Psychological Bulletin

1998, Vol. 124, No. 3, 333-371

Copyright 1998 by the American Psychological Association, Inc.

0033-2909/98/$3.00

The Scientific Legacy of Sigmund Freud:

Toward a Psychodynamically Informed Psychological Science
Drew Westen
Harvard Medical School
and The Cambridge Hospital/Cambridge Health Alliance
Although commentators periodically declare that Freud is dead, his repeated burials lie on shaky
grounds. Critics typically attack an archaic version of psychodynamic theory that most clinicians
similarlyconsider obsolete. Central to contemporarypsychodynamictheory is a series of propositions
about (a) unconscious cognitive, affective, and motivational processes; (b) ambivalence and the
tendency for affective and motivational dynamics to operate in parallel and produce compromise
solutions; (c) the origins of many personality and social dispositions in childhood; (d) mental
representations of the self, Others, and relationships; and (e) developmentaldynamics.An enormous
body of research in cognitive, social, developmental,and personalitypsychologynow supports many
of these propositions. Freud's scientific legacy has implications for a wide range of domains in
psychology, such as integration of affective and motivationalconstraints into connectionist models
in cognitive science.

Freud, like Elvis, has been dead for a number of years but
continues to be cited with some regularity. Although the majority
of clinicians report that they rely to some degree upon psychodynamic I principles in their work (Pope, Tabachnick, & KeithSpiegel, 1987), most researchers consider psychodynamic ideas
to be at worst absurd and obsolete and at best irrelevant or of
little scientific interest. In the lead article of a recent edition of
Psychological Science, Crews (1996) arrived at a conclusion
shared by many: " [ T ] h e r e is literally nothing to be said, scientifically or therapeutically, to the advantage of the entire Freudian
system or any of its component dogmas" (p. 63).
Despite the explosion of empirical studies of unconscious
cognitive processes (see, e.g., Greenwald, 1992; Kihlstrom,
1987; Schacter, 1992), few reference Freud; none cite any contemporary psychodynamic work; and in general, psychodynamic
concepts are decreasingly represented in the major psychology
journals (Robins & Craik, 1994). The situation is similar in the
popular media and in broader intellectual discourse. Publications ranging from llme to the New York Review of Books
periodically publish Freud's intellectual obituary, with critics
charging that Freud's ideas--such as his dual-instinct theory
or his hypotheses about female personality development--are
seriously out of date and without scientific merit (e.g., Crews,
1993).

Many aspects of Freudian theory are indeed out of date, and

they should be: Freud died in 1939, and he has been slow to
undertake further revisions. His critics, however, are equally
behind the times, attacking Freudian views of the 1920s as if
they continue to have some currency in their original form.
Psychodynamic theory and therapy have evolved considerably
since 1939 when Freud's bearded countenance was last sighted
in earnest. Contemporary psychoanalysts and psychodynamic
therapists no longer write much about ids and egos, nor do
they conceive of treatment for psychological disorders as an
archaeological expedition in search of lost memories (Aron,
1996; Gabbard, 1994; Horowitz, 1988; Kolb, Cooper, & Fishman, 1995; Mitchell, 1988; Wachtel, 1993). People do sometimes describe feelings or behaviors in therapy that conform
remarkably to aspects of Freud's psychosexual theories (such
as a patient of mine with erectile problems whose associations
to a sexual encounter led to an image of having sex with his
mother, followed by some unpleasant anal imagery). Nevertheless, psychotherapists who rely on theories derived from Freud
do not typically spend their time lying in wait for phallic symbols. They pay attention to sexuality, because it is an important
part of human life and intimate relationships and one that is
often filled with conflict. Today, however, most psychodynamic
theorists and therapists spend much of their time helping people
with problematic interpersonal patterns, such as difficulty getting emotionally intimate or repeatedly getting intimate with the
wrong kind of person (see Greenberg & Mitchell, 1983).
In psychology, students' introduction to psychodynamic the-

Drew Westen,Departmentof Psychiatry,Harvard Medical School, and

Department of Psychiatry, The Cambridge Hospital/Cambridge Health
Alliance, Cambridge, Massachusetts.
I thank Daryl Bem, Seymour Epstein, Kurt Fischer, David Funder,
Glen Gabbard, John Kihtstrom, Stephen Kosslyn, Dan McAdams, Jonathan Shedler, Robert Steinberg, and Joel Weinbergerfor their comments
on earlier drafts of this article.
Correspondence concerning this article should be addressed to Drew
Westen, Departmentof Psychiatry, The Cambridge Hospital, 1493 Cambridge Street, Cambridge, Massachusetts 02139. Electronic mail may
be sent to [email protected].

