The Behavior Analyst

1995, 18, 1-21

No. 1 (Spring)

Rules, Culture, and Fitness

William M. Baum University of New Hampshire
Behavior analysis risks intellectual isolation unless it integrates its explanations with evolutionary theory. Rule-governed behavior is an example of a topic that requires an evolutionary perspective for a full understanding. A rule may be defined as a verbal discriminative stimulus produced by the behavior of a speaker under the stimulus control of a long-term contingency between the behavior and fitness. As a discriminative stimulus, the rule strengthens listener behavior that is reinforced in the short run by socially mediated contingencies, but which also enters into the long-term contingency that enhances the listener's fitness. The long-term contingency constitutes the global context for the speaker's giving the rule. When a rule is said to be "internalized," the listener's behavior has switched from short- to long-term control. The fitness-enhancing consequences of long-term contingencies are health, resources, relationships, or reproduction. This view ties rules both to evolutionary theory and to culture. Stating a rule is a cultural practice. The practice strengthens, with short-term reinforcement, behavior that usually enhances fitness in the long run. The practice evolves because of its effect on fitness. The standard definition of a rule as a verbal statement that points to a contingency fails to distinguish between a rule and a bargain ("If you'll do X, then I'll do Y"), which signifies only a single short-term contingency that provides mutual reinforcement for speaker and listener. In contrast, the giving and following of a rule ("Dress warmly; it's cold outside") can be understood only by reference also to a contingency providing long-term enhancement of the listener's fitness or the fitness of the listener's genes. Such a perspective may change the way both behavior analysts and evolutionary biologists think about rule-governed behavior. Key words: rule, rule-governed behavior, culture, fitness, evolutionary theory, rule giving, rule making, rule following, bargain

This progression should impress psychologists in general and behavior analysts in particular. As it continues, the interests of evolutionary biologists overlap more and more with those of anthropologists, sociologists, psychologists, and behavior analysts. What will happen to behavior analysis if the evolutionists never hear about it? Works like Richard Dawkins's The Selfish Gene (1989), widely read and acclaimed by biologists and laypeople, discuss operant behavior and culture with no reference to behavior analysis. It is inevitable that behavior analysis will be integrated into evolutionary biology; the question is how this will occur. One possibility is that evolutionists This paper is gratefully dedicated to my teacher, Richard J. Herrnstein. A version was will raid our treasury of concepts and presented at the Association for Behavior Anal- take what they find useful while inysis meeting in Atlanta, May 1994. The author venting or reinventing other concepts thanks P. N. Hineline, A. S. Kupfer, J. A. Nevin, H. Rachlin, M. E. Vaughan, and G. E. Zuriff for for talking about human behavior and culture. If that happens, behavior analhelpful comments on earlier drafts. Address correspondence and reprint requests ysis will become a historical footnote, to the author at the Department of Psychology, University of New Hampshire, Durham, New a short-lived movement within psyHampshire 03824-3567 (E-mail: wm.baum chology just before the Darwinian revolution took over. Another possibility @unh.edu).
1

It took about 70 years before the biological community completely accepted Darwin's theory of natural selection. From the outset, evolutionists discussed species-typical behavior, but it was only a matter of time before discussion advanced beyond fixed-action patterns. After the synthesis of Darwinian evolution with genetics in the 1 930s, biologists talked increasingly about individual organisms' interactions with the environment. In the 1960s, with the rise of behavioral ecology and sociobiology, they began to talk about learned (i.e., operant) behavior and culture.

WILLIAM M. BAUM

is that behavior analysts will explicitly build bridges to evolutionary biology and convince the evolutionists that they may profit by listening to us. I have tried to make a start on this in my book, Understanding Behaviorism: Science, Behavior, and Culture

When Watson founded behaviorism with his 1913 manifesto, "Psychology As the Behaviorist Views It," he tied the study of behavior explicitly to evolutionary theory (Baum, 1994a). Within 10 years, however, he reverted to the earlier anthropocentric viewpoint that had characterized psychology since its beginnings (Logue, 1978, 1994). Instead of continuing to relate human and animal behavior to evolutionary history as well as to individual experience, he focused on learning and individual experience to the exclusion of phylogeny. As a result, behaviorism is described in introductory psychology textbooks today as radical environmentalism (Todd, 1994). The most often quoted excerpt from Watson is his boast about what he could do with a dozen healthy infants. As long as evolutionists get their impressions of behaviorism from such descriptions as these, they will continue to regard behaviorism and behavior analysis as irrelevant. It is time for behavior analysis to come full circle and recognize that Watson was right the first time. Our choices are to reach out, get on the field and play the game, or remain forever on the sidelines.
WHY TIE BEHAVIOR ANALYSIS TO EVOLUTIONARY THEORY?

(1994b).

behavior analysts have been slow to integrate evolutionary theory into their discussions of behavior. Perhaps this is due to a lingering anthropocentrism, left over from behavior analysis's roots in psychology. Whatever the reason, the absence of an evolutionary context represents a weakness and a danger for behavior analysis. I discussed the problem at greater length earlier (Baum, 1994a). In brief, evolutionary biology is making inroads into subject matter traditionally assigned to psychology, topics such as decision making, learning, and, in particular, human social behavior and human culture. The biologists' progress derives from the logic of the Darwinian paradigm, which considers humans and human culture to be just particular instances of more general concerns about how natural selection works. Boyd and Richerson (1985)-a biologist and an anthropologist-adapting a quote from Darwin, argued that "Trying to comprehend human nature without an understanding of human evolution is 'like puzzling at astronomy without mechanics' " (p. 1). It was inevitable that evolutionists would come to be interested in the same subjects as behavior analysts. The only question is how behavior analysts will relate to evolutionary biology. The danger here is not of being wrong, but of becoming irrelevant. Evolutionary biology will push ahead because it has the weight of the Darwinian revolution behind it. To remain relevant, behavior analysis must discard anthropocentrism and embrace

evolutionary biology.
Rules As an Example
In this paper, I shall discuss rules and rule-governed behavior as an example of a topic that requires an evolutionary context for full understanding. Perhaps nothing is so uniquely human as rule-governed behavior. This explains much of its fascination, but also raises the danger of anthropocentrism. When we see rules as uniquely

Logically, there is every reason for behavior analysis to align with evolutionary biology. The two fields' interests increasingly focus on the same subject, their common interest in behavior concerns its function in exchange with the environment, and they share the same historical, selectionist mode of explanation. Yet, with a few exceptions (e.g., Glenn, 1991; Petrovich & Gewirtz, 1991; Staddon, 1983),

RULES, CULTURE, AND FITNESS

human, we may forget to relate them to principles that apply to behavior in general, principles that derive from evolutionary theory. Since Skinner's (1969, 1974) discussions of rule-governed behavior, behavior analysts have written a lot about rules. There is one entire book (S. Hayes, 1989), large portions of other books (e.g., L. Hayes & Chase, 1991) are devoted to the topic, and experimental work appears regularly in the Journal of the Experimental Analysis of Behavior. None of this discussion, however, has placed rules in the larger context that Skinner's ideas logically implied: the context of culture and evolution. Glenn's (1988, 1991) discussion of the relations among behavioral, cultural, and biological evolution represents a step in this direction. It makes sense to discuss rules in this context because the rules of a culture are an important part of its practices. The discussion of rules cannot be complete without an account of their place in culture and their origin, like other practices, in a history of evolution, both cultural and genetic. Some of what I shall say will seem familiar. I shall begin by reviewing some basic concepts about evolutionary theory. Then I shall move to defining rules in a way that places them in a cultural and evolutionary context. In a sense, the whole point will be to define rules in a new way, consonant with evolutionary theory. I shall define a rule according to three criteria. First, a rule is a verbal discriminative stimulus, given by a speaker because of its likely influence on the behavior of a listener. Second, the rule signals a short-term contingency in which some target behavior of the listener may be reinforced by the speaker. Third, the rule is occasioned by a long-term contingency on the same behavior, a contingency with its own discriminative stimuli and long-term consequences that affect the listener's fitness or the fitness of the listener's genes.

