5a
Scott O. Lilienfeld
For in psychology there are experimental methods and conceptual confusion. The existence of the experimental method makes us think we have the means of solving the problems that trouble us; though problem and method pass one another by.
(Wittgenstein, 1958, p. 232)
Philosopher Ludwig Wittgenstein’s remark reminds us that in psychology and allied fields, methodological sophistication cannot substitute for a lack of conceptual clarity. At the risk of painting with an overly broad brush, one might contend with considerable justification that, with a handful of noteworthy exceptions, the field of personality disorders (PDs) has been marked by growing technological sophistication in the absence of concomitant increases in conceptual depth. Even a casual inspection of recent issues of psychology and psychiatry journals reveals a growing number of articles applying rigorous multivariate methods to ascertain the latent structure of PD measures, adopting multilevel modeling to analyze ecological momentary assessment data among individuals with PDs, and using advanced functional brain imaging techniques to examine the neural correlates of personality pathology. I do not intend this statement to be read as a criticism; indeed, as a co-author, I have contributed to several such articles myself (e.g., Rilling et al., 2007).
Still, I worry that the impressive methodological advances in many domains of psychology and psychiatry largely obscure the fact that when it comes to PDs, we are still very much at sea with respect to a plethora of fundamental conceptual questions (Clark, 2007; Lilienfeld & Latzman, 2018). I am reminded of a peer reviewer of an early paper who accused me and my co-authors of “applying Cadillac methods to Chevrolet data” (the reviewer was right, by the way). Similarly, one may legitimately wonder whether many of us in the PD field are guilty of applying high-tech methods to crude diagnostic categories that map poorly onto psychological reality. In fairness, this broader problem surely extends to the field of psychopathology writ large. Nowadays, scholars tend to be rewarded for publishing high-tech studies in top-tier journals and securing large federal research-related grants to conduct these studies, not for authoring incisive theoretical analyses.
As Paris’ (this volume) scholarly and well-reasoned chapter reminds us, the field of PDs remains riven by a host of controversies, some primarily theoretical and others primarily methodological. I find myself in agreement with many of Paris’ conclusions. For example, I concur with him that some PD categories, such as borderline and antisocial PDs, bear significant clinical implications despite notable shortcomings in their construct validity; I also concur that our rudimentary understanding of the etiology of most or all PDs is a formidable impediment to resolving quandaries in their classification. I also agree that most or all DSM and ICD PDs, like other mental disorders, are marked by unclear boundaries, but that such fuzziness does not by itself vitiate their validity or clinical utility. Although Paris refers to problems of “misdiagnosis” with respect to borderline PD and other PDs, an alternative possibility is that these conditions do not reflect essentialist entities in nature that are detectable with high levels of measurement accuracy. Instead, most or all PDs may be fuzzy densifications or configurations of dimensions in multivariate space, with the boundaries of these patterns being partly a matter of theoretical preference (Hopwood, 2018).
Inspired by Paris’ review, I briefly address three largely unresolved conceptual questions in the PD domain that I believe merit considerably more attention than they have received. The issues echo many of those raised by Paris, but go beyond them in several respects.
Personality Traits versus Personality Disorders
Early in his chapter, Paris cuts to the heart of the matter by posing what may be the most vexing question in the PD literature: “Is there any real difference, other than levels of functioning, between personality traits and PD?” (p. 103 in the previous chapter). As he observes, the question remains largely unsettled despite assertions to the contrary. Certainly, burgeoning data demonstrate that omnibus measures of general personality, such as those of the higher-order and lower-order dimensions of the five-factor model (FFM), account for hefty chunks of variance in DSM and ICD PDs (Costa & Widiger, 2013; Ofrat, Krueger, & Clark, 2018). Still, such findings, robust and well-replicated as they are, do not address a fundamental question: What is a PD? Although researchers have demonstrated that they can account for much of the variance in dimensional PD indices using FFM measures, they have generally shown scant interest in the puzzling question of why only certain combinations of FFM traits, but not others, are tied to personality pathology. In many respects, the reasoning here has been asymmetrical: Investigators have demonstrated that they can largely account for PDs in terms of FFM trait patterns, but they have not explained why only a subset of FFM trait patterns are relevant to PDs.
