The concept of narcissism has always been a subject of interest for several theoretical approaches (Freud, 1914; Kernberg, 1975; Kohut, 1971, 1977). Still, discussions around this topic are controversial even within the current scientific community (Campbell & Miller, 2011; Skodol et al., 2014). Attachment theory, on the other hand, has a more defined conceptualization, arguing about the psychological connectedness that occurs within specific relationships (Bowlby, 1982, 1988). The two concepts, are traditionally viewed from different perspectives, as attachment indicates the individual’s need for closeness, security, and care from another person who is perceived as separate from oneself (Silverman, 1991, as cited in Pistole, 1995), while narcissism involves a more egocentric image with undifferentiated or merged aspects of the ego organization (Kernberg, 1975; Kohut, 1977). Nevertheless, some common theoretical elements seem to be shared and the current study seeks to investigate them in the context of romantic relationships. The following sections provide an overview of the narcissistic phenomenology and attachment theory and finally a review of the existing literature on the theoretical and empirical findings of their role in adult romantic relationships.
Narcissism
According to Freud (1914), the term narcissism refers to the love for one’s own self-image. More recent theoretical perspectives on narcissism use the term widely in the context of self-esteem, referring to a personality trait which manifests in both healthy and pathological ways. Additionally, narcissism as a personality trait can be conceptualized in a continuum with relatively adaptive (e.g., extroversion) as well as maladaptive characteristics (e.g., expectation of favorable treatment) (Pincus & Lukowitsky, 2010). This form of narcissism is sometimes referred to as subclinical or normal narcissism, and is most commonly measured in the social-personality literature by the Narcissistic Personality Inventory (NPI; Raskin & Hall, 1979). Higher scores indicate individuals possessing similar elements to narcissistic pathology, but not at a clinical level of personality disorder manifestation. On the other hand, in its pathological nature, it can be manifested as a Narcissistic Personality Disorder (NPD) as described by DSM-5 (Section II) (APA, 2013). As Zeigler-Hill and Besser (2013) note, both normal and pathological forms of narcissism, contain maladaptive features, but they differ in the degree and the intensity of their presentation in various contexts (e.g., self-regulation on affective experiences, self-image, interpersonal effect). In their research, Miller and Campbell (2008) also argue that, although the relations between pathological and normal nature of narcissism are considered heterogenous and diverge, this social-personality construct predicts many of the behaviours appeared in the clinical picture (e.g., entitlement, a tendency to use an antagonistic interpersonal style, desire for admiration, fantasies of success). Moreover, despite the theoretical differences concerning the definition and the diversity in conceptualizations of the construct of narcissism (for review, see Pincus & Lukowitsky, 2010), the majority of researchers agree on the grandiosity observed in individuals with narcissistic personality traits as a facade to cover up their low self-esteem (Kernberg, 1975, 2004; Kohut, 1971, 1977; Zeigler-Hill & Besser, 2013).
Based on the contradictory sense of narcissistic self-esteem, researchers revealed two different types of narcissism: the grandiose and the vulnerable (Akhtar & Thomson, 1982; Dickinson & Pincus, 2003; Hendin & Cheek, 1997; Wink, 1991). Grandiose narcissism, which is represented by NPD, projects a general image of grandiosity, grandiose sense of self-importance, constant need for admiration and lack of empathy (APA, 2013). The second type is represented by a vulnerable narcissistic personality organization, which involves a covert core organized around grandiose fantasies and entitlement (Besser & Priel, 2010; Dickinson & Pincus, 2003) and with oscillations between feelings of superiority and inferiority (Rohmann et al., 2012). In contrast to grandiose narcissism, which is characterized by a positive idealized and exaggerated self-image, vulnerable narcissism is characterized by oversensitivity, defense and insecurity (Miller et al., 2011; Wink, 1991).
As Campbell et al. (2006) aptly argue, narcissists’ reluctancy to identify with and respond to others’ feelings, in combination with their need to regulate and boost their narcissistic self-esteem through them, can reach a potential point of inward conflict between a focus on the self and a focus on others. How does the individual then shift between the narcissistic behaviours and the ability to include another person in the relationship with oneself, especially in their romantic affairs? And how do their narcissistic features influence the emotional bond which is expected to develop in romantic relationships? Since attachment theory supports the idea of the person’s innate drive to form proximity with important others (Bowlby, 1982, 1988), it could be used as a useful model to provide an in-depth understanding, as well as, to provide answers to the raised questions.
