This blog post contains resources related to a recent article published in The Guardian, by Jonathan Haidt and Nick Haslam, on campus censorship. That article explored the ramifications of an academic article by Haslam, on “Concept Creep.”  The citation for the academic article is:

Haslam, N. (2016). “Concept creep: Psychology’s expanding concepts of harm and pathology.” Psychological Inquiry, 27, p. 1-17.

A link to the published article is here. (Taylor and Francis has generously made the article open-access.)
A link to Haidt’s published commentary on the political ramifications of concept creep is here.
A link to an open access version of Haidt’s commentary is here.
To learn more about victimhood culture, read this summary of Campbell & Manning (2014),

Abstract of Haslam’s Concept Creep article:

Many of psychology’s concepts have undergone semantic shifts in recent years. These conceptual changes follow a consistent trend. Concepts that refer to the negative aspects of human experience and behavior have expanded their meanings so that they now encompass a much broader range of phenomena than before. This expansion takes “horizontal” and “vertical” forms: concepts extend outward to capture qualitatively new phenomena and downward to capture quantitatively less extreme phenomena. The concepts of abuse, bullying, trauma, mental disorder, addiction, and prejudice are examined to illustrate these historical changes. In each case, the concept’s boundary has stretched and its meaning has dilated. A variety of explanations for this pattern of “concept creep” are considered and its implications are explored. I contend that the expansion primarily reflects an ever-increasing sensitivity to harm, reflecting a liberal moral agenda. Its implications are ambivalent, however. Although conceptual change is inevitable and often well motivated, concept creep runs the risk of pathologizing everyday experience and encouraging a sense of virtuous but impotent victimhood.

Condensed version of the article

Haslam’s paper gives us one of the most powerful conceptual tools for understanding what is happening on American university campuses in the last several years, and on British campuses in the last year or so. Haslam shows that many of our most important moral concepts—such as bullying, trauma, and prejudice—have been “creeping” downward and outward to encompass so much of life, and such small exemplars, that vast swaths of what used to be considered normal human behavior are now seen either as pathological (requiring treatment) or as morally outrageous (requiring punishment). This is why it is becoming more difficult to live and work on college campuses -– the zone of acceptable speech and behavior is steadily shrinking, which chills free speech. Everyone is walking on eggshells.

The article is long, but it is so important that I (Jon Haidt) have created a condensed version of it, to bring it out to a broader readership. To read the full article as published at Psychological Inquiry, please click here. In what follows, I show the structure of the paper and include long quotes from each of its 13 sections. All text is copied and pasted directly from the published article, [except for comments from me, which are in brackets.] I have also bolded the lines that are most important for understanding some recent trends on university campuses in the USA and UK.


Writing in 1993, Daniel Patrick Moynihan, senior senator for New York, alliterated that his country was “defining deviancy down.” Moynihan argued that in response to rising crime and social disorder in the 1970s and 1980s, the public increasingly normalized behavior that would once have been seen as pathological. Sometimes, he proposed, this process was driven by the worthy goal of social inclusion, countering the tendency to stigmatize people on society’s margins. At other times it merely represented a habituation to ongoing social change. Whatever its cause, phenomena that had once been seen as deviant were redefined as the new normal…

There is nothing inevitable about the progressive expansion of normality that Moynihan documented. Indeed, I argue that in recent decades the opposite process has unfolded: the definition of some forms of deviance has enlarged and normality has contracted. Psychology has played a significant role in this process, as many of the concepts it employs to make sense of undesirable forms of experience and behavior have extended their meanings, encroaching on phenomena that would once have been seen as unremarkable… Moreover, although Moynihan argued that liberals [progressives] resist attempts to pathologize deviance, psychology’s expansionary redefinition of negative phenomena arguably reflects a liberal [progressive or left-wing] social agenda. Instead of defining deviancy down, psychology has ubiquitized it up.


Conceptual shifts can be observed in public discourse, the focus of Moynihan’s attention. They can also be seen in the discourse of the social and behavioral sciences. [Haslam then describes the work of Ian Hacking, who showed how categories about people can cause “looping effects” in which the creation of a label, such as “child abuse” or “multiple personality disorder” can influence people who embrace the label, changing their views of themselves and even their behavior.]

