A celebration of Irving Kirsch

This special issue of the AJCH celebrates the life and work of Professor Irving Kirsch (Figure 1; my gratitude to the contributors, to Journal Editor-in-Chief, Stephen R. Lankton, and to Irving Kirsch, who provided me with information and photographs). I decided to organize this tribute not only because Professor Kirsch has been a towering figure in hypnosis and related areas, but because he has been a friendly, warm, and supportive figure for many of us, besides being a real (to use a word he might enjoy) mensch.

Figure 1. Irving Kirsch, 2007.

Irving Kirsch has changed the trajectory of not one or two, but three complex scientific areas: Hypnosis, placebo, and the (in)efficacy of antidepressants, all undergirded by his focus on and sensitivity to how implicit and explicit communications (i. e., suggestions) contribute to what people expect to be able to do or experience, whether in a context of hypnotic or non-hypnotic suggestions (e.g., for pain relief), or responding to inert pills or fake medical procedures (i. e., placebo), and active pharmacological ingredients (e.g., measuring how much the response to antidepressants contributes beyond that to placebos). Because this is a hypnosis journal I will emphasize his contributions to hypnosis, bearing in mind that his other areas of inquiry relate theoretically and empirically to it.

Hypnosis

Professor Kirsch has contributed in various ways to hypnosis theory, research, and clinical practice, often integrating these branches. Early on (Kirsch, 1985), he extended the foundational work of Rotter’s (1954) social learning theory and behavior-outcome expectancies into the areas of response expectancies in hypnosis, placebo, and psychotherapy in general, and he mentioned that it might also be applied to the effect of antidepressants (Kirsch, 1999).

Anticipated by an earlier statement from White (1941), Kirsch’s most important theoretical work has been the response set theory, which posits that expectancies (or expectations when they are conscious) about suggested behaviors and experiences activate responses consistent with them (Kirsch, 1999). In the longest paper in this issue, “The Response Set Theory of Hypnosis Reconsidered: Toward an Integrative Model,” Kirsch’s most frequent collaborator on hypnosis, and an eminent hypnosis author himself, Steven Jay Lynn, along with Joseph P. Green, Anoushiravan Zahedi, and Clément Apelian (2022, this issue) provide an extensive account of Response Set Theory. They describe how it evolved from a “strong” version in which expectancies fully accounted for the variance of hypnotic responses to a more nuanced one in which expectancies are substantial but not the only contributors to hypnotizability (cf. Spanos, Burnley, & Cross, 1993). In what is likely to be a fruitful proposal, they offer an integrative model that considers multiple variables and includes the recent neuroscientific model of predictive coding in which the brain constantly updates a mental model of reality (see also Kirsch, 2018). Predictive coding was clearly anticipated in the non-brain-centric model of constructivists such as Brunner (e.g., “The organism in perception is in one way or another in a state of expectancy about the environment,” Bruner & Postman, 1949, p. 206), and by neuroscientists (e.g., experience is “an inextricable amalgam of represented anticipation and represented perturbation,” Kinsbourne, 1998, p. 241).

Kirsch has also been a fecund and creative hypnosis researcher whose work has illuminated many areas in hypnosis (see also Cardeña, 2022, this issue). To give but one example, a then-doctoral student and he used the “trick” or bogus item of using subtle color lights consistent with the hypnotic suggestion, resulting in a robust enhancement of hypnotic responses. (Wickless & Kirsch, 1989). Although Benham, Bowers, Nash, and Muenchen (1998) did not replicate their results, Kirsch, Wickless, and Moffitt (1999) did, among participants who did not get cues that might allow them to suspect a manipulation. Cardeña and Terhune (2019) concluded after reviewing those and other studies that manipulations of response expectancies do not consistently increase response to suggestions, but they also asserted that there is considerable evidence that spontaneous expectancies contribute to hypnotic experiences and responses, although they are not their only determinants. Relatedly, I have discussed the importance of self-fulfilling prophecies in performance and shamanism, forms of “truthful trickery” in which “acting as if” may change the experience, behavior, and physiology not only of enactors but also of their audience (Cardeña & Beard, 1996).

To illustrate Kirsch’s important clinical contributions to hypnosis, in his earliest book he provided specific examples of how changing one’s response expectancies could have powerful effects and be applied in psychotherapy (Kirsch, 1990). Furthermore, he and other collaborators edited very substantial tomes that can guide novel as well as experienced clinicians in the practice of hypnosis, no matter what their theoretical affiliation. The massive Handbook of Clinical Hypnosis (Lynn, Rhue, & Kirsch, 2010) includes one section on foundational aspects of clinical hypnosis such as individual differences and prevention of negative effects, one on models of hypnotherapy (including psychodynamic, cognitive-behavioral, Ericksonian, and others), and specific chapters on hypnotic techniques in general, and the treatment of psychological disorders, stress and trauma, and medical conditions. This volume was followed by a Casebook of Clinical Hypnosis (Lynn, Kirsch, & Rhue, 1996) with lengthy chapters discussing specific treatments of adults, children, and adolescents, and a case conference with practitioners of different theoretical stripes discussing how they would treat a specific case.