Attachment
According to attachment theory, this biological need is activated especially when infants are in distress. In an effort to survive, they develop an emotional bond – attachment to a caregiver as a basis to facilitate the monitoring of their feelings and behaviours which are displayed whenever their security is being threatened (Bowlby, 1982, 1988). If there is a response to the child’s needs from the caregiver- acting as an emotional nurturer of comfort and reassurance- the desired closeness is acquired and maintained, the child feels adequately protected and supported and thus develops a secure attachment bond. Conversely, if there is inconsistency and lack of sensitivity to the verbal and non-verbal cues of the child’s attachment system, then an insecure attachment bond develops with the child experiencing the caregiver as unreliable and as unsafe figure (Ainsworth et al., 1978).
These early experiences, are stored as information and are internalized as working models of attachment. As far as caregiving and responsiveness are concerned, internal working models are composed by a positive and a negative model of the self and others and can also be divided into two dimensions: anxiety and avoidance (Bartholomew & Horowitz, 1991). People with higher attachment anxiety, feel that others are reluctant to come emotionally to them as much as they would like, are intensively preoccupied by the thought that others don’t really love them and they can show an excessive tendency for emotional dependence; while being afraid of rejection, as well as, separation or abandonment. On the contrary, people with higher attachment avoidance feel uncomfortable when they are emotionally close to others, find it difficult to trust them completely and let themselves be dependent on them, tend to avoid emotional expressiveness and feel pressured when others seek more emotional intimacy than they feel comfortable offering (Mikulincer & Shaver, 2007).
These internal working models are characterized by a constant activation in the presence of a threat and it is also theorized that they can be transferred to new circumstances and relationships from childhood to adulthood (Bowlby, 1982, 1988). Finally, it is acclaimed that they are also responsible for the usual patterns of responses someone exhibits in his relationships- also known as attachment styles. Individual differences in attachment styles (e.g., secure, preoccupied, dismissing-avoidant, fearful-avoidant) are inferred to reflect fundamental differences in the internal working models of view of self and of view of others (Bartholomew & Horowitz, 1991; Mikulincer & Shaver, 2007). According to Hazan & Shaver (1987), couples’ erotic attraction is similar to the attachment styles they formed as infants. Additionally, they were the first to argue that early attachment patterns can determine the way one chooses, relates and responds to a partner. Given the above, what is the nature of the internal representations/attachment dimensions- which emerged from these early experiences and which contribute to the preservation of these patterns in the current interpersonal relationships- to the different types of narcissism?
Although the concepts of attachment and narcissism are typically perceived in different perspectives, elements of narcissism can be observed in the insecure attachment dimensions (Blatt & Levy, 2003; Pistole, 1995). Nevertheless, in a variety of studies, the empirical analysis of narcissism with respect to attachment theory led to rather conflicting findings and moreover it must be noted that on the level of romantic relationships there is limited research in exploring the association between the two variables. Since, to our knowledge, there is only a small number of existing studies on the association of narcissism and attachment in romantic relationships, the present study could shed more light in investigating the connection between these variables. The following section, begins with an overview of general research findings on the relationship between narcissism and attachment as well as the investigation of narcissism’s role in romantic relationships to help us conceptualize the links between theory and research and finally it presents findings from the two available studies which specifically focus on the association of narcissism and attachment in romantic relationships.
Review of Empirical Studies
Brennan and Shaver (1998), found a possible relation between the narcissistic personality disorder and insecure attachment features. They reported a range of attachment representations of both anxiety and avoidance on individuals with narcissistic traits, while examining the relationship between NPD and attachment styles. Dickinson and Pincus (2003), revealed that individuals with high grandiose narcissism reported having a secure and/or avoidant attachment style, while individuals with vulnerable narcissism reported having the anxious/ambivalent and/or fearful attachment style. Smolewska and Dion (2005) demonstrated that despite the fact that grandiose narcissism does not seem to relate to attachment, vulnerable narcissism was found to be linked to the insecure dimension of attachment anxiety. In line with these findings regarding the relation between attachment anxiety and vulnerable narcissism, is the study of Otway and Vignoles (2006).