One intriguing possibility, which merits further investigation, is that only certain patterns of FFM traits are tied to personality pathology because these patterns, but not most others, are linked to interpersonal dysfunction. This hypothesis is broadly consistent with findings that antagonism (low agreeableness), which is an inherently interpersonal dimension, courses through most or virtually all DSM PDs (Saulsman & Page, 2004). This hypothesis also dovetails with the possibility that at least some DSM PDs partly reflect “folk concepts” of interpersonal dysfunction, or readily recognized configurations of personality traits that are interpersonally relevant to us because we find them to be aversive or otherwise challenging (see also Lilienfeld & Latzman, 2018; Tellegen, 1993). For example, the folk concept of psychopathic personality (psychopathy) and, to a lesser extent, the overlapping operationalization of antisocial PD in the DSM, may partly mirror the folk concepts of the confidence artist, wolf in sheep’s clothing, and two-faced person, stemming from the distinctive and paradoxical admixture of traits (poise, self-confidence, and superficial charm, on the one hand, and callousness, guiltlessness, and interpersonal detachment, on the other) observed in these conditions (Lilienfeld, Watts, Smith, & Latzman, 2018). If the conception of PDs as reflections of folk concepts has merit, it would suggest that investigators should more actively pursue the possibility of multiplicative (interactive) rather than purely additive relations among PD features, as certain patterns of personality pathology may reflect interpersonally confusing configurations of personality traits that are themselves only weakly or even negatively correlated (Grove & Tellegen, 1991; Lilienfeld, 2013).
Basic Tendencies versus Characteristic Adaptations
Although not addressed explicitly by Paris, another reason to doubt the purported synonymy between personality traits and PDs is that many PDs appear to be complex admixtures of basic tendencies and characteristic adaptations. Admittedly, the distinction between these two concepts is probably one of degree rather than of kind, but it is still useful for theoretical and pragmatic purposes. Basic tendencies are underlying personality traits, whereas characteristic adaptations are the behavioral manifestations of these traits, reflecting people’s typical ways of adapting to their own dispositions (DeYoung, 2015; McCrae & Costa, 1995). This distinction is consistent with the longstanding observation that comparable levels of the same personality traits can be expressed in a variety of short-term behaviors and long-term lifestyle choices that differ in their adaptivity versus maladaptivity. For example, levels of sensation seeking are substantially elevated among both firefighters and incarcerated criminals, raising the possibility that this trait can be expressed alternatively in prosocial outcomes, antisocial outcomes, or both, depending on still unidentified moderating variables (Harkness & Lilienfeld, 1997).
Consider the DSM-5 criteria (American Psychiatric Association, 2013) for antisocial PD. A few of these criteria, such as impulsivity, appear to refer primarily to basic tendencies, whereas most others, such as reckless disregard for others’ safety and financial/work irresponsibility, appear to refer primarily to characteristic adaptations. For example, in some individuals, impulsivity can be manifested in antisocial behaviors, such as recklessness, whereas in others it can be manifested in prosocial and perhaps even heroic actions (e.g., Neria, Solomon, Ginzburg, & Dekel, 2000; Patton, Smith, & Lilienfeld, 2018). As a consequence, it is prima facie implausible that general personality traits alone will ever map entirely onto the PD criterion space. Personality traits can be expressed in a myriad of potential behavioral phenotypes as a function of other (presumably interacting) variables, and only a subset of these phenotypes is relevant to personality pathology. This conceptual problem, which has not received the attention it warrants, dovetails with Paris’ point that “it remains unclear as to whether trait dimensions can account for the prominent symptoms seen in some PDs, especially the borderline category” (p. 103 in the previous chapter).
This conceptual framework is also broadly consistent with findings that across the DSM PDs, diagnostic criteria that are more trait-like tend to be more stable over time than criteria that are more behavioral. For example, in borderline PD, affective instability – ostensibly a proxy for the personality trait of neuroticism – tends to be more temporally consistent than self-harming behavior (McGlashan et al., 2005). The former feature is more likely to reflect a basic tendency, whereas the latter is more likely to reflect a maladaptive characteristic adaptation to this tendency.
