part0028

In this chapter, I turn our attention to a major form of sexual offending against children, involving children who are related to the perpetrator. This is a surprisingly understudied area, given the numbers of children involved and the potential consequences for the families of perpetrator and victim. In the previous edition, I highlighted incest offending as a puzzle that was not well explained by the models of sexual offending, reviewed in Chapter 4 (of that edition and this edition); my hope was to call attention to the many gaps in the understanding of incest from a clinical and forensic perspective, but this research is only beginning to appear. Indeed, as I mentioned in the Preface of the current volume, incest is my next major line of research. Here, I review what researchers know about incest offending in the hope that a future book would be able to say more about this form of offending.

Legally and socially, incest 1 refers to prohibited sexual contact between certain kinds of relatives (Bixler, 1983; Wolf, 1995). The boundaries of permitted relatedness vary across jurisdiction and time, although the nuclear family is universally or near-universally prohibited. Laws and norms sometimes distinguish between genetic relatives and sociolegal relatives such as a stepfather or stepsibling. In this chapter, I focus on incest involving a close, genetically related child, although I review evidence regarding incest committed against close, sociolegally related children (e.g., stepdaughter, stepson) as a comparison and for context. I mention but do not delve into incest involving adolescents or adults because much less research has focused on this, with fewer clinical and forensic contacts.

In this chapter, I review the evidence on prevalence of incest, comparisons of incest and extrafamilial offenders, the ubiquity of incest prohibitions across time and cultures, a distinction between genetic and sociolegal incest, and evidence regarding the mechanisms by which incest taboos and avoidance have become human universals. Wolf (2014) cogently argued that both cultural and biological effects are needed to understand human incest avoidance. No culture, now or historically, is known in which, for example, fathers could legitimately beget and raise children with their own daughters. I then discuss the leading intermediate explanation—between the distal explanation of inclusive fitness, discussed earlier, and proximate explanations involving disgust or other psychological mechanisms, discussed later—for incest avoidance, originally proposed by Westermarck (1891/1921) and focusing on the effects of growing up in close proximity (propinquity). I consider other explanations as well, particularly maternal–infant association and detection of relatedness through facial resemblance and smell. My main purpose for this chapter is to link theoretically driven research in evolutionary psychology and anthropology, inspired by the challenge of incest to inclusive fitness theory, with applied research on the characteristics of incest offenders and their offenses. Linking these different literatures can lead to interesting ideas to guide future research on incest.

Stoltenborgh, van IJzendoorn, Euser, and Bakermans-Kranenburg (2011) reviewed global estimates of the prevalence of intrafamilial child sexual abuse, finding an overall rate of 13% from 323 self-report studies (compared with 4% from eight studies based on informants, e.g., child protection authorities). Girls were more likely to have been sexually abused than boys, with an average rate of 18% versus 8%. Intrafamilial child sexual abuse accounts for a large proportion of total child sexual abuse cases, especially for younger children, where relatives have much more access to potential victims than for older children who go to school and participate in routine activities outside the home. Comparative studies find that intrafamilial child sexual abuse cases involve younger victims, proportionally more girls, more psychological harm, and more incidents over a longer time span, again reflecting relative access and opportunity (De Jong, Hervada, & Emmett, 1983; Fischer & McDonald, 1998; Magalhães et al., 2009).

In addition, 34% of juvenile victims of sexual offenses who were reported to law enforcement in the United States had a related perpetrator (Snyder, 2000). Unfortunately, these aggregate crime data do not further distinguish offenses by specific relationship. As is true for sexual offenses against children by unrelated adults, girls were more likely to be victimized by a relative than were boys, and the large majority of incest offenders were male.

A substantial literature compares incest (intrafamilial) and extrafamilial offenders against children. In the previous edition of this book, I reviewed a number of major individual studies. Since then, my colleagues and I have conducted a meta-analysis of studies comparing 78 different samples, comprising a total of 6,605 intrafamilial offenders and 10,573 extrafamilial offenders, disseminated in studies from 1978 to 2013 (Seto, Babchishin, Pullman, & McPhail, 2015). Approximately half (46%) of the samples were unpublished. The majority of samples (58) were from the United States or Canada. Most (95%) of the samples created groups on the basis of criminal records, but a few did have information from self-report. Forty percent of the samples did not report whether the incest offenders also had unrelated victims, but when this information was available, most samples had incest offenders with only related victims (87%). Two thirds of the samples (68%) did not break down the incest offenders further according to their relationship with the victim, which is likely to matter a lot in explaining incest. Fathers or full brothers are more closely related to victims than uncles or grandfathers or cousins, and sociolegal relatives such as a stepfather have a familial but not genetic relationship to their incest victims.

We focused on a number of theoretically relevant domains in the meta-analysis, particularly atypical sexual interests (including pedophilia, other paraphilias, and excessive sexual preoccupation) and antisocial tendencies (as reflected in criminal history, juvenile delinquency, substance use, impulsivity, and offense-supportive attitudes and beliefs). As I discuss in the next chapter , both of these domains are important in the prediction of sexual recidivism. As expected, incest offenders scored lower on measures in both these domains. We also found that incest offenders scored lower on emotional congruence with children (which is correlated with pedophilia) and lower on interpersonal deficits, contradicting the hypothesis that incest offenders sexually offend because they are socially inadequate and therefore have fewer adult partner options; incest offenders were much more likely to be married, in fact, and thereby had relatively greater opportunity to offend against related children. This does not rule out the possibility that incest offenders have fewer adult partner options than their nonoffending counterparts or that they do not have fewer adult partners than extrafamilial offenders for other reasons (e.g., they are less physically attractive).

Incest offenders did not differ from extrafamilial offenders on most measures of psychopathology, such as depression or anxiety, but they were more likely to have experienced sexual abuse, family abuse or neglect, and poor parent–child attachment. It is unfortunate that few studies of family dynamics have been conducted, as dysfunctional family relationships and spousal conflict are often cited in the clinical literature as potential causes of incest. It is easy to imagine that individuals growing up with poor parent–child attachment or childhood maltreatment would have more dysfunctional family relationships, both within their families-of-origin and within their current families.