1The term psychodynamic was once more easily distinguished from

psychoanalytic, because it included under its umbrella a variety of approaches that took as axiomaticFreud's postulationof unconsciousmental processes but eschewed his libido theory. Today,however, the distinction is less clear, because mainstream psychoanalyticthinking is much
more interpersonal and no longer wedded to many classically Freudian
notions.
333

334

WESTEN

ory resembles the view found in the popular media and in scientific journals. I recently asked a research assistant with access
to the latest editions of the major textbooks to conduct a brief
study of the index entries of 10 leading upper-level introductory
textbooks. Two of the 10 cited object relations theory, the major
development in psychodynamic theory over the past 30 years. Of
the 4 most prominent contemporary psychoanalytic theorists-Otto Kernberg, Heinz Kohut, Stephen Mitchell, and Charles
Brenner--Kohut was cited in 3 books, Kernberg and Brenner
were cited once each, and Mitchell was not cited in any. This
is analogous to presenting trait psychology without referencing
Costa and McCrae. My intention is not to lay blame here, which
lies equally with psychoanalysts who have made too much of
their work obscure to psychologists, but simply to demonstrate
the widespread nature of the problem. 2
Psychodynamic thinking has, in fact, continued to flourish
over the past decades and has been progressively converging
with the concerns of personality, social, and developmental psychologists (see Barron, Eagle, & Wolitzky, 1992; Westen,
1990a). This largely reflects three shifts. The first is the cognitive revolution, which brought mental processes, including unconscious processes, back into psychological discourse. Although this revolution began with cognition, the lifting of the
taboo on mental processes imposed by behaviorism has inevitably led to a renewed focus on affect and motivation, the primary
areas of psychodynamic interest. Indeed, as cognitive neuroscience progressively moves into these areas, it will probably have
to be renamed.
The second shift is that psychodynamic theory was not mummiffed and buried along with its founder 60 years ago. Psychoanalysis may have been "deconstructed," but it has not been
decomposing. For example, Freud's dual-instinct theory (involving sex and aggression) is the primary, if not the only,
psychodynamic model of motivation learned by most undergraduate psychology majors, because it is the primary one taught in
textbooks. Yet most contemporary psychodynamic psychologists hold that humans have a number of motives, many of them
rooted in biology but nearly all elaborated upon by culture and
experience. Whereas Freud emphasized the pursuit of sensual
and sexual pleasure, object relations theorists, self psychologists, and relational theorists have added a focus on needs for
relationships and self-esteem (see Aron, 1996; Fairbairn, 1952;
Greenberg & Mitchell, 1983; Kernberg, 1975; Kohut, 1977;
Sullivan, 1953). To reject psychodynamic thinking because
Freud's instinct theory or his view of women is dated is like
rejecting modem physics because Newton did not understand
relativity.
The third shift is epistemologicaL For years, many in the
psychoanalytic community have treated psychoanalysis as a religion and have been more interested in protecting than testing
psychoanalytic dogma. This has clearly contributed to the widespread distaste for psychodynamic concepts in academic psychology. Far too many analysts continue to ignore or disparage
efforts to test their theories as the folly of ignorant infidels
and to dismiss alternative hypotheses, conceptualizations, and
treatment techniques--and they do so at the peril of their patients. However, the infusion of psychologists into psychoanalysis over recent decades (and to some extent the increasing medicalization of psychiatry) has led to a greater appreciation for