More recent studies conclude that pathological narcissism is significantly related to the insecure attachment dimensions of anxiety and avoidance, in both non-clinical and clinical sample (Fossati et al., 2015). In their study, Kealy et al. (2015) examined the relationship between pathological narcissism and the two conceptual structures which represent the internal representations of relationship patterns (quality of object relations and attachment styles). The findings of their study showed that pathological narcissism affects attachment anxiety but not avoidance.
It can be said that the presented findings indicate that the connection between narcissism and attachment exists but it varies. Additionally, it can be noted that the concept of narcissism has a direct impact on both the individual’s object relations mental life and the capacity to engage and maintain committed and healthy relationships throughout adulthood (Kernberg, 2011). Nevertheless, only few empirical studies have extended their interest in investigating the role of narcissistic traits in romantic attachment.
In consideration of the fact that romantic relationships are thought to be the place where the deepest emotional needs for intimacy and relatedness can be expressed, what happens when narcissistic elements are also included? An increasing number of theoretical perspectives suggest that the ability to establish a deep emotional commitment, is limited in individuals with narcissistic traits (Kernberg, 2011). For those individuals, romantic interactions are perceived on a superficial level and are regarded as a game of self-enhancement (Campbell et al., 2002). Romantic relationships end up being shallow and transitional (Campbell et al., 2002), with narcissists seeking to find a “trophy” partner (Campbell, 1999). As Besser and Priel (2009) argue, individuals with vulnerable narcissism, who deal with much higher stress levels during the development of relationships, become alert to indications of rejection from their partner and experience greater anxiety during separation, due to the fragile nature of their self-esteem.
Despite the existing research on the association of narcissism and attachment (Brennan & Shaver, 1998; Dickinson & Pincus, 2003; Fossati et al., 2015; Kealy et al., 2015; Otway & Vignoles, 2006; Smolewska & Dion, 2005), as well as, the detailed and extensive theoretical associations (Campbell et al., 2006; Campbell & Green, 2008) and research (Besser & Priel, 2009; Campbell, 1999; Campbell et al., 2002; Campbell & Miller, 2011) regarding the role of narcissism in romantic relationships, the empirical research on the relationship between narcissism and attachment in romantic relationships is limited. A research study which provides some elements of the relationship between narcissism and attachment is that of Rohmann et al. (2012), in which they examined the interpersonal consequences of the two types of narcissism on attachment and love. Results showed that individuals with a higher score on vulnerable narcissistic traits have higher levels of attachment anxiety compared to individuals with a lower score on vulnerable narcissistic traits. On the other hand, individuals with a higher score on grandiose narcissistic traits seem to have lower levels of avoidant attachment compared to individuals with a lower score on grandiose narcissistic traits. However, the study had a limited number of participants (92 students with average age 24 years), taken from a specific university and with an unclear romantic relationship status. Furthermore, a recent study of Altinok and Kilic (2020), revealed that the insecure attachment styles (fearful, preoccupied and dismissive) play a dominant role in predicting narcissism and infidelity intentions in romantic relationships. Though, once again, the sample consisted only of university students- very young adults (mean age 21.52), who reported having an on-going romantic relationship. The status of this on-going relationship was not clarified and no assumptions could be made; highlighting anew perhaps an important gap in the current research literature.
Given, that internal working models of attachment comprise of the individual’s expectations for themselves, others, as well as the world around them (Bowlby, 1988), the following question arises: To what extent do individuals with narcissistic traits have access to secure/insecure attachment dimensions during their romantic interactions? The present study attempts to examine this question in order to provide insight and fill the gap in the existing literature.