Reflecting these same group differences in atypical sexual interests such as paraphilias or excessive sexual preoccupation (motivations; see Chapter 4 ) and in antisociality (facilitation), incest offenders are less likely to reoffend than other sex offenders against children. Some incest offenders have pedophilia, some are excessively sexually preoccupied, and some are highly antisocial, but these explanations for sexual offending against children are not sufficient to account for the large numbers of incest offenders, especially those who only offend against related children, given incest avoidance and taboo. (Incest offenders who also have unrelated child victims tend to score higher on measures of pedophilia and antisociality; e.g., Porter et al., 2000; Rice & Harris, 2002; Seto, Lalumière, & Kuban, 1999.)

Incest is arguably an even bigger Darwinian puzzle than pedophilia because it reduces one’s inclusive fitness (own plus relatives’ reproduction) to the extent it psychologically or physically harms a genetically related child. 2 For example, incest would reduce inclusive fitness if it caused the child to avoid sexual relationships later in life, harmed his or her reproductive capacity as a result of a sexually transmitted infection or physical injury, or impaired his or her ability to effectively parent their own children. Incest involving insemination of close postpubescent girl victims also carries the risk of inbreeding depression, which increases the risk of physical anomalies and thus morbidity or mortality because of combining deleterious recessive alleles (see Bixler, 1992; Thornhill, 1993). For example, Seemanová (1971) examined Czech women who had children sired by a father or brother or children sired by an unrelated man. Forty-two percent of the children from a father or brother had a serious birth defect or early mortality, compared with a 7% rate for children from a nonrelative. Data from Bittles and Neel’s (1994) survey of inbreeding depression suggested a 4% increase in childhood mortality for children from first-cousin marriages. The evidence for inbreeding depression is even stronger for nonhuman animal species (e.g., Charpentier, Widdig, & Alberts, 2007; Crnokrak & Roff, 1999).

Both ethnographic and historical evidence indicate that incest taboos are ubiquitous, although the type and range of prohibited relationships varies; for example, sometimes the sister of one’s deceased spouse is permitted, sometimes she is not. However, prohibitions are essentially universal for nuclear relationships such as parent and child or full siblings (Maisch, 1972; van den Berghe, 1979; Wolf, 2014). The few exceptions involve special exemptions because of religious or cultural rules, as with royal marriages between brothers and sisters in ancient Egypt, to maintain a family dynasty of “god-like” kings and queens (Bixler, 1982). Indeed, incest taboos were present in the oldest written records, including Persian, Egyptian, Greek, and Roman myths; religious prohibitions regarding incestuous relationships across faiths; and historical records (e.g., Gordon & O’Keefe, 1984).

In the Old Testament, Leviticus 18:6–18:18 begins with a general prohibition of sexual interactions with relatives and then specifically lists prohibited familial relationships from a heterosexual man’s perspective: Sexual relationships with mothers, stepmothers, sisters, stepsisters, aunts, cousins, nieces, and female in-laws (except for a wife’s sister, if the wife had died) were forbidden; curiously, sex with one’s own daughter was not specifically prohibited. 3 The inclusion of nongenetic relationships suggests these prohibitions were also intended in part to prevent family conflict and maintain social order. Incest prohibitions are mentioned in the Torah, with a diminishing list over time, and asymmetric rules for men and women ( https://en.wikipedia.org/wiki/Incest#Other_references ). Incest is specifically forbidden in the Q’uran:

Prohibited to you [for marriage] are your mothers, your daughters, your sisters, your father’s sisters, your mother’s sisters, your brother’s daughters, your sister’s daughters, your [milk] mothers who nursed you, your sisters through nursing, your wives’ mothers, and your step-daughters under your guardianship [born] of your wives unto whom you have gone in. But if you have not gone in unto them, there is no sin upon you. And [also prohibited are] the wives of your sons who are from your [own] loins, and that you take [in marriage] two sisters simultaneously, except for what has already occurred. Indeed, Allah is ever Forgiving and Merciful. (Dukes, 2017)

One of the oldest known laws prohibiting incestuous relations is included in the Code of Hammurabi (c. 1780 BCE), which also represents one of the earliest known legal codes. The Code of Hammurabi consists of 282 laws proclaimed by the eponymous Babylonian king. These laws were inscribed on stone pillars and placed in public areas for reference. Two Hammurabi laws specifically refer to incest (translated by L. W. King): Law 154 states, “If a man be guilty of incest with his daughter, he shall be driven from the place,” and Law 157 states, “If any one be guilty of incest with his mother after his father, both shall be burned.” Incest prohibitions are represented in current laws, including those of Canada and the United States (see the Criminal Code of Canada, and see the National District Attorneys Association review of incest laws across American states in 2010).

Wolf (2014) made an important distinction between incest avoidance , which is shared by humans and other animals, and incest taboos , which are culturally propagated by humans. Westermarck (1891/1921) defined incest avoidance as

Wolf (2014) suggested 12 steps through which incest taboos develop. First, inbreeding causes an increase in death or major defects by 20% to 40% among primary relatives. This then creates selection pressure that affects sexuality such that early association causes sexual indifference. Third, as a result of selection pressure, nuclear family incest is rare. Rare events are startling and seen as related to misfortune; Wolf gave examples of superstitious and religious beliefs that incest can result in drought, famine and other disasters. So, fifth, nuclear family incest is rare and therefore startling. Then a consensus forms to condemn incest because the large majority see it as threatening, and the few who do not see it that way stay quiet to fit in. Seventh, human institutions exist when people agree that something exists and agree to assign it a purpose. Thus, incest taboos are the result of a consensus to condemn incest and to avoid the misfortune it is thought to bring. These taboos have a moral quality because they are products of a general reaction that do not appear to serve selfish interests (concern is about collective fortune, not individual fortune). Tenth, the nuclear family is universal and existed before social creation of larger kinship groups, such as clans. When larger kinship groups were created, they were modeled on the nuclear family (e.g., considering certain peers to be “brothers” and “sisters”) and thus included nuclear family incest taboos as constitutional features. Finally, the scope of the extended kinship incest rules vary because the composition of larger kinship groups modeled on the nuclear family varies. This last point is important because it explains why nuclear family incest avoidance seems to be universal, whereas incest taboos for extended kin vary across cultures.

