, a
random number decide whether they should give birth to live offspring, and whether the sex of
this offspring should be male or female. The simulation proceeds to the mortality loop where all
membersof the regionalpopulationcan be evaluatedagainsttheir age-specificchance of death.
Next, a series of fusion and fission moves can be carried out. Thus, a minimum band of
smaller-than-specifiedsize can be made to join the closest larger-sizedband. Conversely,
subterritory populations which have increased beyond a specified point can be split into 2
sections. The smallerone can be placed into the nearestempty subterritoryor join anothergroup.
At the end of the annual cycle, the basic stochastic variablesare checked for the necessity of
adjustmentsand the annualcensus data are accumulated.
In order to reduce chance noise in the system, a single run should simulate at least 400 yr of
actual time. Out of this total, census data should be accumulatedonly duringthe last 200 yr since
it takes severalgenerationsbefore all tracesof disturbancedue to the initial populationgeneration
have subsided. With most of the cultural assumptions,400 yr of simulated time will requireless
than 4 min of execution time on the CDC3800 computer.
The Outcome of the Simulation
Given the assumption that under Pleistocene conditions a person would tend to "marry"the
closest available suitable mate in order to minimize movement, the MES of Paleolithic social
groups can be derived from our model in 3 ways: (1) as the mean or median radius(in terms of

162 american antiquity [Vol. 39, No. 2, 1974
Table 2. Summary of 40 simulation runs.
P(olygyny)
Run Life table MES-Mean -Median M(onogamy) Fertility Level
1 2540 175.4 100 M 119.1

2 25-40 198.0 100 M 122.8
3 2540 166.7 100 M 127.4
4 25-40 183.3 100 M 124.7
5 2540 198.5 100 M 126.4
6 25-40 220.6 125 M 127.2
7 2540 213.3 125 M 117.2
8 25-40 221.5 125 M 120.5
9 25-40 233.5 150 M 133.5
10 25-40 210.8 125 M 123.3
11 25-40 243.3 150 M 121.0
12 25-40 240.9 150 M 127.2
13 25-40 319.0 175 M 121.1
14 25-40 332.7 200 M 122.3
15 25-40 331.3 200 M 127.7
16 25-40 273.5 150 M 139.4
17 25-40 275.6 150 M 132.8
18 25-40 236.7 150 M 128.0
19 27.545 315.8 175 M 118.0
20 27.5-45 79.3 75 M 148.6
21 27.545 218.2 125 M 122.1
22 27.5-45 74.7 75 M 144.9
23 27.545 201.8 125 M 124.7
24 27.5-45 201.1 125 M 129.0
25 27.5-45 194.3 125 M 124.4
26 25-40 242.8 125 P 115.3
27 25-40 246.1 125 P 121.6
28 25-40 248.2 125 P 124.7
29 2540 248.6 125 P 126.7
30 25-40 254.3 150 P 121.1
31 2540 271.9 150 P 123.8
32 2540 283.5 150 P 121.4
33 25-40 266.3 150 P 103.5
34 25-40 294.3 175 P 157.2
35 25-40 265.6 150 P 157.1
36 2540 254.8 150 P 102.2
37 25-40 240.0 150 P 102.5
38 27.5-45 301.9 175 P 158.3
39 27.5-45 279.1 150 P 120.3
40 27.545 258.7 150 P 101.7

Wobst ] PALEOLITHIC SOCIAL S YSTEMS 163
tiers of minimum band territories)within which marriagesare contracted; (2) as the mean or
median number of minimumbands living inside the radiuswithin which marriagesare contracted;
and (3) as the mean or median number of personslivinginside the radiuswithin which marriages
are contracted.
However, given an equal number of persons within the mating distance, the hexagonal radius
will vary with the location of the marriagepartnersrelativeto the peripheryof the regionand with
the mean size of the minimumbands. The number of minimumbands againwill be affected by
their mean size. Thus, we will focus our discussionon the mean and median numberof persons
living within the mating distance. These 2 values are of differential utility. The mean is very
sensitive to extreme values and thus will better reflect the variance and the rangeof the MES,
while the medianbetter reflects the averageexperienceof the population.
Under the given assumptions these 2 values are obtained as follows: whenever a mating is
contracted, the population sizes of all minimumbandslivingwithin the mating distanceare added
up clock-wise,beginningfrom the territoryof the male, until the territoryof the female is reached.
Addition is terminated here (rather than counting out the full hexagonal tier) in order to avoid
minimumbands in the summationwhich might have providedan equidistantmate. At the end of
the simulationthe simple arithmeticmean and the medianare calculated.
Forty simulation runs are summarizedin Table 2. These runs representa total of 16,000 yr of
simulatedtime and 8000 yr of censusedpopulationsimulation.At first glance, the largespreadof
the MES under different conditions and assumptionsappearsto be ratherdiscouraging(range of
mean 79-332, range of median 75-200 people). If the results are translatedinto a spatialnetwork
of minimumbands with a mean size of 25 members,a maximumband will consist of a singletier
or 2 tiers of hexagonal territoriesand thus will have between 175 and 475 members.However,
once the individualruns are separatedaccordingto the assumptionsunder which they were run,
the MES behaves according to our predictionsand the effect of different strategieson the MES
becomes apparent.
Two factors do not have any effect on the MES: the total size of the regionalpopulation, and
the mean size of the minimumbands. Simulations6 and 8 were run underidenticalassumptions.
As is apparentfrom the valuesof the mean and mode, their MESis practicallyidenticalwhile their
mean population size differs by 45 members(which amountsto 9000 membersover the censused
200 yr duration of the run). If all the simulationsare combined, the rangein mean populationsize
is less than 5%of the maximum populationsize, while the rangein MESapproaches400%for the
mean and 300% for the median, indicatingthat the size of the populationhas no influence on the
MES. The same applies to the mean size of the minimumbands:its range(24.8 to 26.9 members)
is only 8% of the minimal value and no directional effect of this variable on the MES can be
observed.Since changes in population or minimalband size do not changethe composition of the
population,this is to be expected.
The results of changesin adult life expectancy andjuvenilemortality on the MESare shown in
Table 3. In 5 simulations, the 2 different model tables were run under identical cultural
assumptions. While the 2 tables are relatively close to each other, given the range of human
mortality and fertility experiencemodeled by Weiss,the decreasedjuvenile mortality(expressedin
an increasein survivorshipto age 15 from 40%to 45%)and the increasedlife expectancy at age 15
(from 25 to 27.5 years) decreasethe MES. This bears out our predictionthat the MESshould be
positively correlatedwith juvenile mortality and negatively correlatedwith adult life expectancy.
If mortality rates and cultural rules are held stable, changesin the age-specificfertility rates do
not have a lasting effect on the MES. Table 4 shows 5 paired simulationswhich were run with
identical assumptionsand input fertility rates. As in all runs, fertility rates were self-adjustingto
maintain a stable population size. The amount of adjustmentrequiredis indicated in the mean
fertility level which gives the mean value of the age-specificfertility rates relative to those in the
input life table. There does not seem to be any correlation between crude birth rate or mean
fertility level and the mean or median MES. However, the MES will temporarilyincrease or

