Understanding Obsessive-Compulsive Personality Disorder

Problems of Specificity: Polythetic Criteria

As reviewed by Pfohl and Blum (1991), Diagnostic and Statistical Manual of Mental Disorders (3rd ed.; DSM-III; APA, 1980) diagnostic criteria for OCPD were altered by the Diagnostic and Statistical Manual of Mental Disorders (3rd ed., rev.; DSM-III-R; APA, 1987) “Personality Advisory Group,” which “chose to expand the description to nine criteria, of which only five were required” (Pfohl & Blum, 1991, p. 372). Specifying nine criteria, of which only five are required, invited significant heterogeneity of type. In other words, little more than half of the specified criteria were required for making a diagnosis; as such, two relatively distinct combinations of criteria could be developed. Nevertheless, even the most distinct forms of OCPD shared at least one common symptom. Unfortunately, this trend was not only perpetuated but also exaggerated in the subsequent revision such that DSM-IV (APA, 1994) lists eight criteria, the presence of any four of which constitutes a positive diagnosis of OCPD. The polythetic format, according to Pfohl and Blum, “allowed for greater variation between patients in how the personality disorder was expressed” (p. 372). Pfohl and Blum additionally note that the expansion of criteria increased the intersection with other personality disorders. Accordingly, the diagnostic criteria for obsessive-compulsive personality collectively have “poor psychometric properties” (Hummelen, Wilberg, Pedersen, & Karterud, 2008, p. 446) with OCPD showing substantial overlap across all other personality disorders leading 77% of Hummelen’s OCPD participants to earn a concurrent personality disorder diagnosis.

This pattern is not confined to OCPD but is endemic to personality disorder diagnosis generally. For instance, Ryder, Costa, and Bagby (2007) cite “excessive within-disorder diagnostic heterogeneity” to be “well known” in personality disorder diagnosis (p. 626). A brief examination of the borderline personality will illustrate the problem of rampant heterogeneity. Trull and McCrae (2002) cite the plurality of symptom combinations for Borderline Personality Disorder, describing it as a “polythetic format for diagnosing BDL [Borderline Personality Disorder]” (p. 48)

. . . there are literally 93 ways to be diagnosed as having BDL (eight items taken five or more at a time). Because any five of the eight criteria can satisfy the BDL diagnostic decision rule, BDL patients are heterogeneous with respect to clinical symptomology. (Trull & McCrae, 2002, p. 48)

Trull and McCrae (2002) expand upon the weaknesses of the polythetic approach by describing the diverse clinical presentations that can result:

For example, a patient can meet the criteria for BDL by exhibiting symptoms of inappropriate-intense anger, recurrent suicidal threats-behavior, identity disturbance, chronic feelings of emptiness-boredom, and frantic efforts to avoid real-imagined abandonment, without showing the unstable-intense interpersonal relations, impulsivity, and affective instability that many clinicians would consider to be hallmarks of the BDL diagnosis. However, another BDL patient might manifest the latter three symptoms as well as inappropriate intense anger and chronic feelings of emptiness-boredom but not suicidal threats-behaviors, identity disturbance, or frantic attempts to avoid abandonment. A comparison of the two hypothetical BDL patients reveals markedly different clinical pictures and may suggest different treatment approaches. (Trull & McCrae, 2002, p. 48)

Trull and McCrae (2002) conclude that “ . . . the polythetic system for diagnosing BDL is a breeding ground for clinical heterogeneity” (p. 48). For the obsessive, this problematic trend is augmented. While 93 different borderline combinations can be produced using five of eight criteria, Hummelen et al. (2008) find that permutations of eight, yield “163 possible combinations of any 4 OCPD criteria” (p. 451). Of course, simply because a combination of symptoms can be produced, it does not mean that it will be produced. Nevertheless, Hummelen et al. found “80 different combinations” within their sample, concluding that OCPD is “heterogeneous within its borders.”

