GROUP SIZE

Intro

A fundamental aspect of the ecology of any species is the pattern of spatial aggregation or dispersion of its individual members: Do individuals clump into large groups (like starlings) or spread out as solitaries or pairs (like spotted owls)? Do they aggressively defend fixed territories against other members of their species (like gibbons) or do they move freely over the landscape seeking resources wherever they can find them (like caribou)?

Ecologists want to do more than simply catalogue these various patterns of spatial organization; they want to explain them using principles of ecological adaptation

Students of human ecology have asked some of the same questions, and some have begun using the same explanatory principles to answer them; in addition, various social sciences concerned with systems of land tenure and property rules have developed related theory and bodies of data that are influential in ecological anthropology

Here and in the subsequent lecture notes, I survey some of these issues, specifically:

1) What principles govern variation in group size (task groups, settlements, etc)?

2) What are the major forms of land tenure, and why do they vary?

3) What is the "tragedy of the commons," and what ways have societies developed to prevent it?

Ecology and Group Size

One fundamental aspect of spatial organization is the degree to which members of a population are aggregated vs. more evenly dispersed

If we define aggregating as co-residence, we can rephrase this as question of the average settlement size

(Note that settlement size is in principle independent of population size or density:  in a given area, the same 1,000 people could be concentrated into a single town, or they could be rather evenly dispersed into 10 hamlets averaging 100 people)

There are a number of ecological factors that could influence whether larger or smaller settlements are more adaptive (beneficial to the survival and reproductive success of the inhabitants)

For example, spatial distribution of key resources might be concentrated or dispersed, which in turn might strongly influence optimal settlement pattern

This is ecological rather than environmental determinism, because relevant resources will differ according to technology, the social system, etc. (e.g., for same environment, foragers' resources might be dispersed, while arable land might be concentrated)

In many cases, the ecological relationship between the spatial distribution of resources and that of people might be less direct, as in the advantages of information-sharing in locating unpredictable caribou herds, or the need to cooperate in defense against enemy raids, either of which might favor aggregation into larger settlements

Whatever the local balance of costs and benefits, it might be productive to test the hypothesis that these will be shape the settlement pattern found in a given locale or population

This same general cost-benefit logic can also be used to generate testable hypotheses to explain variation in shorter-term groupings, such as task groups (e.g., a group of people who cooperate in clearing a swidden or mounting a communal caribou hunt)

In fact, short-term or special-purpose task groups should be easier to analyze this way, since they are likely to involve fewer different goals (each of which might entail conflicting costs and benefits)

For heuristic purposes, let's consider such a group, for example a hunting party focusing on capture of a single type of prey (e.g., jackrabbits or caribou)

Suppose that the return rates from this type of hunting vary by group size:

Group Size	Group Catch/Day	Catch/Person/Day
1	50 kg	50 kg
2	150 kg	75 kg
3	300 kg	100 kg
4	600 kg	150 kg
5	650 kg	130 kg
6	600 kg	100 kg
7	500 kg	70 kg
8	320 kg	40 kg

(to see these data in graphical form, click here)

Now, suppose we ask the question, "What's the ideal group size for hunting (in this situation)?"

Answer depends on what you assume is the relevant unit for optimization:

1) If it's individual hunter, n=4 is optimal (maximizes per capita return)

2) If it's entire group (including perhaps those who stay at home) optimum = 5 (maximizes total catch: 650 kg)

Let's suppose that the dominant decision criterion is per capita maximization; does it follow that size of hunting groups will actually be 4?

Well, if total number of hunters is 4 (or any multiple of 4), then perhaps yes

But what if there are only 5 hunters in the whole camp?

The first 4 hunters would be best off hunting together in a group of 4, since 150 kg > 130 kg

But the 5th hunter would be much better off joining the group rather than foraging alone (130 kg > 50 kg), even though this would reduce the per capita share of existing members (from 150 kg to 130 kg)

Same logic applies if a 6th hunter comes along; he stands to double his return by joining the group (from 50 kg for a day of solitary hunting to 100 kg as his share in a group of 6), but this depresses the share of other group members by about 25% (130 kg vs. 100 kg)

So we might then ask a different question:  What is the equilibrium group size (i.e., size at which no one can do better by joining or leaving the group)?

Answer = 7, since additional hunters will benefit from joining the group until there are 7, at which point an 8th hunter would do better hunting alone (50 kg > 40 kg) (again, these may be easier to grasp in graphical form)

(In fact, if there are 7-8 hunters, everyone will benefit by splitting into two groups of 3-4, though with 7 hunters there may be conflict over who's in the group of 4 hunters vs. group of 3)

In other words, what we have here is a conflict of interest:  the optimum for any one decision-maker depends on whether you're a member of an existing group vs. a prospective joiner

The details of this model are discussed further by Boone (pp 302-305), and graphed in his Figure 10.1  [Member-Joiner Graph]

This model can't provide a general solution to this conflict of interest, nor can we assume that members will prevail (i.e., be able to keep group size at their optimum), because there are likely to be costs to excluding additional members, and these costs are themselves a collective good (all group members will benefit whether they pay the costs of exclusion or not)

What this model can do, though, is alert us to the logic of conflicting interests, and suggest when they will arise: 1) when n_o+1 gets higher per capita return than single individual (where n_o = optimal group size for an individual member), and 2) individuals value self interest more than group interest (i.e., are not altruists)

One way to reduce or eliminate the conflict of interest is to have a different distribution rule; for example, if rather than sharing the catch among the members of the task group, all producers return to the settlement and then pool their (individual or group) catches before dividing equally, then the member-joiner conflict disappears (see Boone reading, pp. 304-5, for details on this "central-place sharing" model and the additional complications it raises)

I've posed the member-joiner conflict in terms of cooperative hunting, but same general issue arises in a myriad of ecological, economic, and political contexts; in fact, it is one way of formalizing the "tragedy of the commons" concept (on which, see the next set of lecture notes on "Land tenure...")