Before discussing explanations for why incest might occur, an important consideration in thinking about incest is the distinction between genetic and sociolegal relatedness. As I discuss at the end of this section, the grounds for expecting different explanations for these two forms of incest are both theoretical and empirical. From a biological perspective, having sexual contact with a sociolegally related child is not inbreeding. From inclusive fitness theory, one could predict differences between genetic and sociolegal incest offenders for two reasons. First, inclusive fitness theory suggests there has been selection pressure over time for psychological mechanisms to treat genetic kin preferentially in terms of affection, care, and investment of resources and to avoid activities that might harm them (also known as discriminative solicitude ; see Daly & Wilson, 1998). Individuals who did not vary their solicitude toward others as a function of their genetic relatedness would be less likely to pass on their genes to the next generation. Consistent with the notion of discriminative solicitude toward kin, M. Wilson and Daly (1987) reported data showing that genetically related children are much less likely to be physically abused than stepchildren. Looking at official Canadian homicide data, an unambiguous measure of maltreatment that is less vulnerable to reporting biases than physical abuse or neglect, Daly and Wilson (1994) found that stepchildren were 60 times more likely to be killed than genetically related children. Even more germane to the present discussion, having a stepfather greatly increases a girl’s risk of being sexually abused (Finkelhor, Hotaling, Lewis, & Smith, 1990; Gordon & O’Keefe, 1984). In a recent study, Hilton, Harris, and Rice (2015) looked at police reports of men who had been violent toward intimate partners. In the subset of 118 cases where men had opportunity to physically abuse a stepchild or a genetically related child, stepchildren were at approximately twice the risk of violence, even after controlling for antisociality (14% vs 8%).

I emphasize that most stepfathers do not physically or sexually abuse their stepchildren; in fact, most stepfathers show a great deal of solicitude toward their new children. But stepfathers do pose a relatively greater risk than do genetic fathers. In humans and many other species, discriminative solicitude is a function of genetic relatedness, which can be estimated through different mechanisms, including familiarity (Rendall, 2004), early propinquity (Westermarck, 1891/1921), or maternal–infant association (Lieberman, Tooby, & Cosmides, 2007). Sociolegal relatives do not have the same early propinquity, because most remarriages do not occur when the child in question is a brand-new infant, and the sociolegal relatives know they are not genetically related to the child. In contrast, genetic relatives usually have early propinquity and believe they are genetically related to the child (although they may be wrong, as discussed later in this chapter). Most studies of incest offenders have not made this distinction between genetic and sociolegal relatives; two thirds of the samples in Seto et al. (2015) did not distinguish incest offenders according to the nature of their relationship to the victim.

Perhaps sociolegal relatives are significantly more antisocial and more likely to be pedophilic than genetic relatives. This would suggest that sociolegal offenders can still be accounted for by the motivation–facilitation model or by other contemporary theories of sexual offending against children. The available evidence is modestly supportive, as sociolegal incest offenders do score higher on antisocial tendencies than genetic incest offenders, but the effect is mostly small in size (Pullman et al., 2017). The motivation–facilitation model can explain sexual offending by some incest offenders, but I do not believe it is sufficient to explain most cases of biological incest if incest avoidance is intact and incest taboos are strong. Even if at-risk biological fathers were more likely to have atypical sexual interests and were more antisocial than the general population, they would be expected to be unlikely to act on their motivations—even with ample opportunity—because of discriminative solicitude.

The Pullman et al. (2017) meta-analysis focused on 27 samples disseminated between 1984 and 2012, including some samples examined in the Seto et al. (2015) meta-analysis. Here, a total of 4,192 genetic incest offenders were compared with a total of 2,322 sociolegal incest offenders. The majority of studies (63%) defined genetic incest as genetic fathers, but in the remaining studies genetic incest offenders included other relationships as well, such as uncles or grandfathers. Most (85%) of the sociolegal incest offenders were in a parental relationship with the victim. Sociolegal and genetic incest offenders did not differ on measures of atypical sexual interests, except in the domain of sexual self-regulation problems; sociolegal incest offenders were higher in this domain. Genetic incest offenders were higher on measures of psychopathology, but this might reflect a consequence rather than potential cause of their sexual offending (e.g., the emotional impact of being identified as someone who had sexually offended against a genetic relative). Consistent with the idea that degree of genetic relatedness may matter, we found that the proportion of genetic incest offenders who were genetically related fathers affected the magnitude of effects in the few domains with sufficiently large samples for moderator analysis. Curiously, genetic incest offenders had fewer prior sexual offenses when the group was composed only of fathers, whereas the genetic incest offenders had more prior sexual offenses when other genetic relatives such as uncles and grandfathers were included. This suggests that genetic fathers are the least likely to have prior sexual offenses, other genetic relatives are the most likely, and extrafamilial offenders are in between.

Some individual studies have found that genetic fathers differed from other genetic relatives in their pedophilic sexual arousal patterns. This comparison was not possible in the Pullman et al. (2017) meta-analysis, so these studies are described next. In Seto et al. (1999), we distinguished between genetic fathers, other family relatives (e.g., grandfathers, uncles, and brothers), and stepfathers, to test the hypothesis that pedophilia can overwhelm incest avoidance. Contrary to our prediction, the 70 genetic fathers responded less to children than the 73 other family relatives but did not significantly differ from the 87 stepfathers. Other family relatives and stepfathers did not differ from 254 men who offended against unrelated children. These results have been replicated (Blanchard et al., 2006; Greenberg, Firestone, Nunes, Bradford, & Curry, 2005; Rice & Harris, 2002).

Westermarck (1891/1921) suggested that individuals who were raised together as children would not be sexually attracted to each other in later life. This view assumes that kin who grow up together would not be sexually attracted to each other and differs from Freudian (Freud, 1905/2000) or sociological views (e.g., Levi-Strauss, 1969) that assume siblings could become sexually attracted to each other if incest taboos did not exist.

A propinquity-based mechanism would have been effective in preventing incest in ancestral environments because siblings were almost always raised together. It could continue to be effective today in many cases, even with more blended families and varied cohabitation arrangements. If Westermarck’s (1891/1921) theory is correct, then incest would be more likely to occur when early propinquity is interrupted. Two related children who were not raised together could become sexually interested in each other, and two unrelated children who were raised together when very young would develop sexual indifference toward each other.