164 amenRcanantlquity [Vol. 39, No. 2, 1974
Table 3. Paired life tables run with identical assumptions.
Life Table Run Median MES Mean MES
25-40 18 150 236.7

27.5-45 25 125 194.3
25-40 9 150 233.5

27.545 21 125 218.2
2540 19a 125 213.4

27.545 23 125 201.8
2540 31 150 271.9

27.545 39 150 279.1
2540 15 200 331.3

27.545 19 175 315.8
mean
MES 15 15.1
decrease
decreasein response to sudden changes in fertility rates, so that it should be positively correlated
with the standard deviation of the annual crude birth rates. This was illustrated with some
simulation results in Wobst (1971). Thus, a population might find it advantageousto maintain
annual crude birth rates within relatively narrow limits to stabilize and minimize the size of its
mating network.
Fertility will have a strong indirect effect on the MES. Since it can cause large changes in the
absolute sizes of populations,it acts as a stimulusto culturalfeedbackmechanismswhich keep the
Table 4. The effect of age-specific fertility rates on the MES. Pairs of runs
simulated with identical assumptions.
Mean Fertility Level

Run (Percentage of Life Table Value) Crude Birthrate Median MES Mean MES
1 119.1 69.4 100 175.4

2 122.8 69.4 100 198.0
3 127.4 71.1 100 166.7

4 124.7 70.6 100 188.3
5 126.4 71.6 100 198.5

7 117.2 67.8 125 213.3
6 127.2 70.6 125 220.6

8 120.5 68.8 125 221.3
9 122.0 69.8 150 233.5

10 123.3 69.7 125 210.8

population size in check. These cultural mechanismsin turn affect the MES. This will become
more apparentafter the effect of culturalruleson the MEShas been discussed.
The following cultural assumptions govern all runs: (1) the regional population is 100%
endogamous;(2) the nuclear families are monogamousor polygynous as specified for the run;(3)
male and female reproductivelife span lasts from age 16 to 50; (4) within this period remarriage
after widowhood is encouraged;(5) a mate searchesfor the closest availablesuitable partner;(6)
optimally, all persons of reproductiveage are married;(7) the sex ratio is 100%unless otherwise
specified; (8) fertility is determinedsolely by the age-specificfertility rates as defined above;(9)
there are no culturalcauses for mortality and no ecological disastersexcept as might be contained
in the age-specificmortalityrates;(10) membersof a minimumband move only (a) if they become
married;(b) if 2 or fewer productives (between ages 16 and 50) remainin the minimumband
(fusion); (c) if a minimum band increases beyond a specified size (fission). If the first occurs,
dependingon the locality rules, one mate joins the subterritoryof its partner.If the second occurs,
the remainingmembers of the minimumband join
thee nearest band which will not overflow by
this influx of population. If the third occurs, a given numberof couples and their offspringmoves
to the nearest unoccupied territory or, if none is available,joins the nearest group that is of
sufficiently smallsize that it does not have to fission when the new membersare added;(11) in the
simulationsof polygyny, unmarriedmales choose their mates before maleswith 1 or 2 partnersin
order to facilitate programmingfor the small memory of the CDC-3800. While this may not
adequatelyreflect reality, it should not bias the size of the MESsignificantly.
Six sets of mating rules were simulated with monogamous and polygynous societies. These
specificationsshould providea rangefrom least to most restrictive.
Set 1: Minimumbands are endogamous, if possible, and patrilocal. Any age combination is
possible for the marriagepartners.Thereis no incest taboo.
Set 2: Incest taboo specifies that Fa, Mo, FaFa, FaMo,MoFa, and MoMoof the mates have to
be different. If only one of these names is shared at whatever level, marriageis excluded.
Otherwise,like Set 1.
Set 3: Incest taboo now appliesto all namesin the potential partners'genealogy,includingthe
great-grandparents.If only 1 is shared,marriageis excluded. Otherwiselike Set 1.
Set 4: Minimumbands are exogamous and patrilocal.Any age combination is possible among
the marriagepartners.Thereis no incest taboo.
Set 5: Incest taboo as in Set 2; otherwiselike Set 4.
Set 6: Incest taboo as in Set 3; otherwiselike Set 4.
Fig. 2 depicts the outcome of 18 runs with life table MT:25-40; of these, 2 each simulateSet
1-6 assumptions for monogamous societies, while polygynous societies are representedwith a
single run for each of the assumptions. Inspection of this figure confirms our prediction that
cultural restrictions will tend to increase the size of the MES, and that polygyny will affect this
variablein the same direction. While there is a certain amount of random noise inherent in the
MES because of the stochastic variationin the number of males and females of a given age and
status, the MES increases steadily from endogamy and incest to exogamy with a strong incest
taboo. At the same time, for the same set of assumptions,polygyny resultsin a higherMESthan
monogamy. The same is reflected in Table 5, where the percentagesof closest matings(within the
minimumband and its closest neighboringband) are also given.
It is surprisingthat the removalof the incest taboo underSet 1 and Set 4 assumptionshas little
or no effect on the size of the MES. Particularlyin the monogamoussimulations,Set 1 vs. 2, and
Set 4 vs. 5 (incest vs. incest taboo) produce almost identical results. The runsin which the incest
taboo was not enforced appearto be mere variantsof the runswith Set 2 and Set S incest taboos.
This cannot be accounted for by the low frequency of incest. For, if all types of incest are

