‘For men arousal is orientation’: Bodily truthing,technosexual scripts,and the materialization of sexualities through the phallometric test

Bodily truthing, materialization, and technosexual scripts

While STS as a field has not engaged much with sexuality studies, or vice versa, a range of scholars have pioneered approaches that have built bridges between the two fields (examples include Casper and Moore, 1995; Epstein, 2003, 2006; Fausto-Sterling, 2000; Fishman, 2004; Jordan-Young, 2011; Loe, 2004; Mamo, 2007; Martucci, 2010; Moore, 1997; Terry, 1999; Waidzunas, 2011). We seek to strengthen these connections by developing three concepts, which we refer to below as bodily truthing, materialization, and technosexual scripts.

Tracing a technological history with citation analysis methodology

To trace the development, uptake, and evolution of the phallometric test in the West, we built a sample of relevant abstracts from selected scientific articles in the Web of Science database, and then performed a content analysis. To build the set, we used a citation analysis method drawing on Evans (2002: 205–208). While citation analysis has known limitations (Edge, 1979; Hicks and Potter, 1991), it served our purposes by permitting us to treat citation counts as a very rough indicator of an article’s influence. We started with Kurt Freund’s 1963 publication, ‘A laboratory method for diagnosing predominance of homo-erotic or hetero-erotic interest in the male’ (Freund, 1963), his first discussion in English of the methodology of phallometric testing. Taking this widely cited article as a ‘seed’, we first retrieved all articles that cited it. We then screened those abstracts for significance and relevance, retaining in our sample only those that had themselves been cited 10 or more times and excluding articles if they were neither scientific research studies that use the phallometric test as a measurement tool, nor review articles with sufficient commentary on phallometry. Dividing the time span of 1963–2010 into six 8-year-long cohorts, we then identified the five most cited articles in the first cohort (1963–1970). We used these five articles as the next set of ‘seeds’, retrieved all articles that cited them, and reapplied the inclusion criteria in order to add an additional set of articles to the sample. Proceeding sequentially, we continued to expand the sample using the five most cited articles in each cohort as seeds. Altogether, this process generated a set of 1290 abstracts for consideration. After screening them for relevance, we were left with the sample of 410 articles.

From an initial review of selected abstracts, we developed a coding scheme with master categories of ‘Uses’, ‘Diagnostic Categories’, ‘Populations’, and ‘Methodology’. We also developed a list of commonly occurring subcategories for each of these master themes, while using ‘other’ codes to collect residual events that did not fit our scheme. We reviewed the full text of the article and/or consulted as a team when abstracts seemed difficult to code, and we developed new codes for subcategories as needed. ⁵ On the basis of this coding, we constructed bar graphs of each code over time, and we used the heights of bars to indicate dominant trends. The three assemblages that we describe in this article emerged inductively from analysis of these trends. This quantitative analysis was then supplemented by careful reading of many full-length review articles from the sample that specifically elaborated on dominant trends in the research field.

This methodology was designed to examine the trajectory of research emanating from the invention of the original test and penile plethysmograph device through direct citation, and we relied on review articles to confirm major trends. We acknowledge that this methodology provides one particular version of the history of the phallometric test in the sciences in the West, based on our strategy for sample construction and our procedure for coding. While an alternative strategy that increases coverage might identify additional assemblages, we believe those we have uncovered to be most central. ⁶

A diagnostic test based on predominance of sexual interest

During the 1950s, when psychiatrist Kurt Freund worked at Charles University in Prague, the Czech military suspected that men were pretending to be homosexual to evade service. Freund developed the ‘phallometric test’ as a means to identify those soldiers who were coming to psychiatrists feigning homosexual desires in order to get a ‘homosexuality certificate’. By exposing subjects to a series of nude images of men and women and seeing which produced erections, a bodily truthing regime would distinguish homosexuals from heterosexuals (Freund et al., 1958). Struggling for credibility and respectability, Freund could not simply visually observe men’s penises or rely on subjects to report about them. To monitor the genitals more systematically, he invented the ‘penile plethysmograph’, a glass cylinder encapsulating the subject’s penis, sealed and mounted such that changes in penile volume could be transmitted through an air tube and objectively recorded by a pen on a rotating drum (Freund et al., 1965). He took the concept of plethysmography, previously used to measure limb volume through air displacement, and applied it to the penis as the ‘volumetric air chamber’ (VAC) plethysmograph, shown in Figure 1 (Freund, 1977).

OPEN IN VIEWER

Figure 1.

Kurt Freund’s volumetric air chamber (VAC) penile plethysmograph (Freund et al., 1965).