As far as we know, there is no such study published in Greek, hence it is quite possible that the innovation of this study lies in the fact that previous studies were conducted in a different cultural setting. According to Kafetsios and Kateri (2020), each cultural setting defines its relational norms, which could shape attachment orientations. In addition, in modern cultures, like Western societies, narcissism scores are higher due to the individualistic aspects of their rearing (for review, see Foster et al., 2003). Hence, the promotion of an autonomous and independent self-development is usually favored by the cultural orientation, reflecting less collectivistic behaviors and encouraging greater self-focus. The Greek culture is mainly considered as a collectivistic one, that also subsumes individualistic values (Georgas, 1989, as cited in Kafetsios & Kateri, 2020). Therefore, it would be interesting to see whether closeness-attachment in romantic relationships in a sample of Greek speaking participants may threaten individuality for the more independent-narcissistic individuals.
Additionally, due to the limited number of existing studies on the association of narcissism and attachment in romantic relationships, we anticipate that our study will contribute to the broader expansion of research on this topic. Moreover, and therapeutically-wise, the association between narcissistic traits and attachment dimensions in romantic relationships, could be effectively utilized in psychotherapy- regarding the activation of similar insecure attachment patterns both in the therapeutic relationship and in patients’ wider romantic relationships. Consequently, the results of the present study could have a possible psychotherapeutic benefit in the conceptualization of difficulties in romantic relationships and ultimately, they could lay the foundation for future interventions for clients who seek therapy to tackle relationship problems. The main purpose of the current study is to examine the relationship between narcissism (grandiose and vulnerable) and attachment dimensions (anxiety and avoidance) in adult romantic relationships. It should be clarified that, the present study took a specific approach in understanding and defining narcissism as a broader network presented by two phenotypes and it is not studied from a diagnostic or clinical perspective as a psychological disorder. Also, this study is in line with the current position in attachment research, in conceptualizing attachment along two continuous dimensions (anxiety and avoidance), which are found to be an accurate representation of the attachment styles framework (Fraley & Spieker, 2003a; Fraley & Waller, 1998). A dimensional approach underlies the notion that psychological dynamics in attachment system do not necessarily need to be categorized, since the relationship representations can be understood in respect of their functioning processes- to two sources of variation, rather than to four, distinct styles (Fraley & Shaver, 2000). Finally, it is generally recommended not to classify people in categorical terms when using ECR-R, as in this way the precision of measurement is reduced and the statistical power is lowered (Fraley & Spieker, 2003a, 2003b; Fraley & Waller, 1998; Roisman et al., 2007). Hypothesis 1: Both types of narcissism (grandiose and vulnerable) will have a positive and statistically significant correlation with the insecure attachment dimensions (anxiety and avoidance). Hypothesis 2: Vulnerable narcissism will be a stronger predictor of the attachment anxiety dimension, in relation to grandiose narcissism. Hypothesis 3: Grandiose narcissism will be a stronger predictor of the avoidant attachment dimension, in relation to vulnerable narcissism.
Method
Participants
One hundred and fifty-one adults from the general population (31 men, 119 women, 1 other) in Cyprus participated in the present study. The age of participants ranged between 25 and 35 years (M = 28.50, SD = 3.62). This age range was chosen based on Erikson’s (1959) developmental theory of psychosocial stages (sixth stage: intimacy vs isolation), due to the fact that the focus of individuals in this evolutionary period is the formation of intimate-loving relationships with others. The criteria for participating in the research were: good knowledge of the Greek language and currently being in a romantic relationship. There was no restriction on professional status or educational level. The socio-demographic characteristics of the participants are displayed in Table 1.
Measures
Experiences in Close Relationships_Revised (ECR-R; Fraley et al., 2000) – Greek version (Tsagarakis et al., 2007). The ECR-R is a 36-item, self-report measure of adult attachment dimensions, which measures individuals on two subscales of attachment: avoidance (the degree to which people state that they feel uncomfortable being emotionally close to others) and anxiety (the degree to which people state that they fear rejection and abandonment). Items are rated on a 7-point scale, ranging from 1 (strongly disagree) to 7 (strongly agree). Coefficient alpha in this sample showed high reliability for both the anxiety dimension (α = 0.92) and the avoidance dimension (α = 0.90).
