Toward a Motivational Definition and Understanding of Addiction


William Miller, Ph.D.

Reprinted with permission from the Motivational Interviewing Newsletter for Trainers
(1998; Vol 5, No. 3; p. 2-6)


It has seemed intuitively obvious to me that addiction is fundamentally a motivational problem. I created quite a stir, however, when I uttered those words recently in the midst of an expert panel. I was nearly tarred and feathered for "resurrecting the moral model."

It turns out that the discomfort raised by saying that addiction is essentially a motivational problem has to do, in part, with a simplistic understanding of motivation. It is easily equated with the "Just say no" bootstrapping perspective that one should just, as the Brits say, "pull one's socks up" (I love that) and behave sensibly. That view does, indeed, reduce to blaming people for their problem.

Volitional Control

To say that choice is involved in addictive behavior is not to say that it is only a matter of choice. Janice Brown and I have proposed that one could quantify the proportion of variance in a behavior that is amenable to volitional control (Brown, 1998; Miller & Brown, 1991). That volitional quotient would vary within the same person over time, and differ across behaviors for the same person, and across people for the same behavior. Some behaviors, such as speech, are readily controlled by conscious attention (although one thinks immediately of individual variability in this as well). Some, like breathing, are controllable within some clear limits. Other phenomena that are ordinarily beyond volitional control can be brought under partial conscious control through training (such as biofeedback) and practice. The degree of variability among people in the volitional quotient will vary across behaviors. Some people find overeating extremely difficult to control, while others rarely give it a thought. With regard to drug use, a person's volitional quotient may vary widely across classes of drugs.

Within this perspective, addiction can be thought of (at least metaphorically) as a declining volitional quotient for a particular behavior or class of behaviors. The degree to which it can be controlled (stopped or reduced) by conscious intention diminishes over time. Perhaps more accurately, volitional control varies more across time and situations, which is consistent with Mark Keller's redefinition of loss of control as unpredictability. The subjective version of this is "using more than I intended." In natural usage, the term "addiction" implies some reduction in volitional control of a behavior.

Yet diminished volitional control of behavior (sometimes termed "diminished capacity" in criminal law) is a necessary but not sufficient condition for what people refer to as addiction. Over time, a person with Parkinson's disease experiences diminishing control over tremors, but no one would call this addiction. The term requires a further judgment that there are conditions under which the person could exercise volitional control, which brings us a step closer to natural language usage. Studies in the 1960s clearly demonstrated that even chronically dependent drinkers with access to alcohol could regulate their drinking given sufficient incentives to do so, at least under controlled laboratory conditions (Heather & Robertson, 1983). More recent treatment research using monetary payment for drug-free urines similarly illustrates the ability of chronically dependent users in the community to alter their cocaine and other illicit drug use in response to incentive. Even the most extreme "disease model" programs that profess solely neurochemical origins of addiction ultimately rely upon the client's volitional abstinence (Milam & Ketcham, 1981). Neither intoxication nor substance dependence mitigates responsibility for committed crimes in the minds of most juries. Indeed, a majority of America's 1.6 million prisoners are incarcerated for crimes committed under the influence of alcohol or other drugs. This, too, reflects the public view that addictive behaviors may involve diminished but not abolished self-control and concomitant responsibility. Self-regulation is retrievable.

Diminished Deterrence

A diminished but retrievable capacity for self-regulation is still not sufficient to constitute what people mean by addiction. Individuals with Parkinson's disease can often stop or diminish their tremors by conscious effort, at least until more profound deterioration occurs. A further condition for addiction is that it inflicts apparent risk or harm to oneself and/or others, and yet persists. Abstinence from food yields a withdrawal syndrome (hunger, malnutrition) but eating does not constitute an addiction in public usage. Only overeating does, in that it inflicts risk and harm. Nor does persistence despite harm by itself warrant application of the term "addiction." Criminal behavior often persists despite the fact that it inflicts harm at least on others, yet it is rarely regarded as addiction. It is the combination of (1) a behavior persisting despite apparent risk and harm, and (2) diminished but retrievable capacity for self-regulation that constitutes what is meant in social usage of the term addiction, at least in the United States.

It is also what has usually been meant by the term "addictive behavior." These two defining characteristics are not limited to the use of dependence-producing psychoactive drugs. Although the establishment of physiological dependence contributes to diminished capacity for volitional control, some of the most widely abused drugs produce little or no withdrawal syndrome. Further, the same conditions for addiction are met by a variety of "compulsive" behaviors such as pathological gambling, which only by the wildest stretch of imagination can constitute a "brain disease." What does and does not qualify as addictive is determined, at least in natural language, by a very large degree of sociocultural judgment about the above two defining conditions.

