Hill et al. (2007)
deduce from various findings of increased psychiatric problems in the parents of children with GIV that it may be the parents rather than their children who are psychiatrically disturbed (see also Lev 
for a similar argument), and that such parents may bias their reports of their GIV children's psychiatric problems. These authors fail to take into consideration, however, that familiality and heritability are common findings in psychiatric conditions. Moreover, even if psychiatric disorders are caused by environmental circumstances, as it is the case, for instance, in posttraumatic stress disorder (PTSD), it is not meaningful to deny their existence in the individual. In addition, some studies fail to identify parental problems as a major risk factor for GID development (e.g., Wallien, van den Langenberg, Knol, Kreukels, & Cohen-Kettenis, in press). In regard to reporting bias, the multimethod and multiinformant approaches used in the systematic clinical evaluations of children with GIVs (e.g., Meyer-Bahlburg, 2002
; Zucker & Bradley, 1995
) show that the reports by parents about their children usually are largely in line with the findings from children's self reports and clinicians' observational evaluation of these children; they can, therefore, not be attributed to mere parental bias against their children as claimed by Hill et al. (2007)
Pathologic Condition Versus Natural Variation
In general, the demarcation of behaviors that are “pathologic” from those that are not poses a challenge to the clinician. Stedman's Medical Dictionary (1995)
defines “pathology” as the “medical science, and specialty practice, concerned with all aspects of disease, but with special reference to the essential nature, causes, and development of abnormal conditions, as well as the structural and functional changes that result from the disease processes. [In Greek, pathos = feeling, suffering disease; in Greek, logos = study, treatise.]” In line with this definition, “psychopathology” is the “science concerned with the pathology of the mind and behavior.” Yet, the Stedman definition of pathology obviously presupposes a consensus on the definition of “disease,” and does not offer a systematic approach to demarcate psychopathologic from non-psychopathologic for the continua of behavioral domains which at the extreme end are categorized as psychiatric dysfunctions and/or mental disorders.
In regard to GIVs, part of the categorization problem is due to the fact that we do not have a well established detailed theory--let alone a neuroanatomic/neurophysiologic model--of normal gender identity development that gives us clear guidance in distinguishing non-pathologic from pathologic. Apart from the gender assignment at birth on the basis of the appearance of the external genitalia, the developmental psychological processes leading to sex-dimorphic behavior, gender schemas, and a gendered self-concept--presumably in dependence on central-nervous system organization as well as on various mechanisms of social learning--appear to be highly intercorrelated (Ruble, Martin, & Berenbaum, 2006
). Under these circumstances, causal directions among psychological processes are notoriously difficult to establish, which makes the delineation of pathologic processes problematic. Moreover, probably due to differences in study populations, there is little overlap and communication between theorists of normal gender development (e.g., Egan & Perry, 2001
; Ruble et al., 2006
) and theorists of GIV. Biologically oriented investigators of GIVs tend to draw on models of behavior development--starting with the effects of genes and hormones early in development on the sexual differentiation of the brain--from nonprimate mammals, especially rodents, but vary to what extent they extrapolate beyond human gender-related behavior to human gender identity, for which there is no clear animal model at present.
