1 Introduction: The Self-Fulfilling Prophecy as a Flexible Theory

‘Self-fulfilling prophecy’ (henceforth SFP) is one of the most successful sociological theories. In an essay published in 1948, Robert K. Merton coined the phrase and defined the SFP as “in the beginning, a false definition of the situation evoking a new behavior which makes the originally false conception come true.” (p. 195)¹ The theory quickly became a mandatory reference both in sociology and social psychology textbooks. It served as an inspiration for innovative research in many other disciplines and is still at the heart of several contemporary debates (Rosenthal, 2002; Ferraro et al., 2005; Felin & Foss, 2009; Zulaika, 2009; Jussim, 2012). As Merton (1994) himself noted, “the ‘paradigm of the self-fulfilling prophecy,’ which was first applied to the sociological problem of ethnic and racial discrimination, has since led to traditions of theoretical and empirical inquiry in social psychology, political science, anthropology, economics, and public administration” (p. 19). The SFP contributed much to Merton’s fame as a ‘middle range,’ analytically rigorous theorist among professional sociologists, but it was also widely successful outside academic circles. The 1948 article was probably his most spectacular accomplishments in public sociology, what Kalleberg (2010) calls “an exemplar in the Kuhnian sense” of an essay published by an academic in the role of public intellectual (p. 203).² Merton & Wolfe (1995) classified the SFP as a ‘strongly incorporated term,’ with many more citations in the US popular press in the 1990s than terms like ‘glass ceiling’ or ‘ethnocentrism.’

As a sociological theory, the SFP primarily serves what might be called a conceptual or epistemic function: it is used to make sense of puzzling phenomena. But it is also clear that seeing the world through the lens of the SFP carries significant practical implications, as Merton himself made clear in his paper. As a matter of fact, there is ample evidence that the concept has been put to use to inform and justify decisions in a variety of practical settings. The US congressional record contains more than 900 mentions of ‘self-fulfilling prophecy’ for the fifty years between 1967 and 2017. The notorious ‘Pygmalion’ studies, which tested the theory experimentally and concluded that teachers’ expectations greatly affect student achievement (Rosenthal & Jacobson, 1968), played a significant role in several lawsuits related to educational equity (Wineburg, 1987). With such a widespread use, it is not surprising that some of the references are seen as departing substantially from Merton’s original intentions. As Merton (1998) ironically remarked,

Not least, the concept has been put to use (and abuse) in the halls of the American Congress and in presidential documents. Again, not always with discriminating judgement. I recall, for one example, the ambivalence with which I read in President Nixon’s budget message of 1971 that he counted on his optimistic forecast for the economy becoming—the language is his—“a self-fulfilling prophecy.” Like many another American, I preferred the prophesied outcome to a recession but was minded to inform the president that prophecies by publicly significant individuals become self-fulfilling, apart from other special conditions, only when the prophet has acquired widespread credibility. (pp. 299–300)

In this paper, I analyze in depth an episode in the history of the SFP: how it was used in the early desegregation court cases, and notably in the briefs prepared and filed to the Supreme Court for re-argument after the 1954 decision Brown v. Board of Education of Topeka. I show that in that occasion, the concept of the SFP was used to defend highly divergent, even opposing, political projects concerning how authorities should implement desegregation. I then go on to propose an interpretation of these flexible political uses that does not point to an ‘abuse’ of the SFP by some of the actors involved but rather to ‘endogenous’ features of Merton’s theory and to ‘exogenous’ properties of the context in which it was used and of the individuals who used it. In particular, I will argue that the SFP’s political flexibility was made possible by its generality, i.e., the fact that it can be applied to many different phenomena and portions of social reality, and its ambiguity, i.e., the fact that in its original formulation the concept is made of different components that can be activated selectively. While Merton was explicit about the policy implications of the concept, this fact alone was not sufficient to counteract appropriation that pointed to different policy lessons.

Additionally, I will argue that the political flexibility in desegregation court cases was made possible by two exogenous factors: the adversarial logic of U.S. court proceeding, which encourages instrumental appropriation of theories, and the fact that the experts who invoked the theory were embedded in environments characterized by different political cultures.

The overarching goal of the article is to offer evidence and insights to better understand how ideas are put to use outside their immediate context of production. The hypothesis that ideas can be politically flexible, which I have discussed at length elsewhere (Brandmayr, 2021), suggests that ideas, including sociological theories, travel across and are used by different social groups and political formations. Flexibility is not only due to exogenous properties of the contexts in which theories are used but also of endogenous features of the theories themselves. Accordingly, we should expect a great deal of variation between theories on how much political flexibility they allow. Some theories, perhaps like Samuel P. Huntington’s ‘clash of civilizations,’ are relatively rigid and are mostly found on one side of the political spectrum. Others, like the SFP, are relatively flexible and can be appropriated more easily by actors who hold opposite political viewpoints. In what follows, I will analyze in depth an episode in the history of the SFP to explain why the theory allows for such flexibility.

