The implications of recursive self-prediction suggest an answer to the age-old question of what the will consists of: The will to stick to a diet has the same nature as the “will” of the nations in World War II not to use poison gas, or of those since not to use nuclear weapons. This will is a bargaining situation, not an organ. In fact it can be well described in terms of bargaining theory.

The relationship of bargaining agents who have some incompatible goals but also some goals in common is called “limited warfare.” Countries want to win trade advantages from each other while avoiding a trade war; merchants want to win customers from each other while lobbying for the same commercial legislation; a husband wants to vacation in the mountains and his wife wants to vacation at the shore, but neither wants to spoil the vacation by fighting; a person today wants to stay sober tomorrow night and tomorrow night will want to get drunk, but from neither standpoint does she want to become an alcoholic. Whether the parties are countries or individuals or interests within an individual, limited warfare describes the relationship of diversely motivated agents who share some but not all goals.

Motivational theory hasn't paid much attention to recursive decision making, possibly because it's hard to study by controlled methods. If a phenomenon is determined by the interaction of A and B, then studying the influence of each while the other is held constant won't reveal the outcome. People who demand to know the causes of behavior will be unhappy with a recursive theory, one that says that the sum of individual causes explains little – that outcomes aren't proportional to any input or mixture of inputs, but to the volatile results of their interaction.

However, analysis of recursive decision making should greatly broaden the field that can be studied. That's what happened in economics when analysts moved beyond behaviors that were continuous functions of other variables and began studying decisions as the outcomes of bargaining games. However difficult it is to study nonlinear systems, such systems probably determine the most important features of choice. As one chaos theorist remarked, “nonlinear systems” may be about as extensive as “non-elephant biology.”

The best way to study recursive systems is to compare what is known about their behavior with models built of specified mechanisms. Direct experimentation may help, but only in verifying the operation of particular mechanisms. In the case of the will, as with the economy, parts that can be controlled are inseparable from a larger whole that's too complex and weighty to be controlled.

I have described a model of learned interests that compete freely on the basis of the time frames over which their rewards are preferred. The most important implication of such a model is an incentive within each interest to learn strategic behaviors that forestall competing interests.

If a person is a population of these kinds of roommates, each clamoring to control the use of the room, how does she make decisions? An interest can't eliminate a competitor simply by providing more reward than the other does, either at one time or on the average, since the competitor might undo the first choice when it became dominant at a particular time in the future. On the other hand, to continue to exist, each interest has to be the highest bidder at some time or it will extinguish; to achieve this, each may have to constrain others and can't be too constrained by them. Just because an interest is dominant at one moment in time doesn't mean it will get its intended reward; while an interest is dominant it has to forestall conflicting interests long enough to realize the reward on which it's based.

HOW ONE INTEREST BINDS ANOTHER

For long-range interests, this usually means committing the person not to give in to short-range interests that might become dominant in the future.

We're now at the heart of a central human paradox: that the better the will is at getting rewards, the less reward it will finally obtain. The paradox arises because the will only works – given its nature as a bargaining situation, we could say “only forms” – in tasks that have regular, clear-cut steps. This clarity fosters anticipation, which increasingly wastes available appetite through premature satiation and which the will is powerless to prevent in any direct way. Although this mechanism provokes solutions that must disappoint anyone seeking a recipe for rationality, it removes the apparent absurdity from three of the most basic human activities, which I'll now discuss: the construction of beliefs, empathy with other people, and motivated indirection in approaching goals.

THE CONSTRUCTION OF FACT PUZZLE

It's now common knowledge that people's beliefs about the world around them are heavily influenced by their own tacit choices, both “innocent” assumptions and wishful thinking. We have to decide so much about attending to or ignoring information that some “social constructivists” have put fact and fiction on a par, under the name “text.” To a great extent, belief does seem to be a goal-directed activity.

There have been plenty of books and articles that describe how irrational we are – in consuming drugs and alcohol and cigarettes, in gambling, in forming destructive relationships, in failing to carry out our own plans, even in boring ourselves and procrastinating. The paradoxes of how people knowingly choose things they'll regret don't need rehashing. Examples of self-defeating behaviors abound. Theories about how this could be are almost as plentiful, with every discipline that studies the problem represented by several. However, the proliferation of theories in psychology, philosophy, economics, and the other behavioral sciences is best understood as a sign that no one has gotten to the heart of the matter.