MES MONOGAMY POLYGYNY MES
1 2 3 4 5 6 1 2 3 4 5 6
350 -+- -+- 350

I
I
300 -+- -+- 300

I
+
+ I
I
+
250 -+- + + -+- 250
+(2) +
+ 1
+ +(2) 1
1 + I
200 -+- + + -+- 200
I + +
I + I
I
* I
150 -+- *(2) -+-
-+- 100
150
150
* * * * *
*(2)
I
100 -+- *(2) *(2)
50 -+- -+- 50
MES 1 2 3 4 5 6 1 1 2 3 4 5 6 MES
~~~~~~~~MONOGAMY~ POLYGYNY ~I.

MONOGAMY
Fig. 2. Result of simulation runs to determine the effect of cultural restrictions on the size of the mating net-
work. Top and bottom: sets of assumptions introduced in the text; left- and right-hand side: value of the MES in
numbers of people residing in the mating distance; * indicates a median, + a mean value.
permitted, about one-fifth of all matings will be incestuous with the given vital statistics. If only
exogamous incest is permitted, still about 5% of all matings will be incestuous.
The effect of incest is limited for 2 reasons: incest increases the number of potential mates only
slightly and it does not increase the size of the mating pool. At the same time, while incest
decreases the mean mating distance for some marriages considerably, it also can be expected to
increase mating distances for the remaining matings once all incestual matings are removed from

Wobst] PALEOLITHIC SOCIAL SYSTEIMS 167
Table 5. The effect of cultural restrictions on the size of the mating network.
Percentage of Matings
Between Two Closest
Run Assumptions Life Table MES:Mean Median Minimum Bands
1 1 Monog. 25-40 175.4 100

2 1 Monog. 25-40 198.0 100 42.75
3 2 Monog. 2540 166.7 100 44.47
4 2 Monog. 2540 183.3 100 41.53
5 3 Monog. 25-40 198.5 100 38.40
7 3 Monog. 2540 213.3 125 36.97
6 4 Monog. 25-40 220.6 125 24.73
8 4 Monog. 2540 221.5 125 25.77
9 5 Monog. 25-40 233.5 150 22.47
10 5 Monog. 2540 210.8 125 27.05
11 6 Monog. 2540 243.3 150 21.51
12 6 Monog. 2540 240.9 150 22.19
26 1 Polyg. 2540 242.8 125 35.26

27 2 Polyg. 2540 246.1 125 33.47
28 3 Polyg. 25 40 248.2 125 30.42
30 4 Polyg. 2540 254.3 150 21.08
31 5 Polyg. 2540 271.9 150 20.40
32 6 Polyg. 2540 283.5 150 18.91
the mating pool. Therefore,while incest has little impact on the mean mating distance, it affects
the variance of the mating distance. This is illustrated in Wobst (1971).
The increase in variance caused by incest thus introduces instability and unpredictability into a
mating system. If the mating network in which individuals, families, and minimum bands
participate is to function in a stable, predictable, and dependable way, incest would be
counter-productive.It increases the social work effort (a larger communication and exchange
network has to be maintainedto counterbalancethe decreasedpredictabilityin mate choice), and
it increasesdifferentialaccess to goods between the membersof the same minimum band (some
members find mates very close, others have to go much further). Both of these factors would
counteract important cultural processes of the band society cultural system.
If age restrictions are placed on mate choice, the MES increases drastically. This is shown by the
runs below:
Run Assumption Median MES Mean MES Life Table
13 Set 1 175 319.0 MT:25:40

14 Set 4 200 332.7 MT:25:40
15 Set 5 200 331.3 MT:25:40

168 ameRncan antiquity [Vol. 39, No. 2, 1974
In these runs, only 1 assumption was changed compared to the previous series: males have
to find younger marriagepartners.In terms of Pleistoceneminimumbandswith a mean population
size of 25 members each, this is equivalent to an increase in the size of the required spatial
network from 1 tier (7 minimum bands with 175 members) to one with 2 tiers (19 minimum
bands with 475 members). The effect of this rule can be mitigated somewhat, if the time span
during which a male can marry is increased. This was done in Run 17: here the males were
requiredto look for mates 5 or more years younger and the time spanin which they were allowed
to marry was increased to age 55. Comparedto Run 14, which was simulatedunder otherwise
identical assumptions,the mean MES decreasedto 275.6 people, and the median MESdecreased
to 150 people.
Changesin the rules of postmaritalresidence do not have any effect on the size of the MES. In
the following 4 runs, all simulatedwith Model Life Table 27.5-40 underSet 5 assumptions,Run 21
assumes patrilocality, Run 23 matrilocality, Run 24 the partnerfrom the largerminimum band
joins the other band, while in Run 25 the reverseis assumed:
Run No. Median MES Mean MES Number of Moves
21 125 218.2 130
23 125 201.8 146
24 125 201.1 0
25 125 194.3 677
It is interestingto observe the effect of the opportunisticpostmaritalresidencerulesin Run 24

and 25. In Run 24, where the smallerminimalband receivesthe mates (as might be the case where
hunting or gatheringare negatively affected by increasedband size), the frequency of minimum
band fusion and fission moves drops to 0, while in Run 25, with the largerbands receivingthe
mates, the frequency of fusion and fission moves increasesto the maximum of the 40 runs, and
fission of minimumbandsinvolvesa largeproportionof band membershipin addition to those that
have to move due to postmaritalresidence rules. Thus, the patternof Run 25 conceivablymay be
found in areas wliere a particularlystrong stress is placed on food sharingand exchangebetween
minimal bands (for instance, areas that have a high temporal and spatial variance in the
productivityof the resourcesexploited by hunter-gatherers).
Three sets of runs simulated the effect of infanticide on the MES. In order to programthis
assumptionin a way that did not overtax the capacity of the CDC3800 computer, infanticidewas
simulatedby modifying the sex ratio at birth. This is not an ideal reflection of reality, but the bias
it introduces should be minimal. The result of these runs is shown below, paired with identical
runsusing a 100%sex ratio.
Run L.T. Sex Ratio Assumptions Mean Fertility Level Median MES Mean MES
(1) 20 27.5-45 66 Set 5 (monog.) 148.6 75 79.3

22 27.5-45 66 Set 5 (monog.) 144.9 75 74.7
21 27.5-45 100 Set 5 (monog.) 122.1 125 218.2
(2) 32 25-40 100 Set 6 (polyg.) 121.4 150 283.5

33 25-40 66 Set 6 (polyg.) 103.5 150 266.3
(3) 39 27.5-45 100 Set 5 (polyg.) 120.3 150 279.1

40 27.5-45 66 Set 5 (polyg.) 101.7 150 258.7

Male infanticide is modeled above, not for cultural reasons, but because of the way the
simulation programis written. The direction, if not the magnitude,of its influence on the MES
should be similar to that of female infanticide. In a monogamoussociety, infanticide drastically
increasesthe demandson female fertility to keep the populationstable. This is due to the fact that
a large number of females remains unmated (male infanticide), or that there are fewer fertile
females per population size. Thus, the mated femaleshave to increasetheir fertility to make up for
this reproductive loss. At the same time, however, the MES shows the smallest value of all
simulationruns. Ninety-nine percent of the matings take place within the population occupyinga
single tier of hexagons. In a polygynous society, on the other hand, infanticide(of males) greatly
reducesthe fertility requirementson females,while the MESdoes not seem to be affected at all.
Sex-specific infanticide introduces an imbalance in the adult sex ratio. In a monogamous
society, this increasesthe number of mate choices for one sex and, concomitantly, decreasesthe
size of the MES relative to a society without infanticide. In a society which permits polygyny,
sex-specific infanticide cannot have the same effect. Since all membersof the matingpool, with or
without infanticide, will find a mate anyway, there is no changein the numberof potential mates
(per person looking for a mate) and the MESshould stay at the same level. These resultsgive some