2 The Self-Fulfilling Prophecy and Desegregation

A proper history of the SFP is still to be written. Merton had long wanted to write a definitive history of the theory (Sztompka, 1986), but ended up publishing a piece containing limited, albeit very interesting, information about the original essay’s publication process (Merton, 1998). My task here is not to fill this important gap in the literature on the history of sociological ideas, but to provide some background information on the context in which the SFP was developed and to show that it was put to use in support of highly different political projects in the desegregation court cases.

As is well known, Merton published the 7000-word essay titled ‘The Self-Fulfilling Prophecy’ in the June 1948 issue of the Antioch Review.³ He was 37 years old and had just been named full professor at Columbia University. Merton (1998) later wrote that he had decided to publish the paper in a journal “for the general reader rather than in an academic journal” (p. 311) because he believed that “the concept of the self-fulfilling prophecy held direct and significant implications for the conduct of social life.” (p. 328) He wanted “to have it become more quickly and widely known than is ordinarily the case with the diffusion of technical sociological ideas into the public consciousness” (p. 299) and thought of it as a “consequential and operative sociological idea.” (p. 307)

Although the example most often associated with the essay is the bank run ‘parable,’ most of it actually amounts to an explanation of “the dynamics of ethnic and racial conflict” present in the US at the time (Merton, 1948, p. 196). Taking inspiration from what he epitomized in the Thomas theorem, Merton wrote that the self-fulfilling prophecy “is, in the beginning, a false definition of the situation evoking a new behavior, which makes the originally false conception come true (p. 195).”⁴ Such “specious validity,” he said, “perpetuates a rein of error,” for “the prophet will cite the actual course of events as proof that he was right from the very beginning” (p. 195). The definition of the poor black immigrant in an industrial city of the North as unreliable and undisciplined is initially false, but by being accepted by many white people as true (through a “scrambled logic,” a process of “self-hypnosis,” and “a tasteless bit of ethnocentrism, seasoned with self-interest”), it ends up shaping their actions towards blacks (p. 200). Since white people occupy a superior standing in society compared to blacks, they put blacks in a situation (such as excluding them from labor unions and various jobs) in which they are forced to become unreliable and undisciplined. This ‘vicious circle’ can be broken by recognizing the false initial definition of the situation and by abandoning it. This, however, is not done by a “simple act of will,” nor by “a continuing ‘educational campaign’” (p. 197). Instead, a “deliberate and planned halt” is required, applying the right “institutional controls,” enacting new “legislation,” and promoting “deliberate institutional change” (pp. 208–210). To those “amateur psychologists” who fatalistically see ethnic groups as necessarily acting in a hostile manner against each other, Merton retorted that “blind panic and racial aggression are not rooted in ‘human nature’,” but are instead “largely a product of the modifiable structure of society” (p. 209). As proof, he cited the encouraging results of a biracial housing study that he was directing at the time.⁵ Warning that “these changes, and others of the same kind, do not occur automatically,” he concluded that “it is only with the rejection of social fatalism implied in the notion of unchangeable human nature that the tragic circle of fear, social disaster, reinforced fear can be broken” (p. 210).⁶

Merton’s essay was published at a time in which important sections of the US progressive intellectual, legal and political elites were mounting a full-fledged attack against racial segregation. The National Association for the Advancement of Colored People (NAACP) had started to fight educational segregation since the mid-1920s, and gained its first legal battles in the late 1930s, obtaining that state schools equalize facilities and teacher pay (Tushnet, 1987). In the late 1940s, new cases were initiated as part of a general strategy aimed at reaching the Supreme Court in the hope that it would overturn the ‘separate but equal’ doctrine. It was then that NAACP attorneys started to recruit psychologists and sociologists, most famously Kenneth B. Clark, to offer evidence in court on the dangers of segregations and on the falsity of the assumptions on which it was based.⁷ In his authoritative account of the involvement of social scientists in these legal cases, Jackson (2001) points out that many were themselves members of racial minorities and mSany were “the products of liberal/leftist political tradition, often with an overlay of radical Christian theology that emphasized the equality of human beings.” (p. 7). Merton did not testify in court, but was closely connected with many of the experts who did, and was part of a broader network of scholars who were instrumental in disseminating ideas against racial oppression. These included Gunnar Myrdal (1944), the main author of An American Dilemma: The Negro Problem and Modern Democracy, Else Frenkel-Brunswik, who co-authored The Authoritarian Personality (Adorno et al., 1950), Robert M. MacIver (1948), author of The More Perfect Union (1948), Ashley Montagu, who was rapporteur in 1950 of the UNESCO (1950) statement on ‘The Race Question,’ and Kurt Lewin, one of the founders of social psychology.⁸ Having its center of gravity in New York City and strengthened by many intellectuals who fled persecution in Europe, this network was supported by various progressive associations and foundations, such as the American Jewish Congress, the Carnegie Foundation, the Ford Foundation, and the Rockefeller Foundation.