As the study of gender development in persons with somatic DSDs often serves as a bridge from biological animal research to human investigation, I want to highlight some of the dilemmas involved in categorizing GIV as psychopathology with two examples of gender change from the DSD area. Consider the case of a 46,XY child with a severe penile abnormality due to non-hormonal causes (e.g., penile agenesis, cloacal exstrophy of the bladder, or traumatic loss of the penis in infancy), who has been raised female, shows strongly masculinized behavior in childhood, and initiates a change to the male gender later, as enacted by about a quarter of individuals with such conditions who have been documented to date (Meyer-Bahlburg, 2005
). What is considered “pathologic” in a case of penile agenesis, for example? Certainly the non-development of a genital tubercle and, later, the absence of a penis in an otherwise normal-male developed individual (normal testes, normal androgen production, normal male reproductive structures, except for the lack of a penis and the location of the urethral meatus in the rectum), is seen as pathologic. After early castration and female assignment, such a child shows markedly masculinized (“tomboyish”) behavior in comparison to non-DSD girls, but nowadays tomboyish behavior is not seen as pathologic. Bisexuality or gynecophilia emerging in adolescence or adulthood is also not categorized as pathologic, given that homosexuality has been removed from the DSM. When such a 46,XY person later initiates gender change to male, it is seen as a “correction” in medical circles outside of psychiatry, and similarly by lay persons. Yet, in DSM-IV-TR it is labeled a mental disorder, namely GID NOS. On the other hand, if one considers the gender change to male a “correction” instead of a “mental disorder,”what about those with the same condition who develop an identity as female and do not change their gender: Do they now have to be considered as having a mental disorder?
Another DSD example is provided by 46,XX newborns with severe degrees of genital masculinization due to classical congenital adrenal hyperplasia (CAH) who, when raised female, typically show markedly masculinized behavior later; some even initiate gender change to male. In these patients, many medical features would be considered “pathologic”: the deletion or mutation of the 21-hydroxylase (21-OH) gene, the resulting deficiency of the 21-OH enzyme, of cortisol and aldosterone, of negative feedback from circulating cortisol on ACTH release leading to continuous stimulation of the adrenal, to adrenal hyperplasia, and to overproduction of adrenal androgens. However, there is less consensus regarding masculinization of the genitalia. For instance, physicians typically categorize a markedly enlarged clitoris as “pathologic,” but many social constructionists emphasize the “natural” variation of peno-clitoral size along a continuum, which they contrast with the “socially constructed” binary system of gender (e.g., Fausto-Sterling, 2000
; Kessler, 1990
). Finally, the well established increased rate of masculinized gender behavior (Meyer-Bahlburg, Dolezal, Baker, Ehrhardt, & New, 2006
) and sexual orientation (Meyer-Bahlburg, Dolezal, Baker, & New, 2008
) as well as the occasional patient-initiated gender change to male (Dessens et al., 2005
; Meyer-Bahlburg et al., 1996
) in this DSD condition poses questions of categorization that are very similar to those in 46,XY penile agenesis. Thus, in such cases, gender-atypical behaviors that clearly “result from the disease processes” (Stedman's Medical Dictionary, 1995
) are not necessarily categorized as psychopathologic.
Some clinicians might question why GIV in persons with DSD needs to be considered at all in the context of the discussion of GID and the DSM. There are several reasons. (1) Despite misgivings among some participants in the respective work groups at that time, DSM-IV and DSM-IV-TR included GIV in DSD as GID NOS, because of similarities in presentation to non-DSD GIVs. (2) In addition one has to note that also in persons with DSD, gender identity development is a psychological process, not just an outcome determined by biological factors. For instance, we have shown that 46,XX girls with classical CAH show a dose-response relationship (on the group level) of prenatal androgens to gender behavior, but not to (dimensionally assessed) gender identity (Meyer-Bahlburg et al., 2004
), i.e., gender identity is less closely related to biological factors than gender-related behavior. In fact, gender identity can accommodate wide variations in gender-related behavior (Meyer-Bahlburg et al., 2006