Narcissistic Personality Inventory (NPI; Raskin & Hall, 1979) – Greek version (Coccosis et al., 1998). The NPI is a self-report instrument designed to measure individual differences in grandiose narcissism in a non-clinical population. It consists of 30 pairs of forced-choice statements, where individuals are asked to choose the statement that best expresses themselves. In the present study, the instrument showed good reliability since the coefficient was α = 0.79.
Hypersensitive Narcissism Scale (HSNS; Hendin & Cheek, 1997). The HSNS is a 10-item, self-report scale, designed to measure vulnerable narcissism. Participants are asked to state their level of agreement on each item on a 5-point Likert scale (1=strongly disagree, 5 = strongly agree). For the purposes of the present study, the scale has been translated into Greek, through forward-backward translation (Behling & Law, 2000). The value of the coefficient alpha in this sample was 0.60.
Sociodemographic Questionnaire. Data on sex, age, nationality, relationship status and relationship satisfaction level were collected through a sociodemographic questionnaire.
Procedure
Following approval by the National Bioethics Committee of Cyprus participants were approached in two ways: electronically (online invitations were posted on social media leading to Google Forms platform) and in-person (data collection from central points of major cities). Participation in the research was voluntary and anonymous. All the necessary requirements of the European General Data Protection Regulation (GDPR) were taken into consideration.
Results
IBM SPSS Statistics (Version 21.0), was used to perform the analyses. All the necessary checks for extreme values, incomplete responses and normality, were performed prior to data analysis. No outliers or incomplete data were detected.
Correlations Between Narcissism Types and Attachment Dimensions
Pearson correlation coefficient was used to assess the correlation among variables (see Table 2). Grandiose narcissism was positively and significantly correlated to both attachment dimensions, Anxiety (r = .17, p = .03) and Avoidance (r = .17, p = .03). Vulnerable narcissism showed a low positive and statistically significant correlation with both attachment Anxiety (r = .28, p < .01) and attachment Avoidance (r = .23, p < .01). Finally, the two types of narcissism were positively and significantly correlated between them (r = .17, p = .04), while attachment Anxiety and attachment Avoidance were found to have a good positive and statistically significant correlation (r = .59, p <. 01).
Multiple Regression for Attachment Anxiety
Hierarchical, multiple regression analysis was used to test if the two types of narcissism significantly predicted participants’ ratings for anxious attachment. The first step in the regression analysis, consisted of only Grandiose narcissism as the predictor variable, while in the second step Vulnerable narcissism was added (see Table 3).
The analysis in the first step showed a statistically significant contribution of grandiose narcissism (β = .17, p = .03) in the prediction of attachment anxiety, which accounted for 3% of the variance of anxious attachment (R2 = .03). By adding vulnerable narcissism to the second step, vulnerable narcissism seemed to have a significant and stronger effect on the prediction of the dependent variable. In particular, it indicated that grandiose narcissism no longer significantly affected attachment anxiety (p > .05), while vulnerable narcissism became a statistically stronger predictor of that dimension (β = .26, p = .002). More specifically, vulnerable narcissism better predicted attachment anxiety by 9%, than grandiose narcissism, since the R2 is equal to .09. Therefore, there was a significant change in the model- with an increase in the value of the coefficient of determination (ΔR2 = .06), in the second step with the addition of vulnerable narcissism.
Multiple Regression for Attachment Avoidance
Hierarchical, multiple regression analysis was conducted to estimate if the two types of narcissism significantly predicted participants’ ratings for avoidant attachment. Again, the first step in the regression analysis, consisted of only Grandiose narcissism as the predictor variable, while in the second step Vulnerable narcissism was added (see Table 4).
The analysis in the first step showed a statistically significant effect of grandiose narcissism (β = .17, p = .03) on the prediction of attachment avoidance, which accounted for 3% of the variance of avoidant attachment (R2 = .03). However, by adding vulnerable narcissism to the second step, vulnerable narcissism seemed to have a significant and stronger effect on the prediction of the dependent variable. Grandiose narcissism no longer significantly affected attachment avoidance (p > .05), while vulnerable narcissism became a statistically stronger predictor of that dimension (β = .20, p = .01). Vulnerable narcissism better predicted attachment avoidance by 7%, than grandiose narcissism (R2 = .07).
