What are set and setting: Reducing vagueness to improve research and clinical practice

Abstract

Research on the therapeutic potential of psychedelics has surged, prompting a re-examination of the role of set and setting in psychedelic-assisted therapy. Yet, these concepts are vague and typically defined over inclusively. We believe that set and setting research should be methodologically reductionist, focusing on specific components rather than set and setting as such. To that end, we propose the mechanism-first approach, which begins with specific, paradigmatic set and setting components, such as “openness to the psychedelic experience” or calming lighting and music. It seeks to understand the mechanisms through which these components affect psychedelic outcomes. Once the mechanisms in paradigmatic cases are understood, researchers can ask whether other mental and environmental factors play the same or similar mechanistic roles. As the process iterates over time, understanding of set and setting expands to include more components. Setting aside the vague, standard definitions of set and setting and focusing, instead, on specific components of set and setting, the mechanism-first approach encourages productive research agendas, focused on specific projects that are mutually informative. To defend it, we outline the problems with standard definitions of set and setting, describe the mechanism-first approach in detail, illustrate it by considering its implications for selected, active research projects, and respond to objections to our approach.

After decades of regulatory and social barriers, psychedelic research is resurgent on a wide variety of compounds. Classic psychedelics are those that operate (in part) through 5-HT2A activity (such as psilocybin, mescaline, and lysergic acid diethylamide (LSD)) and produce mystical-type experiences (Griffiths et al., 2008). These compounds show promise in treating major depression, substance use disorders, and psychological suffering in terminally ill patients (Carhart-Harris et al., 2021; Goodwin et al., 2022; Griffiths et al., 2016). Other compounds with a shared history of “underground” use and regulatory restriction are also sometimes considered (non-classic) psychedelics despite having distinct effects and mechanisms. Some of these also show promise in treating mental health conditions. For example, recent findings suggest that 4-methylenedioxy-methamphetamine (MDMA) is a promising treatment for post-traumatic stress disorder (PTSD) (Mitchell et al., 2021).

Despite the differences in pharmacological mechanism and treatment targets, researchers and clinicians have thought that optimizing the set and setting of use is critical to improving benefits and minimizing risks of various types of psychedelics since at least the 1960s (Leary et al., 1963). To a first approximation, the term ‘set’ denotes the psychological mindset of the person undergoing the psychedelic experience while the term ‘setting’ refers to the physical, social, and cultural contexts in which the drug is administered. Researchers and practitioners have attempted to optimize set and setting in various ways (Rucker et al., 2018, 2021). Sessions are commonly conducted in aesthetically welcoming, dimly lit, “living room-like” settings. Patients are often instructed to recline on a couch while wearing eyeshades and listening to curated playlists of calming music as they “focus their attention inward.” Typically, two “monitors” jointly facilitate or guide the experience. These and other approaches to optimizing set and setting are now codified in safety guidelines in part because they are thought to reduce feelings of anxiety that can arise during the psychedelic experience (Johnson et al., 2008).

Despite the perceived importance of optimizing set and setting for psychedelic research and practice, set and setting interventions are largely untested apart from a few, primarily observational studies (Aday et al, 2022). We know relatively little about how to fully optimize set and setting. Some of the standard interventions may be less important than thought; perhaps some worsen set and setting. Moreover, critical components of set and setting may be optimized in clinical trials in ways that researchers and clinicians fail to appreciate. If so, those components may be overlooked when psychedelic-assisted psychotherapy (PAT) is scaled into widespread clinical practice or naturalistic use (List, 2022). In turn, the benefits seen in clinical trials may fail to generalize.

One problem is that PAT research, generally, and set and setting research, in particular, are hindered by inadequate methodological rigor and a lack of clarity on mechanistic hypotheses (Golden et al., 2022; Johansen et al., 2022; Kangaslampi, 2020; Muthukumaraswamy et al., 2021). Admittedly, scientific understanding of PAT mechanisms in general is still in its infancy (van Elk and Yaden, 2022), but far less is known about how set and setting influence these mechanisms (Golden et al., 2022; Kangaslampi, 2020). Here we argue that a further problem is that the terms ‘set’ and ‘setting’ are vaguely articulated and overinclusive. On the standard definitions, set and setting include (nearly infinitely) many non-pharmacological, contextual factors that are related to the psychedelic experience merely because these are factors in the patient’s mindset or environment when they undergo a psychedelic experience. In research, the results of vague hypotheses about set and setting are hard to test, and vague models of psychedelic mechanisms lack utility. Perhaps for these reasons, among others,recent National Institute of Mental Health guidance suggests that research into psychological components of PAT must have specifically operationalized psychological mechanisms to be of priority (NIMH, 2022). The standard definition’s vagueness may also affect clinical guidelines; at best, the notions of set and setting are superfluous for these guidelines, and at worst, guidelines that invoke them will be hard to follow.

Scientific progress often involves revising theoretical terms to improve rigor and predictive accuracy. Yet, for set and setting, while prior commentators have mentioned the challenge of operationalizing these concepts experimentally (Bremler et al., 2023; Carhart-Harris et al., 2018; Cavarra et al., 2022), little has been done to improve rigor in operationalization. We believe that attempts to do so show that set and setting as such may not be as important as thought—even though some particular components of set and setting likely are.