). (3) A number of recent findings suggest that GID may perhaps be understood in part as a CNS-limited form of DSD or intersexuality, without involvement of the reproductive tract. This is the implication of the demonstration in male-to-female and female-to-male transsexuals of a sex reversal in terms of volume and cell number of sex-dimorphic brain nuclei such as the central portion of the bed nucleus of the stria terminalis (BNSTc; Kruijver et al., 2000
; Zhou, Hofman, Gooren, & Swaab, 1995
), the interstitial nuclei 3 and 4 of the anterior hypothalamus (INAH3 and INAH4; Garcia-Falgueras & Swaab, 2008
), and the gray matter in the right (and possibly the left) putamen (Luders et al., 2009
), although such findings are characterized by large within-group variability and cross-group overlap. Recently, the neuroanatomic findings have been complemented by the demonstration of gender-atypical brain activation patterns in processing steroid based odors and erotic stimuli (Berglund, Lindström, Hejne-Helmy, & Savic, 2008
; Gizewski et al., 2009
It is also conceivable that there may be genetically based systemic sex-hormone abnormalities that do not cause abnormalities of the reproductive anatomy, but nevertheless influence brain and behavior. This is implied by genetic abnormalities (albeit with very modest effect sizes) in terms of increased trinucleotide (CAG) repeats found in the androgen-receptor (AR) gene of male-to-female transsexuals, which are generally associated with impairment of androgen utilization (Hare et al., 2009
); of an increased prevalence of CYP17 gene polymorphisms in female-to-male transsexuals associated with higher serum and tissue concentrations of both testosterone and estradiol (Bentz et al., 2008
), which may explain some hormonal findings reported earlier (Bosinski et al., 1997
); and of significant combined partial effects of three polymorphisms in male-to-female transsexualism (CAG repeats in the AR gene, tetra nucleotide repeats in the aromatase gene, and CA repeats in the estrogen receptor β gene; Henningsson et al., 2005
; for new negative findings, see Ujike et al., 2009
). Such genetic mechanisms may underlie the demonstration of substantial heritability of gender-related behavior in general and GID in particular in child and adolescent twin samples (Coolidge, Thede, & Young, 2002
; Iervolino, Hines, Golombok, Rust, & Plomin, 2005
; Knafo, Iervolino, & Plomin, 2005
; van Beijsterveldt, Hudziak, & Boomsma, 2006
). The absence of genital abnormalities in such cases suggests dose specificity or tissue specificity of the androgen receptor deficit, or timing effects, the latter because it has long been demonstrated in animal research that the sexual differentiation of the brain during a hormone-sensitive prenatal or perinatal period can be modified independently of the (earlier) sexual differentiation of the reproductive tract (Goy, Bercovitch, & McBrair, 1988
). The application of highly sophisticated new techniques for genome-wide profiling of the transcriptomes of peripheral blood mononuclear cells, which led to the demonstration of a discrete set of transcripts directly correlated with XY or XX genotypes independent of male or female genotype of the external genitalia, and another, larger gene set that reflected the degree of external genital masculinization independent of both sex chromosomes and concurrent postnatal sex steroid hormone levels (Holterhus et al., 2009
), appears to open exciting additional possibilities for genetic approaches to GVs.
One feature of animal models of the sexual differentiation of brain and behavior that has not yet found sufficient consideration in human research on gender development is the observation in rats that males have the neural circuitry of all aspects of female sexual behavior. That circuitry is usually blocked by perinatal sex-hormonal defeminization, but can be activated by the induction of an atypical sex-hormone milieu in adulthood (de Vries & Södersten, 2009
). Perhaps, related mechanisms are involved in the development of such phenomena as late-onset GID or contribute to the sexual-orientation change observed in many trans persons after onset of cross-gender hormone treatment (Bockting, Brenner, & Coleman, 2009
; Lawrence, 2005
An alternative biological model that assumes faulty hardwiring (possibly for other than hormonal reasons) of the gender-specific cortical representation of the genitals as the basis of anatomic genital dysphoria in transsexuals was recently proposed by Ramachandran and McGeoch (2007)
, but fails to explain the broad-band gender-behavior changes seen in most individuals with early-onset GIV and requires more empirical support even for its core assumptions.