Therefore, there was a significant change in the predicting model when vulnerable narcissism was added, with an increase of the variance (ΔR2 = .04). That is, if individuals exhibit characteristics of vulnerable narcissism, they are predicted to have higher levels of avoidant attachment.
Discussion
The present study aimed to investigate the relationship between narcissism (grandiose and vulnerable) and attachment dimensions (anxiety and avoidance) in adult romantic relationships. The main findings initially concern grandiose narcissism as a predictor of the two attachment dimensions, while, after a parallel examination of the two phenotypes of narcissism, vulnerable narcissism had the greatest effect on the two attachment dimensions (anxiety and avoidance), offsetting grandiose narcissism that had previously significantly predicted those dimensions.
As expected (Hypothesis 1), the initial analysis showed that both types of narcissism had a positive and statistically significant correlation with the insecure attachment dimensions. These particular findings are in line with theory (Kernberg, 1975; Kohut, 1971, 1977) as well as with research on the exploration of narcissism and attachment (Fossati et al., 2015; Kealy et al., 2015), despite the fact that these studies have not included romantic relationship as a variable. It is thus once again revealed that the narcissistic core -whether it presents grandiose aspects, or vulnerable- is structured around insecure internal representations which are activated by seeking proximity and/ or distance, anxiety and/ or avoidance, something that seems to be applied in their romantic interactions as well. Nevertheless, despite the statistically significant findings of the present study, the correlation between grandiose narcissism and attachment dimensions was found to be very low, whereas the correlation between vulnerable narcissism and attachment dimensions was low to moderate. This could possibly suggest that individuals who reported higher levels of grandiose narcissism might be less anxious and/or avoidant in their romantic relationships, in relation to individuals who show higher levels of vulnerable narcissism. Finally, it is also likely that individuals with grandiose narcissism might lack the ability to encode information about their traits and might be less likely to report any self-evaluation which could be perceived as weak or negative (Grijalva & Zhang, 2016).
For a more detailed investigation of the prediction of narcissism on attachment dimensions, regression analyses were used and fully confirmed the second research hypothesis but partly confirmed the third. More precisely, using a hierarchical regression analysis of the two narcissism types in the two models, grandiose narcissism which was initially placed in the first step alone was found to be statistically significant and was found to positively predict the dimension of both attachment anxiety and avoidance. However, grandiose narcissism seemed to be less statistically significant, when vulnerable narcissism was added and showed a greater effect on the prediction of the two insecure attachment dimensions.
As expected, vulnerable narcissism was a stronger predictor of the attachment anxiety dimension than grandiose narcissism (Hypothesis 2), a finding consistent with results from previous studies (Kealy et al., 2015; Otway & Vignoles, 2006; Smolewska & Dion, 2005), which investigated narcissism and attachment to other concepts beyond romantic interactions. In the present study the difference between the effect of vulnerable narcissism on attachment anxiety, compared to grandiose narcissism, reveals that the defensive mechanisms of vulnerable narcissism are insufficient to protect the individual from anxiety/stress about the availability and emotional response of their romantic partners. In contrast to individuals with grandiose narcissistic traits who can regulate their sensitivity through self-enhancement (e.g. referring to positive self-representations), individuals with vulnerable narcissistic traits are not able to control their self-esteem, thereby experiencing interpersonal distress (Meyer & Pilkonis, 2011). In addition, despite their oversensitivity, vulnerable narcissistic personalities expect their partners to fully respond and attend to their narcissistic needs (Pistole, 1995). On the other hand, attachment anxiety involves a sense of anxiety and vigilance towards rejection and abandonment from significant others (Wallin, 2007). Hence, combining the theoretical elements of narcissism with attachment theory in the present study, it can be stated that the ambivalence between the narcissistic demands and the hyperactivation of the attachment system, potentially leads those individuals to experience a constant fear in their romantic relationship, which highlights their vulnerability and their sense of insufficiency to meet their expectations. This constant hypersensitivity over acceptance and enhancement from significant others, elevates their attachment anxiety even more, as well as their need for self-affirmation and acknowledgement in romantic interactions.