We should therefore expect psychological concepts to undergo semantic changes, and for these altered meanings to have looping effects on how people make sense of themselves personally and collectively. The conceptual changes that I explore in this paper involve alterations in the semantic ‘extension’ of the relevant concepts; that is, the range of phenomena to which they apply. I propose that these alterations take two forms. The first, which I dub ‘vertical expansion,’ occurs when a concept’s meaning becomes less stringent, extending to quantitatively milder variants of the phenomenon to which it originally referred. For example, a mental disorder has undergone vertical expansion if its new diagnostic criteria encompass less severe and debilitating clinical phenomena than previous criteria. Vertical expansion can occur through a lowering of the threshold for identifying a phenomenon or through the relaxation of criteria for defining it. The second form of conceptual change, which I call ‘horizontal expansion,’ occurs when a concept extends to a qualitatively new class of phenomena or is applied in a new context. For example, the concept of ‘refugee’ has expanded to include people displaced by environmental catastrophe, whereas it originally referred only to those displaced by conflict.


The main contention of this paper is that in recent decades the meanings of several of psychology’s key concepts have changed in a systematic way. I argue that those changes have targeted particular kinds of concept and moved in a particular direction. Specifically, it is psychology’s negative concepts – those that refer to undesirable, harmful, or pathological aspects of human experience and behavior – whose meanings have changed, and these changes have consistently expanded those meanings. The concepts in question continue to refer to the phenomena they denoted at an earlier time, but they now also refer to a horizontally and vertically enlarged range of additional phenomena. This semantic inflation is not widely appreciated by psychologists. When it has been noted it has been discussed in relation to a single concept and the general pattern has been missed. In the body of the paper I illustrate the ‘concept creep’ hypothesis by reviewing changes in six concepts drawn from the provinces of developmental, clinical, and social psychology: abuse, bullying, trauma, mental disorder, addiction, and prejudice.


Classic psychological investigations of abuse recognized two forms, physical and sexual. Physical abuse involved the intentional infliction of bodily harm, whereas sexual abuse involved inappropriate sexual contact, including penetrative sex or non-penetrative molestation. Childhood exposure to these forms of abuse was found to increase vulnerability to adult psychopathology, relationship difficulties, and physical ill-health.

Three changes to the conceptualization of abuse that have occurred within the psychological literature over recent decades represent clear cases of horizontal expansion. First, ‘emotional abuse’ (Thompson & Kaplan, 1996) – sometimes labeled ‘psychological abuse’ – was introduced as a new abuse subtype. It refers to forms of maltreatment that need not involve bodily contact, unlike physical and sexual abuse, but includes verbal aggression and other behavior that is domineering, intimidating, threatening, rejecting, degrading, possessive, inconsistent, or emotionally unresponsive. This form of abuse was commonly studied within intimate domestic relationships. This new focus on behavior exchanged between adults represents a second horizontal extension of the abuse concept from its traditional focus on the behavior of adults towards children. A third horizontal extension of the abuse concept is its incorporation of neglect. Neglect implies a lack of appropriate care and concern, as when negligent parents fail to tend to their children’s basic needs for food, shelter, clothing, physical contact, and affection. Emotional abuse and neglect-as-abuse are ideas that represent horizontal extensions of the abuse concept. The former extends abuse into the realm of non-physical harm, where damage is done indirectly through language or social interaction. The latter extends the abuse concept by including acts of omission…. If deciding whether emotional abuse has occurred depends on the self-identified victim’s perception, abuse can be invoked as a description that might seem innocuous from an independent observer’s standpoint. This reliance on highly subjective impressions is a feature of some methods of assessing abuse, as in the following item from a popular self-report measure: “As a child, did you feel unwanted or emotionally neglected?”

[Haslam’s point is that abuse used to have clear boundaries: it required intentional and harmful or inappropriate physical contact of a child by an adult. But gradually the constraints were loosened so that if anyone – child or adult – feels emotionally neglected, it now counts as “abuse.”]


Olweus, the father of bullying research [in the 1970s], proposed three core elements that define the phenomenon. It involves aggressive or otherwise negative actions that are directed towards a child by one or more other people, where that behavior is intentional, repetitive, and carried out in the context of a power imbalance. The victim has less power – whether in numbers, size, strength, age, status, or authority – than the bully.