To make progress, after detailing the problems with standard definitions of set and setting, we suggest that set and setting research should be methodologically reductionistic. It should focus not on set and setting as such but on the underlying, individual components of set and setting, determining which components influence outcomes and how. Our mechanism-first approach encourages set and setting researchers to first understand the mechanistic roles of specific paradigmatic components of set and setting, such as “openness to the psychedelic experience” or calming lighting and music. Researchers should then determine what other elements play the same or similar roles. Thus, they gradually improve our understanding of set and setting over time by focusing on the mechanisms of an expanding set of paradigmatic cases. We use hypotheses about set and setting from the literature to illustrate how research projects might unfold over time on the mechanism-first approach before highlighting and responding to two objections.

Vagueness and overinclusion: The temporal account of set and setting

Common understandings of set and setting are vague and include many variables that are much harder to operationalize than often acknowledged. For example, social and cultural contexts at large, including a wide array of factors from race to political beliefs, are now commonly considered parts of set and setting (Hartogsohn, 2022; Neitzke-Spruill, 2020; Viña and Stephens, 2023). We suspect that many components of set and setting act through various distinct mechanisms, and it is unclear that they are best considered general components of mindset or environment.

Consider vagueness in the standard definitions. Some define set and setting as “the internal and external contextual factors that accompany drug administration,” respectively (Gukasyan and Nayak, 2022). Others define “‘set’ [as the class of] the expectations, assumptions and any other pre-existing psychological factors (including psychopathology) one brings to an experience and ‘setting’ [as the class of] the environmental [factors] in which it occurs” (Carhart-Harris et al., 2018, our emphasis). Generalizing from these statements, we suggest that the standard definitions hold that set is the class of elements in the psychedelic user/patient’s psychological profile (the “mindset”) that are associated with psychedelic use, and setting is the class of elements from the individual’s physical, social, and cultural environment that are associated with psychedelic use (Bremler et al., 2023; Carhart-Harris et al., 2018; Dupuis and Veissière, 2022; Gukasyan and Nayak, 2022; Hartogsohn, 2017; Pronovost-Morgan et al., 2023). Unfortunately, no one has explicitly explained the association in question. Nevertheless, two common interpretations are implicit in the literature.

Most often, the association seems to be understood temporarily. On the temporal interpretation, set and setting are elements in mindset and environment at (approximately) the time of administration: roughly, those encountered just before, during, or after use. This interpretation has the advantage of excluding events from remote parts of the user’s mindset and environment. Presumably, some of these are important to psychedelic outcomes, but surely, the vast majority are not.

Nevertheless, the temporal interpretation does not eliminate mental and environmental components of no obvious importance. A nod hello from a stranger in the hallway right before a psychedelic session or frustration at a traffic jam on the way to a session need not affect the session’s outcome. Often—though not always—such an event will be an unremarkable, everyday occurrence for a psychedelic user; that event’s presence or absence on any given day might be instantly forgotten and unlikely to affect their psychedelic session. Nor is it obvious that there are some important, specific features that such events share with paradigmatic set and setting interventions.

Moreover, the time-related boundaries of the temporal association have not been specified. How far into the past or future is the temporal association? The time of the administration (perhaps beginning when a certain drug-blood or cerebrospinal fluid level is achieved and ending when it falls below a threshold) seems to provide a clear boundary. However, mental state and environment leading up to and after the administration include paradigmatic set and setting components, such as having a calm frame of mind before beginning the session. Once we include some elements of mental state and environment outside the session, the boundaries become vague: should they include minutes, hours, days, weeks, months, or more before and after the session? Critically, the further we expand the boundaries from the session, the more irrelevant components we may be forced to include. Nearly infinitely many variables may interact with a user’s mind or environment over the months that stem from preparation for the session through the session and into the integration period. The vast majority of these variables will be unimportant.

This overinclusiveness has two further implications. First, the claim that set and setting are distinctively important for psychedelics has been thought fundamental to psychedelic research and practice at least since Leary’s early work (1963). Yet, once we appreciate how expansive the notions of set and setting are on the temporal interpretation, it becomes difficult to interpret that claim in any way that is plausible, empirically supported, and interesting enough to be fundamental to psychedelic research.

To say that set and setting are important for PAT must either mean that their role in PAT is larger than their role in other interventions or larger than the role of other elements in PAT. Yet, mental and environmental factors influence psychotherapy and a broad range of physical health interventions, including pharmacological ones (Kaptchuk et al., 2020). For example, we have long known that psychotherapeutic interventions work (additively or synergistically) with pharmacological interventions, as in the case of cognitive behavioral therapy (CBT) when added to selective serotonin reuptake inhibitors (Cuijpers et al., 2014; Cuijpers et al., 2020; Riggs et al., 2007; Walkup et al., 2008). Findings for the influence of these factors might have been distinctive of psychedelics at the time of early research by Leary and others, but not now. Hence, if we interpret set and setting’s importance qualitatively, so that they are important to psychedelics, but not (many) other interventions, the claim appears false. Perhaps a finding that some set or setting factor has larger effects on psychedelic outcomes than on outcomes of many other interventions would support the suggestion that set and setting’s role for psychedelics is quantitatively distinctive. That hypothesis is plausible, but so far, there is little evidence for it.