Let us assume that in the future one or several of the human biological findings above will be shown to be replicable in GIV samples by independent, reliable laboratories. Would the gender-atypical behavior (including sexual orientation) in such cases then be considered “pathologic”? And what about those who become gender-dysphoric and initiate gender change? Would public opinion and government officials not likely refer to a “correction of wrong gender assignment,” in parallel to the analogous cases with somatic intersexuality? Again, on the other hand, if patient-initiated gender change in such GIV cases is a “correction,” a question arises about the psychiatric status of those cases who develop a lasting identification with the assigned gender. The examples above show that there is no clear scientific solution based on etiology alone to the psychiatric categorization of behavior and identity outcomes of pathological medical conditions.
On the basis of some of the biological studies referred to above, some organizations and quite a few transgender activists have embraced the notion of GIV as firmly biologically grounded (e.g., Gender Identity Research and Education Society [GIRES], 2006
; Winter, 2009
). This is clearly premature for several reasons. (1) Leading investigators have criticized several of these biological studies on methodological grounds (e.g., Herbert, 2008
). (2) Each of these biological findings is in need of replication by independent, high-quality laboratories. (3) The hypothesis of CNS-limited “intersexuality” as the basis of GID development has most plausibility for the early-onset form of GID with its well established cross-gender shift in many gender-related behaviors, including later sexual orientation. It has little plausibility for the explanation of the late-onset form of GID, which in many cases seems to develop in the absence of a history of markedly gender-atypical behavior of childhood.
Apart from the biological theories of GIV, there are a number of other explanatory models. Updating earlier psychoanalytic interpretations, several clinician-researchers have hypothesized from the perspective of developmental psychopathology that the development of GID is based on processes involving temperamental vulnerabilities and particular patterns of parent-child interaction (Coates, 1990
; Di Ceglie, 1998
; Zucker & Bradley, 1995
). Others perceive gender transitions in at least some (non-intersex) individuals “as a solution--a way out of some form of social, psychological, or developmental paralysis” that is initially unrelated to issues of gender (Levine & Solomon, 2009
). In yet another clinically rooted theoretical approach, the root of late-onset male-to-female transsexualism is seen in autogynephilia as a form of transvestitic fetishism (Blanchard, 1989
; Lawrence, 2007
), also conceptualized as an “erotic target location error” (Freund & Blanchard, 1993
; Lawrence, 2009
). This theory has led to particularly acerbic controversies, and its specificity has recently been questioned by new empirical data (Moser, 2009
; Nuttbrock et al., 2009a
Self-system theory has led to different psychological models. For instance, Doorn, Poortinga, and Verschoor (1994)
, in modifying the theory of Docter (1988)
, postulated the existence of two gender identity subsystems of the self, one feminine, the other masculine, which may differ in relative strength and may be conditionally or unconditionally expressed. Bockting (2009a)
introduced stigma as an additional factor that affects the relative strength of subsystem expression. In an extensive study of female-to-male transsexualism, Devor (1997)
developed a detailed process model of transsexual development, which attributes an important etiologic function to certain family dynamics in the context of a bi-gendered patriarchal society. In parallel to the observation I made earlier in conjunction with the biological intersex model, Devor (1997, p. 67)
argued that the psychological formation of transsexualism in reaction to unhealthy family dynamics does not necessarily imply pathology of the resulting identity.
One also needs to take into consideration that there are types of identity formation other than those related to gender, such as in people who identify with amputees to the extent that they request limb amputation (“Body Integrity Identity Disorder”; First, 2005
; Lawrence, 2006
), for which a specific biological basis representing a putative natural variation is hard to imagine. Perhaps it can be better understood as a form of identity development that Wilkinson-Ryan and Westen (2000)
have described as “role absorption” in patients with borderline personality disorder. A similar new case report documents the co-occurrence of a desire for a non-mutilative disability with transsexualism (Kolla & Zucker, 2009
). In any case, it is difficult to justify the term “natural” variation for a condition that compels the respective individual to severely alter a healthy body by gonadectomy with attendant infertility and the replacement of intact primary and secondary sex characteristics with those of the other gender.
