Defining Psychotherapy

With managed care looking over our shoulders and an increasingly sophisticated client base forcing us to make sharper distinctions between what we do and the banalities of pop psychology, it’s the rare practitioner who hasn’t felt the pressure to achieve better, more reliable therapeutic results. In fact, the last 25 years of the history of psychotherapy, commemorated in this special anniversary issue of the Psychotherapy Networker, might well be dubbed the field’s Age of Accountability. Even the remnants of Freud’s adherents have felt the pinch. In his new book, Practical Psychoanalysis for Therapists and Patients, renowned analyst Owen Renik argues that the only way to stop his profession’s downward spiral is to prove that analysts, too, can produce measurable symptom relief. Recently, the clinical psychology faculty at a nearby university felt obligated to drop the two therapy courses in their program that didn’t teach cognitive-behavioral techniques and to focus the curriculum entirely on empirically supported treatments. They told the graduate students that mastering these cutting-edge techniques would put them ahead of the curve, and that they should consider themselves fortunate to be witnessing the end of psychotherapy’s “anything goes” era.

Back in 1993, prompted in part by pressure from managed care, Division 12 of the American Psychological Association (APA) created a task force to develop a list of “empirically supported” treatments (ESTs). The attempt to distinguish hype and clinical zealotry from reliable science was aided by legions of aspiring graduate students and their research-savvy professors, who conducted hundreds of psychotherapy outcome studies. Nevertheless, despite all this scientific hoopla, there’s still no compelling evidence that therapists are achieving better outcomes today than they did 25 years ago. With rare exceptions, virtually all treatments produce about the same (modest) level of results. For example, in The Great Psychotherapy Debate, metanalyst Bruce Wampold reports that meaningful differences between approaches are nonexistent, particularly when the investigators’ allegiances are taken into account. Similarly, therapy critics Barry Duncan and Scott Miller report that the equivalence of methods continues to be the “most replicated finding in the psychological literature, encompassing a broad array of research designs, problems, and clinical settings.” To this finding, social scientist Robyn Dawes adds that therapist credentials—Ph.D., M.D., M.S.W., or no degree at all—make no real difference in outcome. Evidently, whatever works in therapy has little to do with the specialized knowledge people bring to the enterprise. The hope that therapy manuals would increase treatment effectiveness hasn’t been fulfilled either. In fact, according to Duncan and Miller, therapists working from manuals sometimes perform more poorly, perhaps because they’ve developed “better relationships with their manuals than with [their] clients.”

So why, despite all the seeming therapeutic advances and the enormous energy that’s gone into establishing therapy as a legitimate science, do we have so little empirically to show for it? In the past quarter-century, we’ve seen truly stunning advances in medicine, including the completion of the Human Genome Project, the invention of imaging technologies such as the fMRI and the PET scan, the development of drugs to lower cholesterol, the improvement in minimally invasive surgery techniques, the discovery of more specific immunosuppressant drugs for organ transplants, and so on. But, aside from lots of hype and the overheated claims of advocates of the various new methods in the therapy field, what comparable successes can therapists point to?

The 25th anniversary of the Networker offers an opportunity to ponder the cavalcade of developments in this field over the past several decades and examine the efforts to establish the scientific foundations of psychotherapy. What follows is a look at some of the theories and methods of treatment that have been proposed and also discredited through the years. More than an exercise in nostalgia, it’s an attempt to understand why the field’s progress appears to resemble a continuous ramble across the same outcome plateau, rather than the story of dramatic, scientifically documented breakthroughs. But, more important, it poses the question of whether our very conception of psychotherapy has yielded a range of misleading expectations and limited the advance of the field

The Golden Age of Learning Theory

When I was a graduate student in the early ’60s, learning theory was supposed to save the field’s bacon—rescuing the profession from the vagaries of psychoanalysis, the technique-centered orthodoxy of Carl Rogers, and the hot tubs of Esalen. My fellow students and I were taught that psychopathology could and should be understood as a series of conditioned responses. Therefore, instead of fostering insight or searching for underlying causes, our job was to eliminate bad habits and reinforce more adaptive responses. Applying learning theory in the clinic was trumpeted as the beginning of an era of science-based, action-oriented methods that would eventually put the softer “talk” therapies out of business.

As a student, I was duly impressed by the sweeping behaviorist agenda. Joseph Wolpe had just invented systematic desensitization—a method for eliminating human phobias by pitting progressive muscle relaxation against the person’s fear responses. It sounded good to me and, better yet, seemed simple enough for even a neophyte therapist to implement. At my first practicum site, I was eager to put this new methodology to the test. Unfortunately, my practicum clients refused to get with the program. Instead of phobias, they complained of relationship entanglements, work conflicts, school failure, shoplifting episodes, and so on. The “bad habits” approach turned out to be a lot harder to apply than I’d first imagined.