That might suggest a different understanding of set and setting’s importance to PAT, on which they are qualitatively important to explaining PAT outcomes, regardless of their role in other interventions. Yet, on the temporal interpretation, the standard definitions make that claim trivial. If we assume that PAT has any psychological mechanisms, as opposed to purely biological ones, then some elements of one’s mindset and environment (which influences one’s mindset) must influence outcomes. That is not the remarkable empirical finding that it is supposed to be (Carhart-Harris et al., 2018; Cavarra et al., 2022; Leary et al., 1963); instead, it is a trivial outcome of the definitions themselves.

However, if we interpret their importance relative to other elements that affect PAT, the claim is either confusing or trivial. On the one hand, how can we compare the role of set and setting relative to that of the substance itself when there is no experience without the substance itself? Quantitatively, that comparison seems impossible, and we do not understand what it would mean qualitatively. On the other, the comparison between features of one’s current mindset and those in the remote past seems uninteresting. Generally, in therapy, some features of the remote past may be important, but features of the present circumstance almost always are. The present circumstances include the symptoms that bring one to therapy; the past is important because it may influence the present. Thus, suggesting that present mindset is important for PAT relative to past mindset seems trivial.

The second implication is more important for our purposes here. On the temporal interpretation, regardless of how we make the comparison between set and setting in PAT and in other interventions, we must interpret the claim that set and setting are distinctively important to psychedelic outcomes to mean merely that any or some elements of set and setting are important to these outcomes. Given the examples above, the claim is implausible if interpreted to mean that all, most, or some plurality of set and setting factors are important (in any relevant sense).

In turn, the temporal interpretation favors methodological reductionism in set and setting research—and, ultimately, a similar reductionism in clinical guidance about set and setting. Pragmatically, researchers need generalizations about specific elements of set and setting that tend to be important to develop and test research hypotheses, and practitioners need such generalizations to offer useful clinical guidance. We cannot formulate such hypotheses or offer such guidance about set and setting as such, but rather about specific elements. Hence, research investigations and clinical interventions are best understood at the lower level of the particular elements in question rather than the general, unhelpful level of set and setting broadly.

We suspect that many researchers already work intuitively with methodological reductionism. Some have highlighted that guidelines on set and setting in research protocols need greater specificity about specific elements of set and setting (Barrett et al., 2017; Okano et al., 2022). Likewise, there is preliminary research on some of these particular elements, such as types of music (Strickland et al., 2021) and specific types of calming videos (Heinzerling et al., 2023) used during administration sessions. Some of the most influential set and setting research supports the importance of specific elements of set, including having a specific intention going into the experience as well as being “open” or “surrendering” to it (Carhart-Harris et al., 2018; Studerus et al., 2021).

However, clarifying that the terms ‘set’ and ‘setting,’ on the temporal interpretation, are best understood as mere loose talk is critical. On the one hand, researchers can easily revert to focusing on vague variables. For instance, despite focusing on specific variables above, Carhart-Harris et al. (2018) also use the vague notion of a “therapeutic setting.” On the other hand, researchers tend to infer from the effects of specific set and setting elements to set and setting as such. Carhart-Harris et al.’s (2018) findings concerning entering with specific intentions and surrendering to the experience are commonly framed as pivotal support for the importance of set and setting as such.

Overinclusion and low utility: The causal account of set and setting

The second interpretation of the standard definitions emphasizes causal associations. On the causal interpretation, set and setting are, roughly, the class of elements in mindset and environment, respectively, that causally contribute to psychedelic outcomes. In turn, by definition, the causal interpretation avoids including irrelevant factors in the class of set and setting factors.

To illustrate, consider models that attempt to taxonomize explanatory factors for psychedelic outcomes at a very general level. Among others, these include the “triad” model that focuses strictly on set, setting, and dose (Rocha et al., 2023) and Bremler et al.’s (2023) three-factor model featuring individual vulnerability, set and setting, and properties of the psychedelic substances themselves. Often these models include temporally remote factors, such as developmentally traumatic events (Bremler et al., 2023), that causally contribute to outcomes as components of set and setting. Importantly, because each taxonomic group on the triad model or three-factor model is a mere collection of specific factors understood at lower levels of explanation, the causal interpretation also favors methodological reductionism for the reasons highlighted above—at least as that interpretation is used in such models.

While set and setting are not vague on the causal interpretation, like the temporal interpretation, it overextends the notions of set and setting and makes them unhelpful for research and clinical practice. First, given the wide variety of factors that contribute to psychedelic outcomes, the causal interpretation forces the notions of set and setting to subsume overwhelmingly many factors. It is unclear that all causally relevant factors in mindset and environment share any significant explanatory or predictive features. For instance, factors from the remote past do not appear to share explanatory or predictive features with the paradigms of openness to the psychedelic experience and calm environment, which tend to occur during psychedelic use.

