One Brick at a Time

Therapy is More Craft Than Art or Science

Magazine Issue
September/October 2012
Placeholder Image for

Most of us went into the practice of psychotherapy after years of having friends tell us that we were easy to talk to about their personal problems. They didn’t say we were the smartest or the coolest, but we were the ones people turned to when something went wrong with a relationship or their parents were driving them crazy. Taking the hint, we ended up making careers out of being skilled conversational partners to people in emotional trouble.

In this era of medical necessity and evidence-based therapies, it’s easy to lose sight of the basic truth that psychotherapy is a special form of conversation: we heal not through prescriptions and procedures, but through talking and listening. Diagnoses and therapeutic models give us markers and road maps, but we have to put them into words, sentences, silences, looks, and gestures–the stuff of human interaction. However, we can get so focused on assessments and strategies that we don’t pay enough attention to the core craft of our work: the spoken word. Surgeons cut and suture; therapists converse.

In workshops, I like to show a tape of a conflicted couple in which the wife, during a first-session diatribe about how emotionally distant her husband is, slips in these words: “I want my husband back.” I ask the audience what they’d do next–what exactly they’d say, and to whom. Typical answers reveal a reluctance to get specific and a desire to talk strategy: “I’d try to get the wife to get in touch with her hope,” or “I’d help the husband hear what his wife is saying.” I keep repeating: “Whom would you address first and what words would you use?” Sometimes I have to be firm: “The floor is only open to two responses: whom and with what specific words?”

As we unpack the conversational options, it becomes clear that the outcome of the session, and perhaps of the therapy, could depend on what’s said next. Nearly everyone in the room understands conceptually that the angry wife has expressed a vulnerable longing for her husband, and that the therapist’s strategy should be to highlight that longing and the hope that it reflects. I break down the choices: say something to the wife or the husband, make that either a statement or a question, go cognitive or affective. Here are examples that have come from audiences:

Cognitive question to the wife: “Could you say more about what you mean when you say you want your husband back?”

Affective question to the husband: “How did you feel just now when you heard your wife say she wants you back?”

Then there are nonverbal options: for the therapist to be cool (using a calm, neutral voice, keeping the body still and relaxed) or more intense, as in my favorite response, saying softly and with emotion to the wife, “Susan, I just heard you say something that really struck me. In the midst of all the pain and hurt you feel in this marriage, you still long for Jeff. You want him back with you.” The rationale here is that you want to turn up the heat on her hope before you do anything else with what she said. It’d be a mistake to ask the husband to respond before Susan goes deeper into her positive feelings for him; after all, everything else she’d just said was critical of him.

We continue the role play with ideas for how to get the husband involved in this pivotal moment. Once again, we have many options: cognitive statement or question, affective statement or question, cool or hot, having him respond to the therapist, or asking him to say something directly to his wife. It’s like a basketball player leading a fast break: do you pass the ball, to whom, and at what time? Or do you take the ball to the basket yourself? In basketball and therapy, you have microseconds to decide–and your choice can have outcome-determining consequences.

Some choices, like asking the husband to turn to his wife and respond in any way he wants to her remark about wanting him back, carry risks of derailing the therapeutic moment, and perhaps the therapy, if it’s the first session and the couple is on the brink. The husband is apt to respond with something like, “You sure have a strange way of letting me know you still care.” The husband now looks like a jerk, and the wife shuts down, firmer in her belief that he’s emotionally unavailable. On this clinical fast break, the therapist has passed the ball too far out ahead of a teammate, who kicks it out of bounds and looks bad.

Here’s what I tell audiences is a better approach at this moment in the conversation: after the wife has stayed with her softer feelings for a moment or two, the therapist leans toward the husband and says softly, “Jeff, I know this has been a hard time for you in the marriage, and you’ve been feeling Susan’s anger. What’s it like now for you to hear her say that she longs to be connected with you again?” I’d have him respond to me, rather than his wife, because if he mixes positive and negative feelings, I can help him focus on the positive ones before I reengage his wife in the conversation.