In October 1952, Merton signed the ‘The Effects of Segregation and the Consequences of Desegregation: A Social Science Statement,’ which was annexed to the appellate brief filed to the Supreme Court (Clark et al., 2004). In May 1954, in what is known as the Brown v. Board of Education decision, the court unanimously ruled in favor of the plaintiffs by declaring racial segregation unconstitutional, stating that segregation negatively affects children by generating “a feeling of inferiority as to their status in the community that may affect their hearts and minds in a way unlikely ever to be undone,” and citing various social scientific works in a footnote, most prominently Myrdal’s Dilemma (US Supreme Court, 1955b; see also Mody, 2002).

The court requested the parties to present further argument on two key questions: whether desegregation should be implemented forthwith or gradually, and in what manner it should be executed. Over a dozen briefs were filed by the parties and external organizations. The ‘Social Science Statement’ signed by Merton and filed for the first decision already contained a clear opinion on this respect, and suggested that desegregation was more likely to succeed the more it was executed “simultaneously” everywhere and with a “consistent and firm enforcement of the new policy by those in authority” (Clark et al., 2004, p. 499).

However, segregationist institutions had by then started to react by enrolling their own social science experts.⁹ The most significant for our analysis of the SFP is a report authored by Lewis M. Killian for the Attorney General of Florida (where segregation was enforced by law), which was annexed to an amicus curiae brief filed to the Supreme Court in 1955.

Born and raised in segregated Georgia, Killian had obtained a PhD at the University of Chicago with a dissertation on race relations in Chicago’s poor neighborhoods, focusing on white working class ‘hillbilly’ immigrants from the South. He had published in the main sociology journals and had just obtained a professorship at Florida State University. He was recruited by Richard Ervin, the Florida Attorney General, to lead the Research Advisory Committee for the Study of Problems of Desegregation in Florida Schools, composed of twenty people in total, mostly university researchers in departments of education and political science. The report issued from the study was annexed to the amicus curiae brief, presenting Killian as “director of research and coordinator of the project” (Ervin & Odum, 1954, p. 102).¹⁰

Killian decided to conduct a survey on the attitudes and opinions on desegregation of Florida officials and community leaders both white and black, including peace officers, school administrators, judges, newspaper editors, and ministers.¹¹ One of the “basic assumptions” of the study was that “expressed attitudes are to some extent and in certain types of situations indicative of later behavior” (Ervin & Odum, 1954, p. 106), an assumption seemingly consistent with Merton’s claim in the SFP essay that one’s definition of the situation shapes one’s behavior.

Killian reported that about three-quarters of white respondents were opposed to the Brown decision and that one third of white respondents, mostly peace officers, declared their intention to actively oppose the implementation of the ruling.¹² Moreover, over 75% of peace officers were expecting riots and violence if schools were to be desegregated, and half of them believed that they would not be capable of containing violent behavior. In the conclusion, Killian suggested that the respondents’ predictions, especially those made by peace officers, did not reflect an accurate assessment of the situation, but were merely the expression of hostile feelings toward Brown and racial integration. Here again his reasoning seems consistent with the 1948 article and echoes Merton’s ‘scrambled logic’ and ‘self-hypnosis. But in a surprising twist, Killian argued that peace officers’ projections of their feelings on their predictions would have the effect of diminishing their efforts to maintain law and order:

The existence of a positive relationship between the attitudes of peace officers towards the decision and their predictions of the inability of police to control serious violence suggests the existence of a tendency to project their own feelings into situations which might arise and into other persons involved. Such highly subjective predictions are very likely to be indications of what the respondent himself would do, or feels he could do, in the situation. Such predictions may very well take on the character of “self-fulfilling prophecies” [footnote to Merton’s 1948 article]. If police officers do not believe that they or others would not be able to maintain law and order if serious violence occurs, the likelihood that they will attempt wholeheartedly to do so is altogether reduced. (Ervin & Odum, 1954, p. 128)

In the “general conclusions” of the report, which were reproduced verbatim in the main body of the amicus curiae brief, Killian argued that, “while it is true that expressed attitudes are not necessarily predictive of actual behavior,” active and violent opposition to desegregation (especially if desegregation was immediate) was to be expected (p. 128). This violent outburst could escalate uncontrollably, since “opposition of peace officers to desegregation, lack of confidence in their ability to maintain law and order in the face of violent resistance, and the existence of positive relationship between these two opinions indicates that less than firm, positive action to prevent public disorder might be expected from many of the police” (p. 128).

