[Brennan, Timothy (1994), Talking to One Selves: The Social Science of Jon Elster, Journal of Communication 44 (1), p. 73-81]

[ Books | Articles | Reviews | Index | The main menu ]

Talking to One's Selves: The Social Science of Jon Elster
[start of page 73]

A review essay by Timothy J. Brennan, University of Maryland-Baltimore

Ulysses and the Sirens: Studies in Rationality and Irrationality. By Jon EIster. New York: Cambridge University Press, 1979. ix + 193 pp. $17.95 (soft).

Sour Grapes: Studies in the Subversion of Rationality. By Jon EIster. New York: Cambridge University Press, 1983. viii + 177 pp. $59.95 (hard), $18.95 (soft).

The Cement of Society. By Jon Elster. New York: Cambridge University Press, 1989. viii + 311 pp. $59.95 (hard), $17.97 (soft).

Nuts and Bolts for the Social Sciences, By Jon Elster, New York: Cambridge University Press, 1989. viii + 184 pp. $12.95 (soft),

Solomonic Judgments: Studies in the Limitations of Rationality. By Jon Elster, New York: Cambridge University Press, 1989. ix + 232 pp, $47,95 (hard), $ 15,95 (soft),

Those who study communication must draw from a variety of disciplines in the social sciences, including psychology, sociology, political science, and economics, as well as communications itself, As a consequence, com munications researchers are well aware of the many methodological dichotomies that underlie these disciplines. Is human behavior best understood as rationally selected pursuit of cleverly defined ends, or is it more pragmatic, spontaneous, or affective? Do people respond most to external stimuli, such as economic incentives, or do actions follow internalized social or ethical norms? Are macro phenomena such as media, markets, and politics the result of separate individual actions, or are the actions of individuals reducible to the effects of primary social entities such as gender or class? Is the point of social science to uncover reductionistic causation or to discover holistic interpretations?
[end of page 73, start of page 74]
We confront these boundary problems whenever we attempt to push a particular methodological perspective to its limits. These questions also arise when we attempt to explain descriptive or interpretive inadequacies of particular perspectives by their methodological flaws, that is, when we trace errors in substance to errors in style. This is particularly associated with criticisms of the individualistic rational-choice methods that dominate economics and currently play a significant role in much political science and sociology. Gary Becker's recent receipt of the Nobel Prize for economics signifies the degree to which rational choice has come to be applied. But taken at face value, the rational-choice model leaves many behaviors and social phenomena unexplained. Most lists of the unexplained would include weakness of will, altruism, voting, collective political action, and adherence to social norms.

On the other hand, individualistic rational-choice models of behavior have a great deal going for them. While almost all of us would acknowledge the effects of social institutions (family, community, state) and constructions (race, class, gender) on our tastes and aspirations, we apparently live our lives each as an isolated individual locus of conscious experience. In addition, much of our behavior in these isolated worlds is intentional, directed at achieving a goal, satisfying a standard, or resolving some trade-off. We need not fit the microeconomic textbook model of climbing onto the highest indifference curve, at a point of tangency to a budget line, to interpret our behavior as aimed toward some target. Finally, the logical structure of individual intentional behavior enables the social scientist to create theories, offer explanations, and make predictions. For the most part, nonindividualistic, nonintentional portrayals of human action do not lend themselves to the same level of analysis and interpretation. 1

Three strategies leave themselves open to the social scientist concerned with reconciling the phenomenological, psychological and methodological advantages of the rational-agency model with the aforementioned apparent empirical failings. The first is effectively to dismiss the empirical ailings, either by denying their relevance or by trivially adapting rational agency models to swallow them up.2 For example, voting is either not interesting enough to explain, or explained by positing a preference for voting, rationally satisfied by the act of voting (Barry, 1978). The second strategy, already noted, is to reject the rational-choice model, concluding that its empirical failings are inevitable. Those reaching this conclusion may substitute social entities for individual agents as prime causes. A more radical reaction along these lines is to dismiss the scientific enterprise altogether, substituting holistic interpretation for analytical explanation. 3