Therapy as Craft

We rarely talk about therapy on the ground like this: what words at what time and with what tone and body language? Maybe we assume that if therapists are trained in good models of therapy or in the common factors of all successful therapy, they’ll just know how to execute the skills. But what if we think of therapy as a conversational craft that we hone over a career with our clients and with a community of conversational healers? The term craft (a Germanic word for “power” or “ability”) describes a skill set for producing useful things. Crafts are traditions passed down; we learn our craft from apprenticeship with masters. Crafts have communities with standards for whether the work is done well and the product meets practical needs. Carpenters recognize good and bad housing construction. Even therapists with different models can recognize good therapeutic craft when they see it, just as musicians who play different instruments or in different genres recognize good musicianship when they hear it.

The craft idea differs from two other images handed down to us: the therapist as scientist-practitioner and the therapist as artist. The first was born in 1949 with the Boulder model of clinical psychology. To the dismay of many academics, the scientist-practitioner model never took hold in community-based practice. Although few therapists would argue that research in areas such as psychotherapy efficacy, neuroscience, and attachment theory (to name a few) are irrelevant to the practice of psychotherapy, the treatment room is miles away from the science lab. The alternative conception, of the therapist as artist, was popular during my training. The best therapists were conversationally creative: think Carl Whitaker, Salvador Minuchin, and Virginia Satir. But art doesn’t have to be useful, and artists today are expected to thumb their noses at tradition and convention. In Woody Allen’s classic line from Bullets over Broadway, “An artist creates his own moral universe.” Psychotherapy may be a creative exchange, but by contrast, we work as part of a community of healers, and it matters whether our clients get better. We’re not each our own moral universe.

So if psychotherapy is a conversational craft, not mainly a science or an art, what are the tools of the craft? They include words, gestures, statements, questions, listening, pausing, and the nonverbals that go with them. (Of course, there’s a knowledge base, too, for the craft of therapy, but here I’m focusing on the tools.) Then there are metaconversational tools, such as pacing versus leading the client, heating up the conversation versus toning it down, staying with a subject versus shifting focus, keeping the conversation productive, recalibrating when the therapy is no longer productive, and using additional tools for multilateral conversations with couples, families, and groups.

For example, heating up a conversation can take the form of using colorful and challenging metaphors. Instead of pointing out for the umpteenth time that a wife was indulging in indirect, negative communication with her husband, here’s what I said: “Susan, I want to tell you what I just saw. I saw you pick up a hand grenade, calmly pull out the pin with your teeth (I mimicked the motion), and toss it into your husband’s lap–and then you seemed surprised when it blew upBANG!and he shut down once again.” The key to this being effective was the twinkle in my eye and a sense of near admiration in my voice for the offhanded artistry she showed in pulling this off. She had to laugh at herself, and I laughed with her. By contrast, I could have challenged her with more clinical (and blaming) language, such as saying something like, “That was pretty passive-aggressive, Susan,” thereby bringing the wrong kind of defensive heat to the conversation. If this seems overwhelming, well, it is–for beginners. Even as an old-timer, I’m always nervous before seeing new clients, because I don’t take for granted that we’ll achieve a therapeutic level of conversation.

I made a craft mistake yesterday that scared me. I was seeing a couple for Discernment Counseling, a way of working with couples considering divorce in which they spend time deciding whether to continue on the divorce path or devote six months of therapy to an all-out effort to restore their marriage to health. It was the second session, and I was talking with the husband alone. When he launched into what I sensed was going to be a lengthy monologue on his wife’s failings, I tried to interrupt and steer him back to looking at himself. He must have felt that I was telling him his feelings were wrong, because I hadn’t said two words before he cut me off and became enraged. No one, he shouted, was going to tell him he didn’t have the right to feel angry at his wife. Not the last goddam therapist, not his wife or adult daughters, and certainly not me. He was so worked up that I felt a pang of fear for my safety. I’d misjudged his openness to have me structure how he talked about his hurt and anger.