Although Killian did not make explicit policy recommendations, he cautiously suggested that there was a lesson to be learned from the survey, namely that desegregation should be gradual and adjusted to local needs:

In view of white feelings that immediate desegregation would not work and that to require it would constitute a negation of local autonomy, it may be postulated that the chances of developing firm official and, perhaps, public support for any program of desegregation would be increased by a decree which would create the feeling that the Court recognized local problems and will allow a gradual transition with some degree of local determination. (p. 128)

Instead, the main body of the amicus curiae brief was much more explicitly normative. It stated that if Brown were to be implemented immediately and indiscriminately, a “seismic shock” would engulf “a large part of the nation in a tidal wave of hate and inflamed emotions,” carrying away “a public school system which took half a century and billions of dollars to build.” The authors of the brief urged to recognize that “some way must be found to protect the constitutional rights of a minority without ignoring the will of the majority,” thus reaching a decision whereby “rational consideration is given to human frailty.” Brown had to be considered as a “high goal” that “men of good will” should strive after, knowing that “the only answer is time and the patient efforts of those who value democracy” (Killian, 1994, p. 97). Thus, a theory that was originally developed to advocate “deliberate institutional change” and reject “social fatalism” (Merton, 1948, p. 209–210), was here used to support a decision that would slow down and limit, rather than strengthen, the scope of institutional controls.

When the NAACP received the Florida brief, it immediately reacted by recruiting its own experts to criticize Killian’s study and defend a rapid and universal implementation of desegregation: Louis Harris and Norman Rosenberg, two experts in survey and polling methods. Rosenberg’s memorandum is particularly interesting for our understanding of the flexible nature of the SFP, as an additional interpretive twist is here given to the theory. Titled ‘The Ability of the Florida Poll to Foretell Future Action,’ the memorandum raised a number of methodological objections against Killian’s study. For example, one of the questions asked in the Florida poll was “Check the sentence that most nearly describes your feelings toward the Supreme Court decision,” and one of its possible answers was “Firmly against the decision; will actively oppose any attempt to end segregation in Florida schools.” Rosenberg noted that this answer conflated questions about one’s attitude with questions about one’s predicted behavior, and raised doubts about the possibility of inferring the latter from the former.

Citing Richard LaPiere’s (1934) famous study of anti-Chinese sentiments in US hotel proprietors, which showed the priority of actions over attitudes, Rosenberg (1955) argued that “even the most firm declarations of future behavior are un-trustworthy when made without a clear conception of the future situation, and when the promised action is found to entail special effort, inconvenience, hardship or danger” (p. 3). The poll lacked effective countermeasures to reduce its ‘artificiality’ and its remoteness to real life situations, thus making it a bad predictor of future action.

To be sure, Killian’s main argument about integration causing violence to occur in Florida was not based on respondents’ declarations of their own behavior, but on respondents’ predictions of others’ behavior (i.e., violent acts against desegregation and peace officers’ ability to control of violence). For Killian, these predictions were not actually predictive of any real event: they were not factually correct, at least not before the materialization of the SFP. But they were true in the sense that people really believed them, and they would act on their basis, thus producing a SFP. However, Rosenberg argued that his objections applied a fortiori to this argument, since the poll’s questions were formulated in such a way to make it impossible to separate “genuine expectation of violence from a mere readiness to give an anti-integration response.” The respondents’ answers were not “more or less considerate judgments of fact” but “disguised attitude responses.”

For Rosenberg, this was a case “where a poll presents as ‘crystallized’ opinion on a matter of fact answers which are brought into being by the instrument that measures them.” Most survey respondents had probably “no firm opinion” on what would have happened but were simply eager to answer in a way suggesting that desegregation would bring much trouble. Citing Herbert Blumer (1948), he wrote that accepting the answers as genuine beliefs meant endorsing a “narrow operationalist position” according to which “public opinion consists of what public opinion polls poll” (Rosenberg, 1955, p. 12). The SFP was not something that would have occurred spontaneously in Florida were integration to be executed rapidly and universally. Instead, a SFP would have been triggered by the fact that social scientists like Killian were interpreting the answers to such questions as genuine beliefs rather than as expressions of feelings:

the check-answers selected may have been conditioned by the very reasonable notion that the surest way of preventing desegregation is to persuade decision-makers that the results of desegregation, if it occurs, will be disastrous. This may be regarded as a variant of the ‘self-fulfilling prophecy’ which Killian animadverts on in the section of his conclusions stressing the significance of the peace officers’ responses to the ‘will there be violence’ and ‘can you cope with violence’ questions. (pp. 12–13)

From this perspective, the answers of the respondents were given with the more or less conscious hope that, by raising the specter of civil war in surveys, they would have influenced decision-makers to stop, slow down, or limit desegregation. The fact that peace officers predicted uncontrollable violence to occur signified simply “their reluctance as responsible officials to approve a new situation that will undoubtedly cause some derangement in their habitual routine” (p. 13).