empirical work, an acknowledgment that psychoanalytic propositions do not rise and fall solely on the basis of their perceived
clinical utility, and a recognition that "I had a patient once" is
not the firmest of epistemological foundations. Conversely,
many cognitive researchers have begun paying greater attention
to the ecological validity of experimental work on thought and
memory, which is leading to, among other things, a heightened
focus on affect and motivation, the central domains of psychoanalysis (e.g., Ceci & Bronfenbrenner, 1991; Neisser, 1991;
Stein, 1997).
The aim of this article is to reconsider Freud's legacy, not
by arguing about the validity of 1920s psychoanalysis but by
considering the relevance of contemporary psychodynamic theory for psychological science in the 21st century. The article
begins by presenting five core theoretical tenets that define current psychodynamic thinking and briefly considers the empirical
evidence with respect to their validity. Although all of these
propositions trace their lineage to Freud and were important to
the way he understood the mind, Freud undoubtedly would have
identified other propositions he considered central as well, such
as his drive theory. These five tenets, however, are the psychodynamic propositions that have best stood the test of time, as
reflected in their widespread acceptance among psychoanalysts
and psychodynamically oriented clinicians and theorists a century after Freud began writing. The article concludes by illustrating the way psychological science might be enriched by incorporating some of the insights that have emerged from 100 years
of psychoanalytic inquiry, using as examples research in social
cognition and models of parallel distributed processing in cognitive science.
Will the Real Psychoanalysis Please Stand Up?
Five Postulates That Define Contemporary
Psychodynamic Theory
Psychoanalysis was once a single theory, identified with its
founder and his particular ideas, so that summarizing its basic
tenets was once less problematic. Today the situation is different,
with no single theory dominating even the mainstream psychoanalytic journals let alone the thinking of those who consider
themselves more broadly psychodynamic. Nevertheless, all psychodynamic theorists generally adhere to five propositions. First,
and most central, much of mental life--including thoughts, feelings, and motives--is unconscious, which means that people
can behave in ways or develop symptoms that are inexplicable
to themselves. Second, mental processes, including affective and
motivational processes, operate in parallel so that, toward the
same person or situation, individuals can have conflicting feelings that motivate them in opposing ways and often lead to
compromise solutions. Third, stable personality patterns begin
to form in childhood, and childhood experiences play an important role in personality development, particularly in shaping
the ways people form later social relationships. Fourth, mental

2 References to these books are available from the author upon request.
I do not reproduce them here, because I have published an introductory
text myself and do not want to derogate what are otherwise excellent
texts (not consciously, anyway--but more on that later).

FREUD'S LEGACY
representations of the self, others, and relationships guide people's interactions with others and influence the ways they become psychologically symptomatic. Finally, personality development involves not only learning to regulate sexual and aggressive feelings but also moving from an immature, socially
dependent state to a mature, interdependent one.
Before I review the empirical data for each of these propositions, two potential objections require consideration. The first
is/hat these five propositions may not adequately represent contemporary psychodynamic thinking. Psychoanalytic theorists are
not all of one mind, and distilling a common core to a wide
range of often-competing theoretical formulations is no simple
task. Isolation of these five tenets n o doubt reflects the systematizing efforts of the author, who is an active researcher as well
as a clinician and whose thinking is heavily influenced by research in a number of subdisciplines and not just by psychoanalytic theory and clinical observation. Nevertheless, these propositions were all once highly disputed in psychology and were
exclusively associated with psychoanalysis. A t the very least,
they represent five important tenets of psychoanalytic theory
even if not the five most central. Recent survey data suggest,
however, that these propositions are indeed central to contemporary thinking among practicing psychodynamic psychologists
and psychiatrists?
A second objection regards the extent to which the propositions outlined here, and the data that support them, are in fact
distinctively psychodynamic. Many developmentalists, for example, would agree that childhood experiences play an important role in shaping personality, and most cognitive-experimentalists now accept the importance of implicit (unconscious)
processes. Further; many of the studies that support these propositions have been conducted by researchers with little interest
in or knowledge of psychodynamic ideas. Perhaps the propositions that have obtained empirical support are simply analogous
to but not isomorphic with psychodynamic ideas, and little is
to be gained by taking a second look at psychodynamic theory.
To dismiss the approach to the mind that Freud inaugurated
with the assertion that nothing about these propositions is or
ever has been unique to psychoanalysis is, however, to engage
in an impressive act of revisionist history. This can be readily
seen with respect to the proposition that much of mental life
is unconscious. Freud was not the first to notice unconscious
processes; poets and philosophers beat him to it. He was, however, the first to build a systematic psychological theory on this
proposition, which was attacked vociferously by psychologists
of nearly every persuasion for almost a century. Until the mid1980s, psychodynamic psychologists were alone among their
colleagues in arguing for the importance of unconscious
thoughts, feelings, and motives. They remain alone among clinicians in attempting to address these processes systematically in
psychotherapy, a state of affairs that I believe will shift once the
implications of the last decade of research in cognitive neuroscience become more apparent in the clinical literature.
With respect to unconscious (implicit) processes, the historical record is quite clear. As the most comprehensive experimental alternative to psychoanalysis, behaviorism dominated academic psychology (particularly in the United States) through
the 1950s and rejected the notion that unconscious processes
(or even conscious processes) could play any causal role in