[Haslam then shows how the concept has expanded horizontally, to include new forms of bullying, such as cyber-bullying, and to extend it to adult interactions, such as workplace bullying. But it is the vertical creep that causes more difficulties as milder phenomena come under the rubric of bullying. One form of expansion is the loosening of the “repetitiveness” criterion. One-shot interactions can now be seen as bullying. A second is the loosening of the power-imbalance criterion. If one child is more popular or just more self-confident, and acts aggressively toward another child one time, it may be considered bullying. But the biggest change, with enormous implications for how students behave when they get to college, is that intentionality is no longer necessary. All that matters is what the victim feels:]

A third form of vertical creep can be seen in the relaxation of the intentionality criterion in workplace bullying research. As Salin (2003) observes, “intent is typically not part of the definition, but instead the subjective perception of the victim is stressed” (pp.1215-1216). Thus bullying can be said to occur even if the identified bully had no intent to harm the identified victim. This broadens the traditional concept of bullying by including behavior that might be inadvertent. This opening of the definition of bullying to the subjectivity of victims arguably represents a fourth form of vertical creep, and is also observed in school bullying scholarship. Olweus (2013), for example, proposes that “the ultimate “power of definition” must reside with the targeted student” (p. 757) as to when a power imbalance occurs. Similarly, Mishna (2012) argues forcefully that victims’ judgments of whether they have been bullied should take precedence over those of perpetrators and adult observers, such as parents and teachers…

Cascardi et al. note that the expanded definition of bullying is now inscribed in US state antibullying statutes, and can have troubling implications for free speech rights and for schools that could “be required to report and investigate every aggressive transgression, from playground teasing and roughhousing to aggravated assault” (p.255). Equally, they argue, blurred boundaries between bullying, harassment, and peer aggression can lead to inappropriate interventions, as these forms of aggression typically require different therapeutic and legal responses.

[This expansion is crucial for understanding the recent spread of “microaggression” theory. If a person feels marginalized, then someone has committed an act of aggression, an act of bullying, even if the perpetrator had no ill intent, and even if the perpetrator was trying to be helpful, as with Dean Spellman at Claremont McKenna. All that matters is that someone felt offended.]


[Original concept:] A trauma was… seen within mid-20th century psychiatry as a physical agent causing organic brain pathology [as in: Traumatic Brain Injury].

[Horizontal creep:] DSM-III (APA, 1980) was a turning point, recognizing “post-traumatic stress disorder” (PTSD) as a mental disorder for the first time. According to the manual, PTSD was a distinctive cluster of symptoms linked causally to a traumatic event. However, in contrast to the DSM-I understanding of trauma, these symptoms were not understood to spring from an organic injury to the brain but from a psychological injury to the mind, caused not by a physical insult but by a distressing experience. This is a classic example of horizontal creep. [Horizontal creep is not the problem, but it opens the door to…]

[Vertical creep: The initial set of criteria for traumatic events was fairly strict:] a traumatic event “would evoke significant symptoms of distress in almost everyone” (p. 238) and be “outside the range of usual human experience” (p. 238). It stated that “such common experiences as simple bereavement, chronic illness, business losses, or marital conflict” (p. 247) generally fail to meet this requirement, and listed rape, assault, military combat, natural disasters, car accidents, and torture as events that generally succeeded.

[But then, the criteria crept downward, becoming – as in all of the concepts analyzed – more subjective:] A recent definition of trauma produced by the U.S. Government’s Substance Abuse and Mental Health Services Administration exemplifies this lowering:]

Individual trauma results from an event, series of events, or set of circumstances that is experienced by an individual as physically or emotionally harmful or threatening and that has lasting adverse effects on the individual’s functioning and physical, social, emotional, or spiritual well-being.

A traumatic event need not be a discrete event, need not involve serious threats to life or limb, need not be outside normal experience, need not be likely to create marked distress in almost everyone, and need not even produce marked distress in the traumatized person, who must merely experience it as “harmful.” Under this definition the concept of trauma is rendered much broader and more subjective than it was even three decades ago.

[Why the creep matters: Because now, anyone can claim to have been traumatized by anything, and therefore has “standing” to demand that the traumatizer or bully or micro-aggressor be punished.]