Rosenberg reproached Killian for nourishing a SFP instead of breaking the vicious cycle. For Killian, the SFP was a false definition of a future situation (peace officers believing that if desegregation occurs, they will not be able to contain violence), leading to a pattern of individual behaviors (less firm law enforcement by peace officers when desegregation occurs), leading to a situation in which the originally false definition becomes true (when desegregation occurs, widespread violence ensues). For Rosenberg, the SFP was a false definition of a future situation (survey respondents claiming that if desegregation occurs too rapidly people will revolt against it), leading to a false definition of a future situation among members of another group (policy-makers believing that if desegregation occurs too rapidly, violence will ensue), leading to a pattern of individual behaviors (measures taken by policy-makers to prevent or slow down desegregation), leading to a situation in which the originally false definition becomes partially true (desegregation does not occur).

Although not annexed to the brief for appellants, Rosenberg’s memorandum inspired the strategy of the NAACP against the Florida brief and other similar documents that had been filed to the Supreme Court. The brief for appellants argued that “dire predictions” of violence were “unreliable” and cited various sources of “modern psychological research,” including LaPiere’s study, showing that “such polls are not a valid index of how the individuals questioned will in fact act in the event of desegregation” (U.S. Supreme Court, 1955a).

The final decision of the Supreme Court, handed down on 31 May 1955, recognized the value of documents such as the Florida brief and sent back the cases to the district courts urging them to make the necessary decisions to integrate schools. The defendants had to make “a prompt and reasonable start toward full compliance” with the 1954 ruling, and admit students to public schools on a racially non-discriminatory basis “with all deliberate speed” (U.S. Supreme Court, 1955b). This vague formula has been interpreted by many, including Killian himself, as a pronouncement in favor of gradualism. In his autobiography, Killian (1994) also claimed that peace prevailed in Florida in the following years (a fact he attributed to the gradualist approach endorsed by the Supreme Court) so that it was not possible to determine whether a SFP would have taken place or not (p. 81).¹³ But the ruling did not put the final point to the legal vicissitudes of racial segregation, as hundreds of lawsuits were filed in the following years against institutions accused of furthering segregation in one way or another. In many of these cases the SFP continued to be employed, often as part of arguments against the use of tracking systems in schools (see Wineburg, 1987).

The main cleavage in the 1955 debates before the Supreme Court was a gradual versus a radical approach in the equalization of race relations. But it should be noted that the political battle was much broader than the positions embodied by the Florida Attorney General and the NAACP. Many people and organizations in the South, were Killian lived and worked, were hard-line segregationists who abhorred the gradualism of the Florida brief as much as the radicalism of the Supreme Court decision. Killian received harsh critiques both from members of the elite (and mostly Northern) network of sociologists to which he belonged, who accused him of making up pseudoscientific justifications for bigots, and from hard-line segregationists he met daily inside and outside the campus, who accused him of accepting the conditions posed by a “communist-inspired ruling” (Killian, 1994, p. 83). Still, Killian felt that while working for the segregationist establishment meant accepting many compromises, it could also be an opportunity to make people aware of the sociological knowledge available on race relations. “It represents a marked advance in the South,” he wrote shortly after his experience, “when social scientists are regarded as competent in the field of race relations, when they come to be viewed as ‘experts’ rather than as ‘agitators’” (Killian, 1956, p. 214).

In summary, the SFP was used in a legal dispute that was also a political battle that had the potential to alter the balance of power for millions of people living in segregationist states. It is of course difficult to establish whether the SFP had any real impact, or, in other words, whether any actor involved in the controversy changed his mind because he was exposed to the theory. The SFP, like any other theory that was mentioned in the proceedings, may have been used strategically as additional rhetorical ammo to support a pre-existing view, or it may have genuinely convinced some people that things had to be done differently in light of its purported lessons. What we do know is that it was mentioned abundantly in the debates surrounding a historic legal decision, a rare privilege for a sociological theory.