BEHAVIOR ANALYSIS AND EVOLUTIONARY THEORY Evolutionary history affects behavior in a variety of ways, acting on the structure of the body by way of genetic evolution (Baum, 1994b). By this means, it provides species-typical patterns of behavior that emerge as a result of normal development. These are important in their own right, but they also make up the raw material of operant behavior (e.g., Segal, 1972). A striking example in human beings is children's early babbling and approximations to speech. Natural selection assures sensory sensitivity to important environmental events, that certain behavior is easily acquired or easily comes under the control of certain stimuli, and so on. To fully understand rule-governed behavior, probably all of these influences need to be considered, because rules are environmental events associated with conspecifics and come easily to control a variety of operant behavior. To understand the acquisition of rule-governed behavior, we need at the least to remember that natural selection ensures that certain events function as reinforcers and punishers. Reinforcers and punishers that are shared by all the members of a species may be called species-typical reinforcers and punishers. They may result simply from normal development, with no need for experience-the smile, for example, is both a fixed-action pattern and a reinforcer-or they may just be so readily acquired that everyone acquires them in the normal course of events; just as ducklings normally imprint on their mother, so human beings normally acquire reinforcers as diverse as candy, physical affection, and vocalizations of approval. Species-typical reinforcers and punishers, like any traits, are selected by their long-term effects on reproductive success (i.e., fitness). Given that individuals vary and that some of the variation depends on differences in genetic inheritance, the genotypes of reproduc-

WILLIAM M. BAUM

tively more successful individuals will be represented more frequently in the gene pool of the species. It is often convenient to speak of fitter individuals, but strictly speaking, and whenever there is any conflict between the individual and the genes, fitness is considered to be a property of genes. Genes reproduce more or less successfully by virtue of the individuals in which they occur. Relations to fitness are nearly always molar, in the sense that they apply on the average and in the long run, rather than in individual instances. This applies to reinforcers and punishers: A reinforcer usually increases fitness, and a punisher usually decreases fitness; these relations hold on the average and in the long run. Usually eating enhances fitness, but an individual organism that eats well but dies without reproducing cannot be said to have had its fitness increased by eating. Like any molar relation, the relation between eating and fitness cannot be observed in individual instances. The genes that provide the basis for eating to function as a reinforcer are selected despite such unlucky individuals, because those individuals possessing those genes, as a type, reproduce more often than other types. The same holds for genes that make a smile a reinforcer and for genes that make a frown a punisher. This line of reasoning will be important to our discussion of the long-term consequences of rule-governed behavior and their impact on fitness.

then turn to rule giving. We begin with an analysis of rule following in familiar terms, in which a rule is seen as a discriminative stimulus (SD) generated by the (verbal) behavior of the rulegiver (speaker) that increases the likelihood of some operant (rule-governed) behavior on the part of the rule-follower (listener).
Two Contingencies A rule relates to two contingencies (cf. Cerutti, 1989; Zettle & Hayes, 1982). In one, which is short term and social, it participates as the discriminative stimulus. I shall call this the proximate contingency (cf. Alcock, 1993). The other, which is long term and ultimately related to fitness, occasions the giving of the rule and involves the same behavior as the proximate contingency. I shall call it the ultimate contingency (cf. Alcock, 1993). In technical terms, relative to the proximate contingency, the speaker's giving the rule is a mand; relative to the ultimate contingency, it is a tact (Skinner, 1957). Figure 1 diagrams the two contingencies in symbols and gives an example. Each contingency is represented in the basic form SD:B > SR, where SD represents a discriminative stimulus, B represents operant behavior, SR represents reinforcement, the colon represents the occasion-setting function of SD and its effect of increasing the likelihood of B, and the arrow indicates that B makes SR more likely. The two contingencies share the same B, but each has its own SD and its own SR. (Strictly speaking, B is not exactly the same for the two contingencies; it overlaps between them-a point we shall address later.)

RULES AND FITNESS Evolutionary theory offers a way to understand both why people follow rules and why people give rules. We shall take up rule following first, and

Figure 1. The two contingencies of rule-governed behavior. A: The shaded ellipse indicates the proximate contingency. B: The shaded ellipse indicates the ultimate contingency. C: An example, in which the rule (proximate SD), "eat vegetables," occasions the behavior of eating vegetables; this behavior is reinforced in the short run by approval. Availability of vegetables might ultimately occasion the same behavior, which is ultimately reinforced by increased fitness as a result of improved health, resources, relationships, and reproduction (HRRR). Both contingencies conform to the standard pattern SD:B -* SR.

A
Proximate Contingency

IS

,#"Ultimate 0D.
%lb
Ultimate

~~~~~~~Ultimate SY

Contingency

-Proximate -/

~~~Contingency

=S-D *
S

Proxima'te 8k

,Ultimate

S.