The first tactic trivializes the need to reconcile; the second sacrifices the nearly self-evident advantages of rational-agent models. The third response would be to devise modifications of the rational-choice model that preserve those advantages while yielding coherent, reasonable explanations of phenomena seemingly beyond the model's reach. This response re-
[end of page 74, start of page 75]
quires insight to see the subtle, unstated assumptions of conventional version of rationality models. It demands creativity to realize that these assumptions may be unnecessary, enduring only because they had remained unseen. Last and certainly not least, effective responses rely on an impish cleverness to figure out how modifying these unstated yet persistent assumptions can allow us to explain the heretofore inexplicable.

Some renegade economists have taken on this challenge - Robert Frank, Thomas Schelling, and George Akerlof come to mind-but perhaps the most insightful, creative, and clever contributions have come from Jon Elster. He believes that the fundamental method for explaining social phenomena is by describing causal mechanisms based on individ ual human action motivated by desires and constrained by opportunities. (Nuts and Bolts for the Social Sciences, pp. 10-14). For most of us, this methodological commitment is paired with the premise that people have fixed, self-focused preferences that remain stable over time. While such premises can be empirically useful, they are not logically necessary. By separating the empirical claims regarding preferences from the method, Elster shows us what we can understand if, for example, we suppose that one's desires, beliefs, and very identity not only can change over time, but that such changes present both opportunities and objects for desire by perions at any particular time.

For communication theorists, the most explicit application of Elster's insight is that problems of communicating and committing among separate individuals have an influence on how each of us communicates within our selves, as we attempt to manage our lives over time. One of Elster's earliest works, Ulysses and the Sirens (1979), discusses an interpretation of one of these intrapersonal communication settings, in the context of what he calls precommitment. "Suppose that you want to lose weight. (We'll see in a moment how problematic a sentence this is.) In accord with that desire, you eat just a small salad for dinner. You can predict, however, that you will be hungry later that night. Under conventional models of rational action, this presents no conceptual problem. If you really want to lose weight - that is, prefer less weight to the satisfaction of a night-time snack - you just won't eat. However, suppose you can predict that later that night you will value the snack over the weight loss, and that you want to keep yourself from eating later on. Again, under conventional models, this appears either irrational (Why would you act to thwart your desires, if those are really your desires?) or incoherent (What can it mean to act to thwart your desires, if your desires are revealed by your actions?).

This appearance of irrationality or incoherence is caused by the implicit assumption that you share the same preferences after dinner and late at night. Elster asks us to imagine instead that these preferences differ, in a particularly interesting way. After dinner, you can predict that later in the
[end of page 75, start of page76]
evening you will want to eat a high-calorie snack, in accord with your late-evening preferences but contrary to your after-dinner preferences. This perspective converts an "irrationality" into a situation calling for persuasion or strategic interaction imposed by: the after-dinner-you on the late-night-you. The theme of Ulysses and the Sirens is to illustrate the strategy of precommitting, that is acting now to lead oneself to act later in accord with current preferences. Elster's title provides an example. In The Odyssey Ulysses orders his men to tie him to the mast of his ship to prevent him from leaving in response to the lure of the Sirens. In the dieting situation, the after-dinner-you could foreclose the opportunity for the late-night-you to eat the late-night snack by asking your spouse to lock the refrigerator and to refuse to give you the key when you ask for it later.

An extension of this framework underlies Elster's reconstruction of weakness of will. In conventional intentionalist models of behaviour, weakness of will makes no sense - one only does what one wants to do. The implicit assumption here is that there is a meaningful person" to speak of, as an effective guardian of long-term self-interest. For example. Elster suggests that we could regard a regular exercise program as a kind of collective-action problem, where the percentage of days one continues to jog to promote health of the person over time is analogous to the percentage of individuals who cooperate to provide a public good (Elster, 1985, pp. 249, 253). The paradigmatic collective-action problem is that each member has an incentive to free-ride, that is, take advantage of others efforts and pull less of one's own weight (Hardin. 1982). A standard example in communications policy is the incentive not to contribute to public broadcasting. In Elster's example, the free-riding incentive is that on any given day one contributes little to overall health by getting up early jog, but would pain a great deal in the short run by sleeping late.

