Brief Therapy Based on Interrupting Ironic Processes: The Palo Alto Model

Abstract

The model of brief therapy developed by Fisch, Weakland, Watzlawick, and colleagues in Palo Alto is based on identifying and interrupting ironic processes that occur when repeated attempts to solve a problem keep the problem going or make it worse. Formulations of ironic problem-solution loops provide a template for assessment and strategic intervention, indicating where to look to understand what keeps a problem going (look for “more of the same” solution) and what needs to happen for the complaint to be resolved (someone must apply “less of the same” solution). Supporting research is preliminary but suggests this approach may be well suited for change-resistant clients.

This article describes a parsimonious therapy approach based on identifying and interrupting ironic processes that occur when repeated attempts to solve a problem keep the problem going or make it worse. “Brief problem-focused therapy” was developed over a quarter century ago by Richard Fisch, John Weakland, Paul Watzlawick, and their colleagues at Palo Alto’s Brief Therapy Center (Watzlawick, Weakland, & Fisch, 1974; Weakland, Fisch, Watzlawick, & Bodin, 1974). Although these authors have not themselves used the term “ironic process,” it captures well the leitmotif of their work—namely, that problems persist as a function of people’s attempts to solve them, and that focused interruption of well-intentioned solution efforts is sufficient to resolve most problems.¹

The Palo Alto model is “strategic” because the therapist intervenes to interrupt ironic processes deliberately, on the basis of a case-specific plan that sometimes includes counterintuitive suggestions (e.g., to “go slow” or to engage in behavior the client wants to eliminate). Calling it only “strategic therapy” may be misleading, however, because this label gives undue emphasis to intervention style and detracts attention from the more fundamental principle of ironic problem maintenance on which interventions are based.

From the perspective of the Palo Alto group, ironic processes are ubiquitous: They happen, for example, when trying harder to fall asleep keeps a person awake, when trying to suppress an unwanted thought increases the thought’s intrusion, when encouraging a depressed partner to cheer up results in more despondency, or when reasoning with a difficult two-year-old intensifies the child’s temper tantrums. Whether occurring within or between people, these processes persist because problem and attempted solution become intertwined in a vicious cycle, or positive feedback loop, in which more of the solution leads to more of the problem, leading to more of the same solution, and so on (Fisch, Weakland, & Segal, 1982). Most important, formulations of ironic problem-solution loops provide a template for assessment and strategic intervention: They tell us where to look to understand what keeps a problem going (look for “more of the same” solution) and suggest what needs to happen for the problem to be resolved (someone must apply “less of the same” solution).

In contrast to most of the other brief therapies described in this special series, the Palo Alto model of brief therapy is non-normative and complaint based: It makes no assumptions about normal or abnormal functioning; in fact, if no one registers a complaint about a behavior, there is no problem (Fisch & Schlanger, 1999). Nor does this approach attach importance to working through underlying emotional issues, teaching clients new skills, promoting developmental insights, or challenging dysfunctional cognitions: The therapist aims simply to resolve the presenting complaint as quickly and efficiently as possible so that clients can get on with life. Also in contrast to most other therapies, strategic intervention is deliberately minimal, and therapy is brief by design rather than default (Gurman, this issue)—less is best. It is assumed that interrupting current patterns of problem maintenance, even in some small way, can lead to further positive developments. More therapy (even of the same kind) confers no additional benefit once the complaint is resolved and can inadvertently contribute to continuance of the problem.

To accommodate the special-series format, we will first review the historical background of the Palo Alto model, then describe the formulations of problem maintenance that guide intervention, followed by the intervention techniques themselves. Subsequent sections consider selection criteria (whether there are any) and several lines of supporting research, both directly and indirectly relevant to the model. We conclude with some observations about possible future directions, not only for clinical applications of the ironic process idea but also for much needed empirical research.

HISTORICAL BACKGROUND

In the 1960s, long before managed care made brief therapy a prime commodity on the mental health market place, Weakland, Fisch, Watzlawick, Jay Haley, and other colleagues then associated with the Mental Research Institute (MRI) in Palo Alto were investigating therapeutic approaches to rapid problem resolution. There were two main lines of influence on this group: One was the “interactional” or “systemic” view of human problems that grew from Gregory Bateson’s 1952–62 research project, in which Weakland, Haley, and MRI founder Don D. Jackson were coinvestigators (Sluzki & Ransom, 1976); the second was the therapeutic work of Arizona psychiatrist Milton Erickson (Haley, 1973), with whom Weakland and Haley consulted many times during the Bateson project.

The Bateson group applied ideas from cybernetics and systems theory to the study of interpersonal communication, including communication in families of people labeled schizophrenic, and in so doing provided what many regard as the intellectual foundation of the family therapy movement (Hoffman, 1981). After the Bateson project ended, Watzlawick, Beavin, and Jackson brought many of these ideas together in Pragmatics of Human Communication (1967). Around the same time, Fisch, Weakland, Haley, and colleagues formed the Brief Therapy Center at MRI to investigate and evaluate ways of doing therapy briefly. They were especially attracted to the uncommon therapeutic tactics of Erickson, whose penchant for giving homework assignments, engaging patients in active (albeit sometimes metaphorical) dialogue, and getting quick results stood in sharp contrast to the then-dominant paradigm of long-term psychodynamic psychotherapy. Also at variance with the pathology-focused psychodynamic viewpoint was Erickson’s take on why people have problems in the first place: Rather than viewing problems as “symptoms” of some deeper, underlying pathology, Erickson’s implicit assumption was that problems arise from the mishandling of common, everyday difficulties encountered in the course of the family life cycle (Haley, 1973). Taken together, the influences of Erickson and the Bateson group represented a move away from the tradition of individual psychopathology toward understanding and treating problems as an aspect of ongoing social interaction (Fisch & Schlanger, 1999; Jackson, 1967).

Since its inception in 1966, the Brief Therapy Center (ETC) at MRI has treated nearly 500 cases in a consistent format. The group meets weekly as a team and works with unselected cases, representing a broad range of clinical problems, for a maximum of 10 sessions. One member of the team usually serves as the primary therapist while the others consult from behind a one-way mirror. After treatment another team member conducts a telephone follow-up interview with the client(s) to evaluate change in the original presenting complaint. The BTC’s pattern of practice has remained remarkably consistent to this day, and the center’s three core members—Fisch, Weakland, and Watzlawick—all participated regularly until Weakland’s death in 1995.