Ultimate Contingency

_~'eci7

Proximate
_

~~~~~~Contingency

_approal) tEat veai

0.

*
n i

sight anif4
signs of

etables"*aproa veggies
e.

Ivegetables

HRRROs
Oa

WILLIAM M. BAUM control and the short-term social reinforcement maintains the behavior), then the behavior constitutes rule following or may be said to be rule governed. If control shifted to the longterm contingency, then the behavior would no longer be rule governed, because the ultimate discriminative stimuli would control (govern) the behavior and the ultimate reinforcement would maintain it. In the example in Figure IC, eating vegetables would come under control of the environmental conditions that indicate their availability and would be maintained by good health. When one speaks of a rule being "internalized," it means that control over the behavior has shifted from the proximate contingency to the ultimate contingency. Although the mentalistic view has the rule move inside, in the behavior-analytic view the change is from one form of environmental control to another. The behavior remains regular, but has ceased to be rule governed and has switched to long-term control. Much human behavior starts out as rule-governed behavior and switches to long-term control. This usually occurs when one person instructs another, for example. The shift is likely if the ultimate reinforcer is powerful and relatively conspicuous, like money or opportunities to reproduce. A novice salesperson is told, "Always ask for more than you expect to get." The behavior of negotiating initially is reinforced by supervisors but soon comes under the control of the extra money earned in the long run. A young woman might be instructed, "Don't trust a man unless he is willing to make a commitment to you." After some experience with men, her behavior might come under stimulus control by signs that are correlated with reliability and might be reinforced by actual reliability. Still, nothing has been internalized; control has just shifted to the ultimate contingency. When control switches from the short to the long term, the behavior

The proximate contingency. The rule, the SD of the proximate contingency, appears at the upper left of the diagrams in Figure 1. It is produced by the verbal behavior of the speaker (Bv in Figure 3). The listener hears or sees the rule, and if the proximate reinforcement, shown at the upper right, is effective, then the listener's behavior B will be strengthened or maintainedthat is, the listener will "follow" the rule. In the example (Figure IC), the speaker (e.g., a parent) says something like, "Eat your vegetables," which is equivalent to "If you eat vegetables, then you will develop properly and remain healthy." If the listener (e.g., a child) eats vegetables, the speaker provides the proximate reinforcer, which may be approval or simply the withholding of disapproval. The effect of this social contingency is to strengthen the eating of vegetables. The ultimate contingency. The SD of the ultimate contingency, shown at the lower left of the diagrams in Figure 1, may be any of a class of environmental conditions. In the example, this class might be called "sights and signs of vegetables." It includes advertisements, vegetables in the store, and vegetables on the table. The ultimate SR, shown at the lower right of the diagrams, is a result that usually enhances fitness (reproductive success) in the long run. In the example, avoidance of disease or preservation of health in the long run makes it more likely that the listener will survive long enough and remain healthy enough to reproduce successfully. Eating vegetables enhances the listener's fitness. If the speaker is a parent or relative of the listener, it also enhances the speaker's fitness. The long-term nature of the ultimate contingency is symbolized by the longer arrow from B to the ultimate
SR.

"Internalizing." Figure 1 suggests a distinction between control by the proximate contingency and control by the ultimate contingency. When the proximate contingency controls behavior (i.e., when the rule exerts stimulus

RULES, CULTURE, AND FITNESS

may also change. The rule-governed behavior of the novice may resemble later skilled performances, but the two may be far from identical. Indeed, the surest sign of the switch may be the change from more variable, deliberate performances to more coordinated, automatic performances. A new contingency is likely to select new behavior. Although Figure 1 suggests that the behavior involved in the proximate and ultimate contingencies is the same, that is a simplification. At least with the shift from novice to master, it would be more accurate to show two overlapping categories of behavior. The shift to long-term control might be common and often desirable, but it may never occur. Particularly with morals and social conventions, behavior may remain rule governed. This may be because the ultimate contingencies involved are particularly subtle. Someone who was taught to tell the truth early in life may continue to do so until death just because it "seems wrong" to lie. Truth telling might never come under the control of its longterm relation to stable social relationships; one might say it remains conventional or principled truth telling. Principled (rule-governed) truth telling might overlap substantially with ultimately controlled truth telling-it might be good enough for most purposes-but the two would be distinguishable in some circumstances. For example, a principled truth-teller might tell a painful truth in a situation in which the ultimate truth-teller would prevaricate for the sake of friendship.

Culture and Fitness I have discussed culture and its relation to fitness elsewhere (Baum, 1994b, chap. 12 and 13), but a brief summary here will help to establish perspective. From an evolutionary viewpoint, culture consists of behavior that is learned as a result of group membership and is transmitted from member to member within a group. As with other forms of learned behavior, the existence of culture depends on the selection of genes that make it possible and likely. Those genetic variants that make for culture are selected if the existence of something like culture resulted in their enjoying higher fitness than their competitors-that is, if their presence in an individual usually results in that individual's behaving so as to increase the frequency of those genetic variants in future generations. Such a view includes not only individual fitness-an individual's usually producing more surviving offspring as a result of culture-but also genetic fitness that results from altruism toward kin-an individual's usually producing more copies of genes shared with kin by helping kin to produce more surviving offspring. In other words, it includes a person's enjoying better health and more resources-factors that improve that person's reproductive success on the average-and a gene's enjoying better reproductive success on the average (e.g., because a maiden aunt assists her nieces and nephews). Culture exists because it enhances the fitness of the genes that promote it. Practices. A culture is composed of practices. These include all sorts of manufacture, social conventions, rituals, songs, stories, and sayings. Only as a result of group membership do people learn how to assemble television sets, how to greet a friend, or how to sing the national anthem. Just as genes are transmitted from parent to child, practices are transmitted from one member of a culture to another, though often from people other than parents

RULES IN A CULTURAL CONTEXT

When we turn from rule following to rule giving, explanation requires two additional steps. First, we must explain

how rule giving and rule following fit into culture. Second, we must explain how the giving of any particular rule becomes a practice of a culture.

WILLIAM M. BAUM

and to people other than children. We may learn to brush our teeth in childhood from our parents, but we also learn to play games from friends, and as adults we learn skilled performances from instructors. Only some people within a group need to engage in a practice for it to be part of the group's culture. Even if only some people in the United States tell their children the story of Hansel and Gretel, telling the story is still a practice of our culture. Even if only some people use chopsticks to eat, eating with chopsticks may still be a practice of our culture; it just occurs with a lower frequency than eating with cutlery. Many of the practices of human culture consist of giving rules. This is particularly true of conventional morals and instructions like "Thou shalt not kill" or "A penny saved is a penny earned." Sommerville (1982) argued that fairy stories play a similar instructive role. Giving advice and giving instructions tailored to specific situations constitute practices. We shall return to rule giving as a practice after we discuss rules in relation to fitness. Cultural evolution. The practices of a culture change as time goes by. Skinner (1981) and others (Baum, 1994b; Boyd & Richerson, 1985; Dawkins, 1989) have pointed out that cultural evolution can be explained by a process of selection analogous to natural selection and operant conditioning. Various practices compete with one another within the culture, and variants may coexist for a time or one variant may completely replace another. Watching television at home competes with going to the movie theater, but the two practices may coexist indefinitely. The frequency of driving a horse and buggy, however, has dropped almost to zero in competition with driving automobiles. Skinner reasoned that cultural variants are selected by their reinforcing and punishing consequences. Reinforcement and punishment, in turn, reflect effects on fitness (Baum, 1994b;