Most unsettling are those fully disparate types formed of mutually exclusive symptom sets (Samuel, Riddell, Lynam, Miller, & Widiger, 2012). Mutually exclusive types cannot be created when diagnosing BPD. In fact, of the 10 extant personality disorders, the obsessive and antisocial personality disorders are the only 2 that allow for mutually exclusive criteria sets (APA, 2000). Mutually exclusive types engender incompatible types (Gibbs Gallagher, South, & Oltmanns, 2003) as the following illustration demonstrates: Variant I presents as a relentlessly dedicated business owner who is personally involved with every aspect of his company because he is reluctant to delegate tasks to subordinates and when he does so, he insists that the work be executed according to exactly his way of doing things (Criterion 6). Consistent with this pattern, Variant I insists that both he and his employees actualize company aspirational goals and operate according to the letter of state standards (Criterion 4). Variant I’s relationship with his family is marginalized by his late hours and his excessive devotion to work and productivity (Criterion 3), while his employees lament his miserly spending style toward both self and employees (Criterion 7). Refuting these charges, Variant I insists that such frugality enabled him to gather the start-up capital necessary to establish his business; he further insists that this same frugality will ensure the survival of his business through the years of privation that are inevitable in the booms and busts of a free market economy.

In contrast, Variant II is a rigid and stubborn (Criterion 8) isolated retiree. He offers to build furniture for others, but becomes so perfectionistic about details that he rarely completes any pieces (Criterion 2). The majority of the neighbors are disturbed by his propensity to hoard worthless objects that seemingly have no material or sentimental value (Criterion 5). Collections of such useless objects crowd his garage and driveway, ruining the aesthetics of the community and ostensibly bringing down property value. One frustrated neighbor is having trouble selling her house because, in her estimation, Variant II’s house is a deterring eyesore. Finally, Variant II spends most of his time unproductively; he seems preoccupied with details and lists, which, rather than aid in the completion of any project, actually interfere with the major point of the activity (Criterion 1).

In conclusion, Variant I is a driven and conscientious entrepreneur who earns grudging respect while providing jobs to employees, goods to consumers, and revenue to the government; whereas Variant II is simply odious to his neighbors and shades into the odd eccentric spectrum. Variants I and II can conceivably be differentiated by self-esteem, functionality, confidence, and identity: Variant I takes substantial pride in his work, acts powerfully in his environment, and identifies strongly with his role, while Variant II can claim no such competence or corresponding confidence. More importantly, Variant I differs from Variant II in basic personality features, such as conscientiousness, that should show consistency within the category. Accordingly, diagnostic challenges mount: (a) There is an obvious lack of concordance between the two presentations despite the shared diagnosis; (b) Variant I meets four criteria and so qualifies for a diagnosis of OCPD, but remains productive, satisfied, and a valuable member of the community in a manner that is inconsistent with the reduced functionality expected to be present across personality disorders; (c) Variant II’s presenting pattern, with its isolation and eccentricity, overlaps with Cluster A personality pathology. In summary, the polythetic approach sanctions diagnoses that are both incompatible with each other and incompatible with the obsessive archetype.

Problems of Specificity: Lack of a Hallmark Feature

Medin and Ortony (1989) recognize that “the notion of a continuum of centrality linking deeper and more superficial properties,” creates intelligibly constructed classification (p. 185). “In the absence of deeper principles to link more superficial properties, categories constructed only in terms of characteristic properties or family resemblances may not be psychologically coherent” (Medin & Ortony, 1989, p. 185) A hallmark symptom is representative of Medin and Ortony’s deep principle; a coherence granting symptom around which all others can subordinately gather. While the DSM-IV (APA, 1994) Axis II structure does not use hallmark symptoms, some Axis I disorders do use hallmark symptoms to good effect. Of the nine criteria for a major depressive episode, for example, the first two are given primacy. While five of the nine criteria are needed for a diagnosis, it is not sufficient to exhibit the last five, or even the last seven criteria, because one cannot be truly depressed without exhibiting depressed mood most of the day (Criterion 1) or anhedonia (Criterion 2; APA, 2000). As the OCPD criteria list is currently formulated, there is no hallmark symptom, no trait, feature, or quality that hierarchically organizes all others; no psychological, interpersonal, behavioral, or cognitive pattern that is invariably and essentially true of the obsessive personality. In this regard, there is a mismatch between extant literature (as reviewed in the three subsequent paragraphs) and its translation into DSM-IV (APA, 1994) criteria.