[Original concept:]    DSM-I contained seven groupings of mental disorders: acute and chronic brain disorders, mental deficiency, psychotic disorders, psychophysiologic disorders, psychoneurotic disorders, personality disorders (which included addiction), and transient situational personality disorders.

[Horizontal creep:] a comparison of the earliest and more recent editions of DSM demonstrates that successive DSM s not only subdivide existing disorders but also open up new psychiatric terrain…. As a result of this consistent pattern of diagnostic spread, many people whose clinical presentation would not have warranted a DSM-I diagnosis—alcohol abusers, insomniacs, bulimics, Touretters, gender dysphorics, anorgasmic women, dyslexic children, and shy adults—would have received a DSM-III diagnosis. DSM-IV and DSM-5 have introduced further horizontal creep… Phenomena that might previously have been understood as moral failings (e.g., substance abuse) bad habits (e.g., eating problems), personal weaknesses (e.g., sexual dysfunctions), medical problems (e.g., sleep disturbances), character foibles (e.g., shyness), or ordinary vicissitudes of childhood now find shelter under the umbrella concept of mental disorder.

[Vertical creep:] Recent editions of DSM sometimes loosen the criteria for determining where normality ends and mental disorder begins. This quantitative easing allows milder, less disabling psychological phenomena to qualify as disordered. Sometimes this relaxation of criteria takes the form of recognizing less severe “spectrum” conditions, as with cyclothymia, a less impairing variant of bipolar disorder, and Asperger’s syndrome, a less impairing variant of autistic disorder, which has recently been reincorporated in the latter diagnosis, thereby vertically expanding it… By misrepresenting normal sadness, worry, and fear as mental disorders, the mental health professions overmedicate, exaggerate the population prevalence of disorder, and deflect resources away from more severe conditions.

[Why the creep matters:] As a result, the proportion of humanity warranting a diagnosis has risen and the proportion of human experience and behavior that counts as disordered has swelled.


[Original concept:] The meaning of [addiction] within psychology, psychiatry, and general medicine in the first half of the 20th century involved physiological dependence on an ingested psychoactive substance. The pharmacological properties of the substance lead the addicted person to require progressively more of it to attain the desired state (tolerance) and cause the person to experience an unpleasant physiological state when deprived of it (withdrawal). This dependency creates an increasingly joyless pattern of compulsive consumption.

[Horizontal creep:] In recent decades the concept of addiction has been enlarged by the identification of addictions that do not involve substances. So-called “behavioral” or “process” addictions [e.g.,] to the Internet and gambling have been proposed in the mental health literatures…

[Vertical creep:] In addition to expanding horizontally to include behavioral addictions alongside the earlier substance addictions, recent developments in psychology have also opened the concept to milder, less extreme forms of compulsive behavior. This vertical expansion is illustrated well by the concept of “soft” addictions (Wright, 2006), which represent persistent activities that carry some cost in money, time, energy, or intimacy. Soft addictions lack the sense of powerlessness, dependency, and compulsion that is typical of standard addictions and the harm they cause is relatively innocuous. By recognizing them as addictions that concept creeps downward into the realm of bad habits and other repetitive pleasurable behaviors where there is a risk of inappropriate diagnosis.

[Why the creep matters: It extends the trend to pathologize what was previously considered either normal behavior or a moral failing, bringing ever more behavior into the fold of a psychotherapeutic culture]


[Original concept:] Allport understood prejudice to involve intergroup antipathy: the prejudiced person holds hostile attitudes towards members of an outgroup.

[Horizontal creep: The early psychologists…] primarily studied varieties of racism, including anti-Semitism, whereas researchers now also study prejudices based on sexual orientation, gender identity, religion, physical appearance and stature, marital status, and even species. (Some of these prejudices—homophobia, transphobia, Islamophobia—also illustrate the horizontal creep of the concept of phobia from irrational fear to attitudinal aversion.) [Horizontal creep is often quite reasonable; the problems arise mostly from the vertical creep…]