Beyond this device, the phallometric test also exposed subjects to visual erotic stimuli in the form of nude still images of males and females (both adults and children), accompanied by a set of calculating practices. Freund created randomly ordered series of photographs, depicting individual men, women, boys, and girls, to test gender and age preferences. Each tracing was rated and the results compared using a statistical significance test. Freund (1963) took the sum of responses to one category of erotic imagery compared with another (male vs. female, pre-pubescent vs. pubescent vs. adult), assigning a diagnosis via each subject’s statistically predominant arousal pattern. Yet a ‘normal’ diagnosis did not foreclose the presence of deviant arousal, since, in the case of age preferences, ‘normal’ heterosexual men often exhibited low levels of arousal to young girls – here the main principle was statistical predominance (Freund et al., 1973). In short, researchers in this assemblage sought to answer the question ‘How do we categorize a man based on his sexual desire?’ and to do so, Freund’s test scripted a series of assumptions including the representativeness of erotic imagery and the relevance of erections to the identification of male desire. Freund also assumed the ‘external validity’ of the test – that is, that results from a highly scripted and technologically mediated laboratory encounter between a researcher and a (willing or unwilling) subject usefully characterized male sexuality in general in the varied social situations in which it can more typically be found.

The materialization of male desire in Freund’s test was bound to his theory of the ‘sexual functional system’. In this theory, ‘erotic arousal level’, observable through genital arousal measurement, was the product of fixed ‘internal erotic motivators’ brought into effect by ‘external erotic stimuli’ experienced visually. Erotic arousal level was a necessary antecedent to ‘erotic behavior’, or sexual activity. Genital arousal measurement was an indirect yet workable means to access a ‘core’ set of internal erotic motivators configured within the nervous system of the subject. Freund felt that his method was ‘crude’, but the best measure available until directly accessing the reproductive system’s neural circuitry was possible. Yet Freund (1981) argued that phallometry surpassed other measures of general arousal, such as galvanic skin response or heart rate, as such techniques failed to capture the specificity of the sexual.

While the phallometric test could be sold as a way to precisely measure sexual arousal, bodily truthing was always vulnerable to faking if subjects produced ‘false’ erections or suppressed a response. Freund identified ‘mechanical artefacts’, jagged lines indicating that the subject had engaged in ‘pumping’, flexing muscles to produce tumescence. Preventing suppression required different tactics, however, and Freund (1961) experimented with administering testosterone shots the day before testing or giving subjects wine. Later, he developed ‘priming’: showing a socially accepted stimulus, such as a nude female, followed by a socially undesirable stimulus, making the withholding of erection more difficult (Freund et al., 1979). Such modifications to the technology demonstrate how the resistance of subjects played a crucial role in the design of the phallometric test. Meanwhile, as researchers tinkered with the technology, research subjects were rendered more or less trustworthy depending on the use of these auxiliary measures. In other words, contrary to the official understanding of the test by investigators, the test situation, in practice, always and necessarily included a dyadic encounter between the experimenter and the man experimented upon. The trajectory of the test over time reflected the particularities – and peculiarities – of these technosocial interactions.

The diagnostic function of the phallometric test has proven an enduring framework, and over the years it has been incorporated by researchers into a number of laboratory and correctional settings. As revealed by our coding of abstracts, the level of diagnostic uses of the test peaked in the 1990s, when sex offenders became a primary focus. Over time, as researchers utilized Freund’s diagnostic method, the technique gradually deviated from his original design. For example, as we will describe later, many researchers began utilizing an alternative plethysmograph called the ‘circumferential transducer’ (CT) in place of the VAC. In other cases, however, tinkering reflected practical exigencies related to the broader circumstances of testing. For example, in jurisdictions where erotic visual imagery of children was illegal, researchers used audio vignettes as an alternative – yet this innovation prompted debates about whether film, still images, or audio vignettes were the most effective and precise stimuli to elicit a man’s true sexual desire (Abel et al., 1975). Even as these modifications were made to the technology, the fundamental technosexual script remained intact within this assemblage: statistical comparisons of erections in differentiated scenarios of arousal could define the subject under test within a sexuality category.