Boyd & Richerson, 1985). Yet cultural practices like smoking tobacco often seem to run counter to fitness. How can this be? Genetic evolution moves at a snail's pace in comparison with cultural evolution. The genetic bases for reinforcers like nicotine, sweets, and aggression were selected in a different environment from the one in which we live now. The environment has changed far too rapidly for any significant genetic evolution to have occurred. As a result, practices arise that afford short-term reinforcement that conflicts with longterm fitness. Cultural evolution continues, however, and new practices evolve that are selected by their long-term effects on fitness. When the ill effects of tobacco smoke become apparent, laws are passed limiting smoking, and antismoking educational campaigns are launched. Similar activities occur for aggression and eating sweets. Educational campaigns illustrate well where rules fit into culture. Exhortations and explanations to "Just say no" and "Eat right for health" are rules. The practice of giving a rule like this exists (is selected) precisely to offset relatively immediate reinforcement and to bring behavior into contact with long-term contingencies that involve fitness more directly. Comparison with cultural materialism. The views of the anthropologist Marvin Harris (1980), called cultural materialism, resemble the behavior-analytic view of culture in defining culture as behavior. In this, they both contrast with traditional anthropological definitions that rely on abstractions like values, ideas, and beliefs. Glenn (1988, 1991) discussed at length the affinity between Harris's view and the behavior-analytic view (see also Malagodi &

Jackson, 1989). Harris (1987; Harris & Ross, 1987) appeals to evolutionary theory to explain cultural practices, just as a behavior analyst might (Baum, 1994b, chap. 13). Where a behavior analyst would refer to contingencies of reinforcement, Harris refers to economic

RULES, CULTURE, AND FITNESS

considerations. Although his explanations are couched in different terms, Harris (1980) tries to accomplish within anthropology the same task that behavior analysts try to accomplish within psychology-to offer concrete explanations of behavior instead of explanatory fictions or abstractions that on examination explain nothing. The correspondence between Harris's views and behavior analysis is strong, but not perfect. Anthropologists like Harris often focus on the institutions of a culture rather than on the behavior of individuals. Whereas anthropologists assign institutional practices to groups, behavior analysts have steadfastly assigned behavior only to individuals. Behavior analysts might argue that, although an institutional practice may be characteristic of a group, its occurrence in the behavior of any individual is explained by contingencies that affect that individual's behavior. Glenn (1988), however, developed the concept of a metacontingency to bridge the gap between individual practices and institutions, attempting to capture the "interlocking" contingencies among the members of a group. Some evolutionary biologists have argued that, when groups are well defined and compete with one another, selection among individuals may be overwhelmed by group selection (Wilson & Sober, 1994). The analogous phenomenon for culture would require that different institutional practices compete either within a group or between groups. It remains to be seen whether a concept like metacontingency is necessary. We may speak of manufacturing cars as a cultural practice, but the practice of holding a job (which might be in a car factory) may prove to be a more useful unit. In this spirit, rather than calling a rule a practice, I have been calling giving a rule a practice.
An Evolution-Based Definition

According to the view presented here, a rule is a verbal discriminative

stimulus (a) that sets the occasion for a listener's behavior to be reinforced by short-term consequences arranged by the speaker and (b) that is given by a speaker's behavior under the stimulus control of a long-term contingency connecting that same listener behavior ultimately to the fitness of the listener's genes. For example, the Golden Rule, "Do unto others as you would have them do unto you," is equivalent to "If you are nice to people you meet, then they will tend to be nice to you." This is still abbreviated, because history determines what actions are nice. The two contingencies involved are: (a) If the listener acts nicely, the speaker reinforces that behavior in the short run, and (b) if the listener acts nicely toward others, those others will sometimes reinforce that behavior with helpful responses that lead directly or indirectly to better health, more resources, and opportunities to reproduce. The remainder of this paper will explain this definition further. One qualification needs to be made right away. If a person's behavior remains rule governed and never shifts to control by the long-term contingency (i.e., remains conventional or principled) one cannot say that person's rule giving is directly under the control of experiences with the ultimate contingency. The speaker may only be repeating an utterance spoken by someone else. Such secondhand conventional rule giving is under stimulus control, not of the original ultimate contingency but of a contingency between obeying the rules of a community and acceptance by the community (vs. disobeying the rules and ostracism). If group acceptance enhances fitness, this qualifies as an ultimate contingency, even if its connection to fitness is indirect (a point we shall take up in the next section). To be perfectly accurate, conventional rule giving should begin with something like, "I have heard ..." or "They say. . .." Even if a rule spreads like this from person to person, remaining secondhand, still somewhere along the line

10

WILLIAM M. BAUM
The connection between resources and fitness, like the connection between health and fitness, is indirect. Having more resources, or having the right resources, usually leads to more successful reproduction. Money, for example, makes one more attractive to a potential mate and makes successful child rearing more likely. When we come to examples like following directions to get to Boston or to assemble a table, the connection to fitness is less direct, but it is still there. Why do you want to get to Boston? Why assemble the table? Chances are, getting to Boston or having the table is connected with achieving status, fostering a relationship, or making money. The connection between a relationship and fitness can be either direct or indirect. It is most direct in a relationship with a mating partner. Rules in the form of marital counseling are common because of ultimate contingencies between successful marriages and fitness. The connection between a good relationship with a child and fitness is also direct. Reproduction is only successful if one's children are successful; hence the prevalence of rules about parenting. Rules that promote benevolent behavior toward relatives derive from the connection between helping relatives and the fitness of shared genes. Genes that make for altruism toward kin are selected by the increase in frequency of those genes that results from the reproductive success of kin (Dawkins, 1989). Caring for one's child helps the fitness of the genes shared by parent and child twice as much as caring for one's niece or nephew. As a result, relatives are favored over nonrelatives in all cultures. The rule "Charity begins at home" reflects this, as do exhortations like "Take care of your brother" and "Be good to your cousins." The connection to fitness is least direct in a relationship with someone who is neither a mate nor a relative. Because I have discussed such relationships in detail elsewhere (Baum, 1994b, chap. 11), I shall be brief here.

someone must have given the rule firsthand, under the control of experiences with the original ultimate contingency. Because, in the long run, only rules that enhance fitness are selected, rule giving is ultimately reined in by natural selection (Boyd & Richerson, 1985). Moreover, any listener hearing the rule given secondhand might rediscover its original basis, particularly if he or she fails to comply and suffers dire consequences. Having defied convention by going barefoot in town and having received a nasty cut, the speaker now gives the wearing-shoes-health rule from firsthand experience. Someone who quits smoking because of the disapproval of family and coworkers may subsequently discover an improvement in health and give the smoking-illness rule firsthand.
Rules and Fitness: HRRR

The ultimate contingency affects fitness in the long run because ultimate contingencies involve four sorts of consequences: health, resources, relationships, and reproduction (symbolized in Figure 1 by the mnemonic HRRR). An ultimate contingency usually involves more than one of these. Indeed, according to evolutionary theory, in the final analysis all consequences bear ultimately on reproduction. Genes that make for health-promoting mechanisms and practices that promote health are selected because good health usually (on the average and in the long run) makes reproduction more likely. The connection between reproduction and fitness is only the most direct of the four. Rule giving about mating and mate selection is bound to be common in any culture. We take for granted the reasons behind "It takes real caring to get through the hard times" or "If a couple doesn't do well in bed, they won't do well at all" or "Choose someone who shares your interests." Successful reproduction, including good child care (e.g., "Stop at two"), is nearly always valued.