Critical reviews of OCPD criteria have separated central traits from their more peripheral counterparts. These investigations tend to divide the symptom set into roughly equal halves. For example, Fossati et al. (2006) factor analyzed OCPD criteria and found that the following three symptoms best capture the essence of the disorder: “shows rigidity and stubbornness [Criterion 8], is excessively devoted to work [Criterion 3] and is over-conscientious [Criterion 4]” (p. 201). Another research group, Hummelen et al. (2008), find that the first four criteria were robust and diagnostically efficient. Criterion 6 was, to a lesser extent, favored for its sensitivity. Collectively considering the diagnostic criteria, Hummelen et al. summarize their findings by stating that Criteria 5, 7, and 8 performed poorly as compared with Criteria 1, 2, 3, 4, and 6. Amid these short lists of symptoms lies a candidate hallmark symptom. While there are some discrepancies between the reviews, both research groups cite Criteria 3 and 4 as diagnostically efficient. And of course Criterion 3 (is excessively devoted to work) is simply a terse definition of Criterion 4 (is overconscientious). In other words, to say that one is excessively devoted to work is nearly synonymous with saying that one is overconscientious. Samuel and Widiger (2010) assert,

Aspects of conscientiousness are evident among most of the American Psychiatric Association (2000) diagnostic criteria, including preoccupation with details, rules, lists, and order (i.e., the FFM facet of order); perfectionism (i.e., an excessive emphasis on competence); devotion to work and productivity (i.e., excessive achievement striving); and even a criterion that refers explicitly to conscientiousness. (p. 238)

As such, it seems that conscientiousness is, not surprisingly, a potential hallmark symptom.

The work of Morey et al. (2003) suggests that conscientiousness uniquely defines OCPD. Morey and colleagues measure all DSM-IV (APA, 1994) personality disorders using the Schedule for Adaptive and Nonadaptive Personality, which contains scales for “propriety” and “workaholism.” Together, propriety and workaholism encompass the supertrait of conscientiousness. In this study, only the paranoid and obsessive personalities were significantly positively correlated with propriety and only the avoidant and obsessive personalities were significantly positively correlated with workaholism. Accordingly, it is only the obsessive personality that correlated significantly with both propriety and workaholism—the two aspects of conscientiousness. This, at once, suggests that conscientiousness can identify the obsessive personality and distinguish it from other personality disorders.

In their review of eight clinical inventories, Samuel and Widiger (2010) find that five inventories properly represent conscientiousness, and fault those that fail to do so. The Minnesota Multiphasic Personality Inventory (MMPI-II), for example, is rare in that it fully excludes conscientiousness. Consequently the MMPI-II is compared negatively with the Millon Clinical Multiaxial Inventory (MCMI-III), which holds conscientiousness preeminent (Samuel & Widiger, 2010). Furthermore, whereas critical reviews of diagnostic criteria and assessment instruments nominate conscientiousness as a candidate hallmark symptom, trait investigations firmly elect conscientiousness.

Lexically driven trait research has more directly and consistently highlighted conscientiousness as the hallmark feature of OCPD (Deary, Peter, Austen, & Gibson, 1998; Morey et al., 2003; Samuel & Widiger, 2010). Researchers such as Widiger, Trull, Clarkin, Sanderson, and Costa (2002) describe OCPD as “primarily a disorder of excessive Conscientiousness” (p. 97). Moreover, consistently across research studies, OCPD is reliably associated with elevations across all facets of conscientiousness: competence, order, dutifulness, achievement striving, self-discipline, and deliberation (Furnham & Crump, 2005; Lynam & Widiger, 2001; Samuel & Widiger, 2011). No other trait, in the number of facet elevations, in the strength of correlation, or in the reliability of its association, equals the ability of conscientiousness to describe OCPD. Both in retrospect and in prospect, conscientiousness emerges as the sought-after hallmark feature. Diagnostic specificity depends on recognizing its preeminence. A polythetic format with no hallmark feature countenances OCPD diagnosis through the secondary avenues of rigidity, frugality, mistrust, and hoarding, without expression of the essential pattern of conscientiousness. An OCPD criteria set that does not invariably require conscientiousness is like a criterion set for a mammal that does not invariably require lactation or thermoregulation. OCPD diagnosis should consequently hold conscientiousness necessary, though not sufficient.