In general, Freund opposed the use of the phallometric test in the context of treatment of sexual disorders. Because internal erotic motivators were determined early in life through prenatal hormonal factors, attempting to change them was relatively fruitless, in Freund’s view. (These considerations led Freund to support the decriminalization of homosexuality in Czechoslovakia in 1961 and the removal of homosexuality from the DSM in the early 1970s (Kuban, 2004).) Moreover, using phallometry in therapeutic scenarios was considered unwise, according to Freund (1977), because of the excessive pressures on subjects to fake responses in these conditions. ⁷

Phallometry as a test of predominance could, however, be employed in studies of the etiology of sexual abnormalities. For Freund, the study of etiology equated with the study of how internal erotic motivators became permanently crystallized. This emphasis led Freund and his colleagues to propose and demonstrate novel ‘paraphilias’ (disorders of sexual arousal or gratification), including ‘courtship disorder’. Defining this new pathology, Freund proposed that normal courtship for males proceeded through four stages, any of which could go awry, giving rise to aberrant behaviors of voyeurism, exhibitionism, ‘toucheurrism’ (inappropriately touching women in public places), and a preferential rape pattern. Because Freund believed that these behaviors tended to co-occur in the same individuals, he proposed that ‘courtship disorder’ was a unified syndrome. Using phallometry, he and colleagues found that some convicted rapists displayed arousal patterns to voyeuristic sex scenes, purportedly confirming the courtship disorder pattern (Freund et al., 1983).

The use of phallometry in scientific studies of etiology became an important element of this first assemblage. These studies, which, according to our content analysis, peaked in the 1980s and early 1990s, primarily investigated pedophilia, rape proclivity, and impotence. In the latter case, researchers were especially concerned with whether the condition was psychogenic or organic, frequently comparing phallometric testing with studies of nocturnal tumescence during sleep to make this determination (e.g. Janssen et al., 1994). Here again, while the methodology shifted with the incorporation of other types of plethysmographs ⁸ and with the emergence of alternative theories of etiology, the overarching principle of the assemblage remained constant: differential diagnosis of sexual conditions based on predominance testing could be a window into the internal erotic motivators within the male subject, sufficiently so to foster research into the etiological crystallization of those internal motivators.

The search for the etiology of various sexualities inevitably meant treating those sexual subjectivities as deviant. While in principle the diagnosis and study of the etiology of a sexuality might be couched in a neutral or even liberatory fashion (as in the quest for the ‘gay gene’), the search for such causes, as Eve Kosofsky Sedgwick (1991) has argued, is inevitably marginalizing. Thus, in this assemblage that yoked diagnoses with particular arousal patterns through statistical comparison, the technosexual script embedded in these practices relegated those diagnoses to a socially subordinated status. In the next assemblage that we consider, in which phallometry was used in the treatment of various sexual disorders, this normalizing function was even more explicit.

A physiological test of treatment efficacy

In 1966, behaviorist S. Rachman used phallometry purportedly to establish that male sexual arousal could be conditioned to fetish objects with classical conditioning techniques. With three heterosexual male psychologists as research subjects, Rachman produced and later extinguished subjects’ arousal responses to a photograph of a pair of women’s black boots. Analogous to Pavlov’s bell that produced salivation in anticipation of food in the famous dog experiments, the photo of boots became arousing as it signaled the anticipation of photographs of nude women. Utilizing penile plethysmography to demonstrate the production of erection, Rachman (1966) concluded, ‘There seems little question, therefore, that sexual arousal can be conditioned to previously neutral stimuli’ (p. 295).

By establishing that male sexual arousal could be altered, Rachman’s work helped pave the way for the behavioral treatment of sexual disorders. According to our coding of abstracts, treatments peaked first in the 1970s, when homosexuals were primary targets, and again in the 1990s, when sex offenders were central. While the treatments, target populations, and data analysis methods varied across time, the principle of this assemblage remained intact throughout, as researchers grappled with the question: ‘How can we know if a man’s sexual desire has been transformed?’ Despite Freund’s rejection of the idea that phallometry could be useful under treatment conditions because of excessive faking, behaviorists adapted the test to treatment scenarios. Especially in scenarios where failed ‘cure’ may have meant continued incarceration or mandatory treatment, the high stakes of bodily truthing often set up the readings of the phallometer against the word of the subject eager for freedom.

Among the first populations to be treated using behaviorist techniques were homosexuals, considered mentally ill by nosological standards of the time. Australian psychiatrist Nathaniel McConaghy (1969) pioneered the use of aversion treatments for homosexuality, still utilizing Freund’s vacuum air chamber penile plethysmograph and statistical methods for measuring predominance to assess outcomes. To extinguish homosexual arousal in an early experiment, McConaghy compared two aversion techniques, first pairing homosexual stimuli with injections of apomorphine to induce nausea and then applying electric shocks to the fingers. In later experiments, he used noxious chemical odors and superimposed open wounds onto homosexual erotic imagery as aversive stimuli. Throughout these experiments, McConaghy and others repeatedly found that they could diminish homosexual responses but could not induce heterosexual arousal. McConaghy (1999) later claimed that his techniques could be useful for men interested in reducing their same-sex attractions, but the aversive reorientation of homosexuality was a failed project.

