Cultivating Diversity: Field Scattering as Agricultural Risk Management in Cuyo Cuyo, Department of Puno, Peru

By Carol Goland, 1993.

Chapter 2 - Theoretical Foundations of a Risk Perspective

(page 12)

INTRODUCTION

All living organisms must cope with variability in the environmental factors affecting their health, reproduction, and survival, including the availability of important resources. Natural disasters (e.g., droughts, earthquakes, floods) and cultural ones (e.g., civil war) may lead to widespread famine, starvation, and social disruption (see Devine 1988; D'Souza 1988; Franke and Chasin 1980; Garcia and Spitz 1986; Seavoy 1986, Will 1990). Individuals and communities may have few means to counter the effects of sudden catastrophes such as these. But not all environmental variability is so massive in scope and impact. Localized and less severe environmental perturbation is more routine. We should probably consider it normal. But it is no less dangerous to the continued functioning and survival of human communities. Low-level, relatively frequent, environmental disruptions require practices that shield people from the recurrent threat of scarce resources. These practices presumably are routinely incorporated into production decisions and activities (Campbell and Tretcher 1982; Colson 1979; Dirks 1980; Halstead and O'Shea 1989; Richards 1939).

Humans do not simply react to environmental change. Much more so than other species, they play an active role in altering their environment and in mediating its effects before disruption becomes severe and widespread. Events in nature become hazards only in the context of interaction with human systems (Burton et al. 1978). At the same time, human action may be structured to reduce the damage created by environmental variability. Much of cultural behavior has been interpreted in this context of adaptation.

The focus of this analysis is the strategies used by farmers in two peasant agricultural communities to buffer themselves from production shortfalls that may result from routine environmental variability. While not denying the potential importance of other resources, I emphasize the availability of food. This work builds on a well-developed anthropological tradition of concern with cultural adaptation to environmental problems in the subsistence quest.

My discussion proceeds in several stages: (1) I introduce adaptation as the central concept guiding this analysis and consider difficulties surrounding its definition and operationalization; (2) this leads me to a discussion of the analysis of environmental variability; (3) I then review some of the more recent anthropological approaches to environmental variability and the common cultural mechanisms for coping with it; (4) I present precise definitions of risk and uncertainty, and review formal risk reduction models drawn from micro-economics and evolutionary ecology; and (5) with this framework, I focus on the central question of this research, the analysis of agricultural field dispersion as a strategy for coping with risk.

ADAPTATION

(page 13)

The concept of adaptation has been central and integrative in anthropology. Cultural patterns, especially subsistence practices, have been seen as adaptive responses to particular environmental conditions and variables. Despite its centrality, usage of the concept has been problematic. This is due in part to the several meanings it may denote and frequent imprecision about which definition is intended. This problem is found in both biological (Lasker 1969; Mazess 1975) and anthropological (Alland 1975; Kirch 1980) discussions.

Dobzhansky (1968) distinguishes four concepts related to adaptation. An adaptive trait is a specific feature of an organism which is related to its function, e.g., the opposing thumb is an adaptation to grasping behavior. This is a common use of adaptation in biology, where structure a is described as necessary for producing behavior b. Alternatively, the concept of adaptedness may be used to describe the relative fitness of an organism in a particular environment, that is, a state of being: the organism is adapted to its environment. A more dynamic process is reserved for the term adaptation. It denotes the process of change whereby a better fit between the organism and its environment is achieved. In this sense, adaptation is evolutionary, and change is transgenerational (Alland 1975).

Dobzhansky (1968) also identifies adaptability as the ability to become adapted. Adaptability is conferred by the capacity to adjust to environmental change through genetic, behavioral, or physiological mechanisms, and thus is dependent on the maintenance of flexibility. Which of these mechanisms is most effective in the long-term is determined in part by the rate and regularity (predictability) of environmental change relative to an organism's generation length (Lewontin 1966; Slobodkin and Rapoport 1974). Behavioral mechanisms are most important when environment changes several times within a generation. When time between change is lengthier, physiological mechanisms may well be more important. Since they are less reversible, genetic response to change represents an evolutionary commitment--persistence for long periods of time-- that constrains future opportunities. Ultimate evolutionary success, according to Slobodkin and Rapoport (1974), may depend on appropriate selection among these mechanisms in responding to environmental change, in order to avoid over-commitment to more costly responses (i.e., genetic) which may affect capacity for response to future events. Behavioral and physiological changes (what Mazess [1975] terms "aptitudes") are first-line defenses which presumably will be engaged before a genetic response.¹