Several lines of evidence support Westermarck’s (1891/1921) theory about sibling incest avoidance, particularly anthropological studies of childrearing practices that have served as natural quasi-experiments regarding the impact of early co-rearing. Wolf (2014) has provided a recent review, focusing on three sources: (a) bint ‘amm marriages between cousins who are raised together because their fathers (who are brothers) share a household—this was favored in Arab countries, such as Morocco; (b) unrelated children raised together on communes in Israel; and (c) unrelated children in hsiao hun (minor marriages) in southern China and Taiwan.

For the first line (marriages between cousins in shared households), McCabe (1983) found a negative effect on marriage and reproduction for arranged marriages between cousins in a southern Lebanese village. In this society, brothers tended to live closely together, and their children were raised like siblings, sharing meals, sleeping, playing, and attending school together. Marriages between first cousins were relatively common in Lebanon at that time, but marriages between cousins who were raised together were significantly more likely to end in divorce—and produced significantly fewer children—than marriages between cousins who were raised apart, or marriages between unrelated individuals.

For the second line, Shepher (1983) studied children who grew up together in small groups on Israeli communes ( kibbutzim ). All the children were within 1 or 2 years of age of each other, and their daily activities included eating, washing, playing, and sleeping together. Although marriages within the communal group were encouraged by their parents, Shepher found that only 14 of 2,769 marriages recorded across 211 kibbutzim involved couples who were raised in the same childrearing group. Moreover, of these 14 couples who were raised together, nine couples did not live together during their first 6 years of life, and the remaining five couples were together for 2 or fewer years of their first 6 years of life, suggesting a critical window—before the age of 6—when propinquity triggers incest avoidance.

For the third line, Wolf (1995) studied the major and minor forms of arranged marriage in Taiwan. In the major form, the bride lived with the husband’s family after the marriage. In the minor form, a sim pua (little bride) was adopted into the family, usually at a very young age, and raised until she became a bride for one of the family’s sons. Wolf found that minor form marriages had greater rates of extramarital affairs and divorce, and fewer children, than major form marriages. No differences in these outcomes were found between the major and minor forms if the bride-to-be was adopted after the age of 3, further supporting the idea of a critical period during which a propinquity-based sexual indifference develops, but setting it at a younger age: 3, rather than 6.

Fessler (2007) noted two other quasi-experiments that are relevant to the Westermarck (1891/1921) hypothesis but that are less often discussed in the history of incest research. The first involves cross-cousin marriages (mother’s brother’s daughter with father’s sister’s son) among the Karo Batak of Sumatra, where this was the preferred arrangement into the 1980s, but according to Singarimbun (1975, cited in Fessler, 2007), only occurred in 4% of marriages. Cross-cousins ( impal ) were highly likely to grow up together because Karo villages were small and dense, so the children had many interactions; many marriages were locally endogamous; and impal were viewed as being like siblings. It is not surprising then that impal marriages, although culturally preferred, were subjectively experienced as incestuous and thus avoided. The second example Fessler described involved the Oneida commune, founded in 1848 and lasting approximately 30 years. Children were seen as being raised by the collective, rather than by their genetic parents. Unlike the kibbutzim, marriages did take place between children who were raised communally. However, Fessler reviewed genealogical records and first-person accounts and suggested that, because of age- and gender-segregation practices, those commune members who did marry each other did not have extensive early propinquity.

This research supportive of the Westermarck (1891/1921) theory has been criticized, most notably by Leiber (2006). The criticisms include anecdotal evidence that Shepher’s (1983) study of the kibbutzim did not take no-privacy and no-sex-before-marriage rules into consideration, as well as an argument that the Taiwanese marriage evidence did not fully reflect the lower status and thus greater strain for the minor marriage form. Leiber (2006) suggested that inbreeding would have been unavoidable in small, isolated demes (local populations with a distinct gene pool) that would be representative of the ancestral environments, with gradations, for example, breeding with a cousin would be preferred over breeding with a sibling. Moreover, Shor and Simchai (2009) conducted interviews with 60 adults who grew up on kibbutzim. Few of these respondents reported sexual aversion to their peers; instead the common response was one of indifference, although some did report feelings of attraction to kibbutz peers. Shor and Simchai suggested that social factors, such as norms about sexual behavior and a concern about group cohesion, might have been more important than early propinquity. Wolf (2014) cogently responded to these criticisms and discussed additional studies that did not share these limitations; interested readers should see his book.

The anthropological studies that I have described were of nonsiblings who were raised together and therefore might have experienced a Westermarck (1891/1921) effect. Some surveys of students have asked about their siblings, including two studies by Bevc and Silverman (1993, 2000) and a series of more recent studies by Lieberman and her colleagues (e.g., Lieberman, Tooby, & Cosmides, 2003). Bevc and Silverman (1993) examined the impact of early separation—and therefore disruption in early propinquity—on sexual behavior involving siblings, as reported by university students. Respondents who had attempted or actually had sexual intercourse with a sibling were significantly more likely to have been separated for more than a year during their first 6 years of life, compared with individuals who did not attempt or have intercourse with a sibling. This effect was only found for penile—vaginal intercourse, which is reproductively relevant because it can result in pregnancy; the effect was not found when comparing individuals who had engaged in nonreproductive behavior, such as kissing or fondling with those who reported no such behavior. This finding was replicated by Bevc and Silverman (2000) in a mixed sample of students and community volunteers. Contrary to their predictions, individuals who had sexual contact with a sibling were more, rather than less, likely to have seen their sibling nude and had more physical contact with that sibling as a child. This suggests it might be propinquity in and of itself—rather than how that time together is spent—that is associated with incest avoidance.

Lieberman and her colleagues’ work produced evidence supporting the idea that it is propinquity rather than knowledge of one’s genetic relatedness or similarity in appearance (or other cues of relatedness) that is the critical element in the development of incest avoidance. Lieberman et al. (2003) examined judgments of moral wrongfulness as a measure of incest avoidance, arguing that such judgments are less susceptible to socially desirable responding than directly asking questions about sexual contacts or sexual attraction to siblings. They found that judgments of moral wrongfulness for sibling incest were positively related to whether the respondent lived with an opposite-sex sibling as a child. Judgments of wrongfulness were not significantly correlated with cohabitation with a same-sex sibling.