Network Models
An assumption underpinning most models of PDs is that these conditions are reflective constructs, meaning that they lie causally downstream of latent variables, such as personality traits (Coltman, Devinney, Midgley, & Venaik, 2008). Nevertheless, recent theoretical and empirical research not addressed by Paris raises the intriguing possibility that certain mental disorders may be better accommodated by network models. According to these models, mental disorders are not caused by underlying variables; instead, these disorders can be viewed as complex networks comprising features that exert bi-directional influences on one another (Borsboom & Cramer, 2013). For example, in major depression, lack of sleep may contribute to concentration and memory disturbance; in panic disorder, shortness of breath may contribute to paresthesias (numbness or tingling in the extremities). Such models are a key future avenue for research on the etiology of PDs – an overarching direction highlighted by Paris in his concluding comments.
It is implausible that network models can account entirely for the features of all PDs, as such a view would imply, for example, that the latent trait of neuroticism plays no causal role whatsoever in avoidant, borderline, or dependent PDs. Such an extreme position harkens back to radical behaviorism, which denies any causal role for personality traits (Skinner, 1974). This view is also difficult to reconcile with behavior genetic findings demonstrating that neuroticism shares substantial amounts of genetic variance with borderline PD (Kendler, Myers, & Reichborn‐Kjennerud, 2011), suggesting that that this condition is caused by more than the bi-directional relations among its features.
At the same time, it is plausible, if not likely, that network models will help to account for some of the phenomenology of some PDs. For example, in the case of avoidant PD, chronic feelings of social ineptitude seem likely to contribute to fear of intimate social contacts; in the case of paranoid PD, the tendency to read hidden malignant meanings into ambiguous remarks seems likely to contribute to reluctance to confide in others (see also Preszler, Marcus, Edens, & McDermott, 2018 and Verschuere et al., 2018 for the application of network models to psychopathy). In this this respect, hybrid models, which posit that latent variables predispose to certain core features of psychological disorders, but that these features in turn often influence each other bi-directionally (Fried & Cramer, 2017), would seem to be well worth pursuing.
Concluding Thoughts
As Paris notes, continued progress in the PD field hinges at least partly on resolution of ongoing conceptual controversies, some of which can be informed by data. In this respect, it would behoove those of us in this discipline to spend more time thinking about deeper conceptual questions and less time thinking about applied questions, such as the optimal factor structure of, or diagnostic criteria for, DSM or ICD PDs (Grove & Tellegen, 1991). At the very least, we should bear in mind Wittgenstein’s maxim that sophisticated research methods, enormously useful as they are, will not by themselves heal all our woes.
References
American Psychiatric Association. (2013). Diagnostic and Statistical Manual of Mental Disorders (5th ed.). Arlington, VA: American Psychiatric Publishing.
Borsboom, D., & Cramer, A. O. (2013). Network analysis: An integrative approach to the structure of psychopathology. Annual Review of Clinical Psychology, 9, 91–121.
Clark, L. A. (2007). Assessment and diagnosis of personality disorder: Perennial issues and an emerging reconceptualization. Annual Review of Psychology, 58, 227–257.
Coltman, T., Devinney, T. M., Midgley, D. F., & Venaik, S. (2008). Formative versus reflective measurement models: Two applications of formative measurement. Journal of Business Research, 61, 1250–1262.
Costa, P. T., & Widiger, T. A. (Eds.) (2013). Personality Disorders and the Five Factor Model of Personality (3rd ed.). Washington, DC: American Psychological Association.
DeYoung, C. G. (2015). Cybernetic big five theory. Journal of Research in Personality, 56, 33–58.
Fried, E. I., & Cramer, A. O. (2017). Moving forward: Challenges and directions for psychopathological network theory and methodology. Perspectives on Psychological Science, 12, 999–1020.
Grove, W. M., & Tellegen, A. (1991). Problems in the classification of personality disorders. Journal of Personality Disorders, 5, 31–41.
