According to traditional cognitive behavioral theories (e.g., Clark & Purdon, 1993; Rachman, 1998a; Salkovskis, 1985), obsessions and compulsions result from responses to specific intrusive thoughts, images, or impulses. For instance, Rachman (1998a) suggested that the appraisal of intrusions as indicating some undesirable hidden areas of the individual's character, lead to an escalation in dysfunctional behaviors, or cause a more intense use of thought-control strategies (e.g., thought replacement or thought-suppression). Further, according to Rachman (1997), the content of the intrusions (e.g., themes of aggression, sex, and blasphemy) play an important role in this process by serving as the initial trigger for dysfunctional appraisals in particular individuals. Similarly, Clark and Purdon (1993; Purdon & Clark, 1999) proposed that intrusions appraised as inconsistent with an individual's sense of self, or their beliefs and values (i.e., as ego-dystonic) may result in their exacerbation into obsessions. Thus, it is proposed in this thesis and implied in traditional cognitive theories that individuals will respond to intrusions that undermine their perceptions of self.
In a study investigating the role of self vulnerability in OCD, Rowa and Purdon (2003) randomly allocated students into two groups. In one group, participants were asked to report on their most upsetting intrusive thought. In the second group, participants were asked to report on their least upsetting intrusive thought. They found that participants in the former group reported that their thoughts contradicted valued aspects of self to a greater degree than participants in the latter group. Similar results were later found with a clinical cohort (Rowa, Summerfeldt, & Antony, 2005). Rowa and Purdon (2003) argued that these findings provide evidence that the specific content of intrusive thoughts is important in maintaining OC symptoms. More specifically, the distress associated with intrusive thoughts is linked to the degree with which the thought contradicts specific aspects of the self. Thus, the intrusion's valence and importance is judged with reference to perceptions of self.
Additional evidence for the role of self-evaluation in OCD comes from a recent study by Lee and Kwon (2003). Using the Revised Obsessive Intrusions Inventory (ROII, Clark & Purdon, 1993; Purdon & Clark, 1994) with a student sample, these authors identified two subtypes of obsessions that were distinguished by their content, degree of ego-dystonicity, triggers, and type of management and response strategies used. The first of these subtypes (autogenous obsessions) included repugnant intrusive thoughts such as unwanted sexual, aggressive, and immoral thoughts or impulses. Autogenous obsessions were evoked by stimuli that are relatively harder to identify and are more abrupt and ego-dystonic than reactive obsessions. The second subtype (reactive obsessions) was found to be associated with thoughts about contamination, mistakes, accidents, asymmetry, and loss. Reactive obsessions have identifiable triggers, are relatively realistic, and are related in a more "rational" way to the neutralizing behavior. These elicit overt neutralizing behaviors, such as compulsive checking and cleaning. Several other studies have found support for the division between autogenous and reactive obsessions, in terms of their content, appraisal and control strategies (e.g., Lee & Kwon, 2003; Lee, Kwon, Kwon, & Telch, 2005; Lee, Lee, Kim, Kwon, & Telch, 2005; Lee & Telch, 2005; Lee, Zoung-Soul, & Kwon, 2005; Moulding, Kyrios, & Doron, submitted). Thus, these results suggest that the extent of the inconsistency with an individual's sense of self and the content domain of the intrusion influences the degree of distress caused by the intrusion and the management strategies used.
Bhar (Bhar, 2004; Bhar & Kyrios, 2000) also investigated the relationship between OCD and the perception of self; specifically, they investigated the relationship between ambivalent sense of self (i.e., the degree of uncertainty in, dichotomy in and preoccupation with one's self-worth) and OC phenomena. They found that individuals suffering from OCD scored significantly higher on the Self Ambivalence Measure (SAM) than non-clinical university students and community controls, although the SAM did not discriminate between individuals with OCD and those presenting with other anxiety disorders. However, even when the presence of depressive symptoms and other anxiety symptoms was accounted for, self-ambivalence maintained a significant association with OC symptoms. These findings support the notion that an ambivalent sense of self is related to OC symptoms, but may not be specific to OCD. However, considerable research have demonstrated that the self is a multi-dimensional construct (Eccles, Wigfield, Flanagan, Miller, & et al., 1989; Harter, 1990; Marsh, 1990). The SAM fails to distinguish between perceptions of the self in different domains. It is possible that ambivalence in specific domains (e.g., the morality domain) is more specifically related to OCD while overall ambivalence of self is related to a range of anxiety disorders. For instance, Rachman and Hudgson (1980) noted individuals suffering from OCD to be "of tender conscience". In order to establish a clearer understanding of the relationship between self perceptions and OCD, one may need to account for the multifaceted nature of the self and the relationship of specific facets of self to obsessive or compulsive symptoms. The following section further explores this notion.