Problems of Sensitivity: Missing Elements

One has a better chance of diagnosing a disorder if a full complement of signs and symptoms is available. Cognizance of all signs and symptoms of a disorder enhances sensitivity and limits false negatives. Generally speaking, the criteria list compiled by the APA has judiciously winnowed the wheat from the chaff, retaining only the most representative aspects of the type. Nevertheless, Ansell, Pinto, Edelen, and Grilo (2008) suspects that DSM-IV (APA, 1994) OCPD criteria do not “capture the full breadth of the construct” (p. 863). A review of the literature, theoretical and empirical, old and new, suggests that there are three symptoms that, despite impeccable credentials, have not been properly regarded: attention, future orientation, and affect modulation.

The cause of obsessive attention was championed by Shapiro (1965/1999, 2002). It was Shapiro who identified the obsessive as having little “volitional mobility of attention.” It was Shapiro (1999) who described obsessive attention as “sharp,” but correspondingly “ . . . limited in both mobility and range” (p. 27). Shapiro contrasted the focused attention of the obsessive personality, which is drawn to minute physical details, with the impressionistic attention of the histrionic personality, which is open to global social signals (Yovel, Revelle, & Mineka, 2005). Truly, obsessive attention is focused as if it was trained by a telescope on a small patch of sky, or trained by a microscope on a tissue sample. Shapiro’s writings on obsessive attention have continued relevance, as evidenced by Yovel et al.’s (2005) recent article “Who sees trees before forest? The obsessive compulsive style of Visual Attention.” As the title implies, obsessives fixate on details to the detriment of the whole. Yovel and colleagues demonstrated that obsessives had more difficulty than controls in reading large letters composed of smaller letters. In this Stroop-like task of attention, obsessives saw the trees more vividly than the forest. Preferential perception of “small local aspects of stimuli” (Yovel et al., 2005, p. 127) is also expressed by obsessives on the Rorschach (Schneider, 2006). Similarly, a study by Gibbs Gallagher et al. (2003) titled “Attentional Coping Style in Obsessive Compulsive Personality Disorder: A Test of the Intolerance of Uncertainty Hypothesis” finds obsessive attention to be channeled as a function of an information seeking bias. Yet the expression of obsessive attention is not confined to formally administered measures of executive functioning and perception. No, the intensity and rigidity of obsessive attention is an influential outgrowth of compulsivity (Fineberg et al., 2010). Along with traits like anxiety, the conscientiousness and compulsivity of OCPD elicit a corresponding constriction of focus (MacDonald, 1995) that inhibits responsiveness to feedback (Fineberg et al., 2010) and impedes the “hierarchical ordering of information based upon relative importance” (Mallinger, 2009, p. 111). The specifics of obsessive attention, in short, are expressed across an array of cognitive (Fineberg et al., 2010; MacDonald, 1995), perceptual (Schneider, 2006; Yovel et al., 2005), and learning situations (Mallinger, 2009) and so embody an important part of the obsessive archetype that can be used diagnostically.

Future orientation has an extensive history. It is documented in the work of Campos (1966) and Pettit (1969) and highlighted by Pollak (1979). Yet interest in this definitional trait has waned. Nonetheless, the OCPD literature suggests that it can be used as a steady diagnostic marker. Pollak’s (1979) estimate of obsessive future orientation continues true despite decades of neglect: “ . . . [there is a] positive relationship between degree of anality and the tendency to use time in a niggardly, thrifty and cautious manner” (p. 237). Eskedal and Demitri (2006) understand the obsessive to be conspicuously trained on the future (Eskedal & Demitri, 2006). Saltzman devotes a section of his book, Treatment of the Obsessive Personality, to time: “Time and the Obsessional.” Time is the “enemy of all” the obsessives’ “plans and programs”; it is the “enemy to be fought and overcome” (Saltzman, 1985, p. 80). By studying reactions to “ego threatening events,” namely measures of cognition and spatial ability, Gibbs Gallagher et al. (2003) empirically show obsessive personality traits to be positively correlated with (a) information seeking about future trials, (b) preparatory activity, and (c) monitoring. In this way, theory and research combine to suggest that the present moment is habitually used in service of the future need. The obsessive, not being able to hoard time as he hoards other resources, will spend it reluctantly. The pleasure of the present is subjugated to the urgency of the future; a pattern that will be expressed in language and narratives as well as acts and behavior. The obsessive’s habitual future orientation directs compulsive action and spoils, in conspicuous excess, the pleasure of the present.