In another study by the same research group, the number of years a male respondent lived with a sister before the age of 11 was positively correlated with his ratings of disgust at imagining different sexual acts with a sister (Lieberman et al., 2007). Male respondents who did not live with a sister in their first 10 years of life were less disgusted at the idea of sex with a sister compared with those who resided with a sister for any amount of time before age 11. Lieberman et al. (2007) pursued an alternative explanation, besides propinquity, for these findings, as described below.

Lieberman et al. (2007) argued that siblings have two main sources of information about relatedness: childhood coresidence duration (i.e., propinquity and an indicator of shared parental investment) and mother–infant association. Seeing an infant being nursed and cared for by one’s mother should be a robust and important cue of genetic relatedness; this cue, however, is available only to older siblings. Younger siblings must rely on the less robust cue of coresidence duration. These different cues are not additive; Lieberman et al. found that coresidence duration was not significantly correlated with ratings of disgust about incestuous sex when mother–infant association was present. For younger siblings, however, coresidence duration is reliably correlated with their disgust ratings. Lieberman and Smith (2012) suggested that observing mother–infant association should lead to intense sexual aversion, whereas coresidence duration would result in duration-dependent aversion, with a dosage effect of the amount of time spent growing up together.

Consistent with the idea that mother–infant association is critical, Luo (2011) examined mother–infant association and early coresidence among Chinese participants who were presented vignettes depicting an adult brother and sister pair engaging in consensual sex, followed by questions about sibling incest aversion. Luo tested the idea that the critical window in early coresidence of the first 3 years of life was more influential when mother–infant association was absent. Luo found no significant effect of coresidence (based on age difference of respondent and sibling) on either male or female respondents who witnessed mother–infant association of a younger sibling. No effect was found for female participants who did not witness mother–infant association with a younger brother. Luo did find an effect, however, for male participants who did not witness mother–infant association, suggesting an interaction between gender and incest aversion cues: “To put it more plainly, it seems that only MPA-absent [no mother–infant association] males whose age difference with their sister is smaller than 3 years had greater sibling incest aversion scores.” (Luo, 2011, p. 292).

A series of studies by Jan Antfolk and his colleagues further elucidated moral disgust reactions to vignettes describing incest. These studies all used a similar design, presenting respondents with hypothetical vignettes and asking them to rate their disgust, sexual arousal, and other reactions to what was described. The degree of relatedness was experimentally manipulated, as was whether the vignette was written regarding third parties or imagining oneself in that situation. This series of studies has produced a number of findings that advance understanding of the psychological architecture underlying incest avoidance and taboo: Antfolk, Lieberman, and Santtila (2012) replicated the finding that strength of moral disgust was positively related to degree of relatedness. Self and related third-party descriptions did not differ from each other, but both elicited more disgust than vignettes describing incest between unrelated third parties. Antfolk, Karlsson, Bäckström, and Santtila (2012) found that women reported more disgust than men on average, consistent with the idea that the costs of inbreeding are heavier for women. Relatedly, Antfolk, Lieberman, Albrecht, and Santtila (2014) found that women in the fertile phase of their ovulatory cycle reported more disgust for vignettes describing themselves versus related or unrelated third parties. Antfolk, Karlsson, et al. (2012) also found that disgust ratings were higher for genetic versus sociolegal relatives, relatives who lived together versus lived apart, and parent–child incest versus sibling incest.

A third source of information about relatedness would be facial and other physical resemblances. For example, using clever experimental designs involving digitally morphed faces, DeBruine (2005) showed that manipulations that increased resemblance to self significantly increased ratings of trustworthiness but decreased ratings of attractiveness for a short-term relationship; no effect was found for ratings of attractiveness for a long-term relationship. DeBruine et al. (2011) showed that having opposite-sex siblings decreased the self-resemblance effect on attractiveness, whereas the effect on prosocial attributions was not.

On the other hand, Fraley and Marks (2010) reported a positive effect of subliminal preexposure to an image of one’s opposite-sex parent or a digitally morphed face resembling self on ratings of attractiveness but not if the participant was aware of the implied genetic relationship. Fraley and Marks’s conclusion is consistent with other evidence of sexual imprinting, where partner selection is affected by their resemblance to one’s opposite-sex parent, and evidence that people seek an optimal degree of unrelatedness—not too similar but not too different—in major histocompatibility complex matching (Bereczkei, Gyuris, Koves, & Bernath, 2002; Marcinkowska & Rantala, 2012). Conscious awareness of relatedness would overshadow any imprinting effect, presumably through disgust.

A point that I do not think is sufficiently emphasized in discussions of Westermarck’s (1891/1921) hypothesis is that it was created as an explanation for incest avoidance (indifference or aversion) between children raised together. Indeed, the quasi-experiments reviewed by Fessler (2007) and Wolf (2014) involve children raised together, and the survey and vignette studies have focused on sibling incest, except for one study that found parent–child incest was rated as more disgusting. The Westermarck explanation, or Lieberman’s explanation involving mother–infant association, might not apply to parent–child incest, where close proximity during a child’s early life might explain the child’s incest avoidance but not incest avoidance by the parent.

Nonetheless, some evidence suggests a Westermarckian effect for parent–child incest. Parker and Parker (1986) found that both fathers or stepfathers who committed incest were less involved in early child care activities than nonoffending men. L. M. Williams and Finkelhor (1995) replicated this finding in a study of Navy fathers in a study that had the advantage of partly controlling for potential differences between incest perpetrators and other fathers in their willingness to be involved in early child care; in their study, the absences of Navy fathers from home during the early life of their daughters were not necessarily voluntary, for example, if they were assigned to go overseas on a tour of duty.

Any analysis of father–daughter incest or brother–sister incest, in particular, needs to consider the role of opportunity. 4 In Seto et al. (1999), we speculated that some nonpedophilic men might still have sexual contacts with their daughters or stepdaughters, especially if the girls are showing secondary sexual development, because they do not have access to adult partners. This might include sexual or relationship dissatisfaction with their current partner (i.e., the mother of the child victims), high levels of relationship conflict, or a lack of status, attractiveness, or resources to attract other adults. Another related possibility is high sex drive or intense mating effort. Teleiophilic men adjust their sexual behavior as a function of their access and success with potential adult partners (e.g., Landolt, Lalumière, & Quinsey, 1995). Although not preferred by most heterosexual men, both prepubescent and pubescent girls elicit some sexual response (Freund, McKnight, Langevin, & Cibiri, 1972; Seto & Lalumière, 2001). Some men who do not have access to preferred adult partners might instead seek partners who are still potentially sexually interesting, but further down what my colleagues and I have described as a sexual preference gradient (Blanchard et al., 2012; Seto et al., 1999). In other words, some heterosexual men who do not have access to women will seek out underage but sexually maturing adolescents, pubescent girls, and prepubescent girls, in that order, and will show no interest in males of any age. I discussed this idea further in Seto (2017b) with regard to a sexual preference gradient for age and hypothesized that this would explain why hebephilia may be more common than pedophilia and why individuals with interests in more than one maturity category would be more likely to be interested in an adjacent category than a more distant category (e.g., one would expect more men to be interested in both sexually mature adults and maturing adolescents than men who are interested in sexually mature adults as well as infants.)