The description of a dependence syndrome expanded the diagnostic concept of dependence beyond its prior limits of physiological tolerance and withdrawal. It added a set of behavioral components that can now constitute a dependence diagnosis in themselves, even in the absence of physiological adaptation. Those symptoms boil down to an inordinate priority given to the behavior: increased time spent in the (drug-related) behavior, deferent withdrawal from other activities, decreased variability in the addictive behavior, persistence of the behavior despite risk and harm, avoidance of situations where the behavior is inaccessible, high priority given to the behavior when resumed after abstinence (rapid reinstatement), and subjective sense of diminished ability to restrain the behavior.

Said another way, the two defining conditions of usage for the term "addiction" are:

  1. The person is willing to pay a price for the behavior that seems, from a normative social judgment perspective, inordinately high.
  2. The person is judged to have diminished ability to regulate the behavior.

It is the latter of these two conditions that brings with it decreased social blame and censure for the harm caused. In criminal law, there are mitigating conditions such as diminished capacity, absence of general and specific intent, and insanity. In society, the imagery of illness is more often used to describe diminished capacity (mental illness, addictive disease). The beneficial sociologic functions of claiming "sick" status have been well described by Talcott Parsons. The assignation of disease or illness provides the individual with access to well-established forgiveness and restoration rituals. There are, however, other routes to compassionate understanding, and the sick role itself has disadvantages. Ironically, those who endorse beliefs consistent with the disease model of alcoholism also tend to endorse moralistic attitudes such as "alcoholics are liars and can't be trusted" (Moyers & Miller, 1993). The disease model, at least in the U.S., has become intertwined with the very moralism it was intended to overcome.

Too High a Price to Pay

Some fascinating current research is exploring why it is that certain harmful drugs (such as tobacco) and behaviors are so powerfully reinforcing. Some drugs, such as cocaine, are so inherently rewarding that laboratory animals will work to the point or exhaustion or starvation in order to continue receiving doses. Other addictive behaviors acquire their reinforcing properties. Gambling is rewarded on a variable ratio reinforcement schedule, one of the most potent schedules known for producing high rates of behavior that are resistant to extinction.

Besides diminished volitional control, what qualifies high-rate behaviors as addicting is that they persist despite harmful consequences. The person is willing to pay what seems too high a price in order to continue them. Said another way, motivation for the behavior has become more attractive than alternative rewards. Again, it is not the case that the person is completely unable to regulate the behavior. Given sufficient incentive - in extreme, a million dollar reward or a gun to the head - self-control is possible. In fact, much smaller rewards have been shown to regulate the use even of highly reinforcing drugs like cocaine and heroin (Higgins et al., 1993; Stitzer & Kirby, 1991) and of alcohol in dependent individuals (Heather & Robertson, 1983).

Competing Motivations

The problem of addiction, then, is one of competing motivations. The term "motivation" here is understood not simplistically as will power, but as involving complex biopsychosocial factors. A classic psychology course in motivation covers a broad range of determinants of behavior including biological drives, learning and conditioning, cognitive processes, emotion, and social influence. A similarly broad perspective is needed to understand the factors that interact to yield the phenomenon of human addiction, where the motivations favoring continuation of the behavior outweigh its harm and the perceived value of available alternatives. It is unsurprising that addictive behaviors seem irrational, because rational cognition is only one of a host of motivational factors.

The route out of addiction involves finding alternatives that are more motivating. Again, the competing motivations may be multiple in number and kind. The suffering associated with an addictive behavior tends to increase over time, shifting the ratio of pros and cons. This is reflected in the concept of "hitting bottom" and having "suffered enough" for change to occur. There is wide recognition of the concept of "high bottom" individuals whose behavior turned around before negative consequences reached dreadful proportions. Decisional balance models and more recent behavioral-economic analyses (Tucker & Vuchinich, 1998) also reflect this perspective of competing motivations.

Sometimes the shift of balance seems utterly external and obvious. A medical professional is caught diverting opiates, and his or her license and livelihood depend on clean urines. A parent is threatened with the loss of marriage and family. A sizeable estate is inherited, with installment payments contingent upon being a non-smoker. An employer gives a worker one more chance to sober up in order to keep a lucrative and rewarding job. In cases such as these it is plain that contingencies have shifted in the social environment. Staying drug-free becomes more reinforcing than continued use, which is now reliably followed by a loss of significant sources of positive reinforcement.