Adaptedness and adaptation cannot be measured in an abstract sense; they are meaningful only within the context of specific environment. Further, the concept must be accompanied by some criterion for assessing fitness. In biology, reproductive success (or some proxy for it, e.g., energy capture) is used most often. Adaptation is predicted by Darwinian theory as a common (though not necessary) outcome of natural selection. Neo-Darwinian theory cannot predict an organism's specific response to an ecological problem, but only that organisms should be adapted (Slobodkin and Rapoport 1974). The analyst must determine what qualities of the environment create selection pressures and what features of the organism provide the basis for an adaptive response. In anthropology as well as biology, these determinations are both an analytical and theoretical challenge.

At its most general level, the present work is a study in adaptability. It focuses on the routine strategies used by Andean agriculturalists to cope with environmental variability. As

(page 14)

on-going responses to common environmental stresses, these adaptations have two important characteristics: they are highly integrated into the production system, and they allow producers to maintain the flexibility needed to respond to potential future stresses. The effectiveness of such routine strategies, and the flexibility they protect, are essential components of adaptability and resilience.

THE ANALYSIS OF ENVIRONMENTAL VARIABILITY

For successful analysis of adaptation--including aspects of adaptedness and adaptability--recognition of environmental variability is vital, since the factors which affect adaptive processes are unevenly distributed in a given environment. Emphasis on normative properties of environment have hampered anthropological approaches to adaptation. Winterhalder (1980) lists the analytical problems this creates. (1) Observed behaviors may be matched to the wrong cause, given "lag time" between environmental change and adaptation. Thus, observed behaviors may be the result of adjustment to past environmental conditions, not extant ones (cf. Mayr 1983; Lewontin 1966; Slobodkin and Rapoport 1974). (2) Limited temporal sampling and an assumption of environmental stasis may result in underappreciation of the dynamics of on-going adaptations (cf. Wiens 1976). (3) Conversely, normative conceptions may fail to recognize that abrupt environmental change does occur and may be causal in subsistence, technology, or other cultural changes. Rates of environmental change cannot be determined a priori. (4) Environments may be classified in generalized categories and matched to idealized cultural features. While this may be useful for correlational studies, alone it sheds little insight on the processes responsible for given adaptations (cf. Mazess 1975). (5) Use of normative descriptions risks denying the partially historical quality of environmental factors, and their ecological and evolutionary significance (cf. Lewontin 1966; Richerson 1977). (6) The normative approach poses explanations in static functional terms, in part because in the context of a fixed environment, emergent forms cannot be related to environmental selection.

Burton et al. (1978:22-23) consider how to measure and describe environmental events in a manner appropriate to the analysis of human response. They suggest seven important dimensions of a perturbation: (1) magnitude (the size); (2) frequency (how often it happens); (3) duration (how long it lasts); (4) spatial extent (how large a space it effects); (5) speed of onset (time between first appearance and peak); (6) spatial dispersion (pattern of occurrences in space); and (7) temporal spacing (pattern of occurrences in time). These dimensions parallel many of the concepts used by ecologists. To be meaningful, all must be discussed in relation to the particular organisms under study; each is reviewed below. Given the specific focus of this study, I consider these dimensions of environmental variability in the context of human decisions about agricultural production.

Magnitude (1) is perhaps the most difficult of these concepts to define. The most common measure is deviation from the average, in absolute or standardized terms (i.e., standard deviations). While this imparts important information about the range of variability, it says little about the impact of these changes on humans or other species. Evaluation of this impact remains an empirical question. For example, a drought may have minimal effects on urban residents, but devastate rural agricultural producers. This is an instance of the general observation that adaptation--and challenges to adaptedness--are context-specific.