Self concept has long been considered to be central to psychopathology with numerous conceptualizations such as self-complexity (Linville, 1985, 1987), self-discrepancy (Higgins, Klein, & Strauman, 1985), self-clarity (Campbell et al., 1996) and self-ambivalence (Guidano & Liotti, 1983) all having implicated in maladaptive behaviors, cognitions and emotions. However, it is argued in this thesis that a multidimensional conceptualization is particularly relevant to further our understanding of OCD.
Indeed, substantial research has demonstrated that the self is a multi-dimensional construct (Eccles et al., 1989; Harter, 1990; Marsh, 1990). One model of self proposed by Harter (1982, 1996) has been particularly influential in the multi-dimensional conceptualization of self. Harter proposed a hierarchical model where self-evaluations of competencies in specific self-domains (e.g., morality, personal appearance, and job performance) are integrated into an overall self-evaluation. Individual differences are found in the importance attributed to each domain and the individual's evaluation of their own competence in each of these domains. According to this model, an individual's perceived competence in valued domains is strongly related to their feelings of self worth (i.e., global self-esteem). Interestingly, Harter's model (1982, 1996) is consistent with Bowlby's (1969, 1973) suggestion that attachment experiences result in the construction of self representations as being lovable and competent. Harter's model (1982, 1996) also specifies a particular self structure linking perceptions of competence with general self worth.
Indeed, research has shown that an individual will value some domains of self more highly than others and that competence in domains of particular personal importance is more highly correlated with feelings of self worth than competence in domains judged to be less important (Harter, 1998; Messer & Harter, 1986; Neemann & Harter, 1986). The discrepancy between evaluations of competence and significance in self domains has been found to be negatively related to self-worth (Harter, 1998; Messer & Harter, 1986; Neemann & Harter, 1986).
Bhar (2004) examined the relationship between evaluations in specific self-domains and global self-esteem in students scoring either high or low on an obsessionality scale (Padua Inventory -Revised; Burns, Keortge, Formea, & Sternberger, 1996). The evaluation of self-domains was made using a revised version of the Self Attributes Questionnaire (Pelham, 1991) where items assessing ethical/moral standards and acceptance by other people were added in order to rate domains considered relevant to OC phenomena. Self-esteem in the high-obsessional student group was associated with a different number and pattern of self-domains than in the low-obsessional student group. Specifically, students in the low-obsessional group associated their self-worth with academic achievements, athletic gains, physical attractiveness, social skills, and acceptance. In contrast, students in the high-obsessional group related self-worth to moral standing, social skills, social acceptance, and physical attraction. Thus, vulnerability to OC symptoms may be associated with a specific pattern in self structure.
Although the evaluation of an individual's self worth is believed to be relatively stable, it has also been shown to be "context" sensitive. For instance, Markus and Kunda (1986) showed that individuals adjusted and regulated their self evaluations in response to an experimental manipulation of their sense of uniqueness and similarity to others. Marcus and Kunda invoked the construct of the "working self-concept" (WSC) to explain these variations, suggesting that individuals possess both a stable self-concept and a temporary one. The WSC form a subset of one's self-conceptions. It is affected by situational factors and influences feelings of self-worth at any given point in time. That is, situational variables lead to short-term changes in self-evaluation (Markus & Wurf, 1987). Interestingly, the working self-concept is consistent with attachment theories which emphasize that the attachment system is triggered in specific situations associated with stress or anxiety (Bowlby, 1969, 1973, 1988a) and implicates self-vulnerability in the dynamic of anxiety reactions. Thus, there is a significant overlap between the fixed and contextual view of the self, which suggests a less reductive attitude to how we should evaluate individual self-worth.