Bremler et al. (2023) worry that developmentally traumatic events will count as both set and setting factors and factors creating individual vulnerability, impairing their model’s ability to distinguish the two constructs fully. However, they may not appreciate the number of historical factors the causal interpretation includes as parts of set and setting. On the causal interpretation, every causally relevant variable in mindset and environment throughout the lifespan is a set or setting factor, so most factors that create individual vulnerability will also count as set or setting factors.

Consider the rare adverse outcomes (such as psychosis and suicidal behaviors) and more common adverse outcomes (such as anxiety) that Bremler et al. (2023) aim to explain. Outside of psychedelics, there is vast research on a myriad of psychological (Aslam et al., 2024; Lee et al., 2022; Lysaker et al., 2021; Schubert et al., 2015; Urcelay, 2024) and environmental factors (Argabright et al., 2022; Barzilay et al., 2021, 2022; Guloksuz et al., 2018a, 2018b; Moore et al., 2022) that influence these mental health outcomes. Very few psychiatric disorders are determined by genetic factors; instead, most mental health outcomes result from causal interaction between environmental, psychological, and genetic factors. Presumably, many of these factors will be relevant to psychedelics—whether explaining outcomes additively or synergistically. Notably, even if we do not understand how psychedelics specifically interact with these factors, we are beginning to understand how other substances, for example, THC (Barzilay et al., 2022), interact with these factors to produce adverse outcomes.

Second, the notions of set and setting, so construed, are not useful for research or clinical practice. One potential role for set and setting is to help identify factors that may influence psychedelic outcomes—candidates for research. The temporal interpretation illustrates this role; in principle, it identifies a class of set and setting factors and leaves it to empirical investigation to determine whether they are causally relevant. (Unfortunately, given our points above, the temporal interpretation includes too many factors in set and setting to be practical in this role). By contrast, set and setting cannot even play this role in principle on the causal interpretation. On the causal interpretation, the fact that a factor is a set or setting factor is explanatorily and epistemically posterior to the fact that it is causally relevant; we can only know that some factor is a set and setting factor on this interpretation after we see that it makes a causal contribution.

Perhaps the notions set and setting seem useful simply because they organize explanatory factors at a very general level of explanation, as illustrated by the triad model and Bremler et al.’s model. However, because set and setting include all causally relevant features of mindset and environment on the causal interpretation, they appear to be redundant with the notions of psychology and environment. Moreover, not only can the notions of psychology and environment organize relevant variables, but further, specific constructs of psychological factors and environmental factors for mental health outcomes—and their interaction with other variables—are already well-operationalized and well-researched in other contexts as illustrated above. While those models may not have been adequately explored in psychedelic research, there is a strong case to add psychedelic usage as a variable in those models rather than to posit new, poorly operationalized notions of set and setting that appear merely synonymous with psychology and environment in the psychedelic context.

Loose talk (suggested by the temporal account) and synonymy with psychology and environment (suggested by the causal account) need not be a problem as long as researchers are clear about how they operationalize the notions more precisely. However, for the reasons above, we believe that such operationalization will require methodological reduction to isolate precise variables to manipulate in set and setting research and to guide clinical practice. Once we understand the breadth of these notions, it becomes clear that we cannot achieve those goals by focusing on set and setting as such.

The mechanism-first approach details how such research can proceed. The specific elements should be identified and individuated by their role in mechanisms that explain psychedelic outcomes, and researchers should try to determine which further elements play similar roles. If successful, that research agenda would identify and test specific interventions on those variables, producing useful clinical guidance.

The mechanism-first approach

A mechanism for a given phenomenon is composed of parts (e.g., individual neurons in the nervous system or core beliefs in a person’s psychological structure) that interact to produce that phenomenon (Craver, 2007; Craver and Darden, 2013; Machamer et al., 2000). Many scientific advances in biology, neurology, and psychology have resulted from the broad project of identifying potential mechanisms, individuating their components, and charting the interactions and organization of these components that produce observable phenomena. For example, an explanation of a psychotherapeutic intervention might describe the parts involved (i.e., beliefs, desires, etc.) that are targeted by the intervention. By contrast, a psychopharmacological intervention’s mechanism might involve a description of how the intervention targets neurotransmitters and brain states.

The mechanism-first approach is an iterative research agenda, according to which, set and setting researchers should first focus on paradigmatic examples of (interventions on) set and setting that are widely regarded as critical to psychedelic outcomes. These include paradigms that are thought to promote good outcomes, such as “openness to the experience” and creating a calm background environment, and those that are considered bad for psychedelic outcomes, such as being “closed to” the experience. Even if there is debate about whether various other factors should count as set or setting components, and which of these factors are important, these are widely agreed-upon paradigms.

On the mechanism-first approach, after verifying that paradigmatic factors significantly affect outcomes, researchers should clarify the mechanisms through which these factors exert their influence. As the mechanisms are explained, hypotheses about further elements of psychological profile and environmental context that may also influence these mechanisms will emerge for testing. If those further elements are determined to influence outcomes through the same or similar mechanisms, they become new paradigms of set and setting. In addition, as potential mechanisms become clear, psychedelic researchers can draw on how similar mechanisms are understood in other research to elucidate how they work in the psychedelic domain or identify candidate set and setting factors that might engage those mechanisms in the psychedelic space. Notably, many PAT modalities already in use draw on psychological interventions that are well understood in other contexts (Cavarra et al., 2022); often, these can be applied to specific set and setting factors as we illustrate below. Over time, as we understand more about the variety of mechanisms at play and the specific elements of psychology and environment that act through mechanisms, the class of paradigmatic set and setting components and our understanding of the mechanisms through which they act expand.