RULES, CULTURE, AND FITNESS

11

The link to fitness depends on reciprocity. Relationships persist over time because of mutual reinforcement, each person providing reinforcement for the other's actions. This reinforcement takes the form of help with a variety of problems that involve preserving health, gaining resources, and reproducing. A friend may alert you to a new medicine, lend you money, or introduce you to a potential mate. Group membership may be thought of as a relationship between an individual and a group in which conventional behavior, including making sacrifices, is reinforced by other group members who may have no personal relationship with the recipient. The benefits are the same as in more personal relationships. In summary, every rule is given directly or indirectly under the control of a long-term contingency by which the rule-governed behavior preserves health, gains resources, builds relationships, or affords opportunities to reproduce. The ultimate contingency ultimately enhances the fitness of the listener or the listener's genes.
Fitness and Reinforcement

The relation between rules and fitness explains why the listener's compliance to the rule serves to reinforce the speaker's verbal behavior of giving the rule. A rule is given, as Skinner (1971) wrote, for the "good of others"-that is, their good is the speaker's good in the long run. The longterm benefits to the speaker, as for the listener, are in health, resources, relationships, and reproduction. This holds even if the rule giving itself is initially rule governed. An inexperienced manager may instruct salespeople because some supervisor told him to. An inexperienced parent initially may tell a child to wear a coat because that is what good parents do. After some experience, however, the rule giving shifts to control by the long-term contingency between giving instruction and higher profits or between giving

the coat-health rule and the child's continued health. The long-term reinforcement provides the basis for the listener's compliance to function as a (presumably conditional) reinforcer of the speaker's rule giving. The manager instructs a new salesperson because, in the manager's experience, the salesperson's compliance leads to higher profits for the manager in the long run. Without a long-term payoff, however, the manager would offer no instruction. A child's putting on a coat when it is cold outside affects both the fitness of the child and the fitness of the parent. Unless the child grows into an adult who will reproduce, the parent's investment is lost. A child's illness is aversive both to the child and to the parent. If the child's coat donning shifts to long-term control, it will be maintained by avoidance of chill and illness (Herrnstein, 1969). Why the stimuli corresponding to chill and illness should be aversive, however, must be traced to fitness. Those of our ancestors who treated chill and illness with indifference are no longer represented in the population. Similarly, those of our ancestors who treated the signs of illness in their offspring with indifference are no longer represented in the population. Some signs of illness in a child may require experience on the part of the parent, but obvious symptoms like bleeding, crying, coughing, and lethargy probably require no special experience to render them aversive. If giving the coat-health rule is occasioned by the parent's experience of a contingency between children wearing coats and children showing no signs of illness, then giving the rule is maintained by avoidance of signs of illness. Rule Giving, Rule Following, and Rule Making Every culture includes rule giving
on the part of speakers and rule following on the part of listeners. Another

term, rule making, the activity of giving a rule the first time, which is strict-

12

WILLIAM M. BAUM

ly under the control of experience with the ultimate contingency, might be useful as a means for talking about the origins of rules. Rule giving as a practice. There are two senses in which rule giving may be a cultural practice. First, the giving of a particular rule may be a practice. Every culture includes customary sayings and admonitions like "Blood is thicker than water" and "Fools rush in where angels fear to tread." Like the Golden Rule, these sayings are highly abstract but, with appropriate history, become applicable to specific concrete situations. "Blood is thicker than water," for example, may alter a listener's behavior in seeking a promotion if the competition is the boss's nephew or may strengthen the likelihood of helping out a cousin. The second sense in which rule giving is a cultural practice is that a culture may include a general practice of instruction, in which relatively specific rules are given in response to concrete situations. Examples include the vegetables- and coat-wearing-health rules discussed earlier and the giving of rules like "If you want to take good photos, you'll have to get a camera with a good lens" or "The best way for you to go into Boston is Route 93." Instead of being customary sayings, these are specific instances of customary helpfulness. Social intercourse is peppered with instances of this tendency to share experiences; it costs the speaker little and may strengthen highly reinforced behavior in the listener. Low cost to the giver and high benefit to the receiver are the conditions under which altruism toward nonrelatives becomes likely and under which people may be said to behave "for the good of others" (Alcock, 1993; Skinner, 1971). The speaker's rule giving is reinforced, but only some of the timeon the average and in the long run. Rule following as a skill. If rule giving may be a generalized cultural practice, rule following is an even more essential cultural skill. Children are taught to follow commands even be-

fore they learn to speak. Later they learn to follow instructions and advice, generalizing from parents to teachers and friends. Lost in an unfamiliar area, people unhesitatingly inquire directions of total strangers. These wellbrought-up strangers, of course, generally give directions worthy of being followed. Although the importance of skillful rule following is essential to culture, it tends to be taken for granted. Its importance becomes apparent when it fails. Neglected children miss the training and grow up into adults who fail to do what they are told (i.e., commit crimes and resist instruction). Origins of rule making. Like any other operant behavior, rule giving requires a special account of its first occurrence. After the giving of a particular rule has been reinforced, its repetition is understood, but where did the rule come from in the first place? To explain the first occurrence of nonverbal behavior, we appeal to induction and response generalization (Segal, 1972). Operant behavior originates either in nonoperant behavior (induced or elicited; e.g., autoshaping) or in operant behavior that is under other control that transfers to a new situation. Rule making-giving a rule for the first time-arises from other operant behavior. It depends on a history of reinforcement for verbal behavior under the control of regularities in behavior or environment. A statement like "Don't ask Liz questions in the morning; she's always grouchy," a rule in the present sense, occurred for the first time after several morning encounters with Liz, but the very first time it occurs, it depends on a history of reinforcement for such generalizations. This history may go back to a child's early training in naming objects, then in naming events, then in talking about simple relations, and so on. The discriminative stimuli get progressively more complex with further training. The discriminative stimuli that occasion rule making may be firsthand encounters or stimuli generated as a re-

RULES, CULTURE, AND FITNESS

13

RULES AND RULE-GOVERNED BEHAVIOR The word rule has had different meanings in different contexts (Reese, 1989). The vernacular usage differs VERBAL BEHAVIOR from the behavior-analytic usage, and the standard behavior-analytic usage Our concern with rules requires us differs from the one that evolutionary to focus on verbal behavior, because theory would suggest. we are concerned in part with discriminative stimuli and consequences pro- The Standard Behavior-Analytic vided by a speaker for the sake of Definition changing the likelihood of some type Skinner (1969, 1974) and others of action on the part of a listener. (e.g., Blakely & Schlinger, 1987; S.

sult of firsthand encounters-data. For the surgeon general to declare that smoking is bad for health, it may have been unnecessary to have known anyone ill from smoking. It may only have been necessary to have looked at some tables of correlation coefficients. When one is taught to examine data and draw conclusions, one is being trained in rule making. It is good for a culture to include the practice of training citizens in rule making, because from thence arise new practices and more rapid adjustment for the sake of long-term impact on fitness. Long-term effects on fitness offer an evolutionary explanation of the existence of rule giving, rule following, and rule making. This sort of explanation has been rare in behavior analysis (Glenn, 1991). One may wonder whether it is essential, because evolutionary explanations may seem to affect practice and experimentation little or not at all. Such a view is unlikely to be correct, because both practice and experimentation are affected by the concepts brought to bear in inventing new methods and experiments. Without the concept of genetic inheritance, one would never arrive at discussions of genetic diseases and genetic engineering. Evolutionary thinking has already affected behavior analysis by sparking recognition of species-specific behavioral tendencies that interact in various ways with operant behavior (Segal, 1972; Staddon, 1977). In addition, there are more general reasons to unite behavior analysis with evolutionary theory: intellectual completeness and survival as a discipline.