At the same time, one of my friends (and classmates) was asked to see a client named Eugene—intense, wiry, heavily tattooed, and court-referred. Although the police never formally charged him with the crime, the body of a homeless individual with whom he’d been seen drinking the night before turned up floating under a nearby bridge the next morning with multiple stab wounds. As soon as the consulting room door closed, Eugene glared at him and said simply, “I could kill you in a minute.” “Really?” sputtered my friend, “I would have thought you could do it in half that time.”

That response, born of my friend’s survival instinct and natural wit rather than any particular learning-theory dictum, caught Gene off guard, and he remained stock-still for what seemed like an eternity. Then he sat back in his chair and let out a hearty laugh. My friend’s off-the-cuff parry had evidently hit all the right notes, signaling the start of an unlikely but solid friendship. In subsequent weeks, you could hear the two of them chuckling and carrying on halfway down the hall.

In the year that followed, Gene was never late and never missed a session. He spoke candidly and passionately about his life, including his nefarious “business” deals, his hopes for reestablishing contact with his children, his sexual exploits, and his upcoming court battles. He wept openly when recounting the day he saw his best friend shot. During their work together, my friend became Gene’s role model, substitute parent, legal advisor, sounding board, confidant, and fellow traveler. When their sessions ended, Gene reminded him to call, day or night, if anyone should ever “give him trouble.” The offer was heartfelt. There was no doubt that their work together was therapeutically powerful, although I would have been hard-pressed to parse it in learning-theory terms.

In my own initial forays with clients, learning-theory principles took a back seat to my desire to simply get through sessions without becoming tongue-tied or mouthing meaningless platitudes. As I struggled to find my “voice” as a clinician, I found the advice of my psychoanalytically oriented supervisors much more helpful than anything I had read about or heard discussed in my theoretical seminars. This created a bind. I’d already rejected psychoanalysis as bunk and I was fiercely protective of my scientific bona fides as a therapist. Yet I wanted to become an expert clinician like my supervisors. I wasn’t sure exactly what the secret of their talent was, but whatever it was, I knew I wanted some.

What about Science?

The question of whether therapy was or would soon become a science was a hot topic among my classmates. Over beer and pizza, we debated endlessly whether the theories of B. F. Skinner, Albert Bandura, or Julian Rotter had much to offer a working clinician. Could psychoanalysis be researched? Had John Dollard and Neal Miller made Freud scientifically respectable? Would the faculty ever permit a dissertation on Fritz Perls’s empty-chair technique?

We found it ironic and disturbing that the faculty members who talked most about therapy as science never saw clients. One of the exceptions was a psychoanalytically trained game theorist who frequently allowed us to observe his sessions from behind a one-way mirror. He was masterful with clients, but, unlike others on the faculty, openly scoffed at the idea of therapy and science becoming cozy bedfellows. I once asked if he thought “reinforcing” an egocentric client for bringing flowers home for his wife was a good idea. He stared at me as if I was from another planet and said, “You don’t really believe any of that crap, do you?”

At one of our late-night bull sessions, the classmate who was Gene’s therapist concluded that therapy was, at root, just a form of paid companionship. Therapists were there to befriend folks who lacked a social support system of their own or whose existing relationships were too conflict-ridden to be useful. Actually, this notion of therapy-as-friendship was explored in William Schofield’s popular book Psychotherapy: The Purchase of Friendship. The idea had little appeal for us, however, not because it didn’t fit our experience, but because we couldn’t stomach the notion that we were putting in all this hard work in graduate school just to become therapeutic geishas. For the same reason, we were reluctant to accept Jerome and Julia Frank’s proposition from Persuasion and Healing that therapists were merely modern-day medicine men, reawakening hope and capitalizing on the placebo effect.

Over the next several years, my theoretical allegiances were admittedly wobbly. I was disillusioned with behavior therapy in general, and systematic desensitization was fast losing its scientific legitimacy. Investigators had reported that they could omit the presumably active ingredient (progressive relaxation) and still get the same results. Later, I personally collaborated on a study that showed that an entirely bogus “treatment” (that purposely violated the laws of learning) would also work, provided we convinced participants that they were improving.