additionalsupportto the prevalenceof infanticideamonglivinghuntersand gatherers.
Infanticide certainly is one of the mechanismsby which hunters and gatherersmaintain their
population density at a level that can be supported by local resourcesand far below the point
where diminishingreturnsfor productiveeffort set in.
However, sex-specific infanticide allows for a stable and minimal size mating network in a
society that is largely monogamous. It makes mate choice a more predictable event and it
decreasesthe distanceto which one has to extend one's communicationnetwork to facilitatemate
exchange. At the same time, it increases the fertility requirementfor females. This makes it
difficult to accept sex-specific infanticide in areas where females carry a large burden of the
productiveeffort.
Several conclusions can be drawn from my simulations of the minimum equilibriumsize of
Pleistocene societies. The most important of these concerns the range of this parameter.Given a
largevariety of cultural and biological assumptions,the minimalequilibriumsize remainswithin a
rangeequivalent to a maximumband consisting of 7 to 19 minimumbandsof 25 people each. In
terms of spatial distribution,this is equivalent to spatial networks consistingof either 1 or 2 tiers
of hexagonal subterritories.Withinthis narrowrange,Pleistocene populations were able to select
from among a large number of alternativestrategiesthose which suited their particularcultural,
biologicaland spatialadaptiverequirements.
The size of the minimalequilibriumgroupis determinedby population fertility rates,mortality
rates, and sex ratio, as well as by culturalrulesimposed on the matingsystem. Juvenilemortality is
positively related to the size of the MES, while adult life expectancy is negatively related to this
value. The suspectedhigh infant mortality of Paleolithichuntersand gathererswould tend to move
the MES toward the upper part of the observed range. Fertility influences the MES through its
variance, the variance of the fertility and the MES being positively correlated. It would be
advantageousfor Paleolithicpopulationsto maintainan almost constant birth rate in orderto keep
the MES down. Child spacing, infanticide, and family planning in general, as reported in the
ethnographicliteraturefor huntersand gatherers,would tend to work toward this aim.
The imposition of cultural rules on the mating system significantlyinfluences the value of the
MES. The more restrictivethese rules, the largerthe size of the MESwill be. Wemight expect the
imposition of such rules under certain circumstances:in areas which require large cooperative
work parties;in areaswith a very uneven raw materialdistributionrequiringconstant and stable
reciprocitybetween largenumbersof minimumbands;and in areasof high population densities,in
which competition and friction might be lessened through stable and large mating systems and
their cultural corollaries. Under low population densities a minimal number of restrictiverules

would be expected. Rules that will increase the size of the MES include age restrictions on
matings, exogamy, polygyny, and regulations against mating with certain classes of relatives.
Incest is not advantageous to hunting and gathering populations because it introduces an
element of instability and imbalance into the social system. If incest is permitted, mate selection
becomes a less predictableprocess.Incest increasesthe varianceof the distancesbetween the social
units from which the mates derive. At the same time, it has little effect on the mean mating
distance and on the size of the MES.The incest taboo is thus of adaptiveadvantageto hunting and
gatheringpopulations.The same was shown for sex-specificinfanticide.In societies that allow only
a minimum of polygyny, sex-specific infanticide will stabilize the mating network, decreaseits
size, and make mate choice more predictable.
In conclusion, the MES is not a constant, but a variable with a relatively narrow range. This
rangeis controlled by mortality, fertility, sex ratio, and culturalrules on the matingsystem. Placed
into a regional, ecological, and cultural frame of reference, the MES can be used to predict the
spatial extent of Pleistocene societies.
THE MINIMUM BAND
Since "25" has been so frequently used in this paper as the mean size of minimum bands among
Pleistocene hunters and gatherers, I wll discuss the relevance of this or comparable values to
Paleolithic societies. So many ethnographershave recordedminimumband sizes around25 people
amongmodernhuntersand gatherersthat this figurehas recently been labelled "magical"(Lee and
DeVore 1968). The list below is only a selection from those groupsmentioned by the participants
in the symposiumBandSocieties (Damas 1969).
Minimum Source in
Band Society Band Size Damas (ed.) 1969
!Kung Bushmen 25 mean Marshall p. 281

Hadzapi 20-60 range Bicchieri p. 209
Birhor 25 mean Williams p. 146
Semang 20-30 range Gardner p. 211
Andaman Islanders 30-50 range Gardner p. 211
Athapaskans (in general) 20-75 range McKennan p. 104
eastern subarctic hunters 25-50 range Rogers p. 52
Iglulingmiut 35 mean Damas p. 210
Copper Eskimos 15 mean Damas p. 210
If we add Australian aborigines, for whom Birdsell (1957) reports a mode of 25, a range of
20-70 people per minimum band appears to be characteristic for modern hunters and gatherers on
4 continents, in the northern and southern hemispheres, in the tropics and in the Arctic, and at
population densities from .002 to .8 persons/km2. Thus, the "magical range"appearsto be of
general adaptive significance to hunting and gathering societies.
The existence of powerful feedback mechanisms, which keep the size of minimum bands within
this relativelynarrowrange, can be best demonstratedin a negativefashion. For this purpose,the
simulation model was slightly altered so as to prohibit any moves except those requiredby
patrilocality. This removed all negative feedback controlling the size of minimum bands. At the
beginning of the simulation, a regional population of 169 "minimum bands" was created

containing 8 members each. (This simulation was carriedout with a life table derivedfrom the
Illinois Archaic, see Wobst 1971.) The birth rate was allowed to fluctuate in order to keep the
regionalpopulation balanced againstthe population mortality. The resultsof this run by ten-year
intervalsare shown below.
Mean Number of Mean Number of

Ten-Year Period Occupied Territories Occupants
1 168.8 8.67
2 164.0 8.88
3 157.8 9.68
4 149.8 10.44
5 139.0 10.75
6 125.0 11.15
7 120.3 11.84
8 115.0 12.44
9 107.8 13.17
10 102.3 13.96
In the absenceof counterforces,the numberof occupied territorieswill steadily decrease,while

the mean size of the minimumbandswill steadily increase.If this programwere allowed to run for
several centuries, the number of occupied territorieswould decreaseto 1, while the size of the
minimum band would increase to that of the regionalpopulation. This is not an artifact of the
particularsimulation,but a responsepredictablefrom smallsamplestatistics.
All minimumbands are subject to small, annualsize fluctuationsfrom the cumulativeeffects of
the probabilisticvariablesof mortality, fertility, and sex ratio. These population parametersare
binomial in character, that is, they are summary statements of individuals'chances to die or
survive, and to be barren or to bear offspring (the mortality chances of a given individualare
expressed in the binomial equation: [the chance of survivalof individualof given age in given
year] + [chance of death] = 1). This has 2 implications:as the minimum band increasesin size,
the fluctuations due to mortality become proportionally smaller;and, the larger the band, the
smallerits chance to die out due to such fluctuations.