Similarly, some heterosexual males without any access to females will engage in same-sex behavior in places such as all-male boarding schools. Consistent with the idea of a sexual preference gradient from adult to underage adolescent to pubescent to prepubescent child, 11% of the fathers who denied incest in the survey by L. M. Williams and Finkelhor (1995) admitted they had experienced some sexual arousal to their daughters. Given the highly sensitive nature of this question, one can assume that a higher proportion of these fathers experienced this sexual response, with some of them denying it to the researchers. Fathers or brothers living with young girls have more opportunity than men who do not.

A very important avenue for further research on the opportunity explanation of some father–daughter and brother–sister incest is to identify the factors associated with being sexually interested in a daughter or younger sister, whether it is acted on or not, and then the factors associated with action. This opportunity hypothesis about some incest offending could be tested by examining the attractiveness (physical and otherwise) and conventional sexual histories of incest offenders. If it is correct that some men commit incest because they lack other sexual opportunities, one would predict that fathers or brothers who commit incest are less attractive, have a later age of sexual onset, and have fewer adult sexual partners than relevant comparison groups (e.g., other fathers, brothers from the same family who did not commit incest). Analyses would need to control for pedophilia (because pedophilic individuals would also be expected to have less sexual involvement with adults) and for antisociality (because antisociality is positively correlated with number of adult sexual partners and negatively correlated with age of sexual onset). The opportunity explanation might be particularly relevant for stepfathers or stepbrothers, who would not be subject to the same incest avoidance effects as genetically related kin. There could still be incest avoidance, however, even with conscious knowledge of nonrelatedness, because coresidence duration or mother–infant association could unconsciously influence psychological tendencies.

Theories that focus on family dysfunction as a cause of incest might also represent a form of opportunistic offending (e.g., Maddock & Larson, 1995). In these theories, father–daughter incest is more likely to occur when the parental relationship has broken down and the mother is sexually and emotionally unavailable—or perceived to be unavailable—by the father. For example, a father may turn to his daughter to fulfill his sexual and emotional desires because his marital relationship is high in conflict or because his spouse is withdrawn; in essence, his daughter is put into a spousal role. Some evidence is consistent with this view of father–daughter incest. Lang, Langevin, van Santen, Billingsley, and Wright (1990) compared 92 incest offenders (86% were genetic fathers or stepfathers) with 42 nonoffending men and found that the incest offenders reported less communication with their partner, felt more lonely, and were less satisfied with their partner. The incest offenders and nonoffending men did not differ in the length of their current marriages or their number of prior marriages. Lang et al. did not include a comparison group of non–incest offenders, however, and did not further distinguish between genetic and sociolegal fathers in their study. An alternative explanation is that the greater spousal relationship conflict is a result of suspected infidelity, wherein the man is less certain about his paternity of putative offspring (see later). It is also possible that differences were influenced by retrospective biases, where incest offenders recall less satisfaction in their relationships to justify their offending. Prospective research following couples would be needed to elucidate these effects.

Similarly, brother–sister incest is more likely to occur in families that are characterized by parental conflict, lack of parental supervision, or multiple forms of abuse or neglect (e.g., Cyr, Wright, McDuff, & Perron, 2002; Latzman, Viljoen, Scalora, & Ullman, 2011; J. A. Thornton et al., 2008; Worling, 1995). Such families may be more at risk of sibling incest because brothers are more likely to act out, sisters are more vulnerable, and neither child is supervised effectively by parents. Family dysfunction may also play a prominent role in explaining families with multiple incest offenders and multiple child victims, including intergenerational cases where victims of incest grow up and commit incest offenses against their own children (Faller, 1991).

The understanding of human incest avoidance is further complicated by differences on potential fitness costs based on age (including fertility status) and gender. For age, Wolf (2014) noted that many known cases of incest involve older male relatives offending against prepubertal girls; although both psychological and physical harm can be done, inbreeding depression is not a cost because the victim is not yet fertile. Moreover, genetic relatives who perpetrate incest might avoid potentially reproductive behavior. Consistent with this idea, Rice and Harris (2002) found that genetic father incest offenders (36%) were significantly less likely than sex offenders against unrelated children (47%) to vaginally penetrate their victims. This study did not further distinguish incestuous behavior by victim age, however. These results might explain one of the paradoxes of incest offending: Given that incest offenders are unlikely to be pedophilic, one would expect a higher prevalence of postpubertal victims than prepubertal victims, whereas the victimization data indicate incest victims are, on average, prepubertal. It may be the case that the cost of incest is unconsciously appraised as lower for prepubertal children because they are not at risk of reproduction and thus not at risk of the inclusive fitness costs of inbreeding depression.

The direct fitness costs for the victim are potentially higher than the inclusive fitness costs for the perpetrator, suggesting an asymmetry based on both age and gender. Haig (1999) argued that the costs of incest are higher for females than for males because of the sex difference in minimal parental investment (Trivers, 1972). A father who has sex with his daughter does not lose other opportunities to engage in reproductive behavior, whereas his daughter may lose an opportunity to have a healthy child fathered by a nonrelative. A brother is in a similarly asymmetric position with regard to his sister, where the minimal cost to him is lower. Haig’s (1999) thinking about incest leads to testable hypotheses that organize existing evidence and suggest directions for further research. I had hoped in the first edition that this discussion would stimulate further research on the underpinnings of incest but that has not taken place yet; perhaps the second time is the charm?