Other early target pathologies for behaviorists included transvestism and a series of fetishes including sadomasochism and exhibitionism. In an early study with five patients treated for transvestism and fetishism, Marks and Gelder applied electric shocks to the arms and legs of patients as they held, viewed, or wore women’s panties, slips, blouses, and pajamas. Whereas McConaghy continued to use Freund’s predominance testing to evaluate efficacy, Marks and Gelder used ‘before’ and ‘after’ comparisons of erection levels, depicted with bar graphs. Replacing statistical significance tests, appeals to commonsense attributions of low and high levels of arousal became frequent in the literature. To demonstrate cures, researchers converged on 20 percent full erection (PFE) as a standard for maximum post-treatment erection level to deviant stimuli (Howes, 2003). Yet, in such studies, erection itself was rarely used alone to materialize rehabilitated desire, as it was frequently coupled with measures of subjective arousal or the researcher’s assessment of behavior within subjects’ relationships. In the Marks and Gelder (1967) fetish study, for example, for a patient who desired being kicked with women’s shoes to the point where his wife had broken his coccyx to satisfy him, treatment was considered successful not only because of modified arousal patterns, but also because the patient now performed regular sexual intercourse with his wife two to three times per week.

However, the ethical status of aversion therapy was challenged in the early 1970s, leading to significant decline in its use. These treatments came under fire from gay liberation activists during protests to remove homosexuality from the DSM in the United States. The gay liberation movement, which had gained momentum in the late 1960s, targeted the homosexuality diagnosis as a major pillar of social oppression, and by 1970, activists demanding demedicalization disrupted meetings of the American Psychiatric Association. Indeed, when McConaghy presented his work on aversion treatments for homosexuality at the Association’s conference in San Francisco, activists interrupted him with shouts of ‘vicious’, ‘torture’, and ‘where did you take your residency, Auschwitz?’ (Bayer, 1981: 103). Not only was aversion therapy thought to be unethical, but researchers themselves began to realize that it was ineffective, as the technique could not induce heterosexual arousal (Lande, 1980).

Following the demise of aversion therapy, behaviorists moved toward softer approaches involving repetitive conditioning. Habituation, the repetition of a stimulus in order to produce boredom, was one of these techniques. Other methods included ‘covert sensitization’, where a desired heterosexual stimulus was introduced once a man had an erection from deviant stimuli, and ‘orgasmic reconditioning’, where men were directed to masturbate to heterosexual fantasies or imagery. Despite these attempts to maintain its legitimacy, the treatment of homosexuality fell out of favor in mental health professions, and, according to our content analysis, research on homosexuality treatment using penile plethysmography also declined. ⁹

While phallometric testing could be a convenient method for assessment, a less cumbersome pick-up device was needed before the test could be used with larger populations. The VAC, while respected for its accuracy at low levels of sexual arousal, was considered by many researchers to be too difficult to use in treatment settings. Behavior therapists developed different models of ‘CT’ plethysmographs for treatment and research, the most popular one being a ‘strain gauge’ designed by British psychiatrist John Bancroft. Contemporary versions of this device are shown in Figure 2. Bancroft’s device was a thin rubber tube filled with mercury, which would change electrical resistance when it expanded and contracted (Bancroft et al., 1966). ¹⁰ The subject would place the device around the shaft of his penis, and unlike with the VAC, he could wear his clothes normally during the test. Despite its convenience, a number of debates emerged between employers of VACs (such as Freund and McConaghy) and advocates of CT. Comparative experiments in which subjects wore both devices revealed that the VAC was more accurate at low levels of arousal because, at the onset of tumescence, the penis typically increases in length before it increases in girth, and in fact girth decreases slightly before it increases. Opponents of the VAC argued that the bell jar apparatus could itself be sexually stimulating (McConaghy, 1999). Despite arguments in favor of the VAC, the CT became available as a professionally manufactured device, initially marketed by Farrall Instruments. ¹¹

OPEN IN VIEWER

Figure 2.

Circumferential transducer (CT) penile plethysmographs. (a) Mercury-in-rubber strain gauge (Crooks and Baur, 2014: 43) and (b) electromechanical strain gauge (King, 2012: 79).

In the late 1970s and early 1980s, as feminists drew attention to problems of rape and sexual abuse, treatment regimens using the new CTs turned to the population of sex offenders. The Association for the Behavioral Treatment of Sexual Aggressives was founded in 1984 after a series of brown bag lunches, one of which included a presentation by William Farrall of one of his company’s CT plethysmographs (Prescott, 2013). The earliest treatments for sex offenders, primarily child molesters, were based on aversion therapies, much like treatments for homosexuality. However, as aversion therapy declined, researchers sought new means to contend with recalcitrant urges, leading to developments in cognitive behavioral therapy and the introduction of pharmacological treatments such as chemical castration and hormone therapies.