As psychologist Robert Fancher wrote in Cultures of Healing, the behavior therapy movement was based on “high-minded scientific wishes and hot-headed . . . contempt for its predecessors,” but it ended with “the rapid decline of its putative scientific basis.” Everyone except the zealots soon realized that cognitive factors played a central role in all therapies, including those labeled “behavioral.” Perhaps animal suffering could be adequately understood in terms of simple conditioned responses, but human problems involved thoughts, images, beliefs, attitudes, and expectations.

Adding the “C” to CBT

Impressed by the potential importance of cognition, psychologist Michael Mahoney, a staunch behaviorist, began paying regular visits to Aaron Beck’s Center for Cognitive Therapy in Philadelphia and maintaining a running dialogue with the folks at Albert Ellis’s Institute for Rational Living in New York. In 1977, he founded the Cognitive Therapy and Research journal and worked to insure that the missing “C” was added to what soon became cognitive-behavioral therapy (CBT). The purists considered him a traitor. They’d fought hard to take the “talk” out of therapy, and now here was one of their own electioneering to put it back in. They declared the newly coined phrase “cognitive-behavioral” an oxymoron, warning that reseating cognition at the therapeutic table would represent a giant step backward for the profession.

Despite these dire predictions, CBT quickly gained popularity and became the poster child for the melding of science and practice. Almost as soon as it was announced, cognitive scientists and learning theorists disavowed any similarity between their own work and the propositions of CBT. As Fancher noted, Aaron Beck’s theory of psychopathology was “very lovely” but also highly simplistic and fundamentally flawed. Gloomy thoughts don’t necessarily cause depression and, in fact, the reverse is more likely to be true. Furthermore, when people’s moods shift, they automatically tend to think more optimistically, even though neither their logic nor their ability to appraise evidence has improved. In short, the relationship between depressed thoughts and feelings is much more likely to be part-whole than cause-effect. Although CBT fares about as well as any other approach in outcome studies, those findings do little to strengthen the causal claims that form the core of the model. Moreover, the method’s central technique—”disputing” erroneous cognitions—has rarely proven effective in studies that test the component features of a specific approach.

Mahoney eventually left the CBT fold and found himself critiquing the movement he’d helped create. In his later writings, he declared CBT the grand embodiment of the “myth of rational supremacy.” This is the popular but erroneous assumption that irrational thoughts are invariably dysfunctional and that “thinking and reasoning can and should control . . . one’s life.” Given the inadequacies of its theory and the simplicity of its method, the miracle is that CBT works at all, yet it does work moderately well, probably because clients bond with their therapists and accept what they’re told. After all, the CBT message that a bit of rational thinking can work wonders is appealing, wholesome, and as American as apple pie—a fitting companion piece to Nike’s “just do it” philosophy.

A recent incident with a neighbor clarified for me why I have such trouble accepting CBT’s premises. Fred had consulted a local CBT therapist, complaining of depression. At the therapist’s direction, he filled out a “behavioral activation form” listing activities he experienced as pleasant or reinforcing. At the following meeting, Fred and his therapist agreed upon how many points each of these activities was worth. In successive weeks, Fred’s homework assignment was to earn a set number of points by choosing various mood-enhancing projects. I knew about this arrangement because one evening when Fred was having dinner at our house, he mentioned that a meal with us was a featured item on his list, worth, I think he said, 15 points.

Fred felt that this activities approach was working. Therefore, a week later, while I was giving him a lift to the post office, I was surprised to hear him say that he was feeling blue. Remembering the high marks he gave the list exercise, I asked if he was therefore planning to do some items from his list in order to perk up his mood.

“Nope,” he explained, “I already have enough points for this week.”

I found this response puzzling. “So, let me see if I understand this,” I said, groping to follow his logic. “When you do these activities, you feel better. At the moment you aren’t feeling that great. You could easily do a couple of these activities and that might help. However, you don’t intend to do it because you already have your points for the week?”

“You got it,” he said.

“But,” I protested, “if these activities help you feel better, why not do more of them even if you don’t need the points?”

“Well, to tell the truth,” Fred confessed, “I guess I’m just not ready to feel better.”

Somehow, I knew exactly what he meant. Sometimes we want to savor our mood rather than chase it away. Sometimes we experience a deep need to nurse our wounds and rehearse our regrets. Ruptures in the pattern of meanings that connect the past with the present and the future require repair work—missing “connective tissue” has to be restored. Such repairs call for quiet contemplation or empathic conversation rather than a quick change of subject. Thus, it may have been productive for Fred to sit for a while with his mood rather than immediately zap it. In my experience, CBT manuals have no room for these sorts of subtle but important motivational distinctions.