Above we saw that relying on the standard definitions of set and setting to develop research hypotheses would point researchers to consider every component of a user’s mindset and every element of their environment. The mechanism-first approach avoids this problem. Experts should be able to agree on a finite number of interventions as critical to optimizing set and setting (Pronovost-Morgan et al., 2023).Once one of them is found to have a large effect size and its mechanism is illuminated, only a finite number of interventions are likely to be strong candidates for further testing: to see whether they act on that same mechanism.

Of course, individuals’ psychologies vary greatly. Thus, the standard definitions correctly emphasize that, given the heterogeneity of individuals’ psychologies, for PAT (or any type of therapy), many (perhaps infinitely many) specific elements in someone’s mindset and environment may be relevant to their psychedelic outcome. Nevertheless, it is not merely because the individual has them in mind or because they are in the individual’s environment. Instead, it is because, for that individual, they play some specific mechanistic role. That specific mechanistic role makes those elements relevant, and that role is generalizable across individuals, even if different mental and environmental elements play that same role in other individuals. For example, individuals’ specific trauma histories are unique, and so the specific events that are covered in prolonged exposure will not generalize across individuals. But the presence of some trauma history and the mechanism through which re-exposure to that history leads to re-learning and symptomatic improvement does generalize (Foa et al., 2019). There is no evidence that PAT resists such generalizations. However, they will neither be found when considering mindset and environment generally, nor when considering all the (nearly) infinitely many, unique components of each individual’s mindset and environment. Instead, we need to conceptualize set and setting factors at a middle level that includes the more specific individual elements but also provides useful generalizations for research and clinical guidance—as we do with early life trauma.

To further illustrate the proposed research agenda, we detail the implications of the mechanism-first approach for specific hypotheses from the literature on set and setting’s (1) ability to mitigate poor psychedelic outcomes and (2) interaction with placebos. However, first, three general points about how the mechanism-first approach bears on the current literature are important. First, the mechanism-first approach’s focus on specific components of mindset and environment coheres well with current research beyond the cases that we detail below. In addition to research noted above (Barrett et al., 2017; Carhart-Harris et al., 2018; Kaelen et al., 2018; Strickland et al., 2021), it also coheres with research focused on the role of engaging with psychedelics with therapeutic intentions (as opposed to, e.g., purely social ones) and feeling ready (or prepared) for the experience (Haijen et al., 2018). Additionally, a recent Delphi study proposes reporting guidelines for set and setting (Pronovost-Morgan et al., 2024), highlighting paradigm set and setting components that researchers agree on. Such methods are excellent means to determine the paradigms with which the mechanism-first approach should begin.

Second, the mechanism-first approach leaves open the possibility that some paradigmatic set and setting factors might be surprisingly unimportant for psychedelic outcomes. For example, preliminary work with small samples suggests that dissonant music may not negatively affect psychedelic outcomes (Strickland et al., 2021). If replicated with larger samples, calming music may no longer count as a paradigmatic example of setting. More generally, because views about what factors count as paradigms, the notions of set and setting will evolve as we understand mechanisms better.

Third, and relatedly, some of the specific hypotheses that we consider below may not bear out. Our goal is merely to illustrate the mechanism-first approach. Set and setting interventions most likely operate through many mechanisms beyond those we can consider here. Insofar as these mechanisms resemble those in other treatment contexts, the mechanism-first approach advocates extrapolating the evidence from these other contexts—in ways we illustrate. However, we also leave open that set and setting interventions involve some sui generis mechanism(s). If so, on the mechanism-first approach, those mechanisms may be high-priority research topics because we will not be able to understand them by extrapolating from evidence about other treatments.

The mere mitigation hypothesis

The mitigation hypothesis holds that set and setting interventions mitigate potential risks from the psychedelic experience or negative influences that might interfere with it. To illustrate, some have suggested that optimizing set and setting is important because it mitigates negative reactions, such as anxiety or “bad trips” (Johnson et al., 2008; Rocha et al., 2023). The mere mitigation hypothesis is stronger. It claims that set and setting mitigate risks or negative influences and that they have no further way of affecting psychedelic outcomes; for example, they do not play any role in the fundamental mechanisms that produce psychedelic benefits.

Research that might support either hypothesis is underway. Synthetic analogs of classic psychedelics tested in mouse models produce therapeutic effects without the hallucinogenic effects (Cameron et al., 2021; Cao et al., 2022). Other research is exploring the effects of psychedelics under sedation in humans; early-phase work has recently been published on ketamine (Lii et al., 2023), and work is underway on psilocybin (Raison, 2022). Suppose that either such analogs or classic psychedelics administered under sedation produce benefits equivalent to those from standard PAT. Suppose, further, that calm lighting and openness to the psychedelic experience are shown to improve outcomes for those undergoing standard PAT. Together, these findings would support the mere mitigation hypothesis.