The consequences of incest are higher for females than males because of obligate sex differences in minimal parental investment; therefore, first fathers, brothers, uncles, and grandfathers should be much more likely to commit incest than their female counterparts. Indeed, incest offenders are more likely to be male, based on both self-report and criminal justice data (Snyder, 2000; Stoltenborgh et al., 2011). Second, girls should exhibit greater aversion to the possibility of incest, resist incest more than boys, and should experience more negative reactions if incestuous activity takes place. Indeed, women are more likely to report negative reactions to childhood sexual abuse, much of which is committed by older relatives (Rind, Tromovitch, & Bauserman, 1998). Third, girls would experience more negative reactions if the incest occurs after they reach puberty and are potentially fertile, compared with girls who are prepubertal, and to experience more negative reactions when the incest includes vaginal penetration and insemination. Rind et al. (1998) did not find that the occurrence of penetration was a significant moderator of childhood sexual abuse outcome in their meta-analysis, but there were too few studies to distinguish between female and male respondents or between offenses committed by related versus unrelated perpetrators.

Drawing from inclusive fitness theory, the incidence of incest should be inversely related to the actual degree of genetic relatedness between the offender and the victim. Thus, men who are not genetically related to a child (e.g., stepfathers or stepbrothers) should be more likely to commit incest than men who are closely related to the child (fathers, brothers); other family members, who have an intermediate degree of genetic relatedness to a child (cousins, uncles, or grandfathers) would fall in between these two groups in the incidence of committing incest. This gradient should be sensitive to even finer distinctions, such that full siblings are less likely to commit incest than half-siblings, who in turn are less likely than stepsiblings.

Paternity uncertainty refers to the biological fact that, whereas mothers can be certain that they are genetically related to the children that they bear, many fathers cannot be 100% certain, because their putative child may have been the result of the mother having sex with another man. Thus, for some males, their putatively genetic daughter or sister is in fact unrelated to them, making their offending more like stepfather or stepsibling incest. Psychologically speaking, a putative father with a high degree of paternity uncertainty is more similar to a stepfather than a genetic father with high paternity confidence. It follows then that paternity uncertainty should be related to the likelihood of incest, such that fathers who have more uncertainty about the paternity of their putative daughter are more likely to engage in sexual activities with her (Haig, 1999).

This prediction could be tested by comparing offenders who sexually offended against a daughter with extrafamilial offenders who are also fathers but offended against an unrelated child on indicators of paternity uncertainty, including conscious doubts about the paternity of their daughter, absences from the home around the time of the girl’s conception, their spouse’s history of infidelity, jealousy about potential infidelity, and perceived resemblance to one’s child in terms of looks and personality. My colleagues and I are currently conducting a study testing these ideas.

Paternity uncertainty would be expected to create asymmetries in the potential costs of incest involving paternally related versus maternally related family members. Incest is predicted to be more likely to occur with paternally related children than with maternally related children. For example, one would predict that a brother would more likely to commit incest with a paternally related half-sister than with a maternally related half-sister, because he is 100% certain he is related to the second girl but less than 100% certain he is related to the first girl. Other predictions based on matrilineal (through the mother) compared with patrilineal (through the father) descent have been supported for studies of investment by aunts and uncles (Gaulin, McBurney, & Brakeman-Wartell, 1997) and by grandparents (Euler & Weitzel, 1996). Studies of this hypothesis would require very large samples, for example, recruited through online surveys promising anonymity or confidentiality. The infrastructure for such surveys was less developed when the first edition of this book was written, but now it is possible to conduct online surveys about very sensitive topics, producing data of comparable quality to in-person surveys (e.g., Bailey, Bernhard, & Hsu, 2016).

Haig’s (1999) hypotheses about incest suggest that paternity uncertainty could play an important role in the understanding of incest committed by (putatively) genetic fathers, brothers, and paternally related half-brothers. For paternity uncertainty to play a role, the risk of nonpaternity needs to have been sufficiently high to affect Darwinian selection among our ancestors. This hypothesis would be moot if heterosexual couples were entirely monogamous and fathers could be certain about their paternity, just as mothers are always certain of their maternity. The reality is that heterosexual couples are not always monogamous, and fathers cannot be certain about their genetic relationship to offspring. National lifetime prevalence data suggest that 10% to 15% of married women have had sex with someone other than their husband while married (Laumann, Gagnon, Michael, & Michaels, 1994; Wiederman, 1997). In a nationally representative sample of 4,884 married women, the 1-year prevalence rate for infidelity was 1% in face-to-face interviews and 6% in computer-assisted self-report (Whisman & Snyder, 2007).

Infidelity does not equal nonpaternity. What, then, is the extent of nonpaternity? That is, what proportion of men who believe they are the genetic father of a child are wrong? Estimates of nonpaternity are noisy, depending on the population being sampled and the reasons for the genetic testing. One would expect higher rates of nonpaternity when testing is conducted because of suspicion or a dispute about paternity (e.g., in cases involving custody or child support claims) than in studies of families who are seen for medical genetic screening with no concern about nonpaternity. On the other hand, in one survey, a majority (88%) of medical geneticists who prescreen for genetic counseling advise mothers that genetic testing would reveal nonpaternity, thereby biasing the estimates that are obtained in screening (Lisker, Carnevale, Villa, Armendares, & Wertz, 1998). Self-selection by mothers who are aware that the genetic testing might reveal nonpaternity would produce underestimates of false paternity.

The literature on genetic evidence of nonpaternity was reviewed by Anderson (2006). He identified 22 nonpaternity rates from studies that analyzed genetic samples collected as part of medical screening (median 2%) and 30 nonpaternity rates from studies of genetic data collected as part of a paternity dispute of some kind, with a much higher median rate of 30%. One would expect higher rates of nonpaternity among incest offenders who are higher in antisociality than the general population, because they have more unstable marital relationships (see Lalumière, Harris, Quinsey, & Rice, 2005). Nonpaternity rates of 30% or more suggest that paternity uncertainty could account for a substantial proportion of putatively genetic incest cases. On the other hand, a meta-analysis suggests nonpaternity rates are declining (Voracek, Haubner, & Fisher, 2008) and historical analyses of data from multiple societies suggest nonpaternity rates are low, with a 1% estimate (Larmuseau, Matthijs, & Wenseleers, 2016). If it is closer to 1% to 2% than it is to 30%, is this enough to generate effects with regard to incest at the population level?