I switched to a calm, listening stance, focusing my attention on his pain and not allowing my nonverbals to communicate fear or distress. (I recall Jay Haley teaching a nice piece of craft: when you feel threatened by a client, act the opposite nonverbally, open and calmwhich signals safety to the client.) The storm soon passed. I knew it was important for me to pace him better, but I told myself that I shouldn’t overreact and play it too safe. So I challenged him several times, but with better skills. After waiting for him to pause, I’d say, “I have a thought about what you’re saying that you may or may not agree with. I’m wondering if . . . .” I gave him permission to disconfirm my statements and maintain his sense of self-protection.

Near the end of the session, he told me that the angriest he’d gotten in his whole life was at a therapist who, after he said he was feeling guilty about how he’d behaved in his marriage, rubbed his bearded chin and said, “Guilt: I’m not hearing guilt.” If that happened as reported, it was a megamistake by the therapist (the client refused to return to therapy). Mine was a smaller mistake, although potentially serious. Without knowing the client well enough, I’d tried to lead him without enough pacing of his feelings, and I’d done so by interrupting him without waiting for a natural pause.

In this situation, the advantage for me in thinking about my work as a craft is that I knew immediately that I’d made a technical mistake. Of course, I learned a lot about my client from what happened, but too often, when our clients act dysfunctionally in sessions, we veteran therapists focus on them and not enough on ourselves–on what we did or didn’t do, or on what we can learn. How did my client’s previous therapist respond to the therapy-ending outburst? He called the wife to tell her that her husband had narcissistic personality disorder.

Wired for Overconfidence

Recent work in decision-making psychology has shed new light on skills and how we develop them. In his book Thinking, Fast and Slow, Daniel Kahneman, the only psychologist to win a Nobel Prize (in economics), proposes that skills are learned with two human operating systems: System 1 (intuitive, automatic, fast) and System 2 (more deliberate, more logical, slower). New skills are difficult and tiresome to learn because they take so much System 2 work at the beginning. Read a puppy therapist’s mind during a session and you’re likely to hear this: “What should I ask next? Am I talking too much? How am I going to get the husband into the conversation? Should I try something cognitive here, or maybe something behavioral? Now what was she just saying?” Being in flow this is not.

Once this taxing period is past, the therapist’s skills and intuitions come to reside in System 1. A seasoned therapist senses what’s needed in the moment and confidently executes a smooth, readily available skill. No need to calculate each step! I experience this process as having the confidence that I can listen deeply to my clients without knowing where I’ll take the conversation later–what interventions I may make. I trust that the insights and words will be available to me at the right time, and they usually are.

This is the beautiful side of having skills: the feeling of being on your game, of doing something you love, and getting paid for it, to boot. But there’s a dark side to this confidence. Kahneman’s work is mostly about the unreliability of System 1 and the inadequacy of System 2 in correcting us when we go astray. Once our skills reside in a zone of confident intuitions, our reflective, self-corrective abilities come into play only when System 1 generates a warning signal that it’s becoming unreliable. The problem is that System 1 doesn’t easily admit being off base. When a skill isn’t adequate, System 1 generally substitutes another response that comes to mind, often approximately correct, but sometimes quite wrong. System 2 usually endorses and rationalizes these ideas and feelings, because it can’t easily distinguish a genuine skill from a substituted, biased response delivered up with confidence. System 2 is a sucker for confidence, it seems, and as Kahneman says in his sardonic style, System 2 is lazy when asked to work.

We can easily see this happening in other professions. When a patient with diabetes is getting worse, a physician in training assumes that he or she may not be offering the right treatment, and seeks consultation. A seasoned physician using standard treatment protocols is likelier to conclude that the patient isn’t being compliant with the correct treatment regimen (losing weight, watching diet, taking medications as prescribed). This explanation for repeated treatment failures is System 1 talking and System 2 agreeing. For an experienced physician to entertain the possibility that he or she may lack skills for engaging and motivating patients and their families to incorporate a difficult treatment regimen into their lives would take a big shift out of a professional System 1 comfort zone. It would require learning challenging new skills and becoming a clumsy beginner at new ways to communicate with patients and their families–something that System 2 wouldn’t welcome unless there were a pressing need. And if the physician’s community of medical providers embraces the traditional approach to his/her healing craft–make a good diagnosis, inform patients about how to handle their illness, and the rest is up to them–then there’s little likelihood that a particular doctor will say, “Hold on. I lack a skill set with these patients, and I’d better learn it.” In this case, change has to start from the outside, as is happening in outcome-oriented healthcare that has financial repercussions when a clinic’s pool of diabetic patients is below expected levels of metabolic control. Now more providers are signing up for courses in motivational interviewing.