Affect modulation has a similarly distinguished pedigree stemming at least as far back as Segal (1961). Affect modulation was again discussed by Paykel and Prusoff (1973) and then highlighted in Pollak’s (1979) review at decade’s end. Years later, Millon and Davis (1996) revived interest in disordered obsessive affect modulation. Thereafter, much research has ensued (Gibbs Gallagher et al., 2003; Greve & Adams, 2002; Hummelen et al., 2008; Villemarette-Pittman, Stanford, Greve, Houston, & Mathias, 2004). The aggregate of such recent research suggests that obsessives do not spontaneously express emotion with full range, and sometimes even recoil from untrammeled emotional expression in others. Even before such research was undertaken, the authors of DSM-III (APA, 1980) recognized constrained emotional expression as a diagnostic marker and elevated its importance above other clinical features (Samuel & Widiger, 2010).

Overcontrol is the habitual obsessive affective state, but the obsessive manifests affect modulation difficulties in the fullest sense of the term: There are problems both of overcontrol and of dyscontrol. This pattern appears to be reflected in the study by Morey et al. (2003) using the Schedule for Adaptive and Nonadaptive Personality, in which the obsessive personality evidenced a strong negative correlation with “disinhibition” and “impulsivity” and yet a modest positive correlation with “aggression.” Millon and Davis (1996) succinctly capture this dichotomy in the following metaphor: “Appearing deliberate and well poised on the surface, the compulsive sits atop this tightly constrained but internal powder keg” (p. 517). These aberrant lapses from overcontrol into discreet bursts of dyscontrol (Villemarette-Pittman et al., 2004), often result in intense anger and overt aggression (Greve & Adams, 2002). Both in its overcontrol and in its lapses into dyscontrol, obsessive affect modulation is distinctive, prominent, and reliably expressed, suggesting that it will dutifully serve as a diagnostic marker.

DSM-IV (APA, 1994) criteria for the dramatic, emotional, and erratic Cluster B personalities effectively employ problems of affect modulation as diagnostic markers. Obsessive affect modulation is no less marked; it simply resides on the opposite end of the spectrum. For instance, the impulsive and mercurial borderline personality reacts limbically with little cortical oversight, whereas the compulsive and rigid obsessive personality reacts cortically with little limbic expression. The borderline is habitually impulsive, exhibiting a persistent state of affective dyscontrol; the obsessive is habitually compulsive, exhibiting a persistent state of affective overcontrol. Both patterns of affect modulation are remarkably different from that which is observed in the nonpersonality disordered person, and both patterns are important features of their respective disorders.

Problems of Sensitivity: Concrete Expression

This section exclusively addresses Criteria 5 and 7, both of which are described as inefficient by Fossati et al. (2006), a research group that advocates for their exclusion from the next DSM revision:

Additionally, OCPD criterion five and criterion seven did not load on the expected latent dimension and were not efficient in discriminating the OCPD criteria latent dimension from the other two Cluster C PD criteria latent dimensions. This finding strongly suggests that future revisions of the DSM exclude “Is unable to discard worn-out objects” (criterion five) and “Adopts a miserable [miserly] spending style” (criterion seven) from the list of OCPD criteria. (Fossati et al., 2006, p. 200)

Similarly, Seedat and Stein (2002) question the relationship between OCPD and hoarding:

While hoarding is recognized as one of the eight symptom criteria of OCPD and may be linked with other OCPD criteria such as perfectionism, studies have failed to show any significant differences between hoarders and community controls on the OCPD subscale of the Millon Multiaxial Clinical Inventory–II (MCMI-II). (Seedat & Stein, 2002, p. 18)

Likewise, Mataix-Cols and colleagues (2010) state “there is remarkably little empirical evidence to support the inclusion of hoarding as one of the OCPD criteria” (p. 562). They support this assertion by showing hoarding, and to a lesser extent miserliness, to be too weakly related to the OCPD construct. Initially, they describe independent criteria correlations; afterwards, they describe factor analytic solutions: First, while intercorrelations among other OCPD criteria are moderate (.35-.62), hoarding showed poor (.19-.28) correlations with other symptoms. Second, factor analytic studies find three factors—“rigidity” and “perfectionism,” which are respectably intercorrelated (.51), and a third factor comprising hoarding and miserliness that is not strongly correlated with either rigidity (.27) or perfectionism (.35).