At this stage of our knowledge, none of the proposed theories of gender development are sufficiently empirically validated to permit firm conclusions regarding the delineation of psychopathologic from normal processes. In particular, the conceptualization of GIV as a fully biologically based identity that is accidentally embedded in a body of incongruent sex is not easy to ground in empirical evidence (see also Blanchard, 2008
). Also, it seems entirely conceivable that there are more pathways to GIV than one. Perhaps the solution of this issue has to await the application of computational models of normal and dysfunctional brain operations within theoretical neuroscience to the sexual differentiation of brain and behavior (Thagard, 2008
Impairment and Distress
In the development of DSM-III, “impairment” and/or “distress” became the primary criteria for the categorization of a behavioral condition as a “mental disorder”. This is echoed in a paper by Wakefield and First (2003)
who suggested that “GID” without “impairment” or “distress” should just be classified as a “dysfunction”, which attains status as a “mental disorder” only when combined with “impairment” and/or “distress”. It seems to me, however, that the definition of “dysfunction” is the same as, and not less problematic than, that of “pathology”. Also, some authors (e.g., Langer & Martin, 2004
) have questioned the presence of psychiatric dysfunction in individuals with GIV altogether. Moreover, what defines “impairment” and “distress”? DSM-IV and DSM-IV-TR state that the diagnostic features of GID must include “persistent discomfort about one's assigned sex or a sense of inappropriateness in the gender role of that sex.” The text further states:
“Distress or disability in individuals with GID is manifested differently across the life cycle. In young children, distress is manifested by the stated unhappiness about their assigned sex. Preoccupation with cross-gender wishes often interferes with ordinary activities. In older children, failure to develop age-appropriate same-sex peer relationships and skills often leads to isolation and distress, and some children may refuse to attend school because of teasing or pressure to dress in attire stereotypical of their assigned sex. In adolescents and adults, preoccupation with cross-gender wishes often interferes with ordinary activities. Relationship difficulties are common, and functioning at school or at work may be impaired.” (American Psychiatric Association, 2000, p. 577
As a clinician working with such children, I see several problems with this paragraph (all of which are in need of more systematic empirical documentation). (1) When one takes the developmental history of preschool children with GID, the initial features are not gender dysphoria, but gender-atypical temperament and activity preferences, and, in many boys, unusual sensory sensibilities (Coates & Wolfe, 1995
). (2) In non-GID children, “preoccupation” is not limited to gendered activities. For instance, DeLoache, Simcock, and Macari (2007)
found “extremely intense interests” in nearly a third of a sample of 84 boys and 93 girls (aged 11 months to 6 years of age) from predominantly white middle class families in the U.S., with a boy:girl ratio of 3:1. (3) In young children with GIV, gender “dysphoria” appears to develop, when the cognitive development is far enough advanced and if the gender-atypical inclinations are criticized and opposed by the parents and others. (4) Gender segregation in the peer group is normative in childhood and not labeled “impairment” if it is gender-typical. Is it not appropriate for a highly gender-atypical child to affiliate with the peer group that is more compatible with his or her gender behavior, especially if it also offers more acceptance (as shown by Wallien, Veenstra, Kreukels, & Cohen-Kettenis, 2009
)? (5) Many later problems (e.g., school refusal) appear to be secondary to the child's experience of stigmatization of the gender-atypical behavior. (6) In the general population, individuals vary considerably in stress responsivity and emotional coping, and, in my clinical work, I am impressed by a similar variability of individuals meeting criteria for GID or GID NOS. There also seems to be considerable intraindividual variability in gender-related distress over time. I am, therefore, not convinced that--in the absence of systematic documentation of distress--it is appropriate to routinely attribute “inherent distress” to all who want to change gender (Zucker, 2006
). If one postulates “inherent distress,” would one not also have to attribute something like “body dysphoria” to patients with somatic diseases or disorders who decide for surgery (say, of a facial wart) or radiation treatment (of cancer) and thereby label them as having a “mental disorder”? On the other hand, limiting the disorder category of GIV to those with marked distress would imply the exclusion from medical assistance of those without.