The binomial series given below illustrates the different effect of mortality on groups of
different sizes. The mean chance of an individualto die is derivedfrom the crude death rate of
Weiss's25-40 life table.
Chance of "Band" to Die Out

Group Size in Given Year
1 .0566
2 .0030
3 .0001
4 .00001
5 .0000005
A similar series characterizes population fertility so that the larger the number of women in the
local group, the smallerthe proportionalfluctuations in the numberand the sex of birthsand the

smaller the probability of 0 births in a given year. Thus, in the absence of natural and cultural
mechanisms controlling the size of local populations, local groups of small size are strongly
selected against and a strong trend operates towardthe agglomerationof the minimumbands of a
society into a singlelocal group.
A local population aggregateof the size of the MES, however, is not adaptiveto most hunting
and gatheringsocieties. Most naturalenvironmentsdo not permit consistent long-termaggregates
as large as the MES. Given nonrandomly distributed, seasonally and annually varying food
resources, the largerthe size of the co-residentgroup, the higher will be the cost of transportin
relation to the amount of production. A point is reached at which it becomes more
profitable to split the local groupin orderto maximizethe amount of productionand to minimize
transport costs and trampling space. Given Pleistocene population densities, and Paleolithic
exploitative technologies and means of transport,this point must have been far below the size of
the MES.
At the same time, cultural feedback mechanismsprevent the aggregationof large, consistent
local groups. Band societies are characterizedby a lack of formal political inutions
instit such as
courts, councils and chieftainships.Conflicts are usually resolvedthrough fission of the band (Lee
and DeVore 1968:12). This process effectively keeps the band size at low and stable levels since
The higher the number of such conflicts, the greater will be th ae nce or band fission, thus
maintainingthe minimumband at a low equilibriumsize.
The previously mentioned processes of interband food sharingand visiting work in the same
direction. If, throughthe randomvariationsin mortality, fertility, and sex ratio, the size of a given
minimumband increasesto a level far above that of other bandswithin a given society, it is more
profitable for the surplus to move to smallerbands (with smallerdensity of producersand thus
higherchancesof hunting and gatheringsuccess) than to maintainthe largersocial aggregate.These
moves are made more easily since the amount of personalpropertyis kept at a minimumlevel, and
since there is no immovablepersonalpossession such as individualclaimsto land or fishingrights.
Finally, the exploitative technology of hunters and gatherers does not require year-round
maintenance of large cooperative groups in a given place. A minimal band group provides a
sufficient manpowerreservoirfor most task-specificgroups in a huntingand gatheringsociety. In
the model life table population used in most of our simulations (MT.-25-40),there are 11 to 12
adults in a minimumband of 25 people. This is sufficient for manninga range of different task
groups for the day-to-dayexploitation of the environment.The few occasions duringthe seasonal
activity cycle which requirelargergroups,such as game drives,the maintenanceof game fences, or
rituals, can be handled at a lower year-round cost by means of temporary alliances between
neighboring minimum bands, rather than through the year-round maintenance of larger local
groups.
The lower size threshold of minimum bands is determinedby a series of culturaland statistical
considerations.Thus, a minimumband has to be largeenough to offer a reasonablechance for the
unimpaired transmissionof cultural knowledge to the next generation. The collective lifetime
experience of the membersshould providethe individualwith a sufficient rangeof choices to cope
with daily crisis situations and with sufficient input from daily observationsto insurehunting and
gatheringsuccess. There have to be enough membersto effectively carry out the daily activities
requiredfor the group'ssurvival,to decreasethe impact of individualfailurethroughmutual food
sharingand cooperation, and to exploit the environmentmore efficiently through a division of
labor by age and sex.
The half-life, borrowed from nuclear physics, can be used to describethe longevity and relative
stability of social units (in the absenceof culturalfactors counteractingthe stochasticvariablesof
fertility, mortality, and sex ratio). Since populations under equilibrium conditions and on a

Wobst ] SOCIAL SYSTEMS
PALEOLITHIC 173
regional scale never completely die out, there can be no finite value for their longevity. Ratlier,
extinction is an exponential decay process which theoreticallyhas an infinite tail. The half-life is
then simply the number of years requireduntil one-halfof the originallyexisting social units have
ceased to exist.
For minimumbands, this value was calculated by tabulatingall runswith life tableMT:25-40,
and adding the number of minimal bands that had become extinct for each run. These sums were
then totalled. The resultingvalue was dividedby the total length of the runsin orderto determine
the mean numberof bands dying out per year. This figurewas in turn dividedby the mean number
of bands per year, the result being the chance of an individualband to die out in a given year.
Based on 21 simulations,which ran for a total of 4200 yr, I obtained a value of .00389 (as the
chance of a band to die out due to mortality, fertility, and sex ratio fluctuationsin a given year).
The half-life can then be calculatedusing the following equation: .5 = (chance of minimumband
to survive a single year [.996]) (number of years [x] ). Solving this equation logarithmically,we
obtain a value of 177 yr for the half-life of minimum bands. Due to the simple assumptions
governingthe simulations,this value can only be consideredas a roughestimate. Nevertheless,its
magnitude suggests that local groups in the range of 25 members are adaptive to hunting and
gatheringsocieties. This size is sufficient to guaranteethe survivalof minimum bands over many
generations.
This is not to imply that minimumbands are actually this long-lived,or that they are of the
stable composition implied by the model. It only serves to show that minimum bands are of
sufficient size to survive most stochastic fluctuations of mortality, fertility, and sex ratio: in
addition to their cultural adaptiveness, they are statistically and demographicallylong-lived.
Comparedto this, the nuclear family is a statistically short-livedevent. As was shown in Wobst
(1971), the half-life of a nuclear family household ranges between 21 and 29 yr dependingon
the life table utilized. Statistically,only 50%of the nuclearfamilieswill survivea singlegeneration.
Before these "magicalnumbers"can be appliedto specific Paleolithicsocieties, they have to be
adjustedto take into account the specific environmentalconditions to which a given social group
is adapted. As in the case of the MES, the value of "25" establishesonly the lower limit of the
range of mean sizes of minimum bands under equilibriumconditions due to our assumptionof
maximalenvironmentalconstraints.
Since the survival chance and the diachronic stability of nuclear families are both so low,
householdscould not function as independenitculturalor biologicalunits. The culturalunits which
are significant to the survivalof the cultural system are the minimum band with its half-life of
approximately180 yr and the maximumband as the breedingpopulation.
SOMEARCHAEOLOGICAL
IMPLICATIONS