The temporal dimension of environmental variability contains several aspects: how often (2), and how quickly (5) it happens, how long it lasts (3), and how predictably it occurs (7). Frequency is an important determinant in the maintenance of routine responses. If pest infestations are a common occurrence in agricultural fields, insecticides or other available

(page 15)

measures will be used consistently; if they are infrequent, the high cost associated with this input may not be warranted. Frequent events are also more easily remembered (Kirkby 1974), along with the appropriate responses to them. The rapidity of onset constrains ability to respond. If a sudden wind storm strikes, there is little that can be done to protect a crop. In contrast, the slow progression of drought can be met with a series of increasingly serious and irreversible responses, from re-seeding, to hauling water to fields, to eventual abandonment and out-migration. Closely related is the length of time a perturbation lasts. For example, plants may be able to tolerate temperature chills of several hour's duration, but are likely to be harmed by more prolonged exposure over the course of several days.

The final dimension of temporal variability is its statistical regularity or predictability (7). Predictability has two components: constancy, and contingency (Colwell 1974). Here predictability is measured by examining an environmental variable during each interval (e.g., seasons) within a larger time period (an annual cycle). For example, if the state of the variable remains the same during all seasons, complete predictability arises from constancy. On the other hand, if the variable changes during the annual cycle, but returns to the same state seasonally, year after year, then complete predictability is a consequence of contingency. These two aspects of predictability have unique implications for human response. Predictability from constancy implies lack of variability, and in the framework used here often requires no buffering mechanism.² But events which arise from contingency require response, ceteris paribus. Seasonal fluctuations are highly predictable, while fluctuations over longer cycles tend to be less so. When fluctuations are predictable, response may be equally regular and explicit. Most buffering strategies are responses to this type of variability (Dean et al. 1985). In temperate latitudes, reasonably predictable frost-free periods determine the agricultural season. Less predictable occurrences, and those that occur over longer cycles, make response more difficult. When the time scale becomes so great that the periodicity and cyclicity cannot be recognized (a problem especially for societies lacking detailed historical records), it is unlikely that buffering mechanisms or responses will be routinely integrated or recognized (Dean et al. 1985; Slobodkin and Rapoport 1974).

The spatial dimension includes aspects of extent (4) and homogeneity (6). Droughts generally occur over large spatial extents, but microclimatic and microtopographic variation may create heterogeneity or patchiness at a much smaller scale. In contrast, hail storms tend to occur over a somewhat limited spatial area, but are homogeneous within the area of impact. Spatial variation is likewise organism-defined. Its specific dimensions affect its recognition by the organism, and thus whether the organism will be able to respond (see Wiens 1976 for discussion of spatial heterogeneity in ecological theory).

In summary, dimensions of environmental variability have distinct implications for the form and effectiveness of alternative coping strategies. Spatial and temporal scales, magnitude, and predictability are key determinants of how environmental fluctuations will be experienced and countered. Explicit recognition and analysis of these factors is essential for identifying buffering strategies and for understanding why they are incorporated into the behavioral repertoire of a population. This study is situated in an environmental context of extreme spatial heterogeneity and frequent but unpredictable temporal fluctuations. These factors shape key features of the agricultural production system throughout the Andes (Chapter 3) and in the communities studied (Chapter 4).

ANTHROPOLOGICAL APPROACHES TO COPING WITH VARIABILITY

(page 12)

Anthropologists have had few opportunities to document response to large-scale environmental disruption. Only by chance do they find themselves in the midst of disaster (the analytically "lucky" anthropologists include Firth [1959], Schneider [1957], Scudder 1962], and Spillius [1957]). This is in part a consequence of a field work tradition which relies on a single "ethnographic year." The probability of observing disaster in any given place and time is slim. The traditional ethnographic year also creates other problems: there is a tendency to uncritically assume that the year observed is representative of all years, that cultures and environment exist in a fully recurrent pattern (Colson 1979; Winterhalder 1980). It is difficult to gain appreciation for the range of variation with which communities must cope from year to year, and thus difficult to recognize the coping mechanisms that have been routinely integrated into cultural practices. At all levels (individuals, households, communities, etc.) strategies to reduce vulnerability to food shortage presumably are incorporated into the behavioral repertoire (Halstead and O'Shea 1989). Year-to-year environmental variability is normal, and though its precise timing and scale cannot be predicted with certainty, the fact that it will occur can. This paradox of "certain unpredictability" is the basis for the use of buffering mechanisms to protect against the possibility of loss. These are the first-line defenses routinely engaged to counter environmental variability and change. As low-level responses, they attempt to reduce the impact of potential foreseeable shortfalls and solve current problems, while at the same time maintaining flexibility for response to future events (Minnis 1985).