3 Endogenous Factors of Political Flexibility: Generality and Ambiguity

Reading Merton’s 1948 essay, one might think that the policy implication of the SFP is obvious: the reason why members of racial minorities do not act as they should is not that they are naturally predisposed to do so, but because existing patterns of discrimination prevent them from doing so. Discrimination is not based on good reasons but on false beliefs (which are reinforced by existing discriminatory practices) and self-interest. In order to make it possible for racial minorities to act as required or expected, and in order to correct false beliefs in the majority population, policymakers should immediately make discrimination (and its extreme variants, such as segregation) illegal, impractical, or materially impossible.

However, as we saw, in the briefs produced by social scientists for oral arguments before the Supreme Court in desegregation cases, the SFP was used to support not only this policy option but also a much more prudent and gradualist one. It is a legitimate question whether Merton, Killian, and Rosenberg actually used the same theory. One could argue that by employing the same term, these authors meant different things. If that were the case, it would not be correct to say that the SFP was used flexibly, but rather that different theories were used to promote different political project. Instead, I argue that it is more parsimonious and sociologically fruitful to assume, following Barnes (1974), that “any particular belief may be made to serve any particular interest,” (p. 129) and that, in the case at hand, a single theory was used in a politically flexible manner, a fact that was enabled by several factors.

The first source of the SFP’s political flexibility is its generality (as opposed to particularity). Social science theories can have different degrees of generality; the phenomena that they are supposed to make sense of can be more or less particular, more or less concentrated in space and time, and they can deal with bigger or smaller portions of the social world (see Stinchcombe, 1968; Alexander, 1982; Swedberg, 2020). Highly general theories can be applied to a higher range of social groups, events, and phenomena that might have a different political resonance for different people, so that they are more likely to be used in a politically flexible way (see Brandmayr, 2021). Of course, there may well be theories that are more general than the SFP: after all, Merton was a staunch advocate of ‘middle-range’ theories and the SFP certainly qualifies as such. But the generality of the SFP was displayed by Merton himself in his essay, as he offered illustrations taken from disparate situations, such as bank runs and labor union disputes.

In the case in point, Merton, Killian, and Rosenberg applied it to make sense of different things. Merton applied it to racist beliefs of the type ‘black people are inferior to white people.’ The practical recommendation there was to make illegal the acts informed by those beliefs. Killian applied the SFP to beliefs of the type ‘if desegregation is enforced immediately civil war will break out.’ The practical recommendation there was to desegregate gradually. Rosenberg applied the SFP to beliefs generated by polling procedures of the type ‘people believe that if desegregation is enforced immediately, civil war will break out.’ The practical recommendation there was to avoid polling attitudes altogether.

The second source of the SFP’s political flexibility is its ambiguity or multivocality (as opposed to univocality). A theory is ambiguous when it is composed of multiple elements that are relatively independent from each other—they are not merely logical steps in a linear argument—and that can be emphasized selectively when the theory is interpreted and used. By contrast, a univocal theory is one that is highly coherent and where meaning, so to speak, holds together. Ambiguity does not mean that theories are interpreted in different ways by their readers and users. That is rather a consequence of ambiguity; highly ambiguous theories possess an intrinsic multiplicity of voices, which is what allows divergent interpretations and uses (see Levine, 1985; Griswold, 1987; Davis, 1999; Davis, 2008; Reay, 2012). The more ambiguous a theory is, the higher its capacity to accommodate different political uses (see Brandmayr, 2021). The SFP originally appeared as an essay published in a literary magazine, written in a lively style, rich in anecdotes, and using a host of classical rhetorical devices.¹⁴ Different ‘voices’ can be isolated in the essay. Table 1 lists seven of them: anti-fatalism, institutionalism, cognitivism, critique of ideology, moral responsibility, non-intentionality, and positivism.

Table 1: Theoretical components of the SFP

While neither Killian nor Rosenberg explicitly referred to a specific component of the theory, they implicitly emphasized different aspects of it. Killian emphasized the cognitivism and non-intentionality components of the theory, by deciding to give weight to expressed attitudes as indicators of later behavior and by stressing the fact that police officers would not have intentionally neglected to maintain law and order. The positivism component was also strong in his report, as signaled by expressions such as ‘it may be postulated that the chances’ and ‘would be increased by,’’ but these references are more a matter of style than substance and do not seem particularly relevant in determining the line of action recommended by Killian. Conversely, Rosenberg mainly emphasized the critique of ideology component, by stating that declarations of future behavior are un-trustworthy (they are, so to speak, ideological) and that survey respondents seek to influence decision-makers to promote their interests. This claim, in turn, implicitly rested on a strong institutionalist view, according to which the final word and the power to transform society are in the hands of decision makers. Finally, by suggesting that Killian was feeding the vicious circle of the SFP with his survey, Rosenberg implicitly drew on the moral responsibility component.