In the preceding sections we have posited 3 cultural units of predictable and consistent
membershipwithin Paleolithic cultural systems: the maximum band, the minimumband, and the
nuclearfamily. The maximum band, as a distinct manifestation of the cultural system of band
society, is expected to range in size from 175 to 475 people under Paleolithic conditions. The
minimum band, as the maximal local aggregate,is expected to contain around 25 people. A
Paleolithic society, then, must have consisted of between 7 and 19 minimum bands. If their
territories were hexagonally packed, the spatial network of a Pleistocene society must have
consisted of 1 or 2 tiers of hexagonal minimum band territories.The nuclear family identifies a
household group. This unit is the minimal structuralpose of which the membershipis predictable
and recurrent.
Based on modern hunters and gatherers,we would expect to find a largenumberof additional
culturalunits within Pleistocene societies, such as huntinggroups,partiesorganizedto obtain raw

materials, and aggregatesfor rituals. While their presence can be inferred crossculturally,their
membership,size and composition are society-specific. They become predictableonly once the
social and naturalenvironmentof a given society is understood. It is expected that game theory
and Monte Carlo simulationswill advanceour understandingof these task-specificgroups at the
level of the individualsociety.
The longevity of the minimumband and household were determinedin terms of the half-life: if
only demographicand statistical variablesare allowed to enter into the calculations, a minimum
band under the stated assumptionswill have a half-life of around 180 yr while that of the nuclear
family household amounts to around25 yr. Thus, 0.38% of the minimumbands and about 3. 1%of
the nuclear family households will be discontinued every year. Under the assumption of a
stationary population, the same figures express the rate at which new households are formed.
Given life table MT:25-40 and a maximumband with 19 local groupsof 25 memberseach, around
5 nuclear family households will be discontinuedin an averageyear; for minimumbands the same
value should lie around0.07.
The last 2 figures have strong implications for the archaeologicalrecordif the following model
is considered. We can make the (doubtlessly unrealistic) assumptions that a settlement will be
revisitedas long as the culturalunit occupying it remainsfunctional; that the settlement will fall
into disuse if the cultural unit is discontinued;that newly founded culturalunits will occupy new
settlement locations; and that (for illustrative purposes only!) the seasonal cycle of a certain
Paleolithic society involves only household winter camps and minimum band summer camps.
Under the given assumptions,the annual rates of minimum band and household discontinuation
and founding can then be used to calculate the number of nuclear family and minimum band
campsin the area occupied by the assumedPaleolithicsociety.
If the model population remainsin essential equilibriumover 1000 yr, it will produce around
5000 family camps, and around 70 minimum band camps. For hunters and gatherers,these
estimates are clearly too low by severalorders of magnitude.The assumed degree of settlement
stability is not confirmedethnographically(Campbell1968). Nuclearfamilies and minimumbands
function as camping units for a variety of purposes in the course of a year. The numberof camps
utilized during the lifespan of a given cultural unit is heavily influenced by the character,
distribution,and density of resources,the strategyof resourceexploitation, the regionalvegetation
and geomorphology, the proportion of inhabitablespace within the environment,and the ease of
overland travel, as well as by hygienic and epidemiological factors. Depending on the specific
habitat, the numberof camps of the given types should be considerablylarger.
Likewise,the ratio between minimumband camps and household camps(1 to 71) is distorted if
it is evaluatedagainstethnographicdata. The home bases of minimumbandsreceivemorestructural
input than those of households. More ritual is associated with home bases than with household
camps. Locations suitable for the base camps of a minimumband are rare;a far largerproportion
of the territory can be utilized for nuclear family camps. Finally, a minimum band is less mobile
than a nuclear family. Thus, one would expect more family camps per minimumband camp than
is evident from our calculations. Nevertheless,it indicates that the Paleolithic archaeologistneed
not despairabout the lack of potential subjectmatter.
In this context, I would like to introduce a class of archaeological data which is usually
neglected, overlooked, or misinterpreted. Let us consider the annual activity cycle of a
hypothetical Paleolithic society. Year after year, sections of the environmentwill be exploited by
gathering and hunting without leaving traces which archaeologistsusually consider worthy of
investigation, at least if butchering and kill sites are excluded. The hunting and gathering
adaptation demands that a large proportion of the daily work effort be spent outside of
settlements proper. These activities requiretools and produce archaeologicalremainsof potential

interest for the understandingof different procurement strategies. If a Pleistocene society

remainedin an equilibriumstate for a sufficiently long period, we can recoverthese remainsand
incorporatethe activities which they reflect into reconstructionsof the way of life of Paleolithic
societies.
Let us consider that a minimum band exploits a territory of 1250 km2 (at .02 population
density). If 1O%of this territoryis visited regularlyand an averageof 5 artifactsper personper day
is deposited during these visits, after a period of 1000 yr there will be an average of .36
artifacts/m2 within the 125 km2 visited frequently. Under a Poisson distribution (random
dispersalof the artifacts over the area), 125,000 of the meter squares(.21%)would contain 5 or
more artifacts.
My own experiments suggest that we can expect far greaterconcentrationsin certain areasif
models of random walk are fitted to regionalhabitatsand communicationroutes. Let us consider
the following simulationmodel. A gatheringcamp is visited year after year duringa given season.
The materialto be gatheredis spaced randomlyover the surroundingareain patches 10 to 100 m
apart from each other. A collecting trip of 10 km length will provide a sufficient return to the
gatherer.In the actual simulation, the gatheringcamp is the origin of a coordinate system. The
gatheringtrip was segmented into incrementsof randomlength between 10 and 100 m. For each
of these increments, a random direction between 00 and 3600 (in relation to the x-axis) was
generated.Under these assumptions,there is less than 1%chance that the gathererwill touch areas
more than 1 km away from the camp. Even if we rule that a gathererhas to avoid all patches
previouslyvisited on a given collecting trip, and that a group of gatherersgoes on a collecting trip
together, the chances for reachingareas further away from the camp increaseonly slightly. The
more restrictedthe field of movement, the greaterthe resultingdensity of artifactswill be, and the
more the activity area has been exposed to subsequent erosion, the greater will be the
postdepositional concentration of artifacts. It would be well to reanalyze some of the
archaeologicalevidence in this light. Thus, some of the LowerPaleolithic"sites"strungout along
the terracesof the Thames, the Somme, and the Seine may presentonly chance accumulationsof
artifacts deposited outside of occupation areas by repeatedly performed tasks. The dense