From Psyche to System

While the cognitive-behavioral movement chugged along, a group of dedicated investigators, headquartered mainly at the Mental Research Institute in Palo Alto, were moving in an entirely different direction: they were attempting to apply the science of cybernetics and communications theory to the study of human problems. By the early 1980s, their work had inspired the creation of a large array of family, systems, and strategic approaches. Here, at last, was a truly science-based viewpoint that avoided reducing human experience to a collection of disembodied (replaceable) thoughts, a catalogue of individual personality traits, or a series of conditioned responses. These pioneers seemed willing and able to grapple with the complexities of human interaction.

I recall one of the first times I became aware of the potential power of a family focus in therapy. I’d been asked to see a young woman who’d precipitously dropped out of a prestigious college, taken a low-paying job in the community, and had her phone service shut off. When I met with her, she acknowledged smoking a lot of pot and described herself as having poor self-esteem and bad study habits. But she was less forthcoming about the desperate power struggles going on in her family. Little by little I learned, for example, that it was her overbearing father who’d insisted she enroll in the same Ivy League school as her high-achieving sibs, and who was annoyed that she wasn’t following in their footsteps. Although she couldn’t get along without the family’s financial support, she was weary of listening to her father’s lectures about her lackadaisical lifestyle, and she felt humiliated whenever he asked her to explain her mediocre grades. Unable to confront him directly, she did the next best thing: she withdrew from school (without telling him) and had her phone disconnected.

As I began to understand the larger significance of her school withdrawal, I called a family conference during which I explicitly championed the girl’s right to make her own career decisions. An indication of the meeting’s success was that the next day, right after her parents left town, she promptly had her phone service restored. She eventually returned to college, but only after working at a craft shop and living on her own for a number of years. By slightly altering the balance of power in the family, that afternoon’s conference seemed to accomplish more than I might have achieved through weeks or months of individual therapy. In fact, focusing on the young woman’s study habits, pot smoking, and lack of ambition might only have reinforced the family’s oppressive message.

I believed then, and still do, that the family and systems movement had enormous potential, but unfortunately professional hubris hampered what it was able to accomplish. As family researcher Carol Anderson puts it, the field was “rife with overblown rhetoric, overstated claims, and unsupported assumptions,” as well as a “too easy acceptance of anything said forcefully by a leader with strong convictions.” Too often, charismatic figures presided over rival factions, advocating elaborate and expensive formats that required one-way mirrors, reflecting teams, multigenerational sessions, and, in some cases, the suspension of logic.

For instance, it was typical to portray the family system as a clever enemy that resists change and must be outfoxed at every turn by the use of counterintuitive, strategic ploys. Moreover, symptoms were often said to serve disguised system functions. For example, if a boy refused to go to school, the systemic therapist might hypothesize that he was really staying home to protect his mother from unacknowledged spousal abuse. The straightforward possibility that he was just afraid of the schoolyard bully was rejected as hopelessly naive. The systemic therapist was expected to tip the system’s balance using “paradoxical injunctions” and fancy “reframes.” Unfortunately, this often entailed telling untruths to the family (for their own good). Many of us thought that such manipulative practices gave the family-systems approach a bad name, stretching the limits of ethical practice and good science.

A particularly wacky outgrowth of unbridled systems thinking was the so-called “invariant prescription” proposed by Italian psychiatrist Mara Palazzoli Selvini. For example, she advised the parents of psychotic adolescents to leave home for the evening without warning or explanation. They were simply to leave a note saying “we won’t be home tonight.” When they returned, they were prohibited from saying where they’d been or why they’d left. If asked, they’d simply state that “these things concern only the two of us” (which was obviously untrue). This strategy was supposed to destabilize the family structure, restore a healthy hierarchy, and enable the psychotic family member to become symptom free. I’m not making this up. Needless to say, such strategies had neither scientific legitimacy nor a solid research footing.

Things deteriorated further when the field became embroiled in a heady debate involving the relevance of constructivist thinking to family treatment. In a nutshell, the philosophy of constructivism asserts that human beings play an active role in creating their own realities. For some reason, a number of influential family therapists interpreted this to mean that clinicians should immediately stop diagnosing, prescribing, interpreting, and advising. For example, Harlene Anderson and Harry Goolishian argued that therapists should adopt a “not-knowing” stance and avoid presenting themselves as experts. (As I see it, professionals who take the stance of “not-knowing” ought to also take the stance of “not-collecting-fees.”) In any event, this vow of radical neutrality seemed like an unwarranted return to the naive impartiality that Rogerians had been forced to give up years before, after their own researchers had demonstrated that all therapists, regardless of intent, inevitably influence their clients. (This is why Rogers began to emphasize authenticity rather than just reflection.)