As the problem of impotence (or what became ‘erectile dysfunction’) also became a growing medical concern in the 1980s, phallometric testing was linked to treatment regimens in a new way: as a means to test the efficacy of new pharmacological treatments for the condition. While the CT was used in many of these applications, a device called the ‘Rigiscan’ further enabled researchers to examine the strength of an erection, as it used two circumferential loops rather than one, and could differentially measure erection quality at the base and tip of the penis (Mohr and Beutler, 1990). Importantly, it was understood that drugs to treat erectile dysfunction worked not by producing arousal automatically, but by ‘repairing’ men’s abilities to be responsive to sexual stimuli (Loe, 2004). Thus, the logic of pharmacological treatment coincided with the underlying logic of the test being employed to assess its efficacy.

In summary, treatment researchers across the schools of behavior therapy, psychopharmacology, and cognitive behavioral therapy adapted phallometric testing as a means to analyze outcomes of a variety of interventions with a range of populations. Within this assemblage, a technosexual script laid out the boundaries between normalcy and pathology, based on either statistical measures or erection levels arbitrarily set in the laboratory. By materializing healthy sexual desire through measuring subjects’ low levels of arousal to deviant stimuli after treatment, researchers set the stage for the next assemblage, involving the prediction of recidivism on the part of sex offenders. If a threshold erection level could be used to demonstrate ‘cure’, then it followed that it might also be a means to predict whether a convicted sex offender will offend again.

A threshold test to predict sex offender recidivism

Anxieties about sex offenders escalated in the 1990s in what many have characterized as a ‘moral panic’ (Lancaster, 2011). In the United States and elsewhere, widespread concern about violent recidivism and a lenient criminal justice system promoted a number of institutional trends. These included the standardization of sentencing procedures that limited judges’ discretion, the lengthening of sentences for sex offenders, and increased surveillance upon release, including registration and various public notification requirements. Despite its potential utility as a bodily truthing technology in criminal trials, either in favor of prosecution or to absolve the accused, penile plethysmography data have been deemed inadmissible in US criminal courts. The North Carolina Court of Appeals in 1995 and the Fourth Circuit Court of Appeals, in a different case that same year, both deemed the plethysmograph inadmissible under the Daubert standard due to its lack of scientific validity.

In response, researchers have developed a standard for low measures of arousal in an attempt to meet the ‘beyond a reasonable doubt’ test. In effect, researchers sought to materialize criminality as an additional layer of deviant sexual subjectivity. Standardization in this area has also been driven by an acknowledgment that arousal levels in sex offender tests will likely continue to lessen, as increasing regulation requires that erotic stimuli become ‘sanitized and less explicit’, leading to the substitution of computer-generated imagery with ‘less power to evoke full arousal’ (Howes, 2003: 367). However, attempts to standardize the minimum percent increase in penile circumference needed to prove arousal had to overcome the problem of establishing a baseline full erection size in uncooperative subjects. Richard J. Howes (2003), a Canadian psychologist at Stony Mountain Institution (a federal medium-security facility in Manitoba), observed that 20 PFE has emerged as a de facto standard for a meaningful arousal threshold, but as sizes of men’s penises vary, it is difficult to know what this would translate to in units of millimeters. To solve this problem, Howes performed a study that found the mean increase from flaccid to erect for a sample of penises to be 32.6 mm with a standard deviation of 8.8 mm. Using these data, he concluded that a circumferential change score of 10.6 mm is significant enough to pass the 20 PFE test at a 1 percent confidence level, and he proposed that this standardized threshold would pass the ‘beyond a reasonable doubt’ test of the criminal justice system. However, this 10.6 mm figure has yet to gain wide acceptance, and phallometric test data remain inadmissible in courts.

Excluded from criminal trials, purveyors of phallometry in legal contexts have found other applications within the criminal justice system, not only in correctional treatment, but also in post-conviction hearings and monitoring of released offenders. Beginning in 1990 with Washington State’s ‘Community Protection Act’, several US states passed ‘sexual predator’ statutes, which identified offenders eligible for indefinite detention in a ‘civil commitment’ at the end of their sentences. That is, if a sex offender was deemed a ‘predator’ through a risk assessment, he could be sent to a mental institution and kept confined there potentially for the remainder of his life. As defined, ‘sexually violent predators’ committed a crime of sexual violence, had some kind of mental disorder increasing their risk of re-offense, and had been a stranger to their victim prior to the crime or had cultivated a relationship for the purpose of offending (Lieb et al., 1998). The US Supreme Court upheld the constitutionality of such civil commitments in a 5–4 decision in 1997 (Miller et al., 2005).