Therapists aren’t immune to this overconfidence about our intuitions and skills. We share a tendency to rationalize our failures and rely on our current skills instead of upgrading them. When my client blew up at my interruption, my immediate System 1 intuition was that this was a vulnerable, explosive client (true), and that I’d made no mistakes in the session (wrong). If I’d stayed with these intuitions, my System 2 processing would probably have endorsed them–rationalizing, as did his former therapist, that there was something seriously wrong with him–and my work with this man and his wife would have suffered. But over the years, I’ve trained myself to respond to disturbing and surprising developments in therapy with an immediate assumption that I missed something. In this case, I had enough data from the telephone intake and from the client’s wife’s reports about the prior therapy blow-up to anticipate his explosive anger if he felt disconfirmed. Even so, I decided to interrupt him early in our session, rather than let him settle in and feel heard before I moved him in a different direction. Fortunately, my System 2 received enough self-corrective data from System 1 to get to work on a plan to readjust my approach to him.

Getting Off Plateaus

If therapy is a craft at which we become more confident over time, we should get better over time, right? The evidence says otherwise! When it comes to therapy outcomeshow well our clients do–there’s no difference between early-career therapists and later-career therapists. These are averages, of course; some therapists no doubt do get better. But most of us think we’re better because we feel more confident. Scott Miller has demonstrated a related illusion: we nearly all think we’re above average therapists in comparison with our peers. We’re all from Lake Wobegon, it seems.

So why don’t most of us improve over time? I think it’s because we plateau at a certain point and become too confident and comfortable with our skills. So we keep treating clients in the same basic way. In medicine, they call this “clinical inertia,” the tendency to stick with what we know and keep doing it. We help the same percentage of clients as the years go by, and have the same success and failure rates with different problems. Like physicians treating diabetics, we become good at rationalizing our failures as successes in disguise (that couple was doomed to divorce anyway, and we just helped them get there with less hostility), as the fault of the client (not motivated to change, had an Axis 2 diagnosis), or as a problem with the healthcare system (not enough sessions).

My own tendency toward clinical inertia is no different from that of other therapists. After learning my craft, I’ve only seriously shifted my approach when I’ve been propelled by challenging circumstances to upgrade my game. The first time was when I took a job in a family medicine training program and came across people whose medical illness wasn’t just a metaphor for psychological and familial dynamics. I realized that I had no language to talk about bodily pain, blood levels, medication side effects, diabetic reactions, and invasive medical tests that lead to no diagnosis. Like many therapists, I was tongue-tied when clients talked about their medical problems unless I could connect them directly to the psychological or family issues we were working on. I was contributing to the mind-body split in people’s lives: they can talk about their minds with their therapist and their bodies with their doctor.

I was forced to become bilingual by hanging out with medical colleagues and fellow therapists who were making the same discoveries about the limitations of our language. I had no coach, but I had good teammates who shared their successes and failures working on a new therapeutic language. Most of all, I learned by listening to my clients with both ears, not just my psychological ear. Lydia, a client who had multiple sclerosis along with a troubled adolescent daughter, was one of my best teachers. I recall the outset of a session when she immediately teared up. I offered one of my regular System 1 empathic responses: “Sounds like this is a hard time for you.” She calmly replied, “Oh, don’t worry about my tears. That’s my M.S. talking, not me. I’ve actually had a good week, and I’m feeling good today.” I went on to learn how M.S. can make people emotionally labile, and that I shouldn’t assume I knew what was behind a sad look. Through experiences like this and conversations with colleagues, I expanded my craft to be good at conversations that include the biological aspects of the human condition. Nowadays, many therapists are going through the same process in learning the new, multisyllabic language of neuroscience.