The ill-conceived application of constructivism to family therapy sullied the reputation of both. It left clients and therapists baffled and gave critics a grand opportunity to accuse members of the profession of spouting “epistebabble.” Again, this clinical experimentation wasn’t accompanied by any serious research effort. In fact, summarizing the overall state of family therapy research, Carol Anderson puts it this way: “We started with nothing and still have most of it left.”

Back to the Psyche

When the family movement eventually faded, I was disappointed to see that therapists focused once again on the individual. As far as I can tell, interpersonal elements are almost entirely missing from recent methods such as Eye Movement Desentisization and Reprocessing (EMDR), Energy Psychology, Acceptance and Commitment Therapy (ACT), and Positive Psychology. Perhaps it was too much to expect that systems thinking, even if it had been better implemented, could derail the juggernaut of our individualistic culture, the pharmaceutical industry, and the medical model. However, the premature disappearance of a contextual focus is highly regrettable because, as philosopher Wilhelm Dilthey recognized a century ago, meaningful human behavior is inevitably contingent on communal interaction and can’t be successfully studied as a series of individual, decontextualized components. In my own work, I’ve become increasingly alert to contextual issues, although I rarely meet with entire families the way the systems therapists used to do.

Judging by the workshop announcements that come across my desk, emotion-based therapies have become popular again. The return of emotion to center stage is yet another reminder that the cycles of the therapy business rival those of the fashion industry. In 1946, Franz Alexander and Thomas French created a buzz by suggesting that “corrective emotional experience” was the key to successful therapy. Now, six decades later, Susan Johnson makes the same claim in her recent Networker article on Emotion Focused Couple Therapy. Yet, we still don’t have an adequate definition of emotion, not to mention various auxiliary concepts such as emotional dysregulation, affective resonance, broken attachment bonds, and so on.

Gregory Bateson once complained that the social sciences traffic in concepts “so loosely derived and so mutually irrelevant that they mix together to make a sort of conceptual fog” that hinders any scientific progress. That’s why I think Johnson overstates the case when she claims that we now have “a clear, coherent, researched theory of adult love.” In my experience, clinicians still have no real understanding of how behavior, cognition, and emotion are related. Their continued use of the lay definitions of those terms forestalls the development of a more biologically sophisticated roadmap of how human beings function.

As far back as 1958, personality theorist George Kelly warned that parsing experience into those three Aristotelian categories—behaviors, thoughts, and feelings—”confuses everything and clarifies nothing.” Evidently, we haven’t learned that lesson. Therefore, we continue cycling through therapies that variously claim to be behaviorally based, cognitively based, or emotion focused.

Technically speaking, all human activity is “emotional” because it always involves shifting calibrations of the hormonal and musculoskeletal systems. Sleeping, fighting, making love, and studying may not be equally dramatic, but they’re equally visceral. Although the settings are different for every class of behavior, there are no disembodied (nonemotional) activities. Therefore, what unites the themes that interest Johnson and her colleagues—abandonment, hurt, loneliness, and vulnerability—isn’t that they’re “emotional,” but that they involve survival threat (real or imagined, physical or psychological). Isn’t that what all therapies are supposed to be about?

Turning Toward the East

It’s hard to know what to make of the current attraction to any method with “acceptance” or “mindfulness” in the title. Do these Eastern imports represent a rebellion against our ultramechanistic Western practices? I doubt it. My hunch is that we’re witnessing a side effect of the Age of Accountability. Therapists no longer feel comfortable promising clients a complete personality makeover or total symptom remission. Therefore, they’ve adopted more limited treatment goals. The teaching of meditation and mindfulness skills is a perfect strategy for this more modest therapeutic era. If clients can’t get rid of their troubling thoughts, they can at least be taught how to hover lightly above them: they can learn to “float, not fight.” Perhaps combining Eastern philosophy and Western pharmaceuticals isn’t such a bad idea. Anxiety can be rendered less stressful, anger less vexing, and depression less disheartening. Moreover, the improvements are potentially measurable. However, let’s hope that the Western mental health establishment doesn’t turn these noble Eastern traditions into cheap, feel-good gimmicks.