From this work emerged a model of therapy that makes no assumptions about normality or pathology and focuses narrowly on current, observable interaction around the presenting complaint. The first description of the model appeared in a 1974 Family Process article titled “Brief Therapy: Focused Problem Resolution,” by Weakland, Fisch, Watzlawick, and Bodin. The same year, Watzlawick, Weakland, and Fisch also published Change: Principles of Problem Formation and Problem Resolution, a more theoretical treatise that distinguished first- and second-order change and offered many examples of ironic processes. Eight years later, Fisch, Weakland, and Segal published The Tactics of Change: Doing Therapy Briefly (1982), which remains the most comprehensive and explicit description of the BTC’s clinical methods. In 1992, Weakland and Fisch concisely summarized their approach in a book chapter, and more recently, Fisch and Schlanger (1999) have provided another concise description of the model, along with clinical material illustrating Brief Therapy With Intimidating Cases.

Historical connections between the Palo Alto (ETC) model and two related approaches—“strategic family therapy” (Haley, 1980, 1987; Madanes, 1981) and “solution-focused therapy” (Berg & Miller, 1992; de Shazer, 1991; de Shazer et al., 1986)—should be noted as well. Haley collaborated with Weakland during the Bateson project and afterward helped to launch the ETC at MR1 before he moved east, first to work with Salvador Minuchin in Philadelphia and later to establish a training center with Cloe Madanes in Washington, D.C. This collaboration helps to explain why strategic interventions associated with Haley and Madanes are sometimes similar to those practiced by the MR1 group. Yet despite similarities, these approaches have important differences—for example, Haley (1980, 1987) makes assumptions about adaptive relationship structure and the protective function of symptoms that the Palo Alto group does not. Unlike Haley’s model, the “solution-focused therapy” developed by de Shazer and colleagues in the 1980s began as a derivative of the ETC approach. Inspired by the Palo Alto group, de Shazer et al. took “focused problem resolution” (Weakland et al., 1974) as the starting point for a complementary form of brief therapy they initially called “focused solution development” (de Shazer et al., 1986). While the ETC approach remained fairly constant, however, the solution-focused model underwent progressive revision, or evolution (de Shazer, 1991; Miller & de Shazer, 2000), and now appears to have a substantially different emphasis than the parent model (for a detailed comparison of the two models, see Shoham, Rohrbaugh, & Patterson, 1995).

Interestingly, one of the main differences between “focused problem resolution” and “focused solution development,” at least partly semantic, concerns the label “strategic.” Like Haley, the Palo Alto group sees therapy as unabashedly strategic, meaning that the therapist (or therapy team) intervenes deliberately, on the basis of a careful plan, and assumes responsibility for outcome (Fisch et al., 1982; Haley, 1963, 1987). In contrast, de Shazer, Berg, and colleagues now avoid characterizing their solution-focused therapy as strategic, describing it instead as collaborative, co-constructivist, and (by implication) not so manipulative (de Shazer, 1991). This (re) characterization aligns solution-focused therapy with the postmodern tradition of narrative therapy, which rejects the model of therapist-as-expert-strategist for the model of therapist-as-collaborative-partner (Nichols & Schwartz, 2000). Whether narrative and solution-focused therapies are as nonstrategic as their proponents claim, however, is in our view an open question—although calling one’s therapy “strategic” in today’s politically correct climate is probably not a very strategic thing to do (Shoham et al., 1995).

FORMULATIONS OF PROBLEM MAINTENANCE AND CHANCE

Palo Alto-style brief therapy is based on two interlocking assumptions about problems and change:

Regardless of their origins and etiology—if, indeed, these can ever be reliably determined—the problems people bring to psychotherapists persist only if they are maintained by ongoing current behavior of the client and others with whom he interacts. Correspondingly, if such problem-maintaining behavior is appropriately changed or eliminated, the problem will be resolved or vanish, regardless of its nature, or origin, or duration (Weakland et al., 1974, p. 144).

A clear implication in the above quote is that how a problem persists is much more relevant to intervention than how it originated. The Palo Alto group’s central observation, in fact, is that problem maintenance revolves precisely around what people currently and persistently do (or don’t do) in order to control, prevent, or eliminate a complaint. A “problem,” then, consists of a vicious cycle, or positive feedback loop linking a behavior someone considers undesirable (the complaint) to some other behavior(s) intended to modify or eliminate it (the attempted solution)—and this process, of course, is fundamentally ironic.

The Weakland et al. (1974) assumptions quoted above have two other implications: One is that the persistence of complaints, even personal complaints such as feeling depressed or anxious, has much to do with social interaction, in which the behavior of one person both shapes and is shaped by the response of others—yet as we will see, formulations of problem-solution loops can be “self-referential” (individual) as well. The last and most practical implication is that, to resolve a problem, it should not be necessary to understand or change its antecedent “cause,” but simply to break the ironic pattern of problem maintenance by promoting less of the same solution. If this can be done, even in a small way, then virtuous cycles can develop that lead to further positive change. Admittedly, this view of problems and change may strike some readers as too simple, or perhaps misguided if it leads therapists to ignore other important features of the clinical situation. Whatever the merits of this criticism, the Palo Alto model explicitly assumes that parsimonious intervention can be sufficient to resolve many, if not most, of the problems people bring to therapists.

Because social interaction is central to problem maintenance in this approach, relevant ironic processes are more often found between people than within them. Indeed, social contexts ranging from courtship and family interaction to the war on drugs provide countless examples of how well-intended, persistent attempts to influence someone else’s behavior can have consequences opposite to those intended. In Tactics, Fisch et al. (1982) describe interpersonal problem patterns maintained by solutions such as attempting to reach accord through opposition, seeking to gain compliance through voluntarism, or trying to dispel an accuser’s suspicions by defending oneself (and thus confirming them). Case illustrations in the literature also apply interpersonal variations of the ironic process idea to common “individual” problems such as depression (Coyne, Kahn, & Gotlib, 1987), anxiety (Rohrbaugh & Shean, 1987), substance abuse (Rohrbaugh, Shoham, Spungen, & Steinglass, 1995), eating disorders (Fisch & Schlanger, 1999), and child conduct disorder (Patterson, 1982). With depression, for example, the sensible solution of encouraging a despondent person to cheer up or be more active sometimes becomes part of an ironic problem-solution loop in which more encouragement leads to more despondency, leading to more encouragement, ad infinitum (Fisch & Schlanger, 1999; Watzlawick & Coyne, 1980).