A natural temptation is to go the whole nine yards with this conception and view a person as a collective of separate intrapersonal selves over time. Elster (1987. p. 235) asks us to resist this temptation 4 , citing the "asymmetry of time" (i.e., in the intrapersonal setting earlier selves cannot be harmed by later ones) and noting that the indivisibility of persons" others shared memories and coordination possibilities not open to conventional collectives. Despite these differences between collective action and self-control, Elster's reconstruction of personal choices opens the door for communications theorists to construct intrapersonal communication model - talking to yourself - to explain behavior. They may also explain why a person might employ media to persuade later selves to act differently-for example, the dieter might put a picture of a fat person on the refrigerator to discourage the other hungry self from eating. The signals and strategies that persons use to persuade or commit themselves, and the ways in which we intrapersonally interpret and share meanings over time, could be described, measured, evaluated, and modified with insights from communications.

The power of Elster's precommitment idea, however is not merely
[end of page 76, start of page 77]
methodological. By excluding standard notions of identity, Elster's model forces us to reconsider political or social norms such as economic efficiency, in which maximizing options is presumed to be a necessary condition for consumer sovereignty. Media policy analysis teems with potential applications. Imagine that, when reflecting one day at work on my civic responsibility during an election, I believe that I should watch a presidential candidate debate this evening. Suppose I also know that when I get home, I'll feel lazy and opt for a sitcom. I might thus prefer that the sitcom choice he taken from me, that is, that all networks preempt regular programs to carry the debate. Under conventional choice models, forcing networks to carry debates, either by law or moral suasion, makes viewers worse off. (I leave aside First Amendment considerations.) Taking away the sitcom reflects a paternalistic disregard for what the individual wants. By showing us the ambiguities in the phrase "what the individual wants," Elster allows us to construct an argument for limiting choice that follows rather than disregards individual preference. Seeing how the choice to have fewer choices could be rational rather than paradoxical radically attenuates the power of naive consumer sovereignty - to exclude policies from consideration.

A similar example illustrates communications insights from Elster's collective-action model of weakness of will. Instead of using a daily exercise regimen, consider someones interest in following a presidential campaign. This might require spending time each day reading the front page of the newspaper at breakfast and watching the "MacNeil/Lehrer News Hour" in the evening. On any given day, however, that person could well free-ride on the assumption of diligence on other days, and read the sports page instead in the morning and watch "Wheel of Fortune" at night. As with collective action across individuals, this incentive to freeride could cause a person to fail to provide the internal "public good" of becoming politically informed.

In Sour Grapes (1983, pp. 43-53, 109-124), Elster sets out some self- management strategies for dealing with two particular challenges to rational action: willing that which cannot he willed and cognitive dissonance. The former describes desired ends that are inherently by-products of other actions, and cannot be achieved directly. Elster's stock example is sleeping: Wanting to fall asleep keeps one awake. A more pressing example involves political action. Citing de Tocqueville, Mill, and Arendt, Elster suggests that the value in political action is the by-product that it increases civic awareness, not in the political outcome itself (Sour Grapes. pp. 91-99). The crucial point, however, is that this justification is, in Elster's term, "self-defeating"; political action can promote this awareness only if political participants believe that the only purpose of political action is the political outcome. Elster employs a similar argument against using the "best interests of the child" as the standard in custody disputes. Among other difficulties that standard may be self-defeating, in that protracted assessments of this interest can make custody disputes con-
[end of page 77, start of page 78]
tentious and protracted, harming the child in the process (Solomonic Judgments, pp. 145-148).