Then there are those one-time occurrences, even "oddly trivial" events, that people often name when asked about why they quit. These events do not in themselves signal an actual change in social reinforcement contingencies. A cocaine user's eyes slightly shift focus, and he sees himself in the mirror behind the line of powder. A smoker's dog dies of lung cancer. A dependent drinker's beloved pastor stops by for a visit while she is at home intoxicated. Or my favorite example: Premack's (1970) smoker who leaves his children standing in the rain in front of the library while he drives away for cigarettes. What is happening here? Premack called it "conscience." Whatever it is, it seems to involve a sudden shift in meaning, in how the person perceives the pros and cons of the behavior. In one sense, it is as though one or more cons have suddenly become dramatically more salient, taking on a higher value weight. In another sense, it is as if the person steps outside the self for a moment, to see himself or herself from another perspective.

I wonder if this points us to some pieces of the puzzle that has been occupying me for some time: Why is it that motivational interviewing works at all? We seem to know the style and strategies, the operations that evoke change. We know (I think) how to teach people to do it. The efficacy of the approach is replicable across cultures. But why does it work? How can it be that a person who has been persisting for years in a pattern of dependent drinking or drug use despite clear negative consequences, abruptly shifts that pattern after an hour or two of motivational counseling? How is it that having a single session of motivational interviewing before beginning a course of outpatient or inpatient rehabilitation program can double a person's chances of abstinence three months later? The person has learned no new coping skills or conditioned responses, and there have been no changes in the "actual" external contingencies operating in the person's life. The effect of a single counseling session ought to be lost, overwhelmed, drowned by the noise of the social environment. Yet it is not. What theory do we have about what is going on here?

I'm not sure that I can do much better than metaphor at this point, but a picture is starting to emerge in the puzzle. It seems to me that without any overt changes in the external environment, the client nevertheless leaves a successful session of motivational interviewing with a new set of contingencies. From Julian Rotter to radical behaviorism, after all, contingencies have been recognized to be perceived relationships between behavior and consequences. Rule governed behavior is characterized (and sometimes faulted) for its unresponsiveness to "actual" contingencies in the environment. A cognitive map of how things are overrides how things "really" are. It is the perception of consequences, and not only veridical schedules of reinforcement, that shapes behavior. What we are seeing in motivational interviewing may be a sudden shift in how the person perceives the consequences of his or her behavior, and perhaps in the salience of those consequences. It is as if the person leaves with a whole new set of contingencies governing (or beginning to govern) the behavior. Some might describe this as a shift in the stimulus equivalence set to which the behavior belongs.

How does this happen? I have placed a good deal of emphasis on eliciting self-motivational statements from the client. This feels right to me, and is consistent with Daryl Bem's self-perception theory, but I confess I am not certain that this is how it works. Another possibility that occurs to me is that for a brief time in motivational interviewing, we lend clients another perspective, a mirror, a chance to step safely outside of their own frame of reference and to see themselves with new eyes. This is not done by saying, "Listen to me. Here is how I see you," which places the person in the role of a passive listener. It is done by a temporary kind of merging. From the perspective of the therapist we call it empathy, seeking to see the world through the eyes of the client. In a metaphoric sense, we temporarily step inside the client, or better - become one with the client. Naturally, this improves the therapist's understanding of the client, but I think that it also changes the client's perspective. It is as if the client, too, can step into this empathic frame of reference and look back upon himself or herself.

I think that at least two things happen when that is done well. First, I believe that the client is able to see, saliently, some of the consequences of his or her own behavior, as from the perspective of an observer. Call it shame or conscience or hidden observer, there is a conscious process of perceiving in a new way, of seeing, feeling, contingencies. Second, I believe that we also lend clients our perspective of hope for them (Yahne & Miller, in press). It is the magic in my favorite Pygmalion study (Leake & King, 1977), and it is the madness of Don Quixote. From the merged perspective of empathy, the person sees that something is possible, and the seeing begins to make it possible. It was Fritz Perls' definition of teaching: to show a person that something is possible. We refer to it as supporting self-efficacy, but I think it's more than telling a client, "you can do it." It is somehow helping the client see that he or she can do it.

This is quite comprehensible from the perspective of a health belief model and the appraisal theories that have succeeded it, or from self-regulation theory. Protective change occurs when a person sees (1) a serious risk (discrepancy), and (2) the possibility of decreasing it. The four types of self-motivational statements that we have recommended be elicited from clients are statements of perceived problem, concern, intent to change, and ability to change. Turn the model around to promote approach rather than avoidance, and change occurs when a person sees (1) a worthwhile goal (discrepancy), and (2) the possibility of attaining it. The self-motivational statements then become statements of perceived opportunity, value, intent to attain it, and ability to attain it.