In his original essay, Merton included explicit practical recommendations and did not shy away from using loaded language to denounce the workings of those who perpetuate harmful self-fulfilling prophecies. As Kalleberg (2010) remarked, “there can be no doubt that Merton here is a social critic. He unmistakably criticizes the social reality he has first described, emphasized with normative expressions like ‘tasteless ethnocentrism,’ ‘scrambled logic,’ ‘self-hypnosis and ’specious evidence’” (p. 191).

But the theory set forth in the 1948 essay was many other things. It was a highly general explanation that could be applied to many different situations involving very different types of social groups. And it was an ambiguous, multivocal theory, a product of Merton’s peculiar synthesis of scientific and humanistic proclivities.

4 Exogenous Factors of Political Flexibility: Adversarial Legalism and Political Culture

Exogenous sources of political flexibility are all conditions external to a theory itself that make it more likely that the theory will be used in support of opposite political projects. As we saw, such uses either occur in different contexts or, if they occur in the same context, they are performed by different individuals, so that the exogenous conditions should be found in the differences between contexts and between individual theory users. I argue that two main exogenous factors explain the flexibility uses of the SFP in the debates in Brown v. Board of Education. The first is the fact that it was used in a court proceeding in a context characterized by an adversarial system. In the words of Kagan (2003), the American legal system involves a method of dispute resolution, “in which the assertion of claims, the search for controlling legal arguments, and the gathering and submission of evidence are dominated not by judges or government officials but by disputing parties or interests, acting primarily through lawyers.” (p. 9)

This has profound implications on how scientific knowledge is used in court. Kagan (2003) writes:

In complex cases in which expert technical assessments are required, contending American litigants each hire and carefully coach their own expert witness; in a more hierarchical system such as Germany’s, the court appoints a single ‘neutral’ expert witness, who is not coached in advance by anyone. (p. 107)

Many have suggested that this system encourages an opportunistic use of scientific evidence: as Sanders (2009) puts it, “expert knowledge is a partisan resource.” (p. 69) The adversarial system creates “the temptation for purely opportunistic uses of science” (Jasanoff, 1997, p. 68; see also Haack, 2014, Ch. 8).

The adversarial system thus seems particularly conducive to the flexible use of the same theory. If one side in the legal dispute draws on a theory to provide some scientific justification to its case, the opposing side might of course try to refute the theory as pseudoscience. But if the theory is hard to disprove because it enjoys great credibility, then litigants might instead try to claim the theory for themselves and argue that it was wrongly applied in the first place. This is precisely what happened in the Brown v. Board of Education case with respect to the SFP.

The other exogenous factor is that the SFP was used by individuals inhabiting contexts characterized by different political cultures, conceived as “sets of symbols and meanings or styles of action that organize political claims-making and opinion-forming” (Lichterman & Cefaï, 2006, p. 392). Political cultures provide rules and tools to think about what is legitimate and what is illegitimate, what is feasible and what is unfeasible, what is thinkable and what is unthinkable. There is such a thing as political culture because political attitudes do not follow institutional transformations mechanically but have instead a certain persistence over time (Acharya et al., 2016).

There is little doubt that the protagonists of our study were immersed in different political cultures. Merton published the original essay as a professor at Columbia University in New York, in a milieu in which progressive and democratic ideas prevailed and races were not segregated. This was also the environment from which the NAACP hired most of his experts, including Rosenberg. Instead, Killian used Merton’s theory as a professor in racially segregated Florida, where white people, including intellectuals, tended to be much more conservative, especially on racial issues. Although he correctly felt that he was part of the progressive minority of his environment, what he meant by progressivism was not identical to what progressivism was for most people working for the NAACP. Being a progressive in post-war New York meant arguing for the radical dismantling of any form of racial discrimination. The same in post-war Tallahassee meant advocating a much more modest program of reform. White people radically opposed to segregation existed in the South, but the stigmatization and even physical danger that they faced meant that fewer people than in the North identified in and acted according to that position. Similarly, living and being constantly exposed to the reality of segregation in the South meant that it was easier for people living there to visualize the multitude of practical difficulties connected with a universal and immediate desegregation, so that a more modest approach made more sense to them.