In this new legal environment, US states and Canadian provinces needed some way of identifying those most at risk of reoffending. Unlike criminal courts, parole hearings are regulated by legislative statute, allowing state legislatures tremendous leeway to determine the kinds of evidence used by mental health experts to assess risk. For example, Texas allows the practitioner to decide on ‘appropriate assessments and techniques’ (Miller et al., 2005). The phallometric test has qualified under such low standards, with its proponents marshaling evidence of validity drawn from meta-analyses of studies conducted over the years. An increase in recidivism prediction emerging within this legal context began in the 1990s and extends to the present. Researchers using phallometric data in correlation with recidivism rates to determine risk thresholds grappled with the question representative of this assemblage: ‘How can we predict a man’s sexual risk to the community?’

A pivotal meta-analysis cementing phallometry’s place in this context was conducted by R. Karl Hanson and Monique T. Bussière (1998) of the Department of the Solicitor General of Canada in 1998. Yet of all the categories of sex offenders they examined in the meta-analysis, phallometry only predicted recidivism for adult molesters of children, with a correlation measure of r = 0.32, considered strong by research standards. Subsequent studies have supported the conclusion that phallometric recidivism prediction in adults only works with child molesters, leaving open the question of its utility in the assessment of juvenile offenders. Phallometric testing has been widely employed with juveniles in the absence of clear empirical justification (Letourneau and Miner, 2005), and despite the fact that researchers claim that ‘juveniles may be more fluid in their sexual interest and arousal patterns than adult offenders’ (Hunter and Becker, 1994).

Today, phallometric testing also competes with less expensive actuarial methods based on surveys and professional interviews to predict recidivism in sex offenders. Tests such as the Static-99 invoke cognitive factors such as lack of empathy as primary indicators of risk. These instruments bypass a major ethical dilemma facing phallometric testing: the need for sexual images. Yet in tests with violent sex offenders, some actuarial instruments have been developed that incorporate phallometric scores into an overall risk index. For example, the Sex Offender Risk Appraisal Guide (SORAG) includes phallometric test scores as 1 of 14 variables that go into determining a risk score. Similarly, the Sexual Violence Risk-20 (SVR-20) includes phallometric data among its checklist of 20 risk variables. It is noteworthy how SORAG and SVR-20 fold the phallometric test into larger instruments, rendering it invisible to public view. Despite its often-contested status and rejection under Daubert, phallometry continues its performative functions from deep within such actuarial instruments, without risking challenges from external critics.

Legal challenges to phallometric testing on sex offenders help pinpoint some of these performative effects. Critics have characterized the test as a violation of personal integrity, especially when involuntarily required. In the 2006 case US v. Weber, in which the 9th Circuit Court overruled a test requirement in a child pornography case, Justice Berzon (2006) claimed it violated privacy: ‘Plethysmograph testing not only encompasses a physical intrusion but a mental one, involving not only a measure of the subject’s genitalia but a probing of his innermost thoughts as well’. In a concurring opinion, Justice Noonan (2006) went further to declare the ‘Orwellian’ test to be ‘always a violation of the personal dignity …. The procedure violates a prisoner’s bodily integrity by affecting his genitals’.

Within this third assemblage, the technosexual script built into phallometric testing casts offenders meeting low threshold levels of deviant arousal as too risky for release. This occurs within a particular niche of the criminal justice system, where a lower standard of scientific validity has been set for the use of the technology. The technique is much like the polygraph in this regard, as a ‘truth machine’ that has yet to be scientifically demonstrated, but that has performative value in correctional settings nonetheless. The concerns of Justices Berzon and Noonan suggest that phallometric testing may be interpreted as performing a punitive function in addition to the general disciplining of arousal patterns. Yet given the continued panics over the risks of child molesters in many jurisdictions, any arguments that challenge the use of the test in these circumstances would likely face considerable resistance.

Conclusion: the phallometric mosaic and the erotically normal man

Over the course of more than half a century, the technology of phallometric testing has proven remarkably versatile. Table 1 summarizes the key features of the three assemblages we have described. While there are some continuities across the assemblages, the table highlights historical discontinuities. Over time, as the test has done performative work establishing sexual norms and sexual truths, it has been taken up as an adjunct to a wide range of competing psychological theories, and has been adopted and adapted for a diverse set of scientific and social functions. The test’s flexibility has permitted the proliferation of a range of sexual subjectivities characterized variously by innate immutability, conditionability, chemical malleability, and actuarial predictability. Along the way, the test has materialized a shifting array of ‘deviant’ sexualities that emerge as a product of negotiation between technosexual scripts and the unruly desires of research subjects.