How, then, does all this relate to addiction as a phenomenon of natural language? Recall that addiction is perceived by an observer when a behavior appears to be continually pursued at too high a cost, and there is diminished capacity for self-control. When the client becomes that observer, liberation from addiction occurs as he or she sees - in this case through the eyes of an empathic merger - that the cost of the behavior is indeed too high, and that he or she does have the means to change it.

According to Rabbi Dr. Abraham Twerski, the capacity to take the perspective of another and to consider the consequences of one's actions is one of the features that distinguishes human beings from other living things. It is one of those differences between people and animals to which Frank Logan (1993) alluded in saying that we possess good animal models of the acquisition of addiction, but not of recovery. When we temporarily merge with another person empathically, we not only take on that person's perspective but also lend them one. I again raised some eyebrows recently by saying to a scientific audience that this is essentially a form of loving.

In any event, I see an emerging motivational model of addiction. It differs in some ways from, or at least builds upon, the motivational model of behavior change implicit in motivational interviewing. The preceding is a very rough attempt to stimulate some thought (my own, but especially yours) toward a theoretical model that could guide our clinical work through a coherent understanding of the problem we seek to treat. Because the model involves a balance of perceived contingencies, it does not negate and can incorporate a broad range of biopsychosocial motivations that are involved in addictive behaviors. I have illustrated my raw, unpolished thinking process, rambling through studies and stories, law and literature, metaphors and a bit of mysticism. Who has some more pieces to this puzzle?


Brown, J. M. (1998). Self-regulation and the addictive behaviors. In W. R. Miller & N. Heather (Eds.), Treating addictive behaviors (2nd ed., pp. 61-73.)

Heather, N., & Robertson, I. (1983). Controlled drinking. London: Methuen.

Higgins, S. T., Budney, A. J., Bickel, W. K., Hughes, J. R., Foerg, F. E., & Badger, G. J. (1993). Achieving cocaine abstinence with a behavioral approach. American Journal of Psychiatry, 150, 763-769.

Leake, G. J., & King, A. S. (1977). Effect of counselor expectations on alcoholic recovery. Alcohol Health & Research World, 11, 16-22.

Logan, F. M. (1993). Animal learning and motivation and addictive drugs. Psychological Reports, 73, 291-306.

Milam, J. R., & Ketcham, K. (1981). Under the influence: A guide to the myths and realities of alcoholism. Seattle, WA: Madrona.

Miller, W. R., & Brown, J. M. (1991). Self-regulation as a conceptual basis for the prevention and treatment of addictive behaviours (pp. 3-79). In N. Heather, W. R. Miller, & J. Greeley (Eds.), Self-control and the addictive behaviours. Sydney: Maxwell Macmillan Publishing Australia.

Moyers, T. B., & Miller, W. R. (1993). Therapists' conceptualizations of alcoholism: Measurement and implications for treatment. Psychology of Addictive Behaviors, 7, 238-245.

Premack, D. (1970). Mechanisms of self-control. In W. A. Hunt (Ed.), Learning mechanisms in smoking (pp. 107-123). Chicago: Aldine.

Stitzer, M. L., & Kirby, K. C. (1991). Reducing illicit drug use among methadone patients. In R. W. Pickens, C. G. Leukefeld, & C. R. Schuster (Eds.), Improving drug abuse treatment (pp. 65-90). Rockville, MD: National Institute on Drug Abuse.

Vuchinich, R. E., & Tucker, J. A. (1998). Choice, behavioral economics, and addictive behavior patterns. In W. R. Miller & N. Heather (Eds.), Treating addictive behaviors (2nd ed., pp. 93-104). New York: Plenum Press.

Yahne, C. E., & Miller, W. R. (in press). Evoking hope. In W. R. Miller (Ed.). Integrating spirituality in treatment. Washington, DC: American Psychological Association Books.

  [overview - special populations - group approaches]
The Library
[abstracts - bibliography - reprints - ongoing work - minuet newsletter - links]
  [upcoming training - mint trainers - training videos]
Created and maintained by:
Chris Wagner, Ph.D. and Wayne Conners, M.Ed.
Mid-Atlantic Addiction Technology Transfer Center
A CSAT Project

In cooperation with the Motivational Interviewing Network of Trainers (MINT), William R. Miller, Ph.D., and Stephen Rollnick, Ph.D. 

Revised 1/03