On the mechanism-first approach, the mere mitigation hypothesis has two critical implications. First, researchers should determine what paradigmatic interventions play this role, how they do so, and what other interventions might play the same role. For example, suppose that calm lighting and openness to the experience are the first interventions understood to mitigate bad outcomes. If so, the mechanism-first approach advocates that researchers seek to understand how they do so and determine what other interventions employ those specific mechanisms.

Second, several mechanisms through which set and setting might mitigate risks of psychedelics would suggest that some set and setting mechanisms might not be unique to psychedelics. On the mechanism-first approach, research should determine whether these same mechanisms explain other types of treatment and, if so, extrapolate findings from those treatments to improve set and setting research. For example, mitigating attentional distraction is critical to how mindfulness interventions produce benefits—perhaps because doing so enables cognitive reappraisal and alters emotional reactivity (Wielgosz et al., 2019). In PAT, those mechanisms might explain the role of, e.g., blindfolds, quiet settings, or therapeutic calming when anxiety arises. For another example, patients abandoning or avoiding therapeutic interventions in light of challenging/anxious reactions is not unique to PAT. Nor is the thought that “working through the anxiety” is itself therapeutic. The risks of patients aborting intervention in light of anxiety and the emphasis on working through anxiety as part of the mechanism for improvement are critical to many CBT-based anxiety treatments—for example, prolonged exposure therapy for post-traumatic stress disorder. In prolonged exposure, symptoms often increase at initial exposure. Early termination (before exposure results in the learning that leads to symptom improvement on the CBT model) can have adverse effects (Foa et al., 2019). That process appears analogous to theories of psychedelic mechanisms where “bad trips” can result from patients’ lack of openness to the experience and, because of this, foreclose the learning process when anxiety mounts.

The mere mitigation hypothesis might be unwelcome to the tradition on which set and setting interventions are distinctively important for psychedelics, but two points are important. First, some specific set and setting interventions might be distinctively important in degree for PAT—even if the mechanism at play is not unique to PAT. For example, perhaps hyperreactivity or distraction is more disruptive to learning from an intensive (1-2 session) intervention than from interventions that are repeated over numerous sessions.

Second, the mitigation hypothesis might hold, and the mere mitigation hypothesis might fail. Perhaps attempts to reduce standard set and setting interventions result in increased anxiety responses (showing a mitigation role for some set and setting interventions). Perhaps, further, psychedelic interventions that circumvent the characteristic altered states of consciousness produce some, but not all, of the benefits of standard PAT (showing that some set and setting interventions have benefits beyond mitigation). On the mechanism-first approach, that would point to three research agendas. One is to unpack the mechanisms through which the set and setting interventions mitigate the anxiety responses. Perhaps, as suggested above, CBT and other well-studied interventions explain the importance of easing anxiety during the administration session (Studerus et al., 2012) by mitigating the adverse effects of prematurely foreclosing the learning process. Another is to unpack the mechanisms through which the newer psychedelics function to see the degree to which they overlap with those of standard psychedelics; these mechanisms may be primarily pharmacological. A third is to determine the mechanisms through which the standard set and setting interventions contribute to psychedelic benefits over and above those of mitigating negative outcomes.

More generally, we should remember that PAT likely works through a plurality of mechanisms; given the wide variety of their paradigms, set and setting interventions specifically also likely involve many different mechanisms. Even mechanisms that we have begun to study for PAT have not yet been unified (van Elk and Yaden, 2022). Whether they offer distinctive, positive benefits or merely mitigate adverse outcomes, the mechanism-first approach supports extrapolating evidence from other psychotherapies to understand set and setting interventions.

Placebo hypotheses

In discussions of set and setting, psychedelics are often associated with placebos—in part because they appear to display profound placebo effects (Aday et al., 2022; Olson et al., 2020; Szigeti and Heifets, 2024; van Elk and Fried, 2023). In particular, some advocate the placebo hypothesis about psychedelic mechanisms: psychedelics are “super placebos” (or “placebo enhancers”) that work merely by increasing an individual’s suggestibility and susceptibility to “extra-pharmacological” factors, including but not limited to expectation effects (Dupuis and Veissière, 2022; Hartogsohn, 2017). That suggests that set and setting factors just are the extra-pharmacological factors in question. The placebo hypothesis is distinct from, but easily confused with, the hypothesis that set and setting interventions in PAT are merely placebo interventions. On the mechanism-first approach, both hypotheses point research on set and setting toward extrapolating data from placebo research.

Advocates of the placebo hypothesis often seem to imply that psychedelics are placebos because expectancy and suggestibility play critical roles in both placebo and psychedelic mechanisms. However, first, this inference rests on conceptual confusion about placebos. Placebos are, by definition, interventions that play a role distinct from that of some characteristic mechanisms of an (active) intervention, relativized to a specific context of research or treatment (Howick, 2017). A sugar pill is a placebo in many contexts, but not when used to treat hypoglycemia, because there is no characteristic mechanism through which sugar acts in the former cases—in contrast to the clear mechanism through which it acts on hypoglycemia. For similar reasons, placebo mechanisms must be relativized. Transiently increasing blood sugar might be a placebo mechanism in many trials, but not in hypoglycemia trials.