335

human behavior. The successor to behaviorism in experimental

psychology, cognitive science, held to a serial processing model
of cognition, in which memories of past experiences only become significant once they enter short-term memory (a synonym
for consciousness), until the late 1980s. This was fully a century
after Freud's first attempts to describe the influence of unconscious processes on thought, emotion, and behavior. In 1987,
Kihlstrom heralded the existence of the cognitive unconscious.
However, as will be seen, less than 10 years passed before
researchers began to discover that the same principles that apply
to cognition operate with implicit (or unconscious) affective
and motivational processes. So the cognitive unconscious
became the p s y c h o l o g i c a l u n c o n s c i o u s (Kihlstrom, 1990)
and is rapidly becoming the cognitive-affective-motivational
unconscious.
Is this new unconscious nothing but a 1900 Viennese wine
in a 1990s bottle? Of course n o t - - i t comes from a different
tradition and a different set of methods, and it emerged a century
later. Does it bear a strong resemblance to the cognitive-affective-motivational unconscious of contemporary psychodynamic theorists and clinicians (e.g., Brenner, 1982; Sandier,
1987), who have continued bottling a distinctive blend of unconscious processes for decades while decreasing the content of

3To test the centrality of these tenets to contemporarypsychodynamic

clinicians, I recently surveyed 150 randomly selected members of the
American Psychiatric Association and of three clinical divisions (12
[now known as the Society of Clinical Psychology], 29, and 39) of the
AmericanPsychologicalAssociationwho indicated on a previous survey
that their primary theoreticalorientation was primarily psychodynamicpsychoanalytic. Of these, 86 returned the survey, representing a 57%
response rate. All participants were practicing clinicians with an average
of 19.42 years of experience postlicensure.Slightly over half were psychologists, and 34% were psychoanalysts.Respondents were presented
with 10 theoretical statements about personality and psychopathology
and were asked to rate each one on a Liken scale from 1 to 7, where
1 = strongly agree and 7 = strongly disagree, and to rank the statements
in order to indicate the extent to which they agreed with them. The
statements included versions of the five propositions presented here as
well as five major theoretical propositions central to Freudian thinking,
such as the importance of the Oedipus complex, Freud's dual-instinct
theory of motivation,psychosexualtheory, and so forth. For both ratings
and rankings, cliniciansendorsedevery one of the fivepropositionsmore
strongly than every Other one of the propositions, which were taken
from Freudian theory and current textbook coverage of psychoanalysis,
atp < .0001. The same patternsheld when the responsesof (a) psychologists and psychiatrists and (b) psychoanalysts and nonanalysts were
considered separately.As a second index of the extent to which psychodynamically oriented psychologists and psychiatrists share a core set of
beliefs, the ranking data were subjected to Q-factor analysis, a statistical
clustering procedure that transposes the factor matrix so that participants, rather than items, are factored (see Block, 1978). This allows the
researcher to determine whether respondents cluster into groups based
on their response profiles. The results were striking: 45 of the 63 clinicians who provided ranking data loaded on a "contemporary psychodynamic" Q factor defined by the five statementsdescribed here as central
to contemporary psychodynamic theory. This Q factor accounted for
68.1% of the variance. No other Q factor included more than 4 respondents. Thus, the data show a surprising amount of agreementamong
contemporary psychodynamicpsychologists,psychiatrists, and psychoanalysts in their core beliefs about personality and psychopathology.

336

WESTEN

drive theory and sexuality that characterized wine with a more

Freudian bouquet?
Those familiar with both literatures will recognize that it
does, although both approaches to unconscious processes could
be substantially enriched by attention to theory and data produced by the other. On the one hand, psychodynamic theory
continues to contain problematic residues of Freud's model of
the unconscious, which (a) linked unconscious processes inextricably to primitive, instinctual, and emotional processes; (b)
failed to distinguish a variety of different kinds of unconscious
processes (particularly cognitive ones); and (c) failed to recognize the extent to which some unconscious processes can be
adaptive and learned (Westen, in press-b). On the other hand,
the unconscious of cognitive experimentalists tends to be the
comfortable unconscious: automatic, cold (though steadily
warming up), cognitive, and unconscious only by virtue of
mental architecture. As the data to be described below will make
amply clear, this is only a subset of unconscious processes, and
attention to the more affectively "hot" subset described for a
century by psychodynamic theorists and clinicians would provide a more balanced, comprehensive, and ecologically valid
portrayal of how the mind works. Psychoanalytic theorists have
been operating from a theory of parallel processing for a century,
and many aspects of the parallel architecture they have proposed--notably those relating to affect and motivation--have
not yet been incorporated into contemporary connectionist models and will not likely be integrated without explicit attention
to psychodynamic constructs and clinical data. The situation is
no different with the other four propositions to be examined
here, all of which have been central to psychodynamic thinking
since at least 1914, were once vehemently contradicted by competing theories, and now have substantial empirical support.