Work with couples often focuses on the ironic pattern of demand-withdraw interaction (Christensen & Heavey, 1993; Shoham, Rohrbaugh, Stickle, & Jacob, 1998). In Change, Watzlawick et al. (1974) describe this pattern vividly:

In marriage therapy, one can frequently see both spouses engaging in behaviors which they individually consider the most appropriate reaction to something wrong that the other is doing. That is, in the eyes of each of them the particular corrective behavior of the other is seen as that behavior which needs correction. For instance, a wife may have the impression that her husband is not open enough for her to know where she stands with him, what is going on in his head, what he is doing when he is away from home, etc. Quite naturally, she will therefore attempt to get the needed information by asking him questions, watching his behavior, and checking on him in a variety of other ways. If he considers her behavior as too intrusive, he is likely to withhold from her information which in and by itself would be quite harmless and irrelevant to disclose— “just to teach her that she need not know everything.” Far from making her back down, this attempted solution not only does not bring about the desired change in her behavior but provides further fuel for her worries and her distrust— “if he does not even talk to me about these little things, there must be something the matter.” The less information he gives her, the more persistently she will seek it, and the more she seeks it, the less he will give her. By the time they see a psychiatrist, it will be tempting to diagnose her behavior as pathological jealousy—provided that no attention is paid to their pattern of interaction and their attempted solutions, which are the problem (pp. 35–36).

The “solutions” of demand and withdrawal in this example make sense to the participants, yet serve only to confirm each partner’s unsatisfactory “reality.” Asking how the cycle began, or what causes what, would not be relevant in the ironic-process framework, and probably cannot be determined. Describing the pattern would be the therapist’s first concern—and from a systems perspective, the problem-solution loop is its own explanation.

Ironic-process formulations are inherently case specific, or idiographic. In fact, drastically different solution patterns can (and often do) maintain the same type of problem in different cases and situations, and the same solution that helps to maintain a problem in one case may actually work well for another. For example, although demand-withdraw couple interaction plays a role in maintaining many drinking problems, there are other couples for whom a diametrically opposite solution pattern—doing nothing—figures prominently in problem maintenance. We have also seen directly opposite (e.g., permissive vs. coercive) parenting styles maintain the same kind of child behavior problem in different families. This idiographic aspect of ironic processes limits the extent to which any standardized intervention formula can be effective for the same complaint in different interactional contexts.

To interrupt an ironic process, one must first identify the particular solution efforts that maintain or exacerbate the complaint(s), and specify what less of those same solution behaviors might look like. Successful intervention then turns on persuading at least one of the people involved to do less—or better, the opposite—of what he or she has been doing (Fisch et al., 1982). Thus, if the thrust of a client’s solution effort is to push someone else, either directly or indirectly, to change (and this has the ironic effect of making change less likely), the therapist will focus on ways the client could do “less of the same”— for example, by declaring helplessness, demonstrating acceptance, or simply observing. Any of these actions would be incompatible with pushing for change and could suffice to break the ironic pattern.

Interestingly, despite their emphasis on social interaction, the Palo Alto group describes a “self-referential” aspect of complaints such as anxiety states, insomnia, obsessive thinking, sexual dysfunction, and other problems maintained by trying deliberately to “be spontaneous.” These complaints “can arise and be maintained without help from anyone else. This does not mean that others do not aid in maintaining such problems; often they do. We simply mean that these kinds of problems do not need such ‘help’ in order to occur and persist” (Fisch et al., 1982, pp. 136—137). To interrupt within-person ironic processes that maintain problems such as insomnia, anxiety, or intrusive thoughts, a therapist might coach a client to give up deliberate efforts to fall asleep, stave off anxiety, or suppress unwanted thoughts—perhaps by introducing incompatible behavior such as trying to stay awake, approaching feared situations, or scheduling times to deliberately entertain certain disturbing thoughts. The main reason these so-called paradoxical symptom prescriptions can be helpful is that, by attempting to have a symptom deliberately, the client cannot continue in usual ways of trying to prevent it.²

As a practical matter, it is not easy to help clients do less of the same, because this requires that they depart from habitual solution efforts that make sense to them and to which they are often deeply committed. The “art” in brief strategic therapy is to carefully frame suggestions for less of the same in terms consistent with clients’ preferred views of the problem, themselves, and each other. Indeed, grasping and using clients’ views—what Fisch et al. (1982) call “patient position”—is as fundamental to brief strategic therapy as the focused behavioral prescriptions that aim to interdict problem-maintaining solutions. This therapy thus has a crucial cognitive dimension, emphasizing not only what clients do, but also how they view things. The aim, however, is not to challenge or change a client’s views, as in cognitive-behavior therapy, but rather to accept and use those views in promoting less of some “solution” behavior. In other words, while behavior provides the focus for what to change, clients’ views are key to how.

The Palo Alto brief therapy model has several other features that distinguish it from more traditional therapies. First, it emphasizes description, not diagnosis. By requiring a therapist to think in terms of a description of the complained-about behavior and attempted solutions, the model “departs from focusing on what to label and undesired behavior and instead focuses on thinking about how and in what context the undesired behavior is performed” (Fisch & Schlanger, 1999, pp. 6–7; italics original). Thus, there is no assumption that a diagnosis is necessary, or even helpful, for determining an appropriate treatment. In fact, by focusing attention on what a person “has” (depression, attention-deficit disorder, etc.), and on factors in the person that may be causing this (brain chemistry, emotional conflicts, etc.), a diagnostic label tends to detract attention from what that person does in specific social contexts— and how other people react.

Another contrarian aspect is that the model attaches little importance to so-called “common factors,” such as the therapeutic alliance, that are known to contribute to the success of disparate therapy approaches. As if to disqualify the power of therapeutic relationship, the ETC team frequently uses more than one therapist with a given case—and their follow-up results indicate this makes no difference in outcome (Rohrbaugh, Shoham, & Schlanger, 1992). At the same time, as we know from observing live and taped BTC sessions, process researchers would rate BTC therapists very high on connectedness with their clients—as of course they must be in order to competently grasp and use the client’s “position.” Although some writers have proposed that working within the client’s world view is itself a powerful common factor and the main force in any and all therapy models (Miller, Duncan, & Hubble, 1997), this is not the position of the Palo Alto group, whose theory of therapeutic change is quite specific.