Finally, Hummelen et al. (2008) conclude that hoarding and stinginess are not essential elements of OCPD and suggest that the elimination of these criteria would increase the coherence of the OCPD construct. Of these, it was the problematic Criterion 5 that proved to have the lowest “diagnostic efficiency.” Hoarding behavior, according to Hummelen et al. (2008) has dismal “psychometric properties” and fails to reliably relate to OCPD because it is a complex and “multifaceted” symptom that contains the following components: “information processing deficits, emotional attachments, behavioral avoidance and erroneous beliefs about the nature of possessions” (p. 453).

Clearly, research demonstrates that, as they are currently formulated, Criteria 5 and 7 hinder more than they help the diagnostic endeavor. Nevertheless, we must not excise the cancer at the cost of killing the patient. Removing Criteria 5 and 7 would amount to dismantling the anal triad. Truly, the three-legged stool of Freudian theory would fall if deprived of any of its supports; only orderliness and stubbornness would be left to represent OCPD without the aid of parsimony. As Freud understood, parsimony is one of three piers upon which the obsessive edifice rests (Gay, 1989). Yet, parsimony, although it is a defining disposition, need not be expressed concretely as miserliness or hoarding.

As Hummelen et al. (2008) state, hoarding, as is the case for miserliness, is a multiply determined behavior. Other aspects of the person, culture, cognition, experience, perception, education, and past combine with personality to create such specific behaviors. Accordingly, the personality feature of parsimony is correlated with miserliness and hoarding, but the relationship is not unambiguously causal. Furthermore, even when hoarding and miserliness are present, they will be so differentially expressed as to be unreliably identified. For example, an obsessive is just as likely to hoard incommodious objects of no worth as to hoard money which is valued, electronic files which are inconspicuous, or stamps which are innocuous. Similarly, miserliness is not invariably directed at money, but extends to any valued object and even to time and information. The specific culture and identity of the obsessive will prescribe the resource which he jealously guards and which he niggardly spends. This is consistent with Grilo et al.’s (2004) conclusions, who, after studying the stability of OCPD across four intervals spanning 2 years, find OCPD to be structurally stable, not so much in the expression of particular behaviors but in respect to trait dispositions.

It is only the trait of parsimony, with its disposition toward preparation and conservation, which is reliably expressed across person, place, time, and culture. While other OCPD criteria describe dispositions, parsimony was abandoned as a diagnostic marker in its own right in favor of two particular behavioral proxies: hoarding worthless objects and conserving money. After all, as currently formulated in DSM-IV (APA, 1994), conscientiousness is not directed at a particular end, perfectionism does not interfere with a particular task, rigidity and stubbornness are not centered on a particular circumstance—so what makes parsimony any different; why should this disposition be constrained to the expression of hoarding worthless objects and spending money like a miser? Simply because parsimony is not reliably expressed within such confines says nothing about its archetypal importance and diagnostic potential. Truly, we need not be so literal, so rigidly empirical. We must accept covert symptoms where overt signs are wanting (Millon, 2010). Simply observing, while it is always admirably objective, is not always sufficiently inferential (Ryder et al., 2007); some degree of interpretation must be made; some degree of synthesis must be allowed (Medin & Ortony, 1989).

The crux of Medin and Ortony’s (1989) psychological essentialism is that what is deeply ingrained, but largely obscured, can nonetheless dictate what is superficially expressed, but distally definitional:

Central to the position we advocate, which we call psychological essentialism, is the idea that these surface features are frequently constrained by, and sometimes generated by, the deeper, more central parts of concepts. Thus there is often a nonarbitrary, constraining relation between deep properties and more superficial ones. (Medin & Ortony, 1989, p. 180)

Medin and Ortony (1989) illustrate their notion of psychological essentialism by invoking a comparative analysis between a whale and a bear. The whale, to the casual observer relying on superficial form, might be thought a fish. Looking more deeply, observing the suckling calf and anatomical features of the forelimb reveal the mammalian elements that connect the whale, not to his aquatic brethren but to the terrestrial bear. As evolutionary biologists look past superficial features, finding homologous structures in the bat’s wing, the monkey’s hands, the bear’s paw, and the whale’s fin, it is important for psychological diagnosticians to look past superficial features (rigidly specified behaviors such as hoarding), but still require continuity across basic dispositions and traits (parsimony). In applying such logic to the obsessive personality, we recognize that hoarding of worthless objects and miserliness are simply “surface features” that are “constrained by” or arise from a more essential element (Medin & Ortony, 1989, p. 185).