While everyone is in an “accepting” mood, I suggest that temperamental styles be added to the “package” therapists urge their clients to explore and acknowledge. I rarely hear therapists talk about temperament with their colleagues or their clients, yet, I’ve repeatedly found that my clients are grateful whenever I broach the subject. They’re reassured to learn that aspects of their personality—particularly traits they don’t care for—are at least partly rooted in their genetic heritage. Many have been blaming themselves or their parents for being jittery, experiencing stage fright, having temper tantrums, avoiding social gatherings, or developing obsessive thoughts. They’re relieved to discover that solid research evidence, including sophisticated studies of identical twins reared apart and adopted children, links such tendencies to basic biology. Because of the profession’s strong environmentalist bias, clinicians have long ignored evidence of genetic imperatives and uncritically accepted the results of flawed socialization research. For example, many therapists continue to believe that abused children grow up to abuse their own children—the so-called “cycle of violence.” Actually, the vast majority of abused children do nothing of the sort. The aggressivity of the minority seems to be due to their genetic overlap with the abusing parent, rather than to the harshness of their upbringing.

The genetic literature suggests that built-in happiness set points determine how individuals will react to good news and deal with bad news. This information has to be factored into our therapeutic work with anxious and depressed clients.

All temperamental styles have advantages and disadvantages. As I see it, an important role for therapists is to assist clients in discovering how best to work with their natural inclinations. As they say in the East, it’s easier to ride the horse in the direction it’s going!

Accountability Redux

At last count, therapists could choose from among 500 different treatment techniques. This catalogue of possibilities will undoubtedly be winnowed down over the next decade, as the drumbeat for accountability grows louder. Unfortunately, the treatments that remain won’t necessarily be the most effective or the most scientific. It’s worth remembering that one can perform outcome research on anything—advertising campaigns, museum visits, television-viewing habits, and distance healing. Positive results don’t establish or ensure scientific credibility. As I’ve already noted, I personally consider most CBT packages scientifically suspect, even though I realize that they dominate the lists of empirically supported treatments. Because practically every therapy tested succeeds to about the same degree, all that such lists really tell us is that some methods have been studied more than
others. Furthermore, the amount of research support a given method receives seems to be mainly a function of whether it’s popular with academics and easy to manualize (CBT wins on both those counts).

Investigators Barry Duncan and Scott Miller have repeatedly complained that we spend too many of our research dollars searching for the best therapy method, when the preponderance of evidence suggests that therapeutic success is largely a function of client and therapist characteristics, as well as the quality of the relationship that develops between them. They’re right, of course. In fact, most of us endorse a similar position each time we make a referral—we tend to refer to therapists we like and respect, even if their methods or theories don’t jibe with our own.

It’s understandable that those determined to build a science prefer to place their bets on methods rather than personalities. In a sense, the field’s idolatry of the perfect method has been an inevitable byproduct of our commitment to the medical model and the disease-entity approach—an arrangement that former APA president George Albee frequently described as a “pact with the devil.”

Classifying psychotherapy as a medical treatment is an example of what philosophers call a “category mistake”: something has been placed in the wrong conceptual envelope. The term psychotherapy itself (mind-treatment) mystifies what therapists actually do. Thus, we keep attempting to describe, analyze, and study our practices using a vocabulary that’s ill-suited to the purpose. It’s misleading to call meeting with someone a “treatment,” or to label a person’s fervent search for meaning an “anxiety symptom.” Although recent versions of the DSM have wisely substituted the term disorder for disease, the medical connotations remain. Yet even Freud understood that therapy is a verbal interchange and not a medical procedure. Therapists have no salves to apply, no antibiotics to prescribe, and no surgical instruments to wield. Therapy is a rhetorical exercise, not a healing process.

The problem is that when a category error is made, the fallout can persist indefinitely, quietly contaminating every theoretical debate and polluting every research decision. No wonder that despite all the technical innovations, we’re only marginally better off today than 25 years ago. It’s difficult to make progress if you’re deluded about basic principles. Because of our linguistic imprecision, we’re “double blind” (to borrow cyberneticist Heinz von Foerster’s felicitous phrase). In other words, we can’t see what we can’t see. Psychiatrist Thomas Szasz’s proposed solution is to return as quickly as possible to “unadorned, forthright speech” and forgo “the obscurities and fake profundities of professional jargon.” In fact, he warns that the longer therapy is “situated in the context of health care,” the more likely it is “to self-destruct.”

If we look squarely at the fundamentals, it becomes apparent that therapy is neither science nor art—it’s conversation. Conversation is at once the most subtle and complex of all human activities, and our most important problem-solving tool. Some are better at it than others. No one doubts that Bill Clinton has qualities of “attunement” that other politicians don’t share. Similarly, we accept that there are inspiring teachers, crackerjack lawyers, bestselling authors, and top-grossing salesmen. Why are we less willing to acknowledge that some therapists are better than others, and that there are super-therapists whose sterling results can’t be explained by current research?