Finally, the Palo Alto model calls attention to ironic therapy processes, and by doing so provokes critical thinking about the limitations of “therapy” as a panacea for human problems. The popular media have documented this well, with examples ranging from “recovered-memory” and “attachment” therapies to Woody Allen’s interminable analysis. More subtle and mundane illustrations of ironic therapy processes occur, for example, when encouraging a client to “work through” marital complaints in supportive individual therapy makes it possible for the partners to avoid resolving the problem directly, when establishing a “positive therapeutic relationship” with a difficult child undermines the mother’s confidence and ability to parent authoritatively, and when pushing abstinence and change drives a substance abuser out of treatment. In general, one might suspect an ironic therapy process whenever a persistently applied intervention produces little or no improvement and one or more of the people involved fear that suspending the intervention would make the problem worse. Ironic therapy processes may be especially likely to occur when an intervention recapitulates a problem-maintaining solution applied by the clients themselves—a hypothesis supported by at least one empirical study (Shoham et al., 1998).

TECHNIQUES

In a nut shell, this is the Palo Alto group’s formula for doing therapy briefly: (a) define the complaint in specific behavioral terms; (b) set minimum goals for change; (c) investigate solutions to the complaint; (d) formulate ironic problem-solution loops (how “more of the same” solution leads to more of the complaint, etc.); (e) specify what “less of the same” will look like; (f) understand clients’ preferred views of themselves, the problem, and each other; (g) use client position to interdict problem-maintaining solutions; and (h) nurture and solidify incipient change.

Whether seeing one person or several, the therapist’s first task is to obtain a specific, behavioral description of the clients’ complaints. To do this, the therapist asks for examples and seeks to understand where, when, and for how long complaint behaviors typically occur. Because the complaint is not assumed to be a symptom of something deeper (like the tip of a psychological iceberg), the aim is simply to be clear about who is doing what. For example, if someone says they are “depressed” or have a “communication problem,” what exactly do they do or not do that indicates this? And how does the complaint interfere with what they would like to be doing instead? A useful guideline is for the therapist to have enough details to know what the problem might look like on film: “If we had a video of this, what would I see?” If there is more than one complaint, the one clients consider most pressing receives priority, as long as it can be clearly defined in behavioral terms. It is also crucial to know who, besides the complainant, sees the complained-about behavior as a problem—and why this is a problem now. The “why now?” question can provide useful information about the context of the complaint (e.g., ‘My girl friend said if I don’t do something about this she’s out the door”) and suggest where to look for potentially relevant solution patterns.

Eventually the therapist asks what, at minimum, will be an acceptable goal for change. To develop this a therapist might say, “What will you (or he, she, or the two of you) be doing differently that will let you know this (the complaint) is taking a turn for the better?” Besides establishing a criterion for evaluating later progress, questioning clients about their minimum change goals provides a focus for investigating problem maintenance (ironic processes) and conveys that change in the complaint is not only possible but expected.

Given a focused complaint, the next task in brief therapy is to understand the solution patterns that maintain the complaint. Straightforward questions such as “What have you and other people been doing to try to solve this problem?” followed by detailed inquiry about the specifics of those efforts help to illuminate cycles of problem maintenance. The aim is to identify specific interaction sequences in which more-of-the-same solution leads to more of the complaint, leading to more of the solution, and so on—and to do so in a nonblaming way. Often one finds multiple specific solution behaviors that reflect a common theme, or basic thrust, in what people are doing (e.g., trying to control something or someone, talking— or not talking—about the problem). The most relevant problem-maintaining solutions are current, but solutions tried and discarded in the past sometimes give hints about what has worked before—and may again. In one of our alcohol-treatment cases, a wife who in the past had taken a hard line with her husband about drinking later reversed this stance because she did not want to be controlling. Careful inquiry about past solutions revealed that the former hard-line approach, while distasteful, had actually worked—when the wife set limits, the husband had controlled his drinking. Later the therapist relabeled the hardline approach as caring and reassuring, which helped the wife reverse her stance again to break the problem cycle (Rohrbaugh et al., 1995). Thus, although problem-maintaining solutions usually involve someone doing something, what someone doesn’t do (e.g., comment on a loved one’s drinking or drug use) can be important as well. In any case, the therapist’s objective here is simply to recognize problem-maintaining solution patterns; it is neither necessary nor helpful at this stage to point them out to the client.

Based on the formulation of a problem-solution loop, the therapist can develop a picture of what “less of the same” will look like—that is, a picture of what behavior, by whom, in what situation, will suffice to reverse the problem-maintaining solution. Ideally this strategic objective (a term for therapists, not clients) constitutes a 180° reversal of what complainants have been doing. Thus, if a targeted solution is a client’s attempt to control or persuade someone, a reversal of this might involve taking an apologetic, one-down position or acknowledging that the other person may be right. Similarly, if the problem-maintaining solution is attempting to stave off some unwanted personal experience (e.g., anxiety or embarrassment), the objective might be for the client try to bring it on. How, when, and where to implement such reversals depends on identifying specific situations in which the problem-solution sequence occurs. For example, if a parent-child power struggle typically happens at bedtime, or a marital accusation-denial sequence happens when a husband comes home late from work, those would be ideal situations on which to focus an intervention. In addition to informing interdiction strategies, identifying problem-solution loops helps the therapist be clear about what positions and suggestions to avoid—what the Palo Alto group calls the “mine field,” and what we see as the principal breeding ground for ironic therapy processes.

The key element in planning an intervention is stopping the performance of the attempted solution, and in general it is better to ask clients to do something incompatible with the solution than to cease and desist. Some-times, however, a pattern can be interrupted at least temporarily if a client can be induced to observe what is happening in the problem situation, while suspending specific efforts to influence or change it. A therapist might say, “Let’s see what he does this week on his own, without your having to remind him or discuss the matter. It could be worse than you think, but at least then we’ll know what we’re up against. Please keep notes on what happens so we can review them together next time we meet—and if you slip back into being helpful again, be sure to note that too.” If successful, such a “diagnostic intervention” will create a wedge in the client’s usual solution efforts (in this case reminding and discussing) and a corresponding change in the complaint behavior. When positive change happens, even in a small way, the therapist can work to amplify and sustain it—perhaps first by suggesting further observation, and then by reframing the client’s new stance in a manner that encourages its continuation. In our experience, productive compliance with this prescription usually depends on the therapist defining “observation” as something the client must actively do and implying that what the client had been doing (the problem-maintaining solution) was understandable given the circumstances.