Speaker and Listener A rule is generated by the verbal behavior of a speaker. Not all verbal behavior is rule giving, but rule giving is a type of verbal behavior, because it is reinforced by its effects on the listener. We are primarily concerned with auditory rules, but there is no reason a rule cannot be signed, gestured, or written. To say that a rule is a verbal discriminative stimulus is to say only that it is generated by the speaker's verbal behavior and that it exerts stimulus control over the listener's behavior. Insofar as we are concerned with rule following, we are concerned with the behavior of the listener. The listener behaves in response to the rule, in ways that have been reinforced in the presence of such stimuli in the past. In behaving in the way that we call following the rule, the listener may reinforce the speaker's verbal behavior, but more importantly for present discussion, the listener's rule following is reinforced either by some action or nonaction on the part of the rule-giver. The rule-giver might either show signs of approval or withhold disapproval. When nothing happens as a result of rule following, it usually means that following the rule is avoidance behavior (e.g., of disapproval)-that is, it is behavior maintained by negative reinforcement. Much rule following is enjoined by the threat of punishment, as when the speaker says, "Do this or else."

14

WILLIAM M. BAUM the rule refers to the contingency it describes and that the speaker has a meaning, intention, or idea in mind when stating the rule. Strictly speaking, rule giving, like any verbal behavior, has no meaning and refers to nothing, because the explanation of its occurrence implies only a certain history, certain establishing conditions, and certain discriminative stimuli (Baum, 1994b, chap. 7). The notions of meaning and reference are simply foreign to behavior analysis.

Hayes, 1989) define a rule as a verbal description or specification of a contingency or a verbal statement that points to a contingency. Trouble arises because of the terms "description," "specification," and "points to." Calling a rule a description or specification of a contingency is inaccurate and misleading. It is inaccurate because rules, as they are actually given, often fail to describe or specify a contingency. It is misleading because it implies the concepts of meaning and reference used in traditional mentalistic accounts of language. I shall discuss each of these points in turn. Rule giving, like any verbal behavior-like any operant behavior-is defined by its function rather than by its structure. A parent might say to a child, "Put on a warm coat so you won't get sick," which would count as a description of a contingency, but the parent might just as well say, "Put on a coat, it's cold outside," or simply, "It's cold out," or might even just open the closet door or gesture in its direction. Because all of these variants achieve the same effect (the child puts on a coat), they are all members of the same operant, the same functional class. Few members of this functional class actually are descriptions of the contingency between wearing a coat and health. Yet they all function as a rule. Saying that they are implicit descriptions of the contingency might solve this problem, but would require an unusual meaning of implicit, such as "belonging to the same functional category as explicit descriptions." Focusing on variants of a rule that have the structure of descriptions may be misleading, however, by suggesting that the meaning of the rule resides somehow in its structure, rather than in the conditions governing the behavior of giving it. In his 1945 paper on psychological terms, Skinner pointed out that the meaning of a term should be understood as the conditions of its occurrence. The point applies to all verbal behavior, including rule giving. Calling a rule a description implies that

Stimulus Control of Rule Giving

In the vernacular, rules are often spoken of as things that are learned and known. Mentalistic explanations place the rules inside the organism (Hineline & Wanchisen, 1989). Behavioral explanations instead look to past and present environment. When not placed inside the person, rules in the vernacular are often summaries of regular features of the environment or someone's behavior. Statements like, "As a rule, it starts snowing in December" and "As a rule, I get up at 6:00 in the morning" are based on environmental events, but on no one particular event. The stimulus conditions that control their occurrence extend through time over many particular instances. Similarly, many instances over a period of time control a statement like, "As a rule, if I leave the trash cans uncovered, raccoons get into them." This example is occasioned not simply by one event being relatively frequent, but by certain combinations of events being relatively frequent, in particular (a) trash cans uncovered and trash disturbed and (b) trash cans covered and trash undisturbed. Figure 2A illustrates the events involved. Of the four possible behavior-consequence combinations, only two occur frequently. The check marks suggest relative frequency. To the extent that I usually remember to cover the trash cans, the trash is usually undisturbed; those few times when I leave the cans uncovered almost always result in disturbed trash,

RULES, CULTURE, AND FITNESS

15

A
B1

Consecuence:

Cover Trash

B2 Leave Uncovered
VVV

C,=

Trash Disturbed
'V/V

Trash C2 Undisturbed

B
B1 Conseauence: C1 = Lying Approval, V acceptance

B2

Truth-Telling
'VVVv'

C2 = Disapproval, VV rejection
Figure 2. Diagrams of experience with two contingencies. Check marks indicate relative frequency of occurrence of four possible events: B, + C, B, + C2, B2 + C, and B2 + C2. A: The particular contingency illustrated concerns covering trash cans and raccoons disturbing the trash. B: The particular contingency concerns telling the truth, lying, approval, and disapproval.

but occasionally I am lucky and the raccoons don't get into them even though they are uncovered (one check). Technically speaking, the utterance, "If I leave the trash uncovered, then raccoons get into it," is a tact under the stimulus control of a relation between covered and uncovered trash cans and undisturbed and disturbed trash (i.e., a contingency). Figure 2B shows another, social, example. The contingency is between lying and disapproval, on the one hand, and honesty and approval, on the other. The checks suggest that all four conjunctions may

occur, but only two are relatively frequent. Giving the rule, "If you tell the truth, people will generally like you, but if you lie, they will generally dislike you," is under stimulus control of this contingency. Like any discriminative stimulus, a contingency that serves as a discriminative stimulus must be defined by its function. Just as the effective stimulus may differ from an observer's guess, so an effective contingency, as a stimulus, may differ from "reality" or the ostensible reason given by the speaker. The functional contingency that controls

16

WILLIAM M. BAUM

rule giving may be neither veridical nor as described by the rule. Veridicality. Two sorts of reasons may prevent a contingency from being taken as "reality." First, a contingency consists of a limited set of events. Giving a rule sometimes resembles induction on the basis of a small sample. Early in one's life, it might seem that lying rarely results in punishment, but with more experience, one may conclude just the opposite. On the basis of different (limited) experience, one speaker may say buying stocks is good, whereas another speaker may say it is foolhardy. Second, events in a contingency are unequal in their effects. Instances that include a definite outcome appear to have more of an effect than instances in which behavior occurs with no outcome. In experiments on contingency judgment, people judge contingencies to be stronger when the frequency of outcomes is high, even when there is no behavior-outcome contingency at all (Allan & Jenkins, 1983; Alloy & Abramson, 1979). If they are exposed more to the contingency, however, their judgments tend to become more veridical (Dickinson, Shanks, & Evenden, 1984). Although the rules of a culture may be based on long and reliable experience with contingencies, when experience is limited and open to bias, there is no reason to expect everyday rules to be veridical. Ostensive contingency versus functional contingency. Even when a rule appears to be a description of a contingency, the functional contingency may differ from that described. The contingency in Figure 2B, for example, might occasion a rule like, "If you lie you will go to Hell, but if you tell the truth you will go to Heaven." The functional contingency concerns approval and disapproval by members of the community in the here-and-now, not in the afterlife. In the Middle Ages, a physician might give the rule, "If there is fever, then we must let blood." This fever-blood-letting rule probably had little to do with the health of the patients; more likely it was controlled