Overall, self-concept can be defined as comprising several domains of competence. Perceived competence in domains that are attributed greater importance have greater influence on an individual's perception of self-worth (Harter, 1998). In addition, self-worth is sensitive to situational variables through the working self-concept (Markus & Wurf, 1987). Thus, individuals with large discrepancies between their perceived competence and the importance they attribute to specific domains may be vulnerable to stimuli threatening their feelings of competence in these domains ("sensitive" domains). Moreover, if fewer domains are being valued by the individual, this may compound sensitivity to stimuli "endangering" these self-domains. That is, an individual's feelings of self-worth will be over-reliant on perceptions of competence in a limited number of domains.
Following from this, this thesis argues that dysfunctional attachment experiences may increase the likelihood for developing a self-concept comprising relatively few domains that are "sensitive" and that these factors underlie vulnerability in OCD. Intrusive thoughts relating to failure in these "sensitive" domains or to the appraisals associated with them (e.g., "If I have such thoughts, I am immoral, unworthy, and will be abandoned") endanger the individual's sense of self-worth, triggering extreme anxiety. This, leads to maladaptive appraisals and management of intrusions, which in turn exacerbates their effects. In summation, individuals suffering from OC symptoms may be vulnerable to intrusive thoughts that trigger their insecurity regarding their competence in domains they value highly.
While this hypothesis may indicate why particular intrusive thoughts become obsessional, vulnerability of the self does not fully explain the repetitious pattern of behavior associated with OCD. Nor does it fully explain the specific nature of OC symptoms. In the following sections, we will argue that a specific "world-view" may help in assessing how particular symptoms of OCD, including compulsive and repetitive behaviors, manifest themselves in susceptible individuals with a vulnerable sense of self.
The significant role of world-view in the development of OCD has been suggested by several cognitive authors (e.g., Guidano & Liotti, 1983; Menzies, Harris, Cumming, & Einstein, 2000; Riskind, Abreu, Strauss, & Holt, 1997; Salkovskis et al., 1999). The Obsessive Compulsive Cognition Working Group (OCCWG, 1997, 2003) has found that exaggerated evaluation of probability and cost of aversive events (i.e., overestimation of threat) is one of the core beliefs associated with OCD. Riskind et al., (1997) found a higher degree of dynamic threat perception to spreading contamination (i.e., sense of looming vulnerability) in an undergraduate subclinical obsessional group than for those in a control group. These findings illustrate the importance of specific aspects of world view in OC phenomenon. These findings also raise the question of the importance of more general world view assumptions and their relationship with such specific cognitive vulnerabilities.
Guidano and Liotti (1983) argued that a perception of the world as being threatening but controllable translates into active attempts to control the environment in individuals suffering from OCD. According to Guidano and Liotti (1983), such world perceptions stem from early attachment experiences. Others have also related the individual's sense of security in the world to early attachment experiences (e.g., Ainsworth et al., 1978; Bowlby, 1973, 1988; Janoff-Bulman, 1989, 1991). According to attachment theory (Ainsworth et al., 1978; Bowlby, 1973, 1988b), the individual develops a sense of security in the world based on the perception of having a secure base to return to in case of need ("secure-base" phenomena). This sense of security is a reflection of the individual's experience of being nurtured (i.e., perceived degree of feeling loved by the main caregiver and perceived degree of availability and responsiveness of the main caregiver) and later influences the individual's perceptions of the world and of human nature (Bretherton & Munholland, 1999).
Janoff-Bulman (1989, 1991) proposed a useful model whereby feelings of safety or vulnerability in the world are strongly influenced by early experiences such as attachment relationships. Janoff-Bulman argued that an individual's sense of security in the world is reflected in two aspects of "world view" (i.e., perceived benevolence and perceived meaningfulness of the world) and a worthiness-of-self dimension. Perceived benevolence of the world reflects the degree to which the individual judges the world as a positive versus a negative place. This aspect includes two main assumptions: (a) the perceived benevolence of human nature (i.e., whether people are perceived as generally kind or malevolent) and (b) the perceived benevolence of the impersonal world (i.e., whether the world is perceived as a generally a positive or negative place). The perceived meaningfulness aspect involves three main beliefs: justice, controllability and chance beliefs. These three beliefs express the individual's understanding of "who gets what" in the world. For instance, some individuals may assume that their precautionary behavior or actions will determine what will happen to them in the world (controllability-outcome belief). Others may believe that misfortune is least likely to strike "good people" (justice-outcome belief). Others may believe that there is no way to achieve control in the world (chance-outcome belief).