The Existence and Centrality of Unconscious Processes

The most important proposition that has distinguished psychoanalysis from other theoretical systems since its inception is
its postulation of unconscious mental processes. Freud ( 1926/
1953c) considered this the cornerstone of psychoanalytic theory.
The concept of "the unconscious" was not novel to Freud and
was swirling around the intellectual air he breathed in the 19th
century, particularly in Germany (see Ellenberger, 1970; J.
Weinberger, in press). The German idealist philosophers postulated a realm of the unconscious as early as the beginning of
that century, infusing it with mystical elements that were later
appropriated by Jung, such as the view that a single Will is
immanent in all living things and is expressed in unconscious
motives. Schopenhauer (1883/1958) identified this Will with
reproduction of the species, whereas Nietzsche (1956) focused
on power as the unconscious driving force behind human behavior. Freud divorced unconscious processes from their religious
and philosophical underpinnings, schematized them, offered a
theory of their dynamics, and linked this theory to data from
the consulting room. Ironically, the linkage of a theory of unconscious processes to data was one of the main features that distinguished Freud's from earlier views of the unconscious; however,
psychologists who embraced a positivist-empiricist philosophy
of science were never enamored of the kind of data upon which
Freud relied (J. Weinberger, in press).

Until the 1980s, psychoanalysis was alone among psychological theories in its postulation of unconscious mental processes,
although such processes had played some part in the history of
psychology since Helmholtz's ( 1863 / 1971 ) concept of unconscious inference involved in perception. Freud (1900/1953b)
initially emphasized unconscious wishes, but he later distinguished a variety of unconscious processes (Freud, 1915/
1957c), and his followers expanded upon the notion that unconscious fantasies (by which they mean affect-laden beliefs) and
unconscious mental representations play a central role in human
behavior (Arlow, 1991; Sandier & Rosenblatt, 1962; J. Weinberger, in press). Today, the proposition that many cognitive processes are carried out unconsciously (Kihlstrom, 1987) is as
widely accepted by experimental psychologists as the opposite
proposition (i.e., that such processes do not exist) was by their
predecessors less than 2 decades ago. I suspect, however, that
many cognitive scientists would be wary of extending the notion
of unconscious processes to affect and motivation and would
be particularly dubious of the hypothesis that affective considerations can bias the way thought is assembled outside of awareness (the concept of defense). In this section, I examine the data
supporting the proposition that much of mental life, including
thoughts, feelings, and motives, is unconscious. 4

Unconscious Thought and Memory

I will be brief in describing the literature on unconscious
cognitive processes, because it is no longer controversial and
much of it has been reviewed elsewhere (Holyoak & Spellman,
1993; Roediger, 1990; Schacter, 1992; Schacter & Buckner,
1998; Squire, 1987). For years, cognitive psychologists did not
specify whether the processes they were describing were conscious or unconscious, although they assumed a serial architecture in which information entered sequential memory stores and
was retrieved and manipulated in short-term memory (which
was equated with consciousness). The role of consciousness
has received much more attention since the discovery of implicit
memory, which refers to memory that is observable in behavior
but is not consciously (explicitly) brought to mind.
One kind of memory that can be expressed implicitly is procedural memory, "how-to" knowledge of procedures or skills
useful in various activities, such as tying a shoe or driving
a car. Another kind of implicit memory emerges in priming
experiments, in which prior exposure to the same or related
information facilitates the processing of new information,
whether or not the prime was originally even conscious. For
example, presenting the word pair taxi/cab in the unattended
channel using a dichotic listening paradigm renders individuals
more likely to use the less preferred spelling of the auditorially
presented homophonefare/fair, even though they have no recognition memory for the word pair (see Nisbett & Wilson, 1977;
Schacter, 1992). Research in cognitive neuroscience has now
documented that implicit memory and explicit memory (conscious recognition or recall) involve different neural mechanisms (Schacter, 1995; Squire & Zola-Morgan, 1991).
4 For a more exhaustive summary, see Westen, 1998; implications
of these data for psychoanalytic theory are discussed in Westen, in
press-b.