As noted above, a hallmark of the Palo Alto model is framing therapeutic tasks and suggestions in terms compatible with the clients’ own language, world view, or position. This is done to promote compliance, and it is assumed that persuading clients to depart from their usual ways of handling a complaint will not be an easy proposition. Assessing client position depends mainly on paying careful attention to what people say: How do they see themselves and want to be seen by others? What do they hold near and dear? What is their best guess about why the problem is happening, and why the various people involved handle it the way they do? The diversity of client views is enormous: One client may see herself as flexible and open-minded, while another values principle and rigorous commitment. One welcomes direction and advice; another wants to do things his own way. One attributes a child’s problem to inner turmoil; others, to willful disobedience. One wants to make a loving sacrifice; another just wants to get even, and so on.

The strategic therapist not only elicits clients’ views, but also shapes and extends them to set the stage for specific interventions. For example, a therapist might accept a wife’s view that her husband is uncommunicative and unemotional, then extend this view to suggest that his “defensiveness” indicates vulnerability. The extension paves the way for suggesting a different way of dealing with a husband who is vulnerable, rather than simply withholding (Coyne, 1988). Similarly, when working with relatives of a depressed stroke victim, the ETC team accepted and extended the relatives’ wish to encourage the patient by proposing alternative, less direct ways to do this that effectively reversed the prevailing solution patterns of exhortation and overprotection (Fisch et al., 1982).

Two key tactical principles in the Palo Alto model are “work with the customer” and “preserve maneuverability.” The customership principle means simply that the therapist should see the person or persons most concerned about the complaint (the “sweater” or “sweaters”), even when this is not the person whose behavior is being complained about. With child problems, for example, a therapist would work primarily with parents (assuming they are concerned about the problem), but also consult with others (e.g., a grandparent, teacher, or probation officer) if they are “sweating” as well. Similarly, a therapist treating a marital complaint would not require or even encourage the participation of a reluctant spouse, especially if this is what the principal complainant has been doing. The maneuverability principle means that the therapist aims to maximize possibilities for therapeutic influence: “The therapist, to put it bluntly, needs to maintain his own options while limiting those of the patients” (Fisch et al., 1982, p. 23). Accordingly, the Palo Alto group has outlined tactics for gaining (and regaining) therapeutic control, even in initial phone contacts, because “treatment is likely to go awry if the therapist is not in control of it” (Fisch et al., 1982, p. xii). Preserving maneuverability also means that the therapist avoids taking a firm position or making a premature commitment to what clients should do, so as not to constrain what the therapist can do later.

Taken together, the principles of customership and maneuverability explain why Palo Alto-style brief therapists often treat relationship problems, including marital complaints, by seeing people separately. With some marital complaints, as mentioned above, it is better to see a motivated spouse alone than to pursue a reluctant spouse who is not a customer. Seeing people separately also preserves the therapist’s maneuverability, especially when the parties are in heated conflict or hold sharply different views of their situation. Here separate sessions give the therapist more flexibility in accepting each person’s viewpoint and framing suggestions to each of them in a different way. This also helps the therapist avoid being drawn into the role of referee or possible ally, though the main goal is of course to promote change in what happens between the partners.

In addition to focused interventions that target specific problem-solution loops, the Palo Alto group uses general interventions applicable to a broad range of problems and to different stages of therapy. General interventions include suggesting that clients go slow, cautioning them about dangers of improvement, and making a therapeutic U-turn (Fisch et al., 1982). Most of these tactics are variations of therapeutic restraint, a stance highly characteristic of this approach. The most common general intervention is the “go slow” injunction, which a therapist might give to prepare clients for change, to convey acceptance of their reluctance to change, or to solidify change once it begins to occur. According to Fisch et al. (1982), “go slow” messages help to relax the sense of urgency that often fuels clients’ problem-maintaining-solution efforts, and can increase compliance with therapeutic suggestions.

Despite the emphasis given to particular tactics, the outcome of brief therapy rarely turns on a single intervention. Much depends on how the therapist nurtures incipient change and manages termination. When a small change occurs, the therapist’s usual stance is to respond with gentle restraint (e.g., “go slow”) and to build on the interdiction strategy that produced the change. This may require (re)framing the new solution behavior in a way that better approximates the client’s preferred views (e.g., as strong, protective, responsible, etc.) and thus solidifies its continuation. Special tactics may be used with clients who are overly optimistic or overly anxious (e.g., predicting or prescribing a relapse) or show reluctance to depart from their usual solution behavior (e.g., exploring “dangers of improvement”). One way to solidify change, and block reversion to former solution patterns, is to discuss how the people involved could bring the problem back—which of course they don’t want to do. Termination occurs without celebration or fanfare, and sometimes clients retain “sessions in the bank” if they are apprehensive about discontinuing contact with the therapist. If clients ask to work on other problems, the therapist suggests taking time out to adapt to change and offers to reassess the other problems later (Fisch et al., 1982; Rosenthal & Bergman, 1986).

Before closing this discussion of techniques, we should note that the Palo Alto approach has been criticized as manipulative because the therapist typically does not make the rationale for a particular intervention explicit to the client (Wendorf & Wendorf, 1985), and because, to achieve an effective framing, the therapist may say things he or she does not truly believe (Solovey & Duncan, 1992). Defenders, on the other hand, argue that responsible therapy is inherently manipulative (Fisch, 1990), that therapeutic truth telling can be disrespectful (Haley, 1987), and that good strategic therapy profoundly respects the client’s “subjective truths” (Cade & O’Hanlon, 1993).

SELECTION CRITERIA

The ironic-process model applies to all complaints people bring to therapists—so, in theory at least, there are no selection criteria for this therapy approach. Consistent with this, the ETC has from the outset taken unselected cases, requiring only a complaint and at least one complainant. In an archival study of ETC cases seen at MRI from 1967 through early 1991, we found a wide range of problems represented (Rohrbaugh et al., 1992; cf Shoham et al., 1995). The most prevalent complaints (about 45% of those sampled) concerned an interpersonal problem such as marital discord or family conflict, but the ETC also treated many complaints localized in an individual (e.g., anxiety, depression, procrastination, eating disorder, child conduct disorder), and over 15% of the identified patients (IPs) had previously been hospitalized. Regarding the question of selection criteria, treatment outcome in this archival study was unrelated to the nature of the presenting complaint and to whether or not the IP had been institutionalized. Still, because the BTC did not use diagnostic categories or standardized assessment procedures, and had no comparison group(s), it is difficult to know if its therapy was equally effective across types and severity of presenting problems.