The results of the second research hypothesis of the present study support the above view, indicating that attachment in individuals with narcissistic vulnerability could be manifested by feelings of fear of rejection and abandonment when their narcissistic needs are not met by romantic partners. This particular finding is line with results from previous studies (Besser & Priel, 2009; Smolewska & Dion, 2005), which focus on the sense of rejection which seems to be experienced by individuals with vulnerable narcissism and anxious attachment in their romantic relationships. At the same time, the negative self-image which seems to guide individuals to feel rejected in their relationship experiences is also prevalent in the findings. Overall, the findings of this study support the empirical affirmation that the different representations of a narcissistic personality could possibly predict differences in the functions of insecure working models regarding anxiety attachment in romantic relationships. Furthermore, they underline the need for improvement in the clinical distinction between vulnerable and grandiose narcissism.
As for the third hypothesis of the present study, an exclusive examination of the two variables showed that grandiose narcissism predicts attachment avoidance. However, this hypothesis was partly confirmed, since in the second step, vulnerable narcissism showed a greater effect on the prediction of the avoidant attachment dimension in relation to that of grandiose one whose statistical significance was offset. Even though it is generally supported by theory that the “grandiose self” denies dependency on relationships through feelings of arrogance and rejection (Kernberg, 2011), the findings of empirical research (Brennan & Shaver, 1998; Dickinson & Pincus, 2003; Fossati et al., 2015) and those of the present study seem to be rather inconsistent.
The fact that grandiose narcissism predicted the dimension of attachment avoidance, underlines the individuals’ difficulty in establishing deep object relations (Kernberg, 1975, 2004) and the limitations in their capacity for mature love (Kernberg, 2011). The finding from our study is in line with results from the study of Dickinson and Pincus (2003) who claimed, through their examination of both narcissistic types, in general interpersonal problems and attachment, that individuals with grandiose narcissism report having a secure and/or avoidant type of attachment. The additional findings, however, suggest that narcissistic vulnerability plays a more important role on the attachment bond that is formed in romantic relationships, something that was not anticipated in our initial hypothesis. Consequently, the apparent model of grandiosity and the lack of empathy that is clinically observed (APA, 2013), along with the conflicting results of the insecure attachment dimensions that were observed in the present study, are likely to show the unwillingness rather than the inability of individuals with grandiose narcissism to develop emotional intimacy. However, when the grandiose sense of self-significance changes or coexists with vulnerability, a stronger avoidant bond seems to be formed. Therefore, based on the current finding, it is suggested that the presentation of vulnerable narcissism is characterized by negative models of others and controversial models of the self. A negative model of others is parallel to attachment avoidance while a negative model of self is parallel to attachment anxiety (Bartholomew & Horowitz, 1991). Hence, the present finding suggests that, the way whether people with vulnerable narcissistic traits view themselves as worthy of care (model of self), presented inconsistency-controversiality in their level of caregiving (less responsiveness or higher dismissiveness).
Caligor et al. (2015, as cited in Bernardi & Eidlin, 2018) explain that based on their disguised expectations of grandeur, vulnerable narcissists choose to withdraw from social situations when their unfavorable views of their self in relation to others, provoke deep emotions of shame, pain or envy. The same might possibly apply to their romantic interactions. Therefore, despite the fact that their insecure inner representations seek proximity through an anxious attachment (findings of Hypothesis 2), the narcissistic nature of the need for maintaining the delusion of grandiosity often makes it more likely for them to seek to avoid commitment (findings of Hypothesis 3). Hence, the vulnerable narcissistic traits, such as emotional instability and vulnerable self-esteem, might be affected by the oscillation between seeking proximity and avoiding it.
Taking into account the overall findings of the present study, vulnerable narcissism basically covers the grandiose, showing that it constitutes the most central factor for predicting attachment in romantic relationships. Attachment anxiety and avoidance, which were found to have a greater relationship to vulnerable narcissism, than to grandiose, provide further evidence regarding the significance of vulnerable aspects of individuals with narcissistic traits and perhaps underline the need to incorporate different phenotypes of narcissism in psychotherapeutic practice.