Other psychological adaptations are related to the ancestral (and contemporary) risks of infidelity and nonpaternity. Men who are never sexually jealous and therefore watchful regarding potential infidelity would be at risk of investing in offspring who are not genetically related to them and thus would be less likely to pass on their own genes; such men were less likely to be our ancestors than those who were vigilant about the risk of their mate’s infidelity (for a review, see Buss, 2015). Indeed, evolutionary theorists have suggested that a common tendency to report that children resemble their fathers more than their mothers is an attempt to reassure the fathers of their paternity (Christenfeld & Hill, 1995; Regalski & Gaulin, 1993; but see Brédart & French, 1999). Bressan and Dal Martello (2002) reviewed studies showing that people perform little better than chance at matching parent and child, so the perceived resemblance is not real; the biggest influence on resemblance ratings was the respondent’s belief about the genetic relationship between parent and child. Pagel (1997) suggested there would be selection for babies to be generic-looking, to avoid triggering suspicions about paternity because of an obvious lack of resemblance between baby and putative father.

Other studies have found that perceived resemblance to a child significantly influenced males’ hypothetical decisions about important matters, such as adoption, spending time or money on a child, and paying child support, but not female decisions about these matters (Platek, Burch, Panyavin, Wasserman, & Gallup, 2002; Platek et al., 2003). Burch and Gallup (2000) found that a father’s perceived resemblance to a child was significantly and positively related to the self-reported quality of relationship with that child in a sample of men who had committed intimate partner violence. Perceived resemblance to a child was also inversely related to the severity of injuries experienced by the mother from intimate partner violence. These results suggest that another possible cue of paternity uncertainty to study in future research on incest offenders is the occurrence of violence against the victim’s mother, particularly violence that occurs because of sexual jealousy or the suspicion of infidelity.

Because many incest offenders are very unlikely to sexually reoffend, treatment intended to reduce their risk of sexual recidivism is often unwarranted, because there is little room for improvement—they are already very unlikely to sexually reoffend—and treatment has the potential to inadvertently make them worse (see the discussion of the risk principle in offender intervention in the section on Risk, Need, and Responsivity Principles in Chapter 8 , this volume). However, perpetrator treatment might still be needed to help incest victims and to assist in the reunification of the family, when this is possible and desired by victims and other family members. Unlike sex offenders against unrelated children, who usually do not have contact with their victim again, some incest offenders will see their victims again because they return to the family or have some access because of shared custody or visitation. Permanent removal of fathers or stepfathers from the family can have a large emotional and financial impact on the rest of the family and may in fact prevent disclosures by children who want the sexual offending to stop but do not want to break up the family.

Treatment providers may therefore focus on family reunification for incest offenders, a process that typically involves family therapy, therapy and support for the incest victim(s), and monitoring of the offender in terms of potentially risky behaviors and situations (e.g., consuming alcohol when the incest offenses often occurred after the offender became intoxicated). Factors addressed in therapy might include the offender’s acceptance of responsibility and efforts to apologize and make amends, the ability of the nonoffending parent to support the children and monitor the offender, and repairing the family relationships.

Family reunification is complicated by the fact that incest within nuclear families can be associated with marital conflict; physical violence committed by the incest offender against the nonoffending parent; physical abuse or neglect of the victim or other children, and in some cases, the denial of the nonoffending parent, who either does not believe the child disclosing sexual abuse or blames the child for the subsequent upheaval in the family (Deblinger, Hathaway, Lippmann, & Steer, 1993; Paveza, 1988). Unfortunately, systematic evaluations of reunification efforts in families in which incest has occurred have never been reported. We do not know how many families attempt reunification or what impact reunification efforts have on the children, family, or offender. Some relevant literature focuses on family reunification in general child welfare cases (e.g., Fraser, Walton, Lewis, Pecora, & Walton, 1996; Terling, 1999).

Incest avoidance and incest taboos are related but distinct phenomena (Wolf, 2014). Westermarck’s (1891/1921) proposed propinquity mechanism provides an elegant explanation for incest avoidance and has good empirical support from nonhuman animal research and quasi-experiments in humans. Incest taboos are ubiquitous and appear in both historical and cross-cultural accounts of laws and customs regarding permissible sexual relationships. Overall, however, the etiological understanding of incest lags the knowledge being accumulated regarding the origins of sexual offending against children in general.

The gap remains large between (a) theoretical research in evolutionary psychology and anthropology on incest avoidance and taboos and (b) the more applied research carried out by researchers concerning incest offenders and their offenses. This is very unfortunate, because many novel hypotheses could be tested to help improve understanding of not only why incest occurs but why some children may be more vulnerable than others—these hypotheses include those that consider inclusive fitness, parental investment, and paternity uncertainty. Such research could have important implications for prevention programs, particularly in the identification of at-risk families and children.

From this chapter, I hope I have convinced you that the following factors can make major contributions to researchers’ understanding of incest: (a) degree of genetic relatedness, with the likelihood and nature of incest offending expected to vary by relatedness, distinguishing further from matrilineal versus patrilineal relatives; (b) mother–infant association, as another important cue of relatedness in addition to the early proximity identified by Westermarck (1891/1921); (c) perpetrator and victim gender, with perpetrators being much more likely to be male and victims being more likely to be female; (d) type of sexual behavior, with larger effects obtained for incidents involving penile–vaginal penetration than nonreproductive sexual activities, and with differences in the likelihood of penile–vaginal penetration based on genetic relatedness and victim age; (e) family dysfunction, with incest more likely in families characterized by instability and by parental conflict; (f) the attractiveness of the male perpetrator, based on physical and other characteristics, such as status and wealth, which would influence his access to unrelated sexual partners and likelihood of opportunistic offending against a younger relative; and (g) paternity uncertainty, influenced by the characteristics and behavior of the father, mother, and incest victim and connected to what is known about sexual jealousy and intimate partner violence.

Recent evidence supports some of these factors, and a full accounting may eventually find that an important proportion of incest offending can be accounted for by the motivation–facilitation model described in Chapter 4 (this volume; the model in which atypical sexual interests and antisociality can overcome incest avoidance and taboo); by failures of incest avoidance as a result of low or no early propinquity or maternal-infant association; paternity uncertainty; and by opportunistic offending. However, much more work is needed to determine how much incest offending can be explained by these different perspectives, what other explanations may be important, and what researchers and clinicians can do to prevent incest.