Published case reports on strategic therapy often imply that this approach is especially indicated when other, more straightforward approaches don’t work, or are unlikely to work. Representative of this genre is Fisch and Schlanger’s recent book, Brief Therapy With Intimidating Cases: Changing the Unchangeable (1999). Here the authors exemplify how the Palo Alto principles of brief therapy have been successfully applied to problems such as severe depression, assault, self-mutilation, and starvation, as well as cases that present multiple problems or general incapacitation.

Using strategic interventions to deal with resistance or difficult cases is also a theme in the literature on psychotherapy integration. For example, Spinks and Birchler (1982) advised behaviorists facing a therapeutic impasse to take a “strategic detour”; Stanton (1981) proposed using strategic methods when a structural family therapy fails; and O’Hanlon and Weiner-Davis (1989) recommended falling back to MRI tactics (e.g., therapeutic restraint) when solution-focused techniques do not work or appear to become more of the same. A similar “last resort” principle has even been suggested for use of strategic techniques with families of schizophrenics (McFarlane & Beels, 1983).

Interestingly, a number of controlled empirical studies support the claim that strategic interventions are indicated for change-resistant clients—especially when those interventions interrupt an ironic process. These studies come from the tradition of Attribute X Treatment Interaction (ATI) research, which tests hypotheses about which clients benefit differentially from which treatments (Shoham & Rohrbaugh, 1995; Snow, 1991). In general, across problems including substance abuse (Shoham et al., 1998), procrastination (Shoham-Salomon, Avner, & Neeman, 1989), insomnia (Shoham, Bootzin, Rohrbaugh, & Urry, 1996), stress reactivity (Shoham-Salomon & Jancourt, 1985), and marital discord (Goldman & Greenberg, 1992), strategic interventions similar to those used by the Palo Alto group have proven to be superior to straightforward educational, affective, and skill-oriented interventions when clients were assessed to be more rather than less resistant to change.

SUPPORTING RESEARCH

Research on the Palo Alto model in its pure form is sparse. Apart from the ETC archival study and a controlled comparison by Goldman and Greenberg (1992) of emotion-focused and systemic (Palo Alto-style) couple therapy, evidence relevant to efficacy and effectiveness comes either from randomized clinical trials that include strategic/ironic-process components in a broader treatment package, or from controlled studies of specific (usually paradoxical) interventions related at least conceptually to the interruption of an ironic process. In this section, we briefly review research bearing directly or indirectly on treatment outcome, then summarize an overlapping but different line of supporting research documenting the ironic-process model of problem maintenance—including evidence of ironic therapy processes.

Comparative Outcome Studies

Comparative outcome studies of broader treatments that explicitly incorporate strategic techniques and/or ironic-process principles have been done with adolescent substance abuse, mixed child problems, adult depression, and marital discord. With adolescent substance abusers and their families, Szapocznik and colleagues found that adding strategic engagement interventions to a structural family therapy package increased both engagement and treatment retention relative to a community-based, treatment-as-usual comparison condition (Santisteban et al., 1996; Szapocznik et al., 1988). In another study, Szapocznik et al. (1989) compared a similar, strategically enhanced family treatment for mixed problems of childhood to two other conditions: psychodynamic child therapy and a no-treatment (recreational) control group. The two active treatments were both efficacious in reducing behavioral and emotional problems, but at one-year follow-up the families who had had strategic/structural therapy functioned better (and were more intact) than those in the psychodynamic condition. Also with heterogeneous child problems, Szykula and colleagues compared a “strategic family therapy” heavily influenced by the Palo Alto model to behavioral family therapy based on social learning principles, and found the two treatments equally effective (Sayger, Szykula, & Laylander, 1991; Szykula, Morris, Sudweeks, & Sayger, 1987).

Finally, we should note the striking results of a recent British study (Leff et al., 2000) comparing three treatments for adult depression: antidepressant medication, cognitive-behavior therapy (CBT), and a “systemic couple therapy” (Jones & Asen, 1999) that focused on interrupting problematic (and probably ironic) patterns of couple interaction. Systemic couple therapy was clearly superior to the other conditions in terms of treatment retention and outcome, assessed both at the end of treatment and at two-year follow-up. (The CBT arm of the trial was discontinued due to a high dropout rate.) Importantly, the cost of couple therapy in this study was ultimately no greater than that of medication.

A limitation of this comparative outcome research, in the context of our review, is that the independent (treatment) variables are diffusely defined and incorporate much more than the interdiction of ironic processes. A notable exception is Goldman and Greenberg’s (1992) study of couple therapy, which compared a “systemic” treatment based directly on the Palo Alto model and Greenberg’s own emotion-focused couple therapy to a wait-list control condition. The systemic treatment was modeled after the team approach of the Palo Alto group (including use of a one-way mirror) and “focused almost exclusively on changing current interactions, [positively] reframing patterns of behavior, and prescribing symptoms” (p. 967). Relative to the control condition, the two active treatments were both efficacious at termination, but at 4-month follow-up, couples who received systemic therapy reported better marital quality and more change in target complaints than the emotion-focused therapy group. Based on these results and their qualitative (clinical) observations, the authors suggested that the systemic, Palo Alto–style approach may be especially well suited for change-resistant couples with rigidly entrenched interaction patterns. In a subsequent review of empirically supported couple therapies, Baucom, Shoham, Meuser, Daiuto, and Stickle (1998) concluded that “the findings from this single investigation place systemic couple therapy into the category of possibly efficacious [treatments]” (p. 61).

Studies of Paradoxical Interventions

Another line of research indirectly supporting the efficacy of the Palo Alto approach comes from experimental analog studies of so-called paradoxical interventions. Most of this work has focused on individual problems likely (from our perspective) to be maintained by self-referential ironic processes that a symptom prescription would presumably interrupt. The complaints (and possibly relevant solution patterns) in these studies have included mild depression (trying to cheer oneself up), insomnia (trying to fall asleep), procrastination (forcing oneself to get work done), and anxiety (trying to avoid a feared situation, or trying to relax). Experimental manipulations have typically compared some form of symptom prescription either to a no-treatment control condition or to another, more straightforward intervention such as education, skill training, or progressive muscle relaxation.