5 Conclusion: The Limits of Political Flexibility

In contemporary accounts of the struggle against segregation, it is not uncommon to see the SFP listed as a theoretical resource and Merton as an ally. As Weinstein et al. (2004) put it in an influential article, “[i]f the unmet promise of Brown v. Board of Education is to be realized, the ‘social fatalism’ identified by Merton (1948) must be rejected and wider and deeper changes made. Merton argued that negative self-fulfilling prophecies are not inevitable.” (p. 517) In this article, it was of course not my intention to suggest that the SFP did not provide guidance to progressive activists and civil rights organizations, or that Merton should not be considered as someone who contributed to the struggle against racial inequality. Rather, my intention was to show that Merton’s theory of the SFP was also borrowed and put to use by actors who were battling precisely those civil rights organizations on political grounds, and who had starkly different views on racial matters. As I have shown, such political flexibility occurred because of endogenous features of the theory and because of exogenous features of the contexts in which it was used.

Was it possible, though, for the SFP to ‘become all things to all men,’ like Saint Paul? Could it have been used to defend the view of a hard-line segregationist? We have good reasons to think that there are limits to its political flexibility. With all its multivocality, the SFP is still grounded on a vision that puts priority to beliefs over objective constraints, whether such constraints are grounded in social structure or in biological determinism. The SFP clearly expresses what Merton (1994) identified as his “prime theoretical aversion,” namely that “to any extreme sociological, economic, or psychological reductionism that claims to account uniquely and exhaustively for patterns of social behavior and social structure.” (p. 19) In their original borrowings of the SFP, both Killian and Rosenberg still proceeded according to the assumptions that attitudes and beliefs matter, that they make a difference. It is conceivable to imagine people using the SFP to say that only certain beliefs (as opposed to other or all beliefs) matter, or perhaps that only beliefs (as opposed to other factors) matter.¹⁵ On the other hand, it is very hard to imagine someone using it (approvingly) to claim that beliefs do not matter at all.

How is the SFP used in contemporary political disputes? Answering this question would requires a systematic investigation on its own. But if we keep the focus on debates about racial issues in the United States, a cursory glance suggests that the theory is invoked across the political spectrum. Surprisingly, it is on the left that the theory seems to be used more sparingly. The SFP occupies a marginal position in the vast progressive literature on racial justice compared to such concepts as systemic racism or white fragility. In some cases, it is suggested that the structure of racism emerged in the first place as a consequence of white people defining non-white people as inferior, an initially false ‘prophecy’ that eventually became reality for groups subjected to mass incarceration, unemployment and poor health. But overall, the SFP and Merton are almost never mentioned by popular anti-racist writers like Kimberlé Crenshaw, Robin DiAngelo, or Ibram X. Kendi. One may speculate that for many activists on the left, the SFP underestimates the role of strategic and conscious efforts on the part of the white population to maintain their supremacy on non-white people. As we saw, however, ‘non-intentionality’ is just one component of the theory, and other components, such as ‘institutionalism’ and ‘moral responsibility,’ could be emphasized when using the theory in support of progressive goals. Why in the early 21^st century this is not the case is an interesting question that remains to be answered.

On the other side, conservatives of various stripes display a remarkable ingenuity in their uses of the SFP. In the United States, centrists and conservatives worry about what they call the ‘great awokening,’’ the rapid diffusion of radical ideas (often grouped together under the label of ‘critical race theory’) according to which racial inequality and discrimination are ubiquitous and ingrained even within the most revered institutions. Conservatives see these ideas as the false prophecies that, if believed to be true, can have pernicious self-fulfilling effects. For example, the primer on ‘critical race theory’ published by the conservative John Locke Foundation states that the theory is a SFP because it “perpetuates a sense of helplessness among people of color and those belonging to nondominant identity groups,” which “are taught that the entire ‘system’ exists to prevent their success, so why try?” (Balfour, 2021) In other cases, conservative activists claim that ‘critical race theory’ erroneously interprets any social phenomenon through the lenses of power struggles between racial groups, so that people accepting that interpretation are led to see other races as threatening, which in turns exacerbates racial tensions.

Understanding the political power of social science theories calls for a sociology of ideas that focuses not only on how knowledge is produced and evaluated but also on how it is put to use in a variety of contexts, as documented by a growing literature (e.g., Camic et al., 2011; Berman & Milanes-Reyes, 2013; Fligstein et al., 2017). The study of the extents and limits of theories’ political flexibility should be an integral part of such a sociology. While some scholars (e.g., Hallett et al., 2019) acknowledge the multiplicity of meanings, including political meanings, that social science theories can harbor, many accounts still make the implicit assumption that theories are politically rigid: they provide support either to the left or the right, either to progressives or conservatives, either to egalitarians or elitists. As I have shown, the case of the SFP suggests that this assumption is unwarranted. Moreover, different theories likely allow for different degrees of flexibility, depending on such endogenous features as their generality, ambiguity and embeddedness in a normative framework. A better understanding of this variation calls for comparative analyses of theories that focus both on their content and their use.