Clinical Implications
First and foremost, the present study contributes to the collection of empirical evidence about vulnerable narcissism and its possible effect on romantic relationships. Consequently, the study highlights the significance of extending the existing framework of clinical diagnosis, which as is nowadays presented, represents a perhaps oversimplified diagnostic labeling of narcissistic phenomenology without giving emphasis on vulnerable aspects.
In addition, a clearer distinction between grandiose and vulnerable type of narcissism, with the parallel integration of attachment theory, could provide a base to clinicians who work psychotherapeutically with the relational patterns of their patients. Therefore, examining the clinical implications of the finding which suggests that vulnerable narcissism predicts attachment anxiety, vulnerable narcissistic personalities could possibly express an increasing emotional sensitivity in their patient-therapist interactions. Their unrealistic expectations for significant others and their great vulnerability to others’ emotional reactions could possibly be interpreted as personal weakness or rejection during the therapeutic process. Moreover, since avoidant attachment was also found to be significantly predicted by narcissism, it could be argued that narcissistic organizations are in conflict between the deepest fears of lacking a “secure base” and the distorted perception that the sense of security can be restored through avoiding or rejecting relationships. Hence, this internal uncertainty can be transferred into the therapeutic relationship and negatively affect the therapeutic process (e.g., ruptures in the therapeutic alliance, session cancellations, premature terminations).
Based on the present findings, it can be suggested that clinical implications should be designed toward a phenomenological assessment and formulation of the attachment dimensions of individuals with narcissistic traits. More specifically, great emphasis could be given not only on how previous relationship experiences and current expectations could be utilized to evaluate the emotional quality of a patient’s reactions to their relationships, but also on how those reactions could appear in the form of transference in the therapeutic relationship (Bradley et al., 2005). The dynamic of transference has been particularly supported for its usefulness in the therapy of individuals with NPD (Diamond et al., 2013; Kernberg, 2014; Kernberg et al., 2008). Therefore, following assessment and formulation, specific interventions could be designed by working in and on the transference of patients as a co-created process in the therapeutic relationship.
Overall, the findings of the present study could contribute in helping clinicians understand that there is an essential need to provide a “secure base” in the therapeutic process. Thus, the implementation of these research findings in therapy, could be very beneficial for therapists who work with individuals with a narcissistic personality organization.
Limitations and Future Research
Even though the present study provides some promising evidence, it comes with a number of limitations. Initially, the study was conducted with a small sample of the general non-clinical population and with an uneven distribution of gender (female participants outnumbering males). At the same time, the sample did not reflect cultural diversity since the majority of the participants were Greek-Cypriots. In addition, even though the study refers to participants from the general population, those who had finally participated in the study were not clinically assessed and as a consequence this limits the generalization of the findings. Future studies which will make use of a clinical sample could significantly contribute in understanding the relationship between attachment and pathological narcissism. The exclusive use of self-report questionnaires could also be considered as a limitation, since the findings are strictly limited to the subjective evaluation of the participants. Thus, future research could use semi-structured interviews along with self-reporting data. Finally, the reliability of the Hypersensitive Narcissism Scale (HSNS) in the present study was lower than expected, based on previous studies. Future studies could use factor analysis to check for the better classification of the items in the translated Greek version of the questionnaire.
Conclusion
The concepts of narcissism and attachment are traditionally formulated by different theoretical perspectives, with the majority of research studies drawing this distinction in their empirical examination. The present study however, intended to examine narcissism in relation to attachment theory, examining the role that narcissistic traits play in adult attachment in their romantic relationships.
The findings of the present study, support the idea that narcissistic organizations, whether they have vulnerable or grandiose aspects, present insecurities in their attachments in romantic relationships. Nevertheless, despite the fact that both types of narcissism seemed to be significantly related to insecure attachment dimensions (anxiety and avoidance), vulnerable narcissism showed a greater effect on attachment anxiety and avoidance offsetting the initial significant effect that grandiose narcissism showed. These findings draw the attention to the apparent heterogeneity of narcissistic phenomenology, when both narcissistic types are examined together, and also emphasize the need to incorporate a holistic approach in the evaluation of narcissistic presentations compare to the approach which is used in the current clinical diagnostic manuals.
Competing Interests
The authors have no competing interests to declare.