FREUD'S LEGACY
Various literatures on thinking have similarly come to distingnish implicit and explicit thought and learning processes (Holyoak & Spellman, 1993; Jacoby & Kelly, 1992; Kihlstrom,
1990; Lewicki, 1986; Reber, 1992; Seger, 1994; Underwood,
1996). As with research on memory, until very recently psychologists paid little attention to the issue of consciousness when
studying mechanisms involved in problem solving, decision
making, and other cognitive tasks. In the last few years, this has
begun to change as researchers have begun to recognize the
extent to which thought is guided by implicit grammars that
guide performance in tasks ranging from decoding or producing
sentences to standing an appropriate distance from another person (as culturally defined) or producing appropriate notes in
harmony with music playing on the radio. For example, Rubin,
Wallace, and Houston (1993) found that participants asked to
compose a ballad after hearing a series of ballads could follow
twice as many rules used in their composition as they could
consciously articulate.

Unconscious Affective Processes

The existence of unconscious cognitive processes, which were
seen as the province of psychoanalysis just a decade ago, is now
taken for granted by most cognitive scientists. What remains
distinctive about the psychodynamic perspective is the view that
affective and motivational processes can be unconscious as well.
Logically, the assumption that cognition can be unconscious but
affect and motivation must be conscious makes little sense. We
have no reason to assume a parallel architecture for cognition
but a serial architecture for emotion and motivation. Conscious
emotions and motives do not "come out of the blue" any more
than conscious thoughts do, and they often include or reflect
cognitive components (such as attributions or interpretations)
that presumably are assembled outside awareness, as are other
cognitions. From an evolutionary perspective, an organism that
had to reflect on every decision would he at a severe adaptive
disadvantage relative to an animal that could rely on prior learning about the affective significance of various stimuli and produce a rapid response. In fact, several independent lines of research suggest that the cognitive unconscious includes only a
subset of unconscious processes, notably those that are most
automatic, least intentional, and most comfortable to discuss
with one's mother.

Neurological evidence for unconscious emotional responses.

Some of the best documented early examples of unconscious
affective processes came from the study of Milner's famous
patient, H.M. (Milner, Corkin, & Teuber, 1968). Because of
hippocampal damage, H.M. lost the capacity to consolidate
new explicit declarative memories. However, he continued to
demonstrate the capacity for affective learning despite his deficits in explicit memory. For example, following a visit to his
mother in the hospital, H. M. could remember nothing of the
visit but "expressed a vague idea that something might have
happened to his mother" (Milner et al., 1968, p. 216). Johnson,
Kim, and Risse (1985) reported similar findings with patients
suffering from Korsakoff's disorder, who showed a preference
for melodies they had heard five minutes earlier (reflecting the
mere exposure effect, the tendency to prefer familiar stimuli;
Zajonc, 1968) despite their impaired recognition memory for the

337

melodies. Patients with Korsakoff's also had difficulty recalling

information presented to them about two fictional characters but
preferred the one who had been associated a week earlier with
positive attributes. This suggests that the neural circuitry for
affective associative learning is distinct from that for conscious
declarative memory.
Other research with neurologically impaired patients points
in the same direction. Patients with bilateral hippocampal lesions, who have difficulty with explicit memory, can develop
conditioned emotional responses to aversive stinmli even though
they cannot consciously learn the connection between the conditioned stimulus and an aversive unconditioned stimulus. In conWast, patients with an intact hippocampus but bilateral lesions
to the amygdala show deficits in emotional conditioning even
though they are conscious of the link between the conditioned
and unconditioned stimulus (Bechara et al., 1995). In other
words, explicit memory in these patients is intact, but implicit
affective learning is impaired. These data are consistent with
LeDoux's (1989, 1995) finding of two neural pathways for
emotion: one, implicated in conditioned emotiOnal responses,
in Which primitive perceptual information is relayed via the
thalamus to the amygdala (which attaches an affective valence
to the information) without any involvement of consciousness,
and the other, in which the thalamus relays information to the
cortex, which processes the information more deeply before
activating the amygdala. 5
Similarly, individuals with prosopagnosia, who lose the capacity to discriminate faces consciously (a form of explicit
knowledge), may nevertheless show differential electrophysiological responses to familiar versus unfamiliar faces (Bruyer,
1991). Analogous findings have been reported in patients with
visual neglect, who ignore a region of the visual field. In one
study, the investigators presented pictures of two houses to patients with visual neglect (Halligan & Marshall, 1991). In one
of the pictures, the left half of the house was on fire. The patients
denied any difference between the two houses--consciously
neglecting the image on the left--but chose the house without
the fire when asked which one they would prefer to live in.
Research similarly finds that split-brain patients presented with
affectively evocative visual stimuli to the right hemisphere may
show appropriate affective responses but be unable to offer any
verbal explanation for them (Gazzaniga, 1985).
These neurological data are probably of relevance to the argument advanced by Zajonc (1980) that preferences can occur
without inferences. For example, W. R. Wilson (1975) presented
tone sequences to participants in the unattended channel in a
dichotic listening experiment. Although participants showed no
recognition memory for tone sequences they had heard as many
as five times, they reported liking the tone sequences they had
heard better than those they had not (the mere exposure effect ).
In other words, they were developing affective preferences outside of awareness toward stimuli they had never consciously
registered. Murphy and Zajonc (1993) found that subliminal
presentation of positive or negative faces similarly affected participants' liking or disliking of novel stimuli. Experiments with
5 This should not lead to the common assumption that the latter pathway necessarily involves consciousness. Most cortical processes are not
consciously accessible, either.