Later, I went through a similar shift regarding ways to bring a moral dimension into therapeutic conversation. Like many therapists in the 1970s, I was trained to ignore or challenge any moral language a client was unfortunate enough to use in my presence. Fritz Perls became a kind of therapist folk hero in the midst of the cultural ferment of the 1960s, not only by rejecting psychoanalytic orthodoxy, but by challenging guilt-ridden clients to stop “shoulding” themselves. Obligations must be translated into needs and wants, or they’re inauthentic, he taught. When a client on the verge of divorce would say, “I’m worried that I’m trading my kids’ happiness for my own–making them miserable so I can have what I want,” I was trained to keep the conversation focused on the needs of the client’s self: “I think it’s important for you to focus on what’s best for you right now; kids are resilient.” I had no vocabulary to engage in moral discourse that avoided the twin mistakes of trivializing moral intuitions or stoking unproductive guilt and shame.

Although I’d read critiques of “value-free” therapy (feminism nailed that illusion), I didn’t have the craft yet. The breakthrough moment for me was with Bruce, who came for a final session to say good-bye after his wife had kicked him out. After a few minutes, he calmly announced that he was leaving town–and his kids–to start a new life. He’d already abandoned two children after his last divorce. In Kahneman’s terms, nothing in my System 1 skills gave me the language I needed at that moment, but fortunately, my System 2 had taken in enough information about the significance of moral values that I realized I had to try to reflect that knowledge in the therapy somehow. I knew that it wasn’t going to be enough to just encourage him to slow down or to emphasize how much he’d miss his kids. Time was short, and he was about to leave another set of children grief-stricken and confused.

So I broke a taboo by using moral language about the consequences of his actions for others. “Bruce,” I said, “how do you think your leaving will affect your children?” His response: “They’ll be hurt for a while, but they’ll get over it soon.” I replied, “I don’t think so. I’m really concerned that in dealing with your pain right nowwhich I understand feels overwhelming–you’re about to do something that’ll harm your kids permanently.” Bruce paused. He knew I cared about him, and he was absorbing this challenge to his moral integrity. After some back and forth, in which I offered to help him deal with his sense of rejection and hopelessness, I added: “Your children aren’t responsible for this divorce, and I don’t think it’s fair for them to be its casualties.” Bruce eventually decided not to leave town, recommitted to these children, and later reconnected to the set of children he’d left behind after the previous breakup. This experience was a turning point in my career, teaching me that I had to learn the craft of what I later called “moral consultation,” because moral issues are part of the human condition that people bring to therapy, just as medical issues are.

But eventually, I plateaued again in my therapeutic craft, becoming more a teacher of what I knew than an excited learner. Recently, I started on another period of craft development: learning a new way to work with “mixed-agenda” couples, in which one is leaning out of the relationship and the other wants to save it. A couple of decades ago, I learned a basic craft tool from master therapist Betty Carter, which I used in my practice, but never fully explored–like learning a new riff on a guitar without realizing that it could be the basis for a new style. After working with a family-court judge and a group of collaborative divorce lawyers who were looking for places to refer mixed-agenda couples, I felt pushed to develop the tool I’d gotten from Carter into a full-blown protocol for working with a kind of couple that drives therapists crazy. I named the new approach Discernment Counseling. The essence of the protocol was to tell two spouses who were at odds that I’d help each of them accomplish their own goal: one to save the marriage, and the other to figure out whether to divorce or work on the marriage. I’d never imagined couples could tolerate my working two seemingly contradictory agendas at once. The key was to do this mostly in separate conversations with each spouse, along with carefully calibrated summaries at the end of sessions.

Then, with the idea of craft in my consciousness, I decided I needed to seek out enough practice experiences before I began to teach Discernment Counseling to colleagues. In sports terms, I realized I needed enough reps (repetitions) to hone my craft in light of the wide range of mixed-agenda couples who seek therapy. So I asked my lawyer colleagues to send me cases, and I stopped doing regular couples therapy to focus on this new craft. I asked fellow therapists to observe my sessions so that we could process them. I listened to their sessions. I obsessed about things like what words to use to open sessions after the first one, since this is an exploratory, decision-making process, not a traditional couples-therapy process.