concentrationsof Upper Paleolithic remainsalong the riverswhich enter the PannonianBasinmay
be interpretedin the same way. By careful investigationof the habitat of Paleolithicsocieties, we
should be able to define these activity-specificareas. Their remainsdo not constitute clusters of
artifacts and features clearly delimited in time and space, but continuous distributions of
task-specificartifactsand tool kits.
The precedingconsiderationssuggestthat it is possible to apply systematic surveyproceduresto
Paleolithic societies. The largenumberof settlementsoccupied in the course of a givenequilibrium
state of a society, as well as the large number of artifactswhich will accumulatein specific areas
outside the settlements, should enable us to develop sampling models for the areas of given
Paleolithic societies. Such researchcan proceed only in areaswhere the Pleistocene stratigraphyis
fairly complete, such as in the loess areasof France,easternAustria,and southernMoravia.Based
on detailed paleoenvironmentalinvestigations,the areainhabitedby a givenPaleolithicsociety has
to be classified in terms of those factors which are significantto the spatial articulationof the
given cultural network. Stratified random sampling methodology can then be applied to obtain
unbiasedand representativeinformation concerningaspects of the system's operation. Giventhis
approach, the predictions derived from my models for conditions of maximal environmental
constraintscan be put to a test in appropriatehabitats.By meansof game theory and Monte Carlo
simulations, more specific models can be developed to generatepredictionsfor specific societies
and their habitats, so that the models presentedhere can be revisedand improvedin termsof the
archaeologicalevidence.

Acknowledgments. A similar version of this paper, entitled Boundary Conditions for Paleolithic Cultural
Systems: A Simulation Approach was submitted in partial fulfillment of the requirements for the Doctor of
Philosophy (Anthropology) in the Horace H. Rackham School of Graduate Studies at the University of
Michigan. I would like to thank the members of my dissertation committee, Richard D. Alexander, Richard I.
Ford, James B. Griffin, Robert Whallon, Jr., and Edwin N. Wilmsen, for their patience, encouragement,
stimulation, and interest in my progress. My wife Susan, as an interested layman, contributed toward the
inception of this paper by asking questions about Paleolithic society which I was unable to answer with the given
data. Stanton Green, Jim Moore, and Bob Paynter (University of Massachusetts) played devil's advocates and
helped me to retain my sanity while fighting the computer. Part of my research was supported by a predoctoral
fellowship from the Wenner-Gren Foundation for Anthropological Research.
Angeli, Wilhelm
1952 Der Mammutjdgerhalt von Langmannersdorf an der Perschling. Akademie der Wissenschaften, Wien.
Bainesz, Ladislav
1952 Barca bei Kosice. Akademie Vied, Bratislava.
Behm-Blancke, G.
1956 Magdal6nienzeitliche Zeltplatze eines Wildpferdjagerlagers im Kyffhausergebiet bei Bad Franken-
hausen. Ausgrabungen und Funde 13:263-266.
Binford, Lewis R.
1962 Archaeology as anthropology. American Antiquity 28:217-225.
1965 Archaeological systematics in the study of cultural process. American Antiquity 31:203-210.
1972 Contemporary model building: paradigms and the current state of Palaeolithic research. In Models in
archaeology, edited by David L. Clarke, pp. 109-166. Methuen, London.
Birdsell, J. B.
1953 Some environmental and cultural factors influencing the structure of Australian aboriginal
populations. American Naturalist 87:171-207.
1957 Some population problems involving Pleistocene man. In Population studies: animal ecology and
demography, edited by Katherine B. Warren. Cold Spring Harbor Symposia on Quantitative Biology
22:47-70.
1958 On population structure in generalized hunting and collecting populations. Evolution 12:189-205.
1968 Some predictions for the Pleistocene based on equilibrium systems among recent hunters and
gatherers. In Man the Hunter, edited by Richard B. Lee and Irven DeVore. pp. 229-240. Aldine, Chicago.
Bordes, Franqois H.
1953 Essai de classification des industries "Mousteriennes". Bulletin Societe Prehistorique Francaise
50:457-466.
1961 Typologie du pal6olithique ancien et moyen. Publications de l'Universite de Bordeaux, Memoires de
l'Institute Prehistorique, 1.
Bosinski, Gerhard
1969 Der Magdalenien-Fundplatz Feldkirchen-Gonnersdorf, Kr. Neuwied. Germania 47:1-38.
Brandtner, Friedrich
1954 Kamegg, eine Freilandstation des spateren Palaolithikums in Niederosterreich. Mitteilungen der
Prahistorischen Kommission der Osterreichischen Akademie der Wissenschaften (Wien) 7:3-93.
Breuil, Henri, and Raymond Lantier
1965 The men of the Old Stone Age. George G. Harrap, London.
Campbell, John M.
1968 Territoriality among ancient hunters: interpretations from ethnography and nature. In Anthro-
pological archaeology in the Americas, edited by Betty J. Meggers, pp. 1-21. Anthropological Society of
Washington, Washington, D. C.
Casteel, R. W.
1972 Two static maximum population density models for hunter-gatherers: a first approximation. World
Archaeology 4:19-40.
Cernys, Aleksandr Pankrat'evic
1953 Volodymyrivs'ka paleolitycna stojanka. Akademija Nauk, Kiev.
1961 Paleolitycna sto/anka Molodove 5. Akademija Nauk, Kiev.
Christaller, W.
1933 Die zentralen Orte in Suddeutschland. Jena.
Cook, Sherburne F., and Robert F. Heizer
1965 The quantitative approach to the relation between population and settlement size. Report of the
Universitjyof California Archaeological Survey, 64.
Damas, David (Editor)
1969 Contributions to anthropology: band societies. National Museums of Canada, Bulletin 228.
Efimenko, Petr Petrovic
1938 Pervobytnoe obvsestvo. 2nd ed., Leningrad.
1953 Paleoliticeskaja stojanka Borsevo II. Materialy i issledovanija po arxeologii SSSR, 39.
1958 Kosten 'ki I. Akademija Nauk, Moscow.

Freeman, Leslie G., Jr.

1964 Mousterian developments in Cantabrian Spain. Unpublished Ph.D. dissertation. Department of
Anthropology, University of Chicago.
Gibori, Mikl6s
1970 25 Jahre Palaolithforschung in Ungarn (1945-1969). Acta Archaeologica (Budapest) 22:351-364.
Gibori, Veronika
1968 La station du Paleolithique moyen d'Erd-Hongrie. Akademiai Kiad6, Budapest.
Gerasimov, Mixail Mixailovic
1931 Mal'ta-paleoliticeskaja stojanka. Irkutsk.
Higerstrand, Torsten
1968 A Monte Carlo approach to diffusion. In Spatial analysis, edited by Brian J. L. Berry and Owen F.
Marble, pp. 368-384. Prentice-Hall, Englewood Cliffs.
Haggett, Peter
1966 Locational analysis in human geography. St. Martin's Press, New York.
Harrison, G. A., and A. J. Boyce
1972 Migration, exchange, and the structure of populations. In The structure of human populations, edited
by G. A. Harrison and A. J. Boyce, pp. 128-145. Clarendon, Oxford.
Hoernes, Moritz
1903 Der diluviale Mensch in Europa. Vieweg, Braunschweig.
Isard, W.
1956 Location and space-economy. MIT Press, Cambridge.
Johnson, Gregory A.
1971 A test of the utility of Central Place theory in archaeology. In Man settlement and urbanism, edited
by P. J. Ucko, R. Tringham, and G. W. Dimbleby, pp. 769-786. Duckworth, London.
Klein, Richard G.
1969a .Manand culture in the late Pleistocene: A case study. Chandler, San Francisco.
1969b Mousterian cultures in European Russia. Science 165:257-265.
Klfma, Bohuslav
1963 Dolnl' Vestonice. Vflzkum taboriste lovcu mamuti v letech 1947-1952. Akademie Vied, Prague.
1967 Pavlovien a jeho vztahy ve strednf Evrope. Archaeologicke rozhledy 19:558-566.
Kroeber, Alfred L.
1939 Cultural and natural areas of native North America. University of California Publications in American
Archaeology and Ethnology, 38.