OPEN IN VIEWER

Table 1.

Assemblages employing the phallometric test.

Assemblage	Primary calculation technique	Primary model of male sexuality	Central question addressed
Diagnosis	Statistical predominance	Innate, immutable sexual orientation categories	How do we categorize a man based on his sexual desire?
Therapeutic efficacy	Before and after comparison	Malleable sexual arousal and desire, deviant and normal	How can we know if a man’s sexual desire has been transformed?
Prediction	Threshold	Arousal as indicator of recidivism danger or safety	How can we predict a man’s sexual risk to the community?

Table 2 contains many of the categories we have found in our research on the phallometric test, and it shows how the work of diagnosis and differentiation has been a high priority. Phallometric testing has made a significant contribution to what Foucault called the ‘sexual mosaic’, the proliferation of sexualities that characterize modernity. Of course some of these categories have not traveled far beyond the phallometric laboratory. Yet an indicator of phallometry’s power to materialize sexual desire is that claims based on its measurements may sometimes be robust enough to prompt individuals outside the experimental situation to adjust their self-understandings accordingly. Indeed, activist Jim Larsen, chairman of the Bisexual Organizing Project, observed that Rosenthal’s findings affirming the reality of bisexual arousal (described in the introduction) ‘could help bisexuals still struggling to accept themselves’ (Tuller, 2011). Thus, there is a reason to view phallometry as a potentially powerful tool for what Ian Hacking (1986, 1999: 34) has called ‘making up people’, whereby categories of human identity may come to be defined or redefined through technoscientific findings, become popularized and embedded in institutions, and subsequently ‘loop back’ into the beliefs and practices of the categorized in ways that may transform the meanings of those categories.

OPEN IN VIEWER

Table 2.

“Deviant” diagnostic categories materialized through phallometric testing.

Sex offenders	Hypogonadal
Sexual deviants	Homicidal sex offender
Paraphilia	Incest offender
Homosexual	Child molester
Androphile	Nonfamilial child molester
Bisexual	Gynephilic child molester
Hypoactive	Pedophile
Hypersexual	Hebephile
Fetishist (boot, fire (pyrolagnia), underwear, trouser)	Homosexual pedophile
Transvestite	Heterosexual pedophile
Transgender	Courtship disorder
Transsexual	Voyeurism
Fetishistic transsexualism	Exhibitionism
Nuclear transvestism	Obscene telephone calling
Marginal transvestism	Toucheurism
Rapist	Frotteurism
Sadist	Preferential rape
Force-oriented	Nonadmitter
Sado-masochist	Recidivist
Sexually dysfunctional	Homophobic
Impotence: psychogenic, organogenic	Premature ejaculators

In addition to contributing to an array of deviant sexual identities, phallometric practice has also helped to describe a normative male sexuality. Through phallometric research, the ‘erotically normal man’ has emerged with desires arranged in a hierarchy of appropriate arousal: most responsive to adult females, less responsive to female adolescents, and not aroused by men of any age in any capacity whatsoever. To be sure, developments like the abandonment of homosexuality treatment suggest that the category of the ‘erotically normal man’, as constituted through phallometric testing, may have expanded over the years to include normal ‘homosexual’ and even ‘bisexual’ variants. But this expansion of the boundaries of normality remains uncertain. And regardless, many assumptions about the sexuality of a normal man remain unquestioned by, and importantly reinforced by means of, phallometric testing, in ways that remain consistent across the assemblages we describe. What is a normal man essentially like? His sexual reflex is presumptively mechanical, and in the presence of visual or auditory stimuli, he is immediately responsive and prepared for the sexual act, with his state of arousal thereby defining his sexual desire. The explicit model for his mating and sexual behavior is that of other animals, despite the fact that the experimental test situation deploys and requires deeply cultural behaviors, such as the viewing of pornography, that have no analog elsewhere in the animal kingdom.

Yet, as the phallometric test materializes male desire in particular forms of erotic normalcy and deviance, it enacts a series of exclusions from the category ‘man’. As an essentialized category within phallometric testing regimes, ‘men’ must have a traditionally recognizable penis to qualify, yet as research on intersexuality has demonstrated, the delineation of bodies into a sex binary on the basis of externally observed genitals is a culturally derived practice (Fausto-Sterling, 2000). The internal experience of the body as ‘flesh’, as in the case of some transgender men, may be more important for determining sex categories in ways that challenge the idea that one must have a particular anatomical structure that functions in a particular way in order to be ‘male’ (Salamon, 2010). Further exploring the boundaries of male sexuality constituted by this test, future research on phallometry might investigate the contribution to sex binaries and sexuality categories made by researchers as they use their imagination in the selection and design of particular erotic imagery and audio vignettes. Overall, the technosexual script embedded within the phallometric test limits the landscape of the body and mind upon which understandings of male arousal might take shape, promoting a narrow view of the sexual that is unemotionally task driven, and that reinforces a traditional male-embodied sexual subjectivity.