Ethical implications play a role in Elster's analysis of cognitive dissonance, in particular that people may adapt their preferences or beliefs to circumstance, as when the fabled fox claimed the grapes he could not consume were sour anyhow. When preferences or beliefs cease to be independent of circumstance, we can no longer measure freedom as the ability to do what one wants. Here, as when we have to choose between the preferences of the after-dinner dieter and the late-evening dieter, we need to look behind them to their rationality, consistency, and autonomy to judge how these should be weighed in a utilitarian calculation.

A natural extension of Elster's analyses of intrapersonal commitment and instrumental rationality is his recent attention to whether individual rationality can be reconciled with collective action to overcome free-rider problems. One difficulty is that collective agreement requires resolution of bargains. If parties are assumed to be motivated only by narrow self-interest, there are no unambiguous solutions to the "'bargaining problem" in even the simplest, two-person settings (Cement of Society, pp. 64-74).5 Moreover, these bargaining outcomes are subject to manipulation when parties can control their preferences themselves, their strategic positions. or (for the communication theorist) what others know about their preferences (Cement of Society, pp. 82-94).6

A stronger force in ensuring collective action is adherence to internal social and ethical norms. Elster argues that norms are not rational agency in disguise, in that the motivation to follow them does not depend upon sanctions (e.g., ostracizing), or that those sanctions themselves are credible threats only if others are known to follow norms (e.g., to ostracize those who do not keep promises) (Cement of Society, pp. 130-138). Taking norm-motivated behavior as a given leads to some clever descriptions of solutions to collective-action problems. To take one example, assume there are three types of individuals: Kantians who act out of a sense of duty, utilitarians who act to maximize the welfare of the group, and egoists who care only about themselves. From their sense of duty, the Kantian group will support public television regardless of what others do. If there are enough Kantians, the utilitarians may then join in if each of them sees that his or her efforts, combined with the Kantians, adds to social utility. If there are enough of the first two groups, the egoists may join in if, given the Kantian and utilitarian participation, each egoist is better off cooperating than free-riding (Nuts and Bolts, p. 133).

However, despite a clever discussion of the function of a variety of social norms, Elster cannot explain the emergence of norms within the rational-agent paradigm (Nuts and Bolts, p. 123). For rational-agent modelers, the trouble with norms springs from an apparent paradox. On the one hand, norms, and ethical behavior, are willful. The fact of intentional choice makes norms seem characteristically different from preferences or
[end of page 78, start of page 79]
tastes, which in rational-choice models come from elsewhere. A norm is more than an exogenous taste to do good. On the other hand, the ethically interesting aspect of a persons' decision to conform to a moral norm is that it isn't simply a means for achieving these other ends. Norms cannot be interpreted as methods for achieving outcomes valued according to other pre-existing criteria. In rational-choice models, as choices they must be means to an agent's given ends, but norms are noteworthy just because they are not means to those ends (Brennan, 1991).

This survey can give only a glimpse into the realm of topics that Elster treats. The best plaice for the uninitiated to begin is his Nuts and Bolts for the Social Sciences, a short volume presenting brief overviews of Elster's methodology along with the themes presented here. This survey also does not do justice to the flavor of Elster's style. He draws not only on economic, sociological, psychological, and moral theory, as emphasized here, but on all manner of empirical examples, from ancient Greece to our modern foibles.7 Prof. Elster is to rationality as television's Mr. Wizard was to physical science; each showed that their subjects could explain the unexpected.

Potential readers should he forewarned about some aspects of Elster's work. It is likely to strike many as redundant or overblown. Repetition is particularly apparent among Solomonic Judgments, The Cement of Society and Nuts and Bolts for the Social Sciences, all of which appeared in 1989. On the other hand, since Elster's work applies and speaks to many disciplines, what seems crucial to one reader may seem superfluous to another. To those interested in economic methodology, Elster's chief contribution is in taking rational-choice models sufficiently seriously to show what about behavior is genuinely paradoxical or inexplicable from descriptive or moral points of view. After reading Elster, typical lines of criticism of rationality - for example, that irrationality is widespread because people act morally, accept restraints, or go beyond narrow self-interest- seem glib and naive. Once I found this point in each of his works, I tended to skim the rest, looking for illustrative nuances. Others may find his discussions of psychology or social relations the more rewarding aspects, and the methodology the excess baggage.