These theories almost never mention failures of will. This is just not a concept that behavioral scientists used much in the twentieth century. Some writers have even proposed that there's no such thing as a “will,” that the word refers only to someone's disposition to choose. Still, the word crops up a lot in everyday speech, especially as part of “willpower,” something that people still buy books to increase.

It's widely perceived that some factor transforms motivation from a simple reflection of the incentives we face to a process that is somehow ours, that perhaps even becomes us – some factor that lies at the very core of choice-making. We often refer to it as our will, the faculty by which we impose some overriding value of ours on the array of pressures and temptations that seem extrinsic.

All self-control devices can impair your reward-getting effectiveness: If you have yourself tied to a mast, you can't row; if you block attention or memory, you may miss vital information; and if you nip emotion in the bud, you'll become emotionally cold. Unfortunately, personal rules, which are the most powerful and flexible strategy against the effects of hyperbolic discounting, also have the greatest potential for harming your longest-range interests.

SIDE EFFECTS OF WILLPOWER

Suspicions of the will are fairly recent. Recognition of a will-like process that can oppose the promptings of impulse goes back to the classical Greeks, but until modern times it was regarded as an undiluted blessing. For example, Aristotle described not only passions that could overcome people suddenly, but also countervailing “dispositions.” These are forces that develop through consistent choice (habit) in one direction and subsequently impel further choice in that direction. He was clearly rooting for these dispositions to win out.

Lore abounds not only about how people mistrust their own future preferences, but how they sometimes engage in strategic planning to outsmart the future selves that will have these preferences. Here is Ulysses facing the Sirens or Coleridge moving in with his doctor to be protected from his opium habit. We know that the stakes in this intertemporal game sometimes reach tragic proportions. Yet we can't reconcile this game with utility theory's basic meat-and-potatoes notion that people try to maximize their prospects. The irony of smart people doing stupid things – or having to outsmart themselves in order not to – appears in literature again and again, but without an explanation.

This quandary may have been one reason for the popularity of cognitive explanations, which at least stay close to intuition. The problem hasn't undermined utility theorists, but it has cramped their style. They go from success to success in areas like finance and sociobiology, where tough competition selects strongly for individuals who function like calculating machines. However, their attempts to explain self-defeating choice on a rational basis have been unconvincing; the most notable has been the effort by economists Gary Becker and Kevin Murphy to show how a person who sharply devalues the future might maximize her prospective pleasure by addictive behavior. Their proposal is basically that devaluation of the future leads to addictive behaviors, which further increase this devaluation.

Our ideas about deciding divide into two kinds, each of which goes back to ancient times. Theorists from classical Greece to the present have focused on two conspicuous experiences in choice-making, wanting and judging, and have built models around each of them. Models based on wanting say that people weigh the feeling of satisfaction that follows different alternatives and selectively repeat those behaviors that lead to the most satisfaction. Models based on judging take a hierarchy of wants as given and focus on how a person uses logic – or some other cognitive faculty – to relate options to this hierarchy. The weighers, who include the British empiricist philosophers like David Hume, the psychoanalysts, and behaviorists like B. F. Skinner, developed satisfaction-based models that are called “hedonistic” or “economic” or “utilitarian.” On the other side, the German idealists, Jean Piaget, and modern cognitive psychologists like Roy Baumeister and Julius Kuhl have centered their explanations on judgments; this approach is called “cognitive” or “rationalistic.”

Skinner said that not only choice but also deliberation depend on differential reinforcement:

His explanation of addictions is simple: “The effects induced by [addictive] drugs reinforce the behavior of consuming them.” By contrast, cognitivist writers attribute addiction and other “misregulations” to various kinds of interpretive errors, as we're about to see.

I wrote Breakdown of Will in response to Cambridge editor Terry Moore's suggestion that I summarize Picoeconomics. This book is simpler and, I think, clearer. I have also added a great deal, both of research and theory, that I have discovered since Picoeconomics was published in 1992.