Of course, the ability of the test to materialize normative male arousal and to ‘travel’ across research domains has not been without complication or controversy. A variety of problems have plagued researchers and threatened the presumed capacity of the test to characterize and sharply delineate types of sexual individuals. These include disputes about the comparative advantages of different kinds of phallometers, vexing debates about the problem of faking or suppressing of arousal, and questions about threshold effects, as well as many observed anomalies, such as homosexual arousal among heterosexuals and continuities of response between pedophiles and non-pedophiles. Yet, these concerns about internal and external validity have rarely translated into sustained scrutiny of what it means to extrapolate knowledge about sexual subjectivity from this highly constrained experimental setting into a broader world of culture, meaning, and power. What phallometricians ignore, among other things, is the place of the script within the experimental situation. Rather than assume that ‘arousal is orientation’ – that there is an equivalence between a measurable physiological condition, its inner psychological correlates, and its corresponding social identities – we should recognize that it is sexual scripts that create the feedback loops by which men understand that when they are erect they are aroused, that when they are aroused they experience desire, and vice versa. Moreover, it is technosexual scripts – including the material dimensions of the testing apparatus itself – that orchestrate the production of the particular truths that emerge from phallometric testing.

Rather than imagining that such a test can straightforwardly reveal sexual truths, we instead think it may be useful to attend to the performative effects of the test, as well as the processes by which sexual states and social worlds come to be understood on the model of the test. We also propose greater scrutiny of the real-world consequences of the delegation to experts and technologies of the power to characterize inner sexual truths. As Sedgwick (1990) has observed,

To alienate conclusively, definitionally, from anyone on any theoretical ground the authority to describe and name their own sexual desire is a terribly consequential seizure. In this century, in which sexuality has been made expressive of the essence of both identity and knowledge, it may represent the most intimate violence possible. (p. 26)

Through our conjunction of theoretical tools from sexuality studies and STS, we have argued for the importance of close study of processes of materialization, as well as the technosexual scripts that enable them within distinct assemblages of bodies, tools, and knowledge. We also call for expanded attention to truthing technologies (or truth machines) as a class of devices. Comparisons of such technologies may prove particularly fruitful. For example, it would be productive to examine differences according to the target population – such as men versus women (Fishman, 2011) – as well as differences between those truth machines that measure presumptively objective bodily indicators and those – such as Rorschach tests (Galison, 2004; Hegarty, 2003) – that require the expert interpretation of interior mental states. In addition, we might consider the varying role of volition within different truthing technologies. In the phallometric exam, male arousal is deemed to happen automatically, regardless of the subject’s beliefs about the efficacy or purposes of the test, barring the exceptional circumstances that surface in debates about ‘faking’. By contrast, in the case of the polygraph, the subjective beliefs of the person who submits to it are more readily understood to be an integral part of the procedure, because, as Alder (2007a: 126) has emphasized, the device only works if the subject believes that it works and consequently feels anxious about lying. Finally, whatever the similarities among truth machines, the varieties of truth that they produce may certainly differ in ways that merit investigation. In the case of the lie detector, Alder (2007b) adopted a term of comedian Stephen Colbert’s to suggest that what it generates is closer to ‘truthiness’ (p. 271).

In recognition of the range of truth machines that experts may develop, we close with a final research note: new brain scan methods may ultimately supplant the phallometric test as the epistemic window into male sexuality. In 1999, researchers in France began using positron emission tomography (PET) scans to identify brain areas activated in healthy men as they were aroused by visual stimuli, and they found a statistically significant pattern of activity in specific areas of the brain (Stoleru et al., 1999). More recently, researchers have begun using fMRI (sometimes called the ‘no-lie MRI’) to locate sexual truths in the brain (Wolpe et al., 2005). Yet, as Alder (2007a) has observed about the parallel case of the use of the fMRI as a replacement for the lie detector, studies using fMRI ‘squeeze the “body” from the viscera to the brain – as if the brain’s “feelings” were necessarily more authentic, or less easily faked, than those in the lowly heart or stomach’ (p. 265). It remains to be seen whether fMRI will achieve credibility as a form of bodily truthing in the domain of sexuality.