[Vertical creep:] Early social psychological researchers began with an understanding of prejudice as blatant bigotry, examining endorsement of hostile and derogatory statements about African Americans, Jews, and others. However, as rates of endorsement of these statements began to wane later in the 20th century, the understanding of prejudice was broadened. McConaghy (1986) drew a distinction between “oldfashioned” racism, exemplified by endorsement of explicit bigotry, and a subtler and more prevalent “modern” racism. Modern racists, like so-called “symbolic” racists (Sears, Henry, & Kosterman, 2000), do not endorse direct hostility to traditional targets of prejudice but instead denied the continuing existence of racism and expressed opposition to affirmative action policies…. Two new ideas extended the concept of prejudice even further. The concept of aversive prejudice (Dovidio & Gaertner, 2004) applies to liberally minded people who deny personal prejudice but hold aversions, sometimes unconscious, to other-race people. These aversions are not based on hostile antipathy but on fear, unease, or discomfort. The related concept of implicit prejudice (Dovidio, Kawakami, & Gaertner, 2002) referred to people who unconsciously linked negative concepts with racial minority groups more strongly than with majority groups, as demonstrated by tasks such as the Implicit Association Test.

All of the forms of prejudice just reviewed are usually understood from the standpoint of the perpetrator of prejudice… However, some research implies that prejudice exists at least in part in the eyes of the target. Research on microaggressions (Sue et al., 2007), for example, takes the target’s perceptions of prejudice as clear evidence of its existence: If a target perceives a slight as evidence of prejudice, then it is taken as such, even if the slight is ambiguous and its author denies it. Of course, many prejudiced acts are unambiguous, target perceptions may tend to be accurate, and denials of prejudice are frequently not credible. Nevertheless, to count perceived discrimination and ambiguous microaggressions as unqualified instances of prejudice is to subjectivize the concept. In addition to this subjectivity, the concept of microaggression extends the concept of prejudice by encompassing acts of omission and phenomena that reflect anxiety rather than hostility.

[Why the creep matters:] [Haslam adds this reassurance:] It is important to reiterate here that by documenting the expanding meaning of prejudice in recent social psychology I am not questioning the validity of this expansion or advocating a return to a narrower understanding of the concept. Each extension of the concept of prejudice was arguably well justified. [But as Haslam and I note in our essay in The Guardian, creeping concepts of prejudice are behind many of the efforts on campus to shut down speech and regulate behavior.]


In sum, then, conceptual creep has occurred across a diverse assortment of concepts and has commonly involved an increased sensitivity to negative experience and behavior, an increased focus on harmful forms of inaction, and an increased acceptance of subjective criteria for deciding when the concepts apply.


[Raises several possible explanations, but finds them lacking. Raises a more promising approach, a “Darwinian concepts” approach, in which “more successful concepts tend to expand their semantic range…”] Just as successful species increase their territory, invading and adapting to new habitats, successful concepts and disciplines also expand their range into new semantic niches. Concepts that successfullyattract the attention of researchers and practitioners are more likely to be applied in new ways and new contexts than those that do not. For example, the success of the concept of bullying in the developmental psychology literature of the 1970s and 1980s may have made it an appealing concept to apply to analogous behavior observed in workplaces for scholars working in the 1990s.

[Haslam links this approach to Pinker’s (2011) demonstration that violence has been declining in the long run, and especially in the last few decades, and therefore there is a rising sensitivity to “new forms of harm.” He also notes that “psychology has played a role in the liberal agenda of sensitivity to harm and responsiveness to the harmed.” I expand on the notion that concept creep is a rhetorical weapon in my commentary on Haslam’s article, titled Why Concepts Creep to the Left.]


Those drawn to a pessimistic assessment of these changes might argue that the expanding meaning of concepts such as abuse, bullying, and mental disorder is creating a culture of weakness, fragility, and excuse-making, in which everyone is a victim and no one is responsible for their predicament. Those drawn to a more optimistic assessment might applaud the growing sensitivity to suffering and maltreatment. A balanced evaluation of concept creep would be more ambivalent, falling somewhere between conservative reaction and liberal celebration.


[Haslam concludes with a dystopian vision which is, in fact, a description of what life is becoming on some university campuses:] Understanding what drives this trend and evaluating its costs and benefits are important goals for people who care about psychology’s place in our cultures. Equally important is the task of deciding whether the trend should be encouraged, ignored, or resisted. Ultimately this question depends on whether we would be content for most interpersonal frictions to be ascribed to abuse and bullying, for everyday stresses to be described as traumas and habits as addictions, for mental disorder to be more common than its absence, and for prejudice to be seen as a constant undercurrent in social life.