Placebos exert their effects through various pathways. These include activity of µ-opioid and β-adrenergic receptors, cytokines, serotonin and dopamine, and cholecystokinin as well as (potentially overlapping) psychological effects involving expectancy and suggestibility that may be modulated by the relationship between patient and provider or various medical rituals (Kaptchuk et al., 2020). These pathways are also utilized in pharmacological and non-pharmacological treatments across several medical disciplines. Many of these treatments display profound placebo effects, including those in pain medicine, cardiology, neurology, endocrinology, immunology, and psychiatry (Kaptchuk and Miller, 2015). We do not take these other interventions to be placebos because they are associated with placebo effects; instead, they are treatments in the relevant contexts. So why are psychedelics an exception worth calling placebos when these other interventions are not?

Second, mounting evidence suggests that it is naïve to reduce psychedelic mechanisms only to those involving, for example, expectancy and suggestibility. First, there are psychological and neurological characteristic mechanisms through which psychedelics exert their effects without an obvious placebo role (van Elk and Yaden, 2022). Second, psychedelics have positive outcomes in placebo-controlled trials—though admittedly blinding is difficult to ensure in psychedelic trials (Aday et al., 2022; Reiff et al., 2020). Third, preliminary evidence suggests that expectancy effects may influence selective serotonin reuptake inhibitors as much, or more, than psychedelics (Szigeti et al., 2024). These findings suggest that psychedelic mechanisms involve more than expectancy and suggestibility.

Notably, when considering the role of expectations in determining whether some intervention involves a placebo mechanism, we must distinguish between expectancy effects specifically and expectation-related effects generally. CBT involves, among other things, cognitive restructuring of patients’ expectations involving their core beliefs (e.g., the belief that “bad things will happen because bad things always happen to me”) (Beck, 2020). That restructuring is different from merely creating an expectation that, e.g., “therapy will make me better.” Cognitive restructuring operates through a clear mechanism: it changes beliefs that interfere with healthy thoughts, emotions, and behaviors, thereby promoting recovery.

The points above count against the placebo hypothesis, but they do not settle the question of whether set and setting interventions are placebos. In assessing that question, on the mechanism-first approach, we should focus on the paradigmatic cases of set and setting interventions. Insofar as maintaining openness to the psychedelic experience is a paradigmatic set intervention, and interventions that resemble mindfulness or CBT therapies are important to maintaining openness, it seems unlikely that all set and setting interventions are merely placebos. But perhaps some set and setting interventions have no plausible role as therapy themselves beyond inducing expectation effects. Even if, say, calm music is merely a placebo intervention, the mechanism-first approach requires researchers to determine why such music helps and what other interventions might leverage the same placebo mechanisms.

Reductionism and revisionism?

We have encountered two types of skepticism about the mechanism-first approach: either it is a slight, but meaningful, addition to set and setting theory, or it is far too revisionary and minimizes the importance of set and setting. First, some suggest that we have added only slight precision and explicitness to scientific business as usual. Methodological reduction is part and parcel of much biological and medical science (Anderson, 2017), and the type expressed in the mechanism-first approach is consistent with many current projects in psychedelic research. We are happy to make these minor additions, but two points are critical.

First, methodological reductionism is common, but not all biomedical research is methodologically reductionist. When variables can be carefully isolated and controlled, science may merely aim to predict outcomes while maintaining a black box about explanation—for example, in epidemiology or randomized clinical trials. However, we believe that set and setting research should mostly eschew black-box approaches. Since the terms ‘set’ and ‘setting’ are quite vague and include various (seemingly unrelated) components, methodological reduction at least allows us to focus research on specific interventions that may be appropriate, understand how those interventions work, and then to use those findings to generate hypotheses about further interventions of interest.

Second, we think our proposals are, in fact, modestly revisionary. While the mechanism-first approach may appear to be merely methodologically reductionist at first, we believe it has revisionary semantic implications for the concepts set and setting. If the way to clarify set and setting’s implications for psychedelic outcomes is by focusing on their specific components, we struggle to see how useful notions of set and setting could refer to anything over and above those particular components. Given their vagueness, if we can explain everything in terms of the lower-level factors that influence psychedelic outcomes, there is no good reason to postulate the further emergent constructs set and setting. Historically, they are shorthand terms that work well to gesture at vague ideas, but scientific advances and better clinical guidance now require more precision.

We see two semantic options. One is eliminativism: we should drop loose talk about set and setting—at least when formulating careful hypotheses or offering careful clinical guidance rather than merely gesturing at general ideas. (Compare how we now use the terms ‘nature’ and ‘nurture’ in nature-nurture debates.) The other is semantic reductionism: we should redefine set and setting as (nothing more than) the specific components of mindset and environment that play critical roles. On this view, we should carve off irrelevant parts of mindset and environment when we realize that they are not part of the important underlying mechanisms. Not much turns on the choice between these two options; so long as researchers and practitioners are clear about their terms, either route is reasonable.

Thus, to the second objection, we accept that our account is revisionary, but we believe that it is only revisionary in places where set and setting research and clinical practice demand it. Nevertheless, some in the psychedelic community are avowedly antireductionist about set and setting (and psychedelic research more generally) in some vague, but stronger, sense. They believe that the effects of the psychedelic substance cannot, in principle, be disentangled from the effects of set and setting. Doing otherwise is “pharmacologism”: “fetishizing” the substances themselves. Instead, we must resist a “dichotomy between drug and context” (Noorani, 2021). Individuating the mechanisms, as the mechanism-first approach demands, is a mistake.