by a contingency that related blood letting to gaining wealth and status. A much-discussed example is the Jewish dietary laws (e.g., Harris, 1987). The list of forbidden foods is long and varied: camel, swine, shellfish, birds of prey, animals that creep on their bellies, all insects except grasshoppers, crickets, and locusts, and so on. Individual items on the list have been attributed to considerations of health or economics (Harris, 1987). Another explanation is culturally based: The prohibitions were designed to separate the Israelites from the surrounding tribes. The functional contingency might be that if Jews cannot eat with non-Jews they are less likely to marry non-Jews. Whatever the functional contingency, it differs from the ostensive contingency-that keeping the laws prevents God's anger. Verbal behavior that is occasioned not by immediate circumstances but by past experience with a contingency, verbal behavior for which experienced contingencies serve as a discriminative stimulus, is particularly important for arriving at a sound behavior-analytic definition of rules. We avoid the problems with the standard definition by recognizing that all rule giving in the behavior-analytic sense is verbal behavior under the control of experienced contingencies. The reverse, however, would be false; although all contingency-controlled verbal behavior might be rule giving in the vernacular sense, not all such verbal behavior is rule giving in the behavior-analytic sense, because something more is required; rule giving is characteristically reinforced by the listener's behavior.

Reinforcement of Rule Giving

The stimulus control of rule giving tells more about the speaker's behavior than the listener's. Discussion of rules, however, invariably revolves around the idea of rule-governed behavior, which is behavior of the listener, and to which the rule relates as a discriminative stimulus. When a listener hears

RULES, CULTURE, AND FITNESS

17

Speaker :B
Listener :
I

B1 B2 Ci ..SDB

SR

RUIQ-

SD B

~~Rule

Figure 3. Diagram of an interaction in which a speaker's verbal behavior (Bv) generates a rule (SD) governing the listener's behavior. Experience with a contingency, as in Figure 1, serves as the discriminative stimulus for the speaker's rule giving (Bv). The listener's behavior (B) provides reinforcement for the speaker's rule giving.

instructions or advice, the listener's behavior changes. "Turn left at the corner to get to the bank" may be a report of a contingency, but it also affects the listener's behavior, and it is given for the sake of its effect on the listener's behavior. That the listener's following the rule is essential may be seen by speakers' behavior when listeners fail to follow advice or instructions. Speakers then behave as if accustomed reinforcement for their behavior had been withheld. Technically speaking, therefore, the giving of a rule is as much a mand as a tact (Skinner, 1957). It not only is occasioned by the speaker's experience with a contingency, but it has a characteristic reinforcer in the rule following of the listener. The giving of instructions or advice is reinforced by signs of compliance on the part of the listener. Figure 3 illustrates the relations governing the speaker's rule giving. The contingency, indicated as a box, sets the context (SD) for the rule giving (Bv). This produces a discriminative stimulus for the listener's compliance (B), which produces reinforcement for the speaker's rule giving. The listener's compliance also serves as a discriminative stimulus for the speaker to supply reinforcement for it, a point to which we shall return shortly.

We might amend the standard definition of a rule to avoid the problems with "description" and "specification" by saying that a rule is a verbal discriminative stimulus that is functionally equivalent to a contingency statement in the paradigmatic form of "if B, then C," where B is behavior and C is a consequence. We also need to add that the giving of the rule is reinforced by signs that the listener will engage in B and that this common reinforcement defines the phrase functionally equivalent. This is good as far as it goes, but it is too inclusive, because we restrict the label rule only to some of these if-then statements.
Bargain versus Rule

Some if-then statements govern behavior only on a specific occasion and only in the short run. A price posted in a store constitutes an implicit if-then statement like, "If you will give $1.69, then you may take away this loaf of bread." Another example would be, "If you'll let me look at your comic book, then I'll give you a piece of my donut." To distinguish such verbal discriminative stimuli from rules, I shall call them bargains. In everyday terms, giving a bargain is an offer of short-term mutual benefit. In behavior-analytic terms, it is speaker

18

WILLIAM M. BAUM

behavior with characteristic reinforcement-the listener's compliance (i.e., a mand; Skinner, 1957). It has an additional property, however, in that the reinforcement for the listener's compliance also is characteristic. Seen this way, bargains would usually be called promises or threats in everyday usage. For example, "Your money or your life" is a bargain, because it is equivalent to the if-then statement, "If you give me your money, then I will spare your life." In contrast with giving a bargain, giving a rule is occasioned by a longterm contingency that benefits the listener on no particular instance of compliance, but on the average and in the long run. Wearing a coat in cold weather may afford no immediate benefit to a child, but the parent gives the coathealth rule and reinforces the behavior because wearing a coat is generally good for health. The behavior that is strengthened by the rule may never be directly reinforced; it may remain forever under the control of social contingencies. It may also shift to control by the long-term contingency (i.e., it may become internalized in the sense described earlier). Although the distinction between bargains and rules might seem clearcut, it is easy to conceive of ambiguous examples. When a teacher tells a student, "If you do this page of arithmetic, then I will give you a sticker," that might appear to be a bargain. We might also construe it as a rule if we ask why the teacher would say this in the first place. The teacher's behavior is reinforced with money and other social reinforcers, but other people arrange reinforcement for the teacher's behavior because, on the average and in the long run, it is good for the child to complete arithmetic problems and other educational tasks. All the bargains we have considered allow the addition of some long-term benefit to the listener; bread is a resource, and sharing a comic book may cement a friendship. However fuzzy the distinction between bargain and rule, it remains use-

ful anyway, for two reasons. First, it reminds us of the importance of longterm contingencies to understanding rules. Second, it reminds us of the need to explain the behavior of both the listener and the speaker. The difference between rules and bargains is a matter of degree, rather than an absolute dichotomy, partly because it lies in the degree of importance we attach to long-term contingencies, and partly because it lies in the extent to which there is clear benefit to the speaker. A bargain implies obvious benefit both for the listener's accepting the bargain and for the speaker's offering it, whereas a rule implies benefits that are obvious only for the listener's rule following. That the benefits of the speaker's rule giving are less obvious means only that, when we are done explaining rule following, we still need to explain rule giving.

Comparison with Zettle and Hayes (1982)

Zettle and Hayes (1982) criticized the definition of rules as "contingencyspecifying stimuli," arguing that it is "excessively narrow in some respects, and excessively broad in others" (p. 77). They defined rule-governed behavior as "behavior in contact with two sets of contingencies, one of which includes a verbal antecedent" (1982, p. 78). This resembles the present analysis if Figure 1 captures the meaning of "in contact." Zettle and Hayes distinguish what they consider to be two types of rule-governed behavior, pliance and tracking. They define pliance as "rule-governed behavior primarily under the control of apparent speakermediated consequences" (1982, p. 80). They offer as an example handing over one's wallet in response to a thief's saying, "Your wallet or your life." This example and their discussion suggest that Zettle and Hayes (1982) explain pliance entirely by history and short-term contingencies; it requires no reference to long-term contingencies. Their definition remains ambiguous

RULES, CULTURE, AND FITNESS

19