In Social Intelligence, psychologist Daniel Goleman suggests that differences between communicators may be largely innate and traceable to differences in brain structure and chemistry. He writes that “deep listening seems to be a natural aptitude; good listeners rarely have that ability because they have been taught.” Goleman doesn’t rule out the possibility that a person’s attunement skills can be improved through training, but it remains to be seen how substantial or long-lasting such tutored gains might be. In fact, those are the kinds of issues that deserve research attention. In the meantime, it’s a detriment that we’ve implicitly endorsed the managed care assumption that all “providers” are equivalent and equally capable of dispensing some sort of homogenized therapeutic product. The evidence shows that even two therapists trained on the exact same manual obtain different results.

An Alternate Vision

When I was in graduate school, I was too intent on becoming a “doctor” to pay close attention, but I recall George Kelly arguing that a therapist was a kind of specialized research consultant. I think this proposal is worth a second look. It illustrates one possibility for escaping the category mistake that’s ensnared the rest of us. Kelly was always clear about his opposition to the disease-entity approach. Whereas others talked of curing symptoms, he spoke only about helping clients resolve their pressing life concerns. Kelly believed that when clients attempted to solve their problems themselves, they’d keep rattling around in the same unproductive slots—they needed help reformulating their questions so that they could find better answers. In that connection, he loved to tell the story about the drunk who was asked by a passerby why he was on his hands and knees under the lamppost. “I’m searching for my keys,” was the reply. “Is that where you lost them?” asked the good Samaritan. “No, I lost them in that alley, but the light’s much better over here.”

For Kelly, the therapist’s job was to get the client’s inquiry process back on track. He often described the therapy session as a “protected laboratory where hypotheses [could] be formulated, test-tube sized experiments [could] be performed, field trials planned, and outcomes evaluated.” His model was neither restrictive nor scientistic. He considered all forms of personal inquiry legitimate.

The therapy “lab” had to be a place where clients could count on hearing the truth. Otherwise, how could they get an accurate “read” on the life propositions they were testing? Kelly also understood that because the client’s quest required fresh input, it was disadvantageous for the therapist’s ideological precommitments to overlap those of the client. That might result in both of them ending up in the same investigatory ditch, hemmed in by their own assumptions. To help therapists avoid such traps, he urged them to keep in mind that “whatever exists can be reconstrued.” In other words, that no solution is ever the only solution.

Are We There Yet?

David Orlinsky, psychotherapy investigator and coauthor of How Psychotherapists Develop, admits that he reads the psychotherapy research literature as little as possible: “The language is dull, the story lines are repetitive, the characters lack depth, and the authors generally have no sense of humor.” On a more serious note, he complains that the current paradigm distorts “the actual experience and lived reality of what it presumes to study.” In Orlinsky’s view, it does an injustice to the craft of psychotherapy to reduce the process to a set of specific procedures that can be mechanically “taught, learned, and applied.” Although a novice store clerk will make a few sales by following the company manual, the top-grossing salespeople have long ago thrown away the manual, recognizing that the key to effective sales isn’t to be found in a book. Similarly, therapists working from a treatment manual will have their successes, but are they practicing the kind of therapy one could recommend to a relative? If we ourselves were clients, would we want our anxiety level “authenticated” by a Beck Anxiety Inventory or would we insist instead on telling our story to someone who appreciates life’s complexities? As Orlinsky correctly points out, the dominant research paradigm sanctifies and perpetuates a highly constricted and overly concrete conception of clients, disorders, therapists, and the change process. Therefore, the kinds of symptoms and methods academicians study bear little resemblance to the practitioner’s world. Although the studies reported in the literature have the “trappings of normal science,” they contribute only minimally to the field’s progress.

I’d be delighted if researchers came up with a reliable method that guaranteed surefire results, even for novices, but because psychotherapy is a conversation rather than a treatment, it simply isn’t amenable to the kind of science that requires scripted protocols. Continu-ing to force-fit what we do into the medical model adds a layer of mystification that profits no one and keeps researchers stuck barking up the wrong trees. Instead of conducting yet another round of outcome evaluations on the same few mediocre methods, we should shift gears and consider focusing on how the therapists who consistently get the best results actually accomplish their goals. Perhaps such studies would yield useful commonalities.

Kelly’s therapist-as-researcher model is only one option for those of us who see a need to break away from the world of third-party payers and pseudomedical treatments, but his approach is instructive because it builds a form of scientific accountability right into the definition of the task. His theory is that the therapist’s job involves helping clients test their personal hypotheses. When they’ve found the solutions to their conundrums, the project is complete. This is small-scale, client-centered accountability, as opposed to our grander attempts to get new therapy “brands” added to the APA’s list of approved methods.