Often, antireductionists about set and setting appeal to vague terms that we believe should be revised away to make scientific and clinical advances. We have already criticized the view that psychedelics are super placebos. Another view advocates understanding psychedelics as, roughly, things which are “mind-manifesting”—even suggesting that meditation without any pharmacological intervention is psychedelic (Lyon, 2024). That strikes us as stretching the term ‘psychedelic’ beyond recognition. Others claim that psychedelics are “non-specific catalysts and amplifiers” of conscious experience (Grof et al., 2008). We think those terms are vague at best—and confused at worst. It is unclear what it would mean to amplify (all of) one’s consciousness. Perhaps the idea is that psychedelics only amplify some aspects of conscious experience. But the account is vague on which ones—indeed, intentionally so in its claim that psychedelics act non-specifically.

However, proponents of these views tend to share a further view: they advocate extremely broad, nonspecific mechanisms and oppose the idea of distinguishing, e.g., neurological effects from the effects of set and setting as an aim to the broader goal of decomposing components of mechanisms in the way we advocate here. They suggest that thinking of the mechanisms involved in set and setting and those related to the substances themselves requires that we believe the two effects make equal, additive contributions across contexts and that we neglect synergistic effects (Pronovost-Morgan et al., 2023).

Three points are critical. First, nothing we say above denies the possibility of synergistic effects between psychological and pharmacological mechanisms. While we think it unlikely, the mechanism-first approach even allows that there is a single, complex, unified mechanism with many submechanisms that can be studied individually. Part of the appeal of the mechanism-first approach depends on the possibility of clarifying distinctions between the processes through which various elements work. That is compatible with complexity of these mechanisms as well as interaction and synergy between them. For example, one mechanism might moderate another. The point is that, by individuating PAT mechanisms, one can better test ways of intervening on them and, eventually, improve patients’ outcomes.

Second, if one genuinely wants to explain psychedelic outcomes holistically, without disentangling individual components and their contributions, several challenges arise. How can antireductionists describe the synergies between the drug, the user’s mind, and the context they aim to emphasize? If the processes through which the individual components act independently cannot even be described in principle, what can be said about why they work—beyond that they are complicated and involve lots of elements? Similarly, what, if any, clinical guidance can one offer without disentangling the different entities and their processes? What, if any, distinctive activities can the client or the therapist perform to promote favorable psychedelic outcomes?

Some may accept our methodological reductionism but object to its semantic implications. We see no reason to accept the first but not the second. Semantic reduction is sometimes inappropriate. Everything we say here is compatible with emergent properties in general (Beckermann et al., 1992; Gillett, 2016). We just see no reason to think that set and setting are among them—given the vagueness, overextension, and explanatory redundancy of those concepts.

Third, antireductionism might be tempting as a means to include the vast diversity of factors that might matter for individuals undergoing PAT—rather than for generalization across individuals. Perhaps a nod in the hallway is important to some specific person’s outcomes. But, to reiterate a point above, we do not need to resist generalization to explain that. Scientific and clinical practice depend on generalizations that apply across individual cases that can be used to explain some process and predict interventions’ results. Each individual, with their unique psychology, may have a distinctive set of factors that will influence their outcomes. Nevertheless, we need generalizations about these factors to begin to explain how they fit together for some specific person and, in turn, to explain or predict how psychedelics affect that person. Presumably, there is a plausible psychological story that explains why that individual benefited from (or was harmed by) the nod in the hallway. We do not see how the terms set and setting on these accounts add any explanation beyond the ability to say that one’s context matters, and our account allows us to say that.

Conclusion

We have argued against currently influential conceptions of set and setting as overly expansive and pragmatically unhelpful in developing research into or guiding practice on set and setting. By contrast, we have offered a conception of set and setting on which they are constituted by specific mental states and environmental factors that play the specific mechanistic roles that paradigmatic cases of set and setting, such as “openness to the experience” and “calm background environment,” do. As we better understand these roles, we can ask what other aspects of mental states and the environment might play that role and then test hypotheses about those claims—thereby better understanding set and setting and the mechanisms that target them. The mechanism-first approach is reductionist, but not in any pejorative sense, and to the extent it is revisionist, the only revisions are to carve away vague terms that do not explain psychedelic processes.

Ultimately, by focusing on mechanisms, we can improve psychedelic research and psychedelics’ benefits to patients. Moreover, perhaps because of its role in surging psychedelic research, there is growing interest in the role of set and setting in influencing treatment and substance use more broadly (Ataiants et al., 2020; Hartogsohn, 2017; Holmén et al., 2023; Kozak and Miller, 2024; Pronovost-Morgan et al., 2023). A more precise, mechanistic approach to set and setting might also inform research in these other areas and the development of psychiatric care models that leverage these insights. However, as with psychedelics, other domains should first clarify their conceptual frameworks before extending set and setting research indiscriminately—lest they replicate the same vagueness we have critiqued here.