A more compact introduction to clinical hypnosis is the Essentials of Clinical Hypnosis (Lynn & Kirsch, 2006), and for those eager for a deeper immersion into the cognitive-behavioral approach there is Clinical Hypnosis and Self-Regulation (Kirsch, Capafons, Cardeña-Buelna, & Amigó, 1999), which includes session transcriptions using techniques such as sensory recall (Sachs & Anderson, 1967), which will benefit particularly those not very suggestible and may be used by anyone regardless of theoretical orientation.

An applied area to which Kirsch has contributed substantially is the study of how suggestions affect the response to pain stimuli. A coauthored investigation (Milling, Kirsch, Meunier, & Levine, 2002) found that participants who underwent hypnotic analgesia, stress inoculation training, or a combination of both had significantly less experimental pain that the control group, and greater suggestibility was correlated with lower pain scores, although in the hypnosis conditions such relation might have been mediated by expectancies. A recent meta-analysis reported evidence both for the strong effect of hypnotic analgesia and for its association with hypnotizability (Thompson et al., 2019).

Placebo

Kirsch and collaborators have shown that hypnosis can produce similar effects to those of placebos but without deception (see “Clinical hypnosis as a nondeceptive placebo,” Kirsch, 1999, reprinted in this issue). For instance, Milling, Kirsch, Allen, and Reutenauer (2005) evaluated the effect of compound treatments (imaginative and then hypnotic suggestions either preceded or followed by a placebo) with multiple probes for pain. The effects of placebo and analgesia suggestions did not differ significantly when placebo followed the suggestions, and expectancies, but not hypnotizability, predicted pain reduction. More generally, Locher and Koechlin (2022, this issue) in “Do Expectations Influence Pain? Recognizing Kirsch’s Contribution to our Understanding of Pain” review Kirsch’s pioneering research on the effects of hypnosis and placebo on pain manipulations, and conclude that the response to pain is not automatic but partly depends on cognitive processes.

This issue does not contain a general review of Kirsch’s work on the placebo effect, but one of his recent papers accomplishes that very successfully (Kirsch, 2018). In it, he distinguishes between stimulus (what to expect from external stimuli) and response (what to expect, consciously or unconsciously, from one’s response to them), and describes how there is not one placebo effect but a number of them, depending on the neurotransmitter systems affected, the condition being treated, and the physical and social characteristics of the placebo. He also offers a list of verbal instructions to maximize the placebo effect (which are clearly applicable to hypnotic and other therapeutic procedures): Increase the certainty of a response while minimizing the initial magnitude of the (gradual) change, and prepare for setbacks.

Antidepressants

The final area covered in this special issue covers the effect of antidepressants beyond that of placebos . Although at first blush this work would seem to be unrelated to hypnosis, that is not the case. Kirsch has pointed out that hypnosis and placebos are forms of suggestion (e.g., Kirsch, 1994 (reprinted in this issue), 1999a; see also Cardeña, 2022, this issue), hypnotic techniques increase the effectiveness of psychotherapy (Kirsch, Montgomery, & Sapirstein, 1995) and are very effective for depression, with a success rate equivalent to that of other evidence-based psychological interventions (Milling, Valentine, McCarley, & LoStimolo, 2019). More generally, psychological interventions either alone or in combination with pharmacotherapy are more effective than pharmacotherapy alone (Cuijpers et al., 2020).

Although the pharmaceutic industry has forcefully communicated (suggested, if you will) the “chemical imbalance theory” to justify prescribing selective serotonin uptake inhibitors (SSRI), there is conflicting evidence for a link between serotonin and depression (Moncrieff et al., 2022). Thus, a careful weighing of any specific antidepressant effect above and beyond that of placebo is mandatory because the response expectancies generated by medication may generate the anticipated result regardless of the contribution of the pharmacological agent. In “Irving Kirsch Opens a Window on Antidepressant Medications,“ Chen, Oliver, and Raz (2022; this issue) provide a historical overview of Kirsch’s original meta-analysis showing that the effect of antidepressants was not substantially larger than the placebo one, which produced a storm of responses. A recently published patient level meta-analysis by Kirsch and others of all the antidepressant data ever submitted to the USA FDA (Federal and Drug Administration) (Stone et al., 2022) should lay to rest the dismissal of Kirsch’s earlier reports with smaller date samples, as it replicated the same small drug-placebo difference as reported by him and colleagues in their earlier analyses (Kirsch et al., 2008; Kirsch, Moore, Scoboria, & Nicholls, 2002).

When I started organizing this special issue, I thought that I had a good knowledge of Professor Kirsch’s contributions, but during the process I have found out how much his contributions exceeded my knowledge of them and how much I continue to learn from him. Borrowing some of his favorite terms, I predict that the more readers come to realistically expect that they will encounter clear thinking and novel research and ideas in his work, the more they will enjoy and learn from his foundational work.