Waddell (1975, 1989) presents an insightful analysis of the strategies used by the Enga (Papua New Guinea) to cope with frost. Frost is a recurrent environmental perturbation. Minor frosts occur almost annually while the frequency of a major killing frost is about once every 30 years. The Fringe Enga, occupying lands above 2300 m, are exposed to the most severe frosts, while the Central Enga, with territories at elevations between 1500 m and 2200 m, experience only mild and infrequent frosts. The buffering strategies used by the Enga reflect a set of hierarchical responses. At the local level, responses to the potential for frost are incorporated as a routine aspect of agricultural production for both Fringe and Central Enga. All agriculture is based in large mulch mounds. These mounds protect the crops from radiation frosts. Further, gardens are spread between two ecological zones. Valley bottoms are valued for their high productivity, but are more vulnerable to frost due to cold air inversion. Other gardens, maintained on slopes, are more protected from frosts but cannot be as intensively cultivated. This spatial arrangement of productive activities protects the population from total disaster when minor killing frosts strike, once every few years. These low-level responses are in continuous operation.

Higher level responses are engaged only as conditions worsen. The next level identified by Waddell is the intraregional level. Fringe Enga clans hold lands widely dispersed over the landscape, and families of each clan have access to several fields in more distant locations. Although these lands are still within the frost zone of the highlands, topographic differences moderate the severity of frosts. In any given year, produce from these distant gardens may be critical if local yields are reduced by frost. These far gardens may also be used as the source of seed stock in case of extremely damaging local frosts. The final level of response is extraregional. When frosts are long and severe, the Fringe Enga population moves to lower altitudes. Ties to lower-altitude dwellers are constantly reinforced through exchange. The Fringe Enga have several commodities highly valued by Central Enga. The regular exchanges between individuals in the two groups are obligating to both sides. The exchanges insure the Central Enga with commodities they value highly, and they underwrite the hospitality which provides the Fringe Enga places of refuge during periods of extreme stress.

(page 12)

These strategies provide the Fringe Enga with diverse responses to environmental perturbation. Local-level responses are continuous and highly integrated aspects of production strategies. The higher-level responses are engaged only as the stress intensifies. Importantly, use of the lower-level responses does not constrain the options available if conditions worsen.

Although there are a growing number of empirical studies that touch on individual and community response to environmental variability, synthetic treatments are less common. Two notable exceptions are presented by Colson (1979) and Halstead and O'Shea (1989). In both studies, the authors describe cultural responses to food scarcity or its threat. Both explicitly consider the dimensions of environmental variability (time scale, spatial scope, and magnitude) as key determinants of the effectiveness of alternative responses. Both recognize hunger (or the potential for it) as a recurrent and formative feature of human adaptation.

Colson (1979), drawing on her long field work experience in the Pacific Northwest and central Africa, reviews responses to resource scarcity. She considers five routine devices to lessen vulnerability to environmental disruption: (1) diversified economies; (2) food storage; (3) preservation and transmission of knowledge about appropriate responses to hunger; (4) conversion of food surpluses into more durable goods which can be traded for food in emergencies; and (5) maintenance of social relations entailing obligations for assistance in times of need. Colson also recounts a set of hierarchical responses that may occur when these routine buffering mechanisms fail, and as food scarcity spreads³ (in addition see Dirks 1980).

Halstead and O'Shea (1989) present a similar classification of responses to variability: (1) mobility; (2) diversification; (3) physical storage; and (4) exchange. In their scheme, diversification and storage correspond with Colson's usage. Halstead and O'Shea's "exchange" corresponds in part to Colson's fourth and fifth devices: exchange is considered as converting present abundance into future obligation, using a social currency. Halstead and O'Shea's use of mobility embodies in part Colson's view of social relations, since mobility may be predicated on the maintenance of kin and social networks. Each of these strategies is considered more fully below.

Transmission of Knowledge

Information encoded in ritual and oral traditions provides important clues about the condition of the environment and appropriate response to variability (Rappaport 1971b, 1979:97-144; Minc 1986; Minc and Smith 1989). This knowledge may include information about famine foods, appropriate social behaviors (for example, pooling resources), and alternative resource areas. The value of this form of information storage is its potential longevity: it overcomes the limited experience of living individuals, and in this sense represents a collective and enduring memory. This may be critical when environmental perturbations occur at intervals greater than generation length.