A meta-analytic review of 12 randomized comparative designs found that paradoxical interventions compared favorably to no-treatment control conditions, with an overall effect size comparable to that reported for psychotherapy in general (Shoham-Salomon & Rosenthal, 1987). The meta-analysis also suggested that paradoxical prescriptions may be most effective when the therapist reframes the symptom positively before prescribing it— and in a recent study with insomnia patients (Shoham et al., 1996), we tested this hypothesis directly. The prescription to “stay awake,” when positively framed in a manner consistent with the insomniac’s own “position,” was superior to a wait-list control condition. Moreover, for clients high on reactance level, the positively reframed symptom-prescription led to better (shorter) sleep latencies at 6-month follow-up than a straightforward, progressive muscle relaxation intervention. In a related study, Akillas and Efran (1995) found that socially anxious men improved more when a prescription to “be anxious” was presented with a positive frame (rationale) than when it was not.

Although the paradoxical-intervention studies were well controlled, they too provide only indirect evidence of efficacy for the Palo Alto model of brief therapy. As noted above, paradoxical intervention is by no means synonymous with the Palo Alto brief therapy approach—and the technique emphasis in most of these studies precludes our knowing whether, if at all, a given paradoxical intervention was focused on interrupting some specific ironic process. With few exceptions, the interventions were applied in a uniform way across cases, with little attention given to the idiographic assessment of problem-solution loops and use of client position so characteristic of this therapy approach. We are left, then, with a single study (Goldman & Greenberg, 1992) that seems to provide an adequate test of the model’s efficacy. Nevertheless, based on the body of indirect but converging evidence accumulated so far, we believe that the Palo Alto brief therapy model will prove efficacious when properly tested.

The Brief Therapy Center Archives

Apart from the question of efficacy, some data relevant to the effectiveness of the Palo Alto approach are available from the BTC’s own archives. Its procedure, from the outset, has been for a team member other than the primary therapist to conduct a telephone follow-up interview with clients approximately 3 and 12 months following termination to evaluate changes in the presenting complaint, and whether additional problems had developed or further help was sought elsewhere. Based on the answers to these questions, the team then classified each case as attaining success (substantial or complete relief of the presenting complaint with no new problems), significant improvement (clear but not complete relief of the complaint), or failure (little or no change, negative change, or further treatment for the presenting complaint). Weakland et al. (1974) reported one-year rates of success, significant-improvement, and failure of 40%, 32%, and 28%, respectively, for the first 97 BTC cases studied. These outcome figures are comparable to success rates reported in the literature for other forms of psychotherapy (Smith, Glass, & Miller, 1980).

Eight years ago, in collaboration with BTC staff member Karin Schlanger, we updated the tabulation of outcome data for BTC cases seen through 1991 and looked for potential correlates of success (Rohrbaugh et al., 1992; cf Shoham et al., 1995). The outcomes at follow-up for 285 cases with complete interview data were consistent with the previous report, with rates of 44%, 24%, and 32% for success, partial success, and failure, respectively. To study correlates of outcome more closely, we identified subsamples of “clear success” cases (N = 39) and “clear failure” cases (N = 33) for whom one-year follow-up data were complete and unambiguous. Then, after coding a number of clinical, demographic, and treatment variables from each case folder, we examined these as potential predictors of outcome. The results were most striking in what they did not show: Outcome bore no direct relationship to type of complaint; to the age, gender, or education of either the main customer or the identified patient (IP); to whether the IP had had prior therapy, been hospitalized, or carried a psychiatric label; to the individual therapist, or whether more than one therapist had seen the case; or even to the number of BTC sessions. Only one of the variables we examined clearly differentiated the success and failure groups: In successful cases more different people were seen! Further analysis revealed that this last finding applied mainly to interpersonal (mostly marital) complaints—which does not support the Palo Alto group’s claim that marital complaints can be treated effectively by intervening through one spouse (Shoham et al., 1995).

One other notable finding from the archives project concerns the question of how brief effective brief therapy can be. In the mid-1970s the Palo Alto group undertook an experiment to test the feasibility of shortening treatment to five sessions. For over a year, new cases were randomly assigned to either a 5-session limit (N = 13) or the usual 10-session limit (N = 14). As it turned out, cases treated with the 5-session limit fared substantially worse than those allowed up to 10 sessions (p <.01), and when this pattern became clear clinically, the BTC abandoned its “five-session experiment.” Interestingly, over the full 24-year period, the average BTC case was seen only six times, and in neither the full cohort nor our smaller (success/failure) sample did number of sessions correlate with outcome. Thus, while most problems at the BTC were resolved with brief intervention, there was apparently a limit (i.e., 5 sessions) below which enforced brevity was counterproductive.

Although the in-house success/failure statistics from the MRI archives are well within the ballpark of therapy outcome research, there are good reasons why data such as these must be interpreted cautiously. The case records do not reveal, for example, how precisely the resolution of target complaints was defined in the interview, how goal attainment was scaled, whether collateral interviews were conducted in couple and family cases, or how demand characteristics might have colored the clients’ responses.³ In addition, the classification of outcome was done by consensus of the clinical team, which leaves open the possibility of a positive bias. Still, some outcome data for this approach are better than none—and we hope more rigorous research will soon begin to fill the gap.

Research on Ironic Processes

A different line of empirical support for the Palo Alto model comes from research documenting the role of ironic processes in problem maintenance and change. For example, ingenious laboratory experiments by social psychologist Dan Wegner, who coined the term “ironic process,” demonstrate what the Palo Alto group calls “self-referential” problem-solution loops in the realm of mental control. In one set of studies, Wegner and associates found that deliberate attempts to suppress an unwanted thought (e.g., trying not to think of a white bear) often lead, ironically, to increased thought intrusion, at least under stressful conditions of cognitive load (Wegner, 1994). Similar ironic effects have been found in laboratory studies of other mental-control tasks, such as putting golf balls and using Chevreul’s pendulum. While Wegner’s work focuses on ironic processes within people, other research demonstrates how ironic processes occur between people, especially when interpersonal influence is at issue. Ironic-like effects have been noted in studies of psychological reactance, where threats to a person’s behavioral freedom motivate efforts to restore the threatened freedom (Brehm & Brehm, 1981), as well as in naturalistic studies of parent-child and couple communication. For example, a “Romeo and Juliet effect” links parental interference to enhancements of adolescents’ romantic love (Driscoll, Davis, & Lipetz, 1972), and research on demand-withdraw couple interaction, described above, shows how one partner’s attempts to influence the other can have consequences opposite to those intended (Christensen & Heavey, 1993; Shoham et al., 1998).