338

WESTEN

subliminal procedures have shown comparable effects on attitudes toward various stimuli, including the self (see Bargh,
1997; Eagle, 1959). More recently, Murphy, Monahan, and Zajonc (1995) have combined these lines of research (exposure
and priming), demonstrating that subliminal mere exposure effects and subliminal priming effects are additive: Increasing
exposure leads to increased ratings of liking for Chinese ideographs, and priming with a positive or negative stimulus (a
happy or sad face) simply adds or subtracts a constant (that is,
increases or decreases liking ratings the same amount at each
level of exposure). Interestingly, explicit (1 s) priming does
not have the same effect as implicit (4 ms) priming, because
participants recognize the manipulation and appear to counteract
it. Perhaps the most important contribution of this research is
that it demonstrates unequivocally that affective evaluations can
develop unconsciously and that multiple influences on affective
associations can be combined outside of awareness. As we will
see, this finding fits well with a psychodynamically informed
connectionist model that integrates affective and motivational
processes with the more familiar cognitive and perceptual ones.
Conditioning and unconscious affect. A second, and related, source of data on unconscious affect comes, paradoxically,
from research with conditioning paradigms. Lazarus and
McCleary ( 1951 ) paired nonsense syllables with a mild electric
shock and then presented the conditioned stimuli to participants
subliminally. The conditioned stimuli reliably elicited a galvanic
skin response (GSR) even when presented below the threshold
of conscious recognition. Thus, a conditioned stimulus can elicit
affect, as assessed electrophysiologically, even when presented
outside of awareness. Numerous other studies have produced
similar results (see Ohman, 1994; J. Weinberger, in press; Wong,
Shevrin, & Williams, 1994) with dependent variables ranging
from skin conductance to facial electromyography (EMG,
which measures facial muscle indicators of emotion) to evokedrelated brain potentials (ERPs).
Other studies show that classically conditioned emotional responses can be not only elicited but acquired without consciousness (e.g., Bunce, Bernat, Wong, & Shevrin, 1995; Esteves,
Dimberg, & Ohman, t994). For example, faces presented subliminally can become associated with an aversive stimulus (electric shock). The conditioned response is evidenced in facial
EMG and evoked potentials even though the person reports no
conscious awareness of the contingency between conditioned
and unconditioned stimuli (Bunce, Bernat, Wong, & Shevrin,
1995; Wong, Bernat, Bunce, & Shevrin, in press). In another
study, a conditioned auditory stimulus paired with electric shock
while rats were unconscious (anesthetized) produced a conditioned response 10 days later (Weinberg, Gold, & Sternberg,
1984; for similar research on nausea, see Garcia & Rusiniak,
1980). In yet another study, dental patients showed affective
learning while anesthetized with nitrous oxide but not with Novocain (Hutchins & Reynold, as cited in Leventhal & Everhart,
1979). Whereas Novocain effectively blocked the experience
and encoding of pain signals, nitrous oxide apparently blocked
consciousness of pain but did not prevent implicit memory for
the experience. Subsequent research has demonstrated other
kinds of learning that can occur under general anesthesia (e.g.,
Cork, 1996).
Behavior therapists in the 1970s who believed they had