Lee, Richard B.
1972 Population growth and the beginning of sedentary life among the !Kung Bushmen. In Population
growth, edited by Brian Spooner, pp. 329-342. M.I.T. Press, Cambridge.
Lee, Richard B., and Irven DeVore (Editors)
1968 Man the Hunter. Aldine, Chicago.
Losch, A.
1954 The economics of location. Yale University Press, New Haven.
Lozek, Vojen
1964 Die Umwelt der urgeschichtlichen Gesellschaft nach neuen Ergebnissen der Quartargeologie in der
Tschechoslowakei. Jahrbuch fur mitteldeutsche Vorgeschichte 48:7-24.
Mania, Dietrich, and Volker Toepfer
1973 Koenigsaue. Deutscher Verlag der Wissenschaften, Berlin.
Maruyama, Magoroh
1963 The second cybernetics: deviation-amplifying mutual causal processes. American Scientist 51:164-179.
Murdock, George P.
1968 The current status of the world's hunting and gathering peoples. In Man the hunter, edited by Richard
B. Lee and Irven DeVore, pp. 13-20. Aldine, Chicago.
Obermaier, Hugo
1912 Der Mensch der Vorzeit. Allgemeine Verlags Gesellschaft, Berlin.
Okladnikov, Aleksej Pavlovic
1940 Buret'-novaja paleoliticeskaja stojanka na Angare. Sovetskaja Arxeologija 5:290-293.
1941 Paleoliticeskaja zilisca v Bureti. Kratkie soobcenija Instituta Istorii Material'noj Kul'tury 10: 16-31.
Prosek, Frantisek
1953 Szeletien na Slovensku. Slovenskd Archeologia 1:133-164.
Rogers, Edward S.
1969 Band organization among the Indians of eastern Subarctic Canada. In Contributions to anthropology:
band societies, edited by David Damas. National Museums of Canada, Bulletin 228:21-50, 52.
Rust, Alfred
1937 Das eiszeitliche Rentierlager Meiendorf. Karl Wachholtz Verlag, Neumunster.
1943 Die alt- und mittelsteinzeitlichen Funde von Stellmoor. Karl Wachholtz Verlag, Neumunster.
1958 Die jungpaldolithischen Zeltanlagen von Ahrensburg. Karl Wachholtz Verlag, Neumiinster.

Sahlins, Marshall
1972 Stone age economics. Aldine, Chicago.
deSonneville-Bordes, Denise, and J. Perrot
1953 Essai d'adaptation des methodes statistiques au Pal6olithique superieur. Premiers resultats. Bulletin de
la Societe Prehistorique Francaise 50:323-333.
1954-1956 Lexique typologique du Paleolithique sup6rieur. Bulletin de la Societe Prehistorique Francaise
51:327-335; 52:76-79; 53:408412, 547-559.
Spuhler, J. N.
1961 Migration into the human breeding population of Ann Arbor, Michigan, 1900-1950. Human Biology
33:223-225.
Steward, Julian H.
1969 Postscript to bands: on taxonomy, processes, and causes. In Contributions to anthropology: band
societies, edited by David Damas. National Museums of Canada, Bulletin 228:288-295.
Thomas, H. A., Jr.
1967 Human ecology and environmental engineering. Department of Anthropology, University of
Massachusetts. Manuscript.
Vertes, Laszlo
1955 Uber einige Fragen des mitteleuropaischen Aurinacien. Acta Archaeologica (Budapest) 5:279-290.
1956 Problemkreis des Szeletien. Slovenskd Archeol6gia 4:328-340.
1958 Beitrage zur Abstammung des Ungarischen Szeletien. Folia Archaeologica (Budapest) 10:3-15.
1964 Tata. Eine mittelpalaolithische Travertin-Siedlung in Ungarn, edited by L. Vertes. Akademiai Kiado,
Budapest.
1968 Rates of evolution in Paleolithic technology. Acta Archaeologica (Budapest) 20:3-17.
Weiss, Kenneth M.
1973 Demographic models for anthropology. Society for American Archaeology, Memoir 27.
White, Leslie A.
1949 The science of culture. Grove Press, New York.
1959 The evolution of culture. McGraw-Hill, New York.
Wilmsen, Edwin N.
1970 Lithic analysis and cultural inference: a Paleo-Indian case. Anthropological Papers of the University of
Arizona, 16.
1973 Interaction, spacing behavior, and the organization of hunting bands. Journal of Anthropological
Research 29:1-31.
Wobst, H. Martin
1971 Boundary conditions for Paleolithic cultural systems: a simulation approach. Unpublished Ph.D.
dissertation. Department of Anthropology, University of Michigan.
1974 Computer programs for the simulation of cultural systems. To appear in Computer contributions in
anthropology. Museum of Anthropology, The University of Michigan.
Woodburn, J. C.
1971 Ecology, nomadic movement, and the composition of the local group among hunters and gatherers. In
Man settlement and urbanism, edited by P. J. Ucko, R. Tringham, and G. W. Dimbleby, pp. 193-206.
Duckworth, London.
Yellen, John, and Henry Harpending
1972 Hunter-gatherer populations and archaeological inference. WorldArchaeology 4:244-253.


Society For American Archaeology

Uploaded by

Copyright:

Available Formats

Society For American Archaeology

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Society For American Archaeology

Uploaded by

Copyright:

Available Formats

Society for American Archaeology

Boundary Conditions for Paleolithic Social Systems: A Simulation Approach

This content downloaded from 200.54.94.206 on Fri, 10 Jan 2014 11:02:23 AM

I WAS STIMULATEDinto writing this paper by 3 observationson the state of Paleolithic

The ProblemWithina CulturalFramework

This content downloaded from 200.54.94.206 on Fri, 10 Jan 2014 11:02:23 AM

Paleolithic research has only gradually reoriented itself from an antiquarianpreoccupation

This content downloaded from 200.54.94.206 on Fri, 10 Jan 2014 11:02:23 AM

archaeology to classes of archaeologicalevidence such as living structures, portable art, and

This content downloaded from 200.54.94.206 on Fri, 10 Jan 2014 11:02:23 AM

This content downloaded from 200.54.94.206 on Fri, 10 Jan 2014 11:02:23 AM

Assumptionsabout Pleistocene Social Organization

This content downloaded from 200.54.94.206 on Fri, 10 Jan 2014 11:02:23 AM

The Territorialityof Pleistocene Huntersand Gatheres

This content downloaded from 200.54.94.206 on Fri, 10 Jan 2014 11:02:23 AM

The HexagonalAbstractionof Pleistocene SpatialUnits

This content downloaded from 200.54.94.206 on Fri, 10 Jan 2014 11:02:23 AM

The SpatialExtent of Pleistocene Societies

This content downloaded from 200.54.94.206 on Fri, 10 Jan 2014 11:02:23 AM

intensity of communication without impairingthe probability of survivalor without the social

The Time Depth of PleistoceneCulturalSystems

Archaeologists frequently generate a model of a given society on the basis of a specific

This content downloaded from 200.54.94.206 on Fri, 10 Jan 2014 11:02:23 AM

manifestationsof the same self-regulatingequilibriumsystem (see Maruyama1963).

Choice of Approachand Predictions

This content downloaded from 200.54.94.206 on Fri, 10 Jan 2014 11:02:23 AM

This content downloaded from 200.54.94.206 on Fri, 10 Jan 2014 11:02:23 AM

This content downloaded from 200.54.94.206 on Fri, 10 Jan 2014 11:02:23 AM

This content downloaded from 200.54.94.206 on Fri, 10 Jan 2014 11:02:23 AM

Table 1. Model life tables utilized in the simulations.

1) MT:25-40 (Weiss 1973:144)

0-1 .3330 4.4%

2) MT:27.5-45 (Weiss 1973:154)

0-1 .3000 3.9%

This content downloaded from 200.54.94.206 on Fri, 10 Jan 2014 11:02:23 AM

The Outcome of the Simulation

This content downloaded from 200.54.94.206 on Fri, 10 Jan 2014 11:02:23 AM

Table 2. Summary of 40 simulation runs.

1 2540 175.4 100 M 119.1

This content downloaded from 200.54.94.206 on Fri, 10 Jan 2014 11:02:23 AM

This content downloaded from 200.54.94.206 on Fri, 10 Jan 2014 11:02:23 AM

Table 3. Paired life tables run with identical assumptions.

Life Table Run Median MES Mean MES

25-40 18 150 236.7

25-40 9 150 233.5

2540 19a 125 213.4

2540 31 150 271.9

2540 15 200 331.3

Mean Fertility Level

1 119.1 69.4 100 175.4

3 127.4 71.1 100 166.7

5 126.4 71.6 100 198.5

6 127.2 70.6 125 220.6

9 122.0 69.8 150 233.5

This content downloaded from 200.54.94.206 on Fri, 10 Jan 2014 11:02:23 AM

This content downloaded from 200.54.94.206 on Fri, 10 Jan 2014 11:02:23 AM

MES MONOGAMY POLYGYNY MES

350 -+- -+- 350