because of the word "primarily." If we focus only on the short-term speakermediated contingency (the proximate contingency), their idea of pliance fails to conform to the definition of rulegoverned behavior as involving two distinct sets of contingencies. The thief's utterance is only under shortterm control. Blakely and Schlinger (1987) interpreted pliance this way, arguing that a listener's response to a verbal discriminative stimulus under short-term control is simply a discriminated operant. On such a view, pliance (or compliance) would fail to qualify as rule-governed behavior. Ambiguity remains, however, because Zettle and Hayes offer other examples (e.g., a parent says, "It's cold outside," and the child puts on a jacket) that would qualify. The distinction Zettle and Hayes sought to make remains useful, however. In present terms, if we consider pliance to be behavior entirely under short-term control, then what they called a ply is what I have called a bargain. It is a verbal discriminative stimulus equivalent to an if-then statement involving only speaker-mediated consequences. Requiring no long-term contingency for its explanation, it would not be called a rule. Zettle and Hayes define tracking as "rule-governed behavior under the control of the apparent correspondence between the rule and the way the world is arranged" (1982, p. 81). As an example, they offer following directions. Their discussion suggests that they probably intended something like what I am calling simply rule-governed behavior. Ambiguity arises from attributing behavior to an apparent correspondence. In present terms, rule-governed behavior is under the control of the rule given by the speaker. The speaker's verbal behavior is under the control of past experiences with the long-term contingency of Figure 1, as shown in Figures 2 and 3. In this sense, the speaker's behavior might be said to be under control of an apparent correspondence, but as indicated earlier

such a locution raises the problems of reference and intentionality. Further ambiguity arises because Zettle and Hayes emphasize that in tracking "the speaker does not mediate compliance" (1982, p. 81). True, when a listener follows directions like, "The way to get to Greensboro is to follow 1-85," the speaker usually cannot supply consequences, and getting to Greensboro is a consequence itself. These considerations, however, give no basis for distinguishing direction following from other forms of rule following. Direction following originates in a history of speaker-mediated (proximate) consequences; no one would follow directions unless trained to do so, and children usually receive extensive training to follow all sorts of rules, including advice and directions. The origins would make following a particular set of directions rule-governed behavior like any other. That consequences like getting to Greensboro may maintain direction following suggests that direction following, as a general functional class transcending particular sets of instructions, is internalized in the sense discussed earlier. It is possible that Zettle and Hayes intended tracking to refer to a listener's tendency to follow rules even in the absence of speaker-mediated consequences because, in the listener's history, following rules generally paid off.

The Need to Explain Rule Giving

According to the standard behavioranalytic definition, the giving of a rule is explained by its being occasioned by the long-term contingency and reinforced by the listener's compliance. This, however, leaves open the questions of why the long-term contingency should occasion rule giving and why compliance should reinforce it. In addition to its other shortcomings, the standard definition allows no straightforward answer to these questions. What about the coat-health contingency should induce a parent to exhort a

20

WILLIAM M. BAUM
cy: The role of selective attribution. Quarterly Journal of Experimental Psychology, 36A, 29-50. Glenn, S. S. (1988). Contingencies and metacontingencies: Toward a synthesis of behavior analysis and cultural materialism. The Behavior Analyst, 11, 161-179. Glenn, S. S. (1991). Contingencies and metacontingencies: Relations among behavioral, cultural, and biological evolution. In P A. Lamal (Ed.), Behavioral analysis of societies and cultural practices (pp. 39-73). New York: Hemisphere. Harris, M. (1980). Cultural materialism. New York: Vintage Books. Harris, M. (1987). Foodways: Historical overview and theoretical prolegomenon. In M. Harris & E. B. Ross (Eds.), Food and evolution (pp. 57-90). Philadelphia: Temple University Press. Harris, M., & Ross, E. B. (1987). Death, sex, and fertility. New York: Columbia University Press. Hayes, L. J., & Chase, P. N. (Eds.). (1991). Dialogues on verbal behavior. Reno, NV: Context. Hayes, S. C. (Ed.). (1989). Rule-governed behavior: Cognition, contingencies, and instructional control. New York: Plenum. Herrnstein, R. J. (1969). Method and theory in the study of avoidance. Psychological Review, 76, 49-69. Hineline, P N., & Wanchisen, B. A. (1989). Correlated hypothesizing and the distinction between contingency-shaped and rule-governed behavior. In S. C. Hayes (Ed.), Rulegoverned behavior: Cognition, contingencies, and instructional control (pp. 221-268). New York: Plenum. Logue, A. W. (1978). Behaviorist John B. Watson and the continuity of the species. Behaviorism, 6, 71-79. Logue, A. W. (1994). Watson's behaviorist manifesto: Past positive and current negative consequences. In J. T Todd & E. K. Morris (Eds.), Modern perspectives on John B. Watson and classical behaviorism (pp. 109-123). Westport, CT: Greenwood. Malagodi, E. F, & Jackson, K. (1989). Behavior analysts and cultural analysis: Troubles and issues. The Behavior Analyst, 12, 17-33. Petrovich, S. B., & Gewirtz, J. L. (1991). Imprinting and attachment: Proximate and ultimate considerations. In J. L. Gewirtz & W. M. Kurtines (Eds.), Intersections with attachment (pp. 69-93). Hillsdale, NJ: Erlbaum. Reese, H. W. (1989). Rules and rule-governance: Cognitive and behavioristic views. In S. C. Hayes (Ed.), Rule-governed behavior: Cognition, contingencies, and instructional control (pp. 3-84). New York: Plenum. Segal, E. F (1972). Induction and the provenance of operants. In R. M. Gilbert & J. R.

child to wear a coat? What good accrues to the parent if the child does wear a coat? The definition suggested here permits an explanation in a larger perspective-the cultural and evolutionary context.

CONCLUSION The virtue of defining rules and rulegoverned behavior in terms of longterm contingencies of fitness is that it integrates the behavioral analysis with evolutionary thinking. It makes the strength of the behavioral analysis available to evolutionists. At the very least, it offers a richer, more detailed account of how culture and individual human behavior work, but potentially it might change the way evolutionists talk about behavior. For behavior analysts, integration with evolutionary theory affords a large and powerful context within which to understand individual behavior and culture.
REFERENCES
Alcock, J. (1993). Animal behavior (5th ed.). Sunderland, MA: Sinauer Associates. Allan, L. G., & Jenkins, H. M. (1983). The effect of representations of binary variables on judgment of influence. Learning and Motivation, 14, 381-405. Alloy, L. B., & Abramson, L. Y. (1979). Judgment of contingency in depressed and nondepressed students: Sadder but wiser? Journal of Experimental Psychology: General, 108, 441-485. Baum, W. M. (1994a). John B. Watson and behavior analysis: Past, present, and future. In J. T Todd & E. K. Morris (Eds.), Modern perspectives on John B. Watson and classical behaviorism (pp. 133-140). Westport, CT: Greenwood Press. Baum, W. M. (1994b). Understanding behaviorism: Science, behavior, and culture. New York: HarperCollins. Blakely, E., & Schlinger, H. (1987). Rules: contingency-specifying Function-altering stimuli. The Behavior Analyst, 10, 183-187. Boyd, R., & Richerson, P. J. (1985). Culture and the evolutionary process. Chicago: University of Chicago Press. Cerutti, D. T. (1989). Discrimination theory of rule-governed behavior. Journal of the Experimental Analysis of Behavior, 51, 259-276. Dawkins, R. (1989). The selfish gene (new ed.). Oxford: Oxford University Press. Dickinson, A., Shanks, D., & Evenden, J. (1984). Judgement of act-outcome contingen-

Millenson (Eds.), Reinforcement: Behavioral analyses (pp. 1-34). New York: Academic
Press.

RULES, CULTURE, AND FITNESS

Skinner, B. F (1945). The operational analysis of psychological terms. Psychological Review, 52, 270-277. Skinner, B. F (1957). Verbal behavior. New York: Appleton-Century-Crofts. Skinner, B. F (1969). An operant analysis of problem solving. In Contingencies of reinforceinent: A theoretical analysis (pp. 133171). New York: Appleton-Century-Crofts. Skinner, B. F (1971). Beyond freedom and dignity. New York: Knopf. Skinner, B. F (1974). About behaviori.sm. New York: Knopf. Skinner, B. F (1981). Selection by consequences. Science, 213, 501-504. Sommerville, C. J. (1982). The rise andfill of' childhoocl. Beverly Hills, CA: Sage. Staddon, J. E. R. (1977). Schedule-induced behavior. In W. K. Honig & J. E. R. Staddon (Eds.), Handbook of operant behavior (pp. 125-152). Englewood Cliffs, NJ: PrenticeHall.

21

Staddon, J. E. R. (1983). Adaptive behavior and learnin7g. Cambridge: Cambridge University Press. Todd, J. T. (1994). What psychology has to say about John B. Watson: Classical behaviorism in psychology textbooks, 1920-1989. In J. T. Todd & E. K. Morris (Eds.), Modern perspectives on John B. Watson and classical behaviorismn (pp. 75-107). Westport, CT: Greenwood. Watson, J. B. (1913). Psychology as the behaviorist views it. PsYchological Revievv, 20, 158-177. Wilson, D. S., & Sober, E. (1994). Re-introducing group selection to the human behavioral sciences. Behavioral and Brcain Sciences, 17, 585-654. (Includes commentary) Zettle, R. D., & Hayes, S. C. (1982). Rule-governed behavior: A potential theoretical framework for cognitive-behavioral therapy. In P C. Kendall (Ed.), Advances in cognitive-behavioral resecarch and therapy (Vol. 1, pp. 73118). New York: Academic Press.