Harkness, A. R., & Lilienfeld, S. O. (1997). Individual differences science for treatment planning: Personality traits. Psychological Assessment, 9, 349–360.
Hopwood, C. J. (2018). Interpersonal dynamics in personality and personality disorders. European Journal of Personality, 32, 499–524.
Kendler, K. S., Myers, J., & Reichborn‐Kjennerud, T. (2011). Borderline personality disorder traits and their relationship with dimensions of normative personality: A web‐based cohort and twin study. Acta Psychiatrica Scandinavica, 123, 349–359.
Lilienfeld, S. O. (2013). Is psychopathy a syndrome? Commentary on Marcus, Fulton, and Edens. Personality Disorders: Theory, Research, and Treatment, 4, 85–86.
Lilienfeld, S. O., & Latzman, R. D. (2018). Personality disorders: Current scientific status and ongoing controversies. In J. N. Butcher & J. M. Hooley (Eds.), Psychopathology: Understanding, Assessing, and Treating Adult Mental Disorders (pp. 557–606). Washington, DC: American Psychological Association.
Lilienfeld, S. O., Watts, A. L., Smith, S. F., & Latzman, R. D. (2018). Boldness: Conceptual and methodological issues. In C. J. Patrick (Ed.), Handbook of Psychopathy (2nd ed., pp. 165–188). New York: Guilford Press.
McCrae, R. R., & Costa, P. T., Jr. (1995). Trait explanations in personality psychology. European Journal of Personality, 9, 231–252.
McGlashan, T. H., Grilo, C. M., Sanislow, C. A., Ralevski, E., Morey, L. C., Gunderson, J. G., … & Stout, R. L. (2005). Two-year prevalence and stability of individual DSM-IV criteria for schizotypal, borderline, avoidant, and obsessive-compulsive personality disorders: Toward a hybrid model of axis II disorders. American Journal of Psychiatry, 162, 883–889.
Neria, Y., Solomon, Z., Ginzburg, K., & Dekel, R. (2000). Sensation seeking, wartime performance, and long-term adjustment among Israeli war veterans. Personality and Individual Differences, 29, 921–932.
Ofrat, S., Krueger, R. F., & Clark, L. A. (2018). Dimensional approaches to classification. In W. J. Livesley (Ed.), Handbook of Personality Disorders (2nd ed., pp. 72–87). New York: Guilford Press.
Patton, C. L., Smith, S. F., & Lilienfeld, S. O. (2018). Psychopathy and heroism in first responders: Traits cut from the same cloth? Personality Disorders: Theory, Research, and Treatment, 9, 354–368.
Preszler, J., Marcus, D. K., Edens, J. F., & McDermott, B. E. (2018). Network analysis of psychopathy in forensic patients. Journal of Abnormal Psychology, 127, 171–182.
Rilling, J. K., Glenn, A. L., Jairam, M. R., Pagnoni, G., Goldsmith, D. R., Elfenbein, H. A., & Lilienfeld, S. O. (2007). Neural correlates of social cooperation and non-cooperation as a function of psychopathy. Biological Psychiatry, 61, 1260–1271.
Saulsman, L. M., & Page, A. C. (2004). The five-factor model and personality disorder empirical literature: A meta-analytic review. Clinical Psychology Review, 23, 1055–1085.
Skinner, B. F. (1974). About Behaviorism. New York: Vintage.
Tellegen, A. (1993). Folk concepts and psychological concepts of personality and personality disorder. Psychological Inquiry, 4, 122–130.
Verschuere, B., van Ghesel Grothe, S., Waldorp, L., Watts, A. L., Lilienfeld, S. O., Edens, J. F., … Noordhof, A. (2018). What features of psychopathy might be central? A network analysis of the Psychopathy Checklist-Revised (PCL-R) in three large samples. Journal of Abnormal Psychology, 127, 51–65.
Wittgenstein, L. (1958). Philosophical Investigations (3rd ed., trans. G. E. M. Anscombe). Englewood Cliffs, NJ: Prentice Hall.