My current opener is: “I’d like each of you to say something about your frame of mind as we begin our second Discernment Counseling session.” Here I’m avoiding asking for stories of what happened during their week or even their goals for the session (we’ll already have agreed on the goals of Discernment Counseling in the first session). Instead, I’m inquiring about what attitude they’re bringing to the session today. By mentioning what session it is (two out of five), I’m reminding them that there’s an urgency to get to work. For couples who’ve had traditional couples therapy, this opening signals that we’re doing something different from what they tried before. How I open these sessions will no doubt evolve over time in a community of Discernment Counselors, but I’m committed to paying attention to this level of detail.

If therapy is all about pacing the conversation, I’m learning when to be emotionally present but low-key in Discernment Counseling (when the couple is in the room together), and when to be more intensely supportive and challenging (when talking to each partner separately), and how to set up the likelihood of a positive but realistic end to the session (when the couple is again back in the room together). It feels like learning the craft of an orchestra conductor: how to open, how and when to contain or unleash intensity, how to close–all complicated by the fact that the spouses don’t play well together, or they wouldn’t be in my office! It takes lots of reps–doing it over and over while looking for mistakes or better ways to do what’s merely OK now. It’s all about paying attention to details and practicing your skills. I push the therapists I train to work on specific aspects of their craft, such as how to make a smooth transition from supporting a complaining spouse’s feelings to challenging that person’s own contributions to the problem.

Beyond Our Natural Gifts

When it comes to deep conversation about what’s most important in life, therapists are naturals. Unlike many people listening to intense emotional suffering, our instinct isn’t to run or to fix. We know how to stay with pain until our client feels heard and is ready to move. If we were musicians, I’d say that most of us start out with natural good pitch, rhythm, and timbre. If we were athletes, I’d say that we can run, throw, and kick balls better than the other kids. But these natural advantages are only the thin edge of competence and far from genuine mastery. As the old saw goes, “How do you get to Carnegie Hall? Practice, practice, practice.” The equivalent might be said for athletes, surgeons, potters, carpenters, architects, chefs, healers, or anybody else whose occupation requires a high degree of practical skill, along with a body of theoretical knowledge.

These experts—including therapistsdon’t learn how to do their craft once and for all: it’s a process that continues throughout life and is never finished. Not only that, but keeping fit and staying on game means we can’t continue practicing blindly, by rote, what we already know how to do. We have to stretch ourselves, make ourselves uncomfortable by practicing what we don’t know how to do very well. We need to learn and repeat over and over exactly those skills that don’t come naturally, that make us feel like awkward beginners. In fact, a continual willingness to begin all over again may, paradoxically, be characteristic of the acknowledged masters of any skilled practice.

But practicing by ourselves won’t cut it. We can’t remain fully competent in our craft, much less grow and become better at it, without the support and challenge of our colleagues. No decent athletes, much less world-class champions, practice in isolation. They have coaches, whose job is to give them honest feedback, point out their weaknesses, and keep after them to work on the weak links in what, to us mortals, might seem to be total perfection. The same is true of musicians. If they don’t have individual coaches with them every day of their lives, they hear and feel the “coaching” of their conductor, their colleagues, the critics, and ultimately the audience–which will let them know soon enough that they aren’t performing up to par. Psychotherapy is virtually unique in not having built-in coaching/feedback opportunities, and yet, arguably, we need it more than most craftspeople. In the intimate world of the therapy encounter, we truly are unseen, unjudged, unchallenged by any except the client–whose dissatisfaction is only too easy for us to dismiss.

Regular feedback helps us practice and learn more effectively, and it keeps us honest. We can’t easily fool ourselves into thinking that it’s all the client’s fault when five or six others in our consulting group are telling us–nicely, we hope–just how badly we screwed up. In fact, it’s just this kind of sounding board that enables us to become aware of what we aren’t doing right, what we don’t really know (and often don’t know that we don’t know). Feedback not only makes new learning possible, but contributes to one of the most important traits in a therapist, or maybe in any highly skilled craftsperson: a sense of humility and the certain knowledge that we don’t have all the answers.