While the above three beliefs refer to what individuals should expect in the world (i.e., what the world is generally like), the worthiness-of-self dimension refers to the individual's personal evaluation of what they deserve in the world. This aspect includes three dimensions of self evaluation that correspond with the three meaningfulness beliefs. For example, individuals that perceive themselves as competent in their actions when preventing misfortune and who believe that actions determine what happens in the world (i.e., controllability-outcome belief) will have higher perception of worthiness-of-self than individuals holding the same outcome beliefs and perceiving themselves as incompetent in their actions for preventing misfortunes. Similarly, individuals believing in their worth as moral beings will feel more secure in a world where "good things happen to good people" than people believing in the same outcome principle, but perceiving themselves as immoral beings.
Janoff-Bulman's multi dimensional world-view model is similar to Harter's (1982, 1996) model of self worth in that competencies in specific domains and the degree of importance attributed to these domains play an important role in a self evaluative process. Harter suggests that the individual's self evaluation of valued self-domains determines their feelings of self-worth. Janoff-Bulman suggests that the individual's self evaluation process of valued world-view assumptions determines the degree of their feelings of vulnerability in the world.
The examination of differences in world perceptions may have important ramifications for the understanding of OCD etiology and maintenance. While overestimation of threat has been found to be associated with OC symptoms (Jones & Menzies, 1998; OCCWG, 1997, 2005), the underlying personal theory or world-view of threat perception is yet to be fully understood. Individual differences in world-view may influence both vulnerability to specific types of intrusions and reactions to such intrusions. For instance, it could be argued that individuals who rely on their own behaviors to increase their sense of control in the world will be particularly threatened by "preventable" intrusions, such as those involving contamination and accidents. In the same vein, these individuals will tend to use overt neutralizing behaviors, such as compulsive cleaning, to regain their sense of control. Conversely, individuals who associate outcomes in the world with "being a good person" may be likely to be threatened by ego-dystonic repugnant intrusive thoughts such as sexual, aggressive, and immoral thoughts. These individuals will be more inclined to use covert neutralizing behaviors such as thought-stopping and thought-suppression in order to reduce the threat to their perception of self (e.g., as a moral and/or social being). Hence, the differentiation of OC symptom subtypes (e.g., washers versus obsessionals) or distinct OC symptoms (e.g., autogenous versus reactive obsessions) may be helped by consideration of such individual differences.
It could be further argued that any attempt by an individual to control outcomes in the world relies on the basic assumption that the world is, at least partly, controllable (Guidano & Liotti, 1983). An individual may not feel responsible (Salkovskis, 1985, 1999) for the consequences of their actions (e.g., harm to self or others) and may not be likely to attribute exaggerated importance to controlling their thoughts (Clark & Purdon, 1993; Purdon, 1999) without assuming some possible control over them. In addition, it is argued that the need to achieve control is exacerbated in a socially and physically threatening world (Guidano & Liotti, 1983). This is because, an individual's sense of security is likely to be influenced by their perceptions of others (Janoff-Bulman, 1989, 1991) and the likelihood that negative events will occur (OCCWG, 1997).
It is argued here that a wider examination of the individual's world view assumptions will lead to a better understanding of both their general vulnerability to OC symptoms and their vulnerability to specific OC symptoms or OCD subtypes. Consistent with Bandura's (1986, 1997) differentiation of the roles played by beliefs of personal competence (i.e., self-efficacy) and beliefs about the likelihood of desired outcomes (i.e., outcome expectancies), it is argued that cognitive vulnerability to specific presentations of OCD will be exacerbated by an individual's assumptions about the malevolence of the world, and the particular nature of their symptoms will be influenced by their adherence to beliefs about the justice-outcome principle or the behavior-outcome principle. That is, the judgment of one's personal ability to engage in a behavior and their judgment of the likely consequence (Bandura, 1986) may play a differentiating role in the dynamic of OC symptoms. While self-vulnerability and world-view may be a key to understanding an individual's vulnerability to the development of OC symptoms and cognitions, the origin of such internal representations is not well understood. Epidemiological data (e.g., Johnson, Cohen, Kasen, Smailes, & Brook, 2001) and a developmental perspective implicates early parenting experiences which are discussed in the next section.