I've assumed no familiarity with hyperbolic discounting or intertemporal bargaining, so readers of Picoeconomics will find some repetition. However, if you've read the earlier book, you shouldn't assume that this book will therefore be a rehash of ideas you've seen before. In everything I've written I've thought it best to build from the ground up, rather than referring the new reader to works that may be hard to get; drafts of parts of this work have appeared not only in Picoeconomics but also in articles in Jon Elster's and Ole-Jorgen Skog's Getting Hooked and Elster's Addiction: Entries and Exits, The Journal of Law and Philosophy, and a precis in Behavioral and Brain Sciences. However, Breakdown of Will pulls these works together and goes beyond them.

You may be surprised by the conversational style I use. I've adopted this style partly for readability – as a discipline against too many subordinate clauses – but also from a belief that the supposed benefit of an impersonal voice (“the language of scholars”) is false. The fact that someone uses formal language doesn't mean she's objective, and formal language makes it harder to guess at her actual thought processes.

People's patterns of making self-defeating choices have seemed paradoxical from Plato's time down to the present. A patchwork of lore has accumulated to explain each particular paradox, but every local solution has been inconsistent with the solution that some other piece in the puzzle has seemed to require. As in the harder sciences, increased precision of measurement has revealed the possibility of a more comprehensive solution, which I present under the name picoeconomics (micromicroeconomics).

Choice experiments that were sensitive enough to test the difference between exponential and hyperbolic discount curves provided the necessary advance. Hyperbolic discounting confronts conventional utility theory with the likelihood that the conventional theory was not describing elementary principles of choice, but a higher-order cultural invention that doesn't necessarily operate in all people or in all situations. By demonstrating the basic instability of choice, this finding has promoted the problem of estimating value from a trivial matter of psychometrics into the crucial element of motivational conflict. Preferences that are temporary aren't aberrations anymore, but the starting place for a strategic understanding of functions that used to be thought of as organs: the ego, the will, even the self.

However much it has inconvenienced utility theory, the temporary preference phenomenon finally lets it explain self-defeating behavior. Furthermore, although hyperbolic valuation seems complex when compared with the exponential kind, it fits so many aspects of motivational conflict that it promises to simplify that subject substantially.

Our relative overvaluation of nearer experiences does a lot more than make us prone to addictions. It literally throws a curve into many experiences for which our norms are linear. In a culture where one of the basic properties of rationality is consistency, it makes us irrational at the outset.

I've discussed how the seemingly mystical idea of a will – free, more or less powerful, and somewhat brittle – describes the crux of our strategic response to temptation. The elusiveness of the will as a concept has historically come from the fact that it isn't an organ but a bargaining situation. Its brittleness comes from the often perverse inventiveness of sequential negotiators – each one the self, evaluating prospects from a shifting perspective – who are trying to maximize their prospects in a never-ending prisoner's dilemma. But however complicated the mechanism of willpower may seem, it's a neat little package compared to the other expectable consequences of intertemporal bargaining.

The most important departures from conventional utility accounting probably don't come from preference in the addiction range of durations – the urges that last for minutes to days and create the need for personal rules – and they don't come from the side effects of those personal rules themselves, even though these are considerable.

If people temporarily prefer shortsighted alternatives on a regular basis, how do they talk about the experience? It's not something that's supposed to be happening. It not only makes us ineffective in following our own long-term plans, it puts us at risk of exploitation by people who find out what our temporary preferences are. Therefore we might be expected to try to keep it from happening, and when we can't, to conceal it, perhaps even from ourselves. How we try to prevent it is the main subject of this book, and I'll get to it in the next chapter. First we should examine how temporary preference actually feels, since this discussion may otherwise seem rather removed from real life. It will turn out that many diverse experiences that have been thought to require special mechanisms can be explained instead by hyperbolic discounting.

ZONES OF DURATION OF TEMPORARY PREFERENCES

First of all, a temporary preference probably produces different experiences, depending on how long it lasts. Very short ones might not be noticed as preferences, while very long ones might seem wholehearted and not temporary at all.

Addictions

If we start roughly in the middle – not seconds or years but hours or days – we can see the clearest examples, which lead to the common clinical tragedies as well as personal frustrations.