A more important shortcoming is an underemphasis of concrete policy analysis. For many, the interest in social science is motivated by the desire to understand, from a variety of ethical standpoints, how well our institutions work and whether public policy can make them work better. Since our dominant policy paradigm, economic efficiency, is grounded in rationality and consumer sovereignty, Elster's insights go right to the core of these policy foundations. While Elster briefly discusses tax evasion, labor-management relations, and corruption in The Cement of Society, he by and large has not fully exploited his insights in this regard. This omission, of course, presents an opportunity for those with policy concerns to draw on Elster's arguments and metaphors to come up with critiques of
[end of page 79, start of page 80]
current policy. By appealing to an expanded conception of individualism and rationality, such critiques would be more difficult to dismiss than critiques grounded in social functionalism or personal irrationality. Such critiques nevertheless need to describe just how citizens or their government could tell which restraints paternalistically restrict freedom and which facilitate desirable precommitments.

Reading Elster is a bit like a visit to a social science theme park. Not everyone will enjoy all the amusements. Some will adamantly insist that even an afternoon on the rational-agency roller coaster is a waste of time at best, and nauseating at worst. Others, however, are likely to find the twists and turns invigorating, entertaining, and enlightening. As one who continues to think that rational-agency models are useful in interpreting behavior, and that pushing them to their limits is the best way to understand normative objections and descriptive paradoxes, I encourage you to buy the "E" (for Elster) ticket and enjoy the ride.

1. Marxism may offer the best alternative to intentional, methodologically individualistic, rational-choice models. It may therefore not be surprising that Elster his looked at Marxism in his Making Sense of Marx. It would take me beyond the themes of this review to examine this work of Elster's. but I will observe that one reviewer thought the book should be titled "Making Mincemeat of Marx" (Ryan, 1991).
2. An influential example is Stigler and Becker (1977).
3. For a review of this last option, see Rosenberg (1988).
4. In later work (Cement of Society, 1989, pp. 23, 103). Elster seems more comfortable with the successive selves metaphor.
5. Repeated game theory (Myerson, 1991, pp. 323-331), a possible way out, is notorious for its multitude of potential solutions.
6. Schelling (l963) clearly explained how one could improve one's negotiating position by eliminating other outcomes, that is, how "burning one's bridges" to prevent retreat can force an enemy to accept one's advances.
7. Despite the multidisciplinary approach of his work, Elster strongly believes that one should become expert within a discipline. Such expertise he argues is necessary to learn the "ethics of science" (Elster in Swedburg, l990. p. 240).


Barry, B. (1978). Sociology, economics, and democracy. Chicago: University of Chicigo Press.
Brennan, T. (1991). The trouble with norms. In K. Koford & J. MilIer (Eds.), Social norms and economic institutions (pp. 85-94). Ann Arbor: University of Michigan Press.
Elster, J. Weakness of the will and the free-rider problem. Economics and Philosophy, 1, 231-265.
Hardin, R. (1982). Collective Action, Baltimore, MD: John Hopkins Press.
[end of page 80, start of page 81]
Myerson, R. (1991). Game theory. Cambridge, MA: Harvard University Press.
Rosenberg, A. (1988). Philosophy of social science. Boulder, CO: Westview Press.
Ryan, A. (1991, October 10). When it's rational to be irrational. New York Review of Books. pp. 19-22.
Schelling, T. (1963). The strategy of conflict. Cambridge, MA: Harvard University Press.
Stigler, G., & Becker, G. (1977). De gustibus non est disputandum (You can't argue about taste). American Economic Review, 67, 76-90.
Swedburg. R. (1990). Economics and sociology. Princeton, NJ: Princeton University Press [end of page 81]

[Brennan, Timothy (1994), Talking to One Selves: The Social Science of Jon Elster, Journal of Communication 44 (1), p. 73-81]

[ Books | Articles | Reviews | Index | The main menu ]