Especially relevant to interpersonal ironic processes are studies of situations in which a therapist intentionally tries to eliminate some problematic or health-compromising behavior. In one of the procrastination studies cited above, Shoham-Salomon et al. (1989) found that clients actually engaged in more procrastination when they received direct therapeutic suggestions following a reactance-inducing experimental manipulation that threatened their freedom of choice. When the investigators measured clients’ reactance potential in a correlational replication study, those high on measured reactance likewise responded less favorably to straightforward self-control instructions than to a paradoxical symptom prescription, while an opposite pattern obtained for clients with low reactance potential. These results suggest that well-intended therapeutic “solutions” that limit people’s freedom to engage in behavior even they consider problematic can easily backfire, at least when the reactance potential of a person or situation is high.

We observed a similar ironic therapy process in a study comparing two treatments for couples in which the husband abused alcohol (Shoham et al., 1998). One of the treatments (CBT) took a strong stance about compliance with therapeutic instructions and used breathalyzer tests to ensure abstinence from drinking; the other treatment (family-systems therapy, or FST) was more permissive in this respect and used less direct, MRI-style strategies to deal with clients’ resistance. Before therapy began, we obtained observational ratings of demand-withdraw interaction, focusing on the pattern of wife-demand/husband-withdrawal, while the couple discussed the husband’s drinking. Based on the ironic-process model, we hypothesized that couples high on the demand-withdraw pattern would respond less favorably to high-demand CBT than to low-demand FST because CBT replicates a key part of the problem-maintaining solution: An alcoholic husband, in other words, may resist a demanding therapist in the same way he resists a demanding-wife. The outcome results clearly supported this prediction: When couples high on the wife-demand/husband-withdraw pattern were assigned to CBT, they attended fewer sessions, were more likely to drop out, and tended to have poorer drinking outcomes; for FST, on the other hand, demand-withdraw levels made little difference.

We have thus demonstrated ironic effects in two clinical contexts—one where an individual client resists influence from a therapist and the other where a male drinker resists persistent influence from his spouse and therapist. In both contexts, treatments that exert different levels of direct pressure for change appear to fail or succeed to the extent that they activate or avoid interpersonal ironic processes. Taken together, these results add further weight to the argument that systemic/strategic therapy is relatively better suited than many other approaches for difficult, change-resistant clients because it helps therapists avoid getting entrapped in ironic processes. The results also imply that the ironic-process model can be useful for predicting who will most and least benefit from different treatments, based on the match (or mismatch) between how a treatment attempts to exert influence and how clients participate in ironic influence processes in their own lives.

FUTURE DIRECTIONS

The hallmark of the Palo Alto model (to us its main strength, but to others perhaps a weakness) is conceptual and technical parsimony. Accordingly, we prefer to evaluate current and possible future directions according to whether they sharpen or dilute the model’s essential focus on identifying and interrupting ironic processes.

In the clinical area, one trend has been to integrate or, more often, incorporate elements of this model with other approaches. As noted above, several efficacious family therapy packages include strategic techniques that interrupt or avoid problem-solution loops, and the integration literature includes recommendations that MRI-style tactics be used when other, more straightforward approaches fail. Other authors have commented on how the Palo Alto model and other approaches might complement each other. For example, Duncan and Parks (1988) suggested that the “best ingredients” of strategic and behavioral approaches—techniques for compliance enhancement and techniques for skill acquisition—might profitably be combined.

Mixing ideas and techniques from different models is a common and probably reasonable way to approach therapy, but whether this can be done coherently and systematically is another question. Apart from the problem of how to blend interventions based on different theoretical assumptions about problems and change, there is the practical difficulty of specifying when, or under what circumstances, a therapist would use principles or tactics from one approach rather than another. Systematic integration is most feasible for approaches based on similar assumptions about problems and change, and several credible attempts to integrate the Palo Alto model with solution-focused therapy have been offered along these lines (Quick, 1994; Saggese & Foley, 2000). Another promising integration is Eron and Lund’s (1996) “narrative solutions” therapy, which retains the Palo Alto group’s focus on interrupting ironic problem-solution loops, but does this mainly by accentuating inconsistencies between clients’ solution behavior and their “preferred views.” Compared to most integrative and eclectic therapies, these approaches have the virtue of minimally diluting the Palo Alto model’s narrow focus on ironic problem maintenance and change.

In our view, one of the main impediments to advancing the Palo Alto model has been its historical association with constructivism and the more recent rise of postmodern therapeutic thought. At the level of practice, the idea of deliberate (strategic) influence by an expert therapist is anathema to the antirealist, antihierarchical ideology of current narrative and “collaborative” therapies, including solution-focused therapy. Although postmodernists credit the Palo Alto group for embracing constructivist conceptions of reality and rejecting normative views of problems, they fault any “objective” formulation of a problem-maintaining pattern (e.g., a problem-solution loop) that imposes a separation of therapist/observer from the observed individual/couple/family system (de Shazer, 1991). When applied to therapy, the postmodernist mind set actively undermines any theory that implies lawfulness in how problems develop or how people change (Held, 1998)—except possibly the “common factors” theory that itself rejects the relevance of specific therapy models (Miller, Duncan, & Hubble, 1997). Worse, this viewpoint promotes skepticism about the validity of knowledge gained through the “objectivist” methods of social science and dampens enthusiasm for much-needed rigorous research.

The most pressing priority for the Palo Alto model is systematic empirical research. This, after all, is a therapy that justifies itself as pragmatic, efficient, and effective—yet evidence to support that claim is far from overwhelming. Despite the model’s complaint-based, non-normative emphasis, an encouraging development is that some of its leading practitioners now advocate the routine use of standardized process and outcome measures (Johnson, 1995). We are optimistic that collaborations between researchers and strategically oriented clinicians will not only enhance the credibility of this important therapy approach, but also enrich our understanding of how it works (cf Goldfried, Borkovec, Clarkin, Johnson, & Perry, 1999; Lambert, Okiishi, Finch & Johnson, 1998).

Finally, in view of current interest in cost-effective treatments and a “stepped-care” approach to therapy (Haaga, 2000), it is important to identify the clients, problems, or clinical situations for which aiming to interrupt ironic processes will be sufficient and those that necessitate more extensive intervention. The minimalist approach we have outlined here may seem naive if one believes that effective treatments must address deeper client dispositions or how problems originate, yet the sufficiency of parsimonious treatments focused only on problem maintenance and complaint resolution is ultimately an empirical question.