Not long ago, I had Kelly’s model in mind when I saw a couple who were dealing with an affair the husband was having. He’d promised several times to stop seeing his paramour but didn’t seem able to keep those promises. She still hoped to preserve the marriage but was being torn apart by his indecisiveness. When I met with the husband, he said that although he loved his wife, the thought of leaving his lover forever seemed unbearable. Over the previous weeks, the couple had consulted two different counselors, but they were unhappy with the results. One therapist seemed intent on exploring Imago themes about childhood wounds but was less able to offer the couple any practical advice about their current situation. The second took a moralistic stance toward the husband’s transgressions that made both spouses uncomfortable.

I felt there were a number of encouraging factors in this case. Both spouses sought help together and seemed earnest about finding a way to move forward. The husband hadn’t actively sought to have an affair—he’d met the woman, who worked for another branch of his firm, at a training conference. The couple had an apparently well-adjusted 14-year-old son, whom they both adored. And I was impressed that they were able to speak civilly about the situation, perhaps because they’d both known others who’d experienced infidelities. However, the wife indicated that she was fast reaching the end of her rope.

As I met with the couple—both separately and together—a single question kept popping into my head: “Given their circumstances, what option are they unable to contemplate?” The answer sprang to mind between sessions—in fact, while I was in the middle of a lecture. It was sufficiently outlandish that I hesitated to share it with the couple. When I finally decided to propose it to them, I did so tentatively, ready to back off immediately if either of them took offense. In the best Kelly tradition, I presented it as an “experiment” that might simply be worthy of some discussion. If it didn’t work out, we would go back to the drawing board and devise some other approach.

The plan I proposed was that they separate for three months and, during that time, she grant him full permission to live with his girlfriend. At first, they both thought this idea was slightly preposterous, but as I predicted, they were open-minded enough to give it a try, perhaps because they couldn’t think of a better alternative and they were both exasperated by the status quo. As a team, we negotiated the details. For example, he agreed to be discreet about his relationship to avoid embarrassing his wife in front of their mutual acquaintances. At the end of the three months, they’d decide whether to continue the marriage or file for divorce.

Of course, now that the affair was no longer “forbidden fruit,” its appeal waned (as I suspected it might). Sharing an apartment with someone is far different from meeting them for furtive dates in romantic settings. The husband discovered that the girlfriend had habits he disliked, such as leaving her clothes everywhere, and he saw that he’d misjudged the level of compatibility between them. By the end of the three-month trial period, he’d made up his mind: if his wife would still have him, he’d return home.

This wasn’t an “off-the-rack” intervention. It was tailor-made for this particular couple, taking into account their background, their level of sophistication, the advice they’d received from others, and the way they’d reacted to the topics we discussed. It was a plan I’d never thought of before and may never recommend again. It worked for this couple—to the best of my knowledge, they’re still together. Yet it can’t be empirically validated and isn’t apt to appear on the APA list of ESTs anytime soon. There was no artificial separation of emotional, cognitive, and behavioral components. Our discussions were conducted in plain English: no “disorders” were treated, no “symptoms” were eliminated, and no Beck Depression Inventories were administered.

In Kelly’s terms, the couple field-tested a crucial hypothesis—that life with the paramour offered more possibilities for long-term happiness than continuing with the wife did. The couple performed the experiment, drew the necessary conclusions, and moved on with their lives. Despite the idiosyncratic intervention, or perhaps because of it, this case has all the hallmarks of psychotherapy as conversation. We may have to get used to the fact that the process, like all conversations, has improvisational elements that can’t be preprogrammed and routinized.

So, if therapy can’t be done by the book, how are we to train graduate students or help professionals improve their craft? With my own students, I discuss a few basic principles, such as how to form clear therapy contracts and how to listen for a client’s hidden assumptions. I suggest that they avoid reifying human predicaments into “symptoms” and refrain from buying into their client’s causal explanations. All of this is mostly a matter of cleaning out enough psychological “debris” so that trainees can sit comfortably, listen carefully, and think creatively. When I supervise, I give example after example of interventions that challenge clients’ suppositions and help them explore new terrain. Then I hope for the best.

Perhaps Marsha Linehan, the inventor of Dialectical Behavior Therapy, put it best when she advised new students to stop trying to act like therapists: “If they would act like themselves, they would [be better off]. . . . All you are trying to be is simply one human being trying to help another human being. That’s all this is.” Unfortunately, the category mistake obscures that fact.

Michael Lukens, Ph.D., is in private practice in Jupiter, Florida. Mitchell Greene, Ph.D., is in private practice in Wayne, Pennsylvania. The authors wish to thank Robert Fauber and Mark Schenker for their assistance with this article.