For those of us who’ve been in this business for a long time, the idea of therapy as a set of hard-won skills doesn’t fit the glamorous vision of being a therapist we derived from watching the buccaneers of family therapy: Salvador Minuchin, Virginia Satir, Murray Bowen, Carl Whitaker, Alex Haley. It’s hard now to imagine those icons of clinical starpower and shimmering self-confidence (not to mention, the even grander, more august company of the psychoanalytic masterminds of yesteryear) simply as hardworking craftspeople. Yet, there’s something good and solid in the thought that no matter how much prestige any of us acquires, no matter how many clients we see, how big the workshop audiences we attract, how many books and articles we write, we still have a lot to learn. Maybe, since the glory days of the psychotherapy magic shows, we’ve grown up a little as a profession and as individual therapists, and gained some salutary humility about our powers in the process. Maybe by thinking of ourselves more as craftspeople on a noble but futile quest for perfection, we’re becoming better therapists, even as we demystify the glamorous idols of therapy.

Making Consultation Groups Work

Too often, consultation groups—the main venues for discussing our craft and getting feedback—reinforce clinical inertia by focusing on the limitations of our clients and not on our own limitations as therapists. Our colleagues take our side just as physicians do when a patient crashes: you did all you could, given the patient’s issues. Sometimes, of course, this is true; not everyone is a good candidate for what we offer, and we can’t unilaterally control the outcome of therapy. But faulting the client should be way down on the list of explanations when therapy doesn’t go well. Most carpenters wouldn’t blame the wood if a wall collapsed. Our own insight and skill should be where we look first.

Partly, the problem is that we don’t give our consultants and colleagues enough details about our therapeutic conversations. Early in my career, I recall a therapist in case consultations saying from time to time that a client was triggering his “competency issues,” a revelation we all honored as a vulnerable self-of-the-therapist disclosure and not a revealing comment about his real competency. Looking back, I think his competency issues were real: he wasn’t good yet at the craft of therapy, especially with difficult clients. He paced too much with clients and didn’t lead them well enough. He was a good listener, but often ended sessions without giving clients somewhere to go. If he were a singer, his range was too narrowly in the middle–not enough high notes and low notes to have a strong effect on his clients. I hope he went on to improve his craft, but we certainly didn’t help him much.

Given clinical inertia and the tricks that confidence plays on us, how can we push ourselves to improve our craft? The best place to start is with our pressure points–areas of practice that we find stressful or unsatisfying, where our System 1 skills aren’t working well enough, but neither are our familiar rationalizations. A pressure point might be a mode of therapy; couples therapy is notably hard and unsatisfying for many therapists. It might be a type of client problem; one of my pressure points is borderline personality disorder in couples therapy. Or it might be a challenge across modalities and client problems, as when clients express frustration or disappointment with our work. One indication that it’s a pressure point is finding ourselves involved in what my colleague C. J. Peek calls “clumsy conversations,” moments when the words don’t come easily, our confidence waivers, and we’re off our game.

What we lack are venues for turning our fumbles into greater expertise. Psychologist and author Daniel Kahneman asserts, “true intuitive expertise is learned from prolonged experience, with good feedback on mistakes.” While we get plenty of prolonged experience if we practice long enough, we usually don’t get enough specific, honest feedback on our work to improve our craft. After all, most of us practice alone, nearly invisible to our peers. For the most part, our colleagues know only what we tell them about what happens in our sessions, and what they can intuit from how we describe what went on.

Most therapists learn new approaches to treatment in workshops, and, while workshops may show off the presenter’s skills, they don’t necessarily enhance the skills of passive participants. I try to counter this passivity by asking participants to become more active–write down what they’d do and say in a particular clinical moment in a taped session or role play. Then we parse the advantages and disadvantages of potential moves the participants have suggested. We get to learn from mistakes without anyone’s being harmed.

Of course, some therapists go beyond workshops to formal externships in a therapy model. These learning opportunities are better if they involve direct feedback on role plays and taped sessions, but model-based trainers sometimes just assume that trainees come with a good set of basic microskills common across models. I recall a training tape in which a psychologist learning Emotionally Focused Therapy (EFT) showed a couples video session to Sue Johnson and an audience. When the couple ignored him and flared at each other, it was clear that the therapist lacked the skill of preventing escalation. He was attempting a higher-level skill in empathic connection without an important microskill in couples therapy. Sue Johnson then coached him on the craft of preventing escalation before going on to demonstrate the special skills of EFT. Feedback on tapes of sessions like this can help therapists learn both basic and advanced skills. But even after advanced training and credentialing, these therapists are at risk of plateauing again if they don’t find settings for ongoing feedback.

Other than intensive externships, where can we get feedback to improve our skills in therapeutic conversation? One approach I’ve been using is to ask colleagues in case consultation and other conversations what specifically they do and say in situations that lead me to have clumsy conversations or just a vague sense that there may be a better way. Here are my tips for doing this kind of craft consultation: don’t ask for a case consultation, or you’ll get lost in the details of the case, and bring up a challenge that cuts across cases. Then ask for actual words your colleagues use when they deal with it.

I asked my case consultation group what they say to clients who are considering walking out of a marriage because they don’t feel “in love” with their spouse anymore. Before long, I had actual words and phrases from five seasoned therapists, along with their reasoning behind the words. I told them what I generally say in these situations and how clients sometimes respond. Then my colleagues helped with the next level in these conversations, for example, how to respond when the client says, “Yeah, other therapists have told me the same thing, that loving feelings can follow if I change my behavior, but it feels fake.”

Another time, I asked my colleagues how they open sessions of therapy. Do they begin the same way with each client, or do they vary it? Do they let the client start, or do they begin themselves? And what words do they use? We talk so little in our field about things so basic to our work. We have no recognized range of session-opening practices that we can examine for advantages and downsides with different kinds of clients. Carried over into the medical field, this would be like a situation in which every surgeon decided on his or her own where to make the initial incision for a common procedure–surely there must be some better and worse places to start!

Another approach to making case consultation more useful for continued craft learning is to ask for role plays on clumsy moments with actual clients. Identify the conversational exchange that makes you feel stuck and play it out, requesting feedback on how you’re handling it. Then invite colleagues to play your role as the therapist, with you being the client, and observe their language, intonation, and nonverbals. It’s one thing to hear the feedback, “I think you may be pushing too hard; I’d be curious with the client about what’s keeping her stuck.” It’s another thing to see your colleagues demonstrate how to do that. One of my colleagues gave me elegant words for an intervention with a client who’s no longer progressing: “We all have in us a pull toward change and a pull toward stability–toward keeping things the way they are. We’ve been working on the change part for a while; maybe we should talk about the stability part now.”

What if your case consultation group isn’t up for this way of working? The culture of many groups is pretty much set on putting out clinical fires, rather than promoting clinical improvement. Maybe a subset of your consultation group would be willing to meet at other times to try a new approach. But we need new kinds of groups: developmental communities of practice. They wouldn’t be for help with specific cases, at least directly, but for learning to improve our craft by pooling the best practices of the group and searching out other practices from experts and admired colleagues.

These groups could gather face-to-face, online, or through phone bridges. Every group member would identify pressure points, clumsy conversations, and areas of curiosity about whether his or her current practices could use refurbishing. Although clinical models and strategies could be part of the conversation, the norm would be that everything gets driven down to the level of specific exchanges with clients.

Every session would involve a tape or a role play, not just general discussion. Expert tapes would be watched not mainly for demonstrating clinical models, but for the conversational craft that cuts across models and therapeutic styles. Someone would take notes to accumulate the lessons coming from the group. If these developmental communities of practice were linked in a larger network, there could a web repository of collective challenges and lessons. The vision would be that therapists would get better, not just more confident, with experience.


Photo © Hisham Ibrahim / Corbis

William Doherty

William Doherty, PhD, is professor of family social science and director of the Minnesota Couples on the Brink Project at the University of Minnesota. He’s the author of the forthcoming book, The Ethical Lives of Clients: Transcending Self-Interest in Psychotherapy