"cured" homosexuality in male patients subsequently discovered that their patients had suppressed the conscious response
while remaining physiologically aroused by pictures of naked
men, as demonstrated by genital plethysmography (which measures sexual arousal in males by measuring degree of erection;
McConaghy, 1976). These findings echoed other research demonstrating that suppression of the subjective experience of emotion does not eliminate the psychophysiological component of
the emotion (Rachman, 1978). Presumably, this component
could continue to exert an impact on behavior despite the absence of conscious affect.
Unconscious affect and attitudes. Social psychologists
studying attitudes and prejudice now recognize the importance
of distinguishing conscious and unconscious attitudes (see
Greenwald & Banaji, 1995). Social psychologists have traditionally defined attitudes as including not only a cognitive component but an affective-evaluative component (and often a set
of behavioral tendencies as well, such as the tendency to vote for
candidates with particular ideological agendas). Unconscious
attitudes include unconscious affective dispositions, which may
be activated automatically and without conscious awareness.
Fazio (1990) has found that when people are consciously focusing on their attitudes, these attitudes heavily influence their behavior. When they are not focusing on their attitudes, only chronically activated, often automatic, and unconscious attitudes do
so. This finding is directly parallel to a similar finding in the
motivation literature, reviewed below, that conscious motives
guide consciously chosen behavior; whereas implicit or unconscious motives guide behavior over the long run, when
consciousness is not directly focused on goals (McClelland,
Koestner, & Weinberger, 1989). Petty and Cacioppo's (1986)
work on attitude change distinguishes a central route to attitude
change, which involves conscious attention to reasoned arguments, and a peripheral route, which is usually based on heuristics, automatic processes, affective appeals, and primitive evaluative processes that require minimal attention and effort 6 (see
also Chaiken, 1980).
Some of the most convincing research on unconscious affect
comes from studies of prejudice (Devine, 1989; Dovidio &
Gaertner, 1993; Katz, 1981; Katz & Hass, 1988), which demonstrate that people in the United States who consider themselves
nonracist often have two conflicting sets of attitudes that influence their behavior: one explicit and the other implicit. The
latter, which are more negative, bias people's information processing when their attention is not focused on their conscious
values and are likely implicated in relatively automatic behavioral responses, such as when people check their wallets after
a Black man has passed on the subway.
Of most relevance is research demonstrating that individual
differences in unconscious or implicit negative racial attitudes
are often uncorrelated with individual differences in conscious
6 The fact that attitude models have begun to converge with psychodynamic models is in some respects not surprising, because some prominent models of attitudes (e.g., Devine, 1989; Fazio, 1990) are based on
the distinction between controlled and automatic information processing
(Schneider & Shiffrin, 1977), which is similar in certain respects to
Freud's distinction between secondary and primary process thought (i.e.,
rational versus associational and unconscious thought).

FREUD'S LEGACY
racial attitudes. In a landmark study, Fazio, Jackson, Dunton,
and Williams (1995) measured implicit attitudes by presenting
to participants a series of Black and White faces followed by
either a positive or negative adjective. The participants' task was
to press a key indicating the affective connotation of the word
(good or bad). Because negative attitudes toward Blacks should
facilitate (prime) responses to negative adjectives, response latency to negative words following priming with a Black face
can be used as a measure of the affective quality of implicit
attitudes--that is, the association of affects with representations
of Blacks. The investigators found that this measure of individual differences in implicit attitudes predicted an implicit behavioral index--the extent to which a Black confederate of the
experimenters rated participants as friendly and interested in
their interaction with her when she debriefed them about the
study--but did not predict conscious attitudes about current
events involving racial issues. In contrast, a measure of conscious racism predicted participants' responses to these conscious attitude questions but not their behavior with the Black
confederate. 7 Similar findings emerge with gender stereotypes:
Implicit and explicit attitudes toward gender are often only minimally correlated, if at all (Banaji & Hardin, 1996).
Other studies documenting unconscious emotional processes.
A number of other literatures document unconscious affective
processes (see, e.g., Niedenthal & Kitayama, 1994). Women
high in self-reported sexual guilt report less arousal but show
greater physiological arousal (as assessed by genital plethysmography) than those low in sexual guilt while watching an
erotic videotape (Morokoff, 1985). Research on hypnotic analgesia often finds that physiological indexes of pain show persistence of pain despite conscious self-reports and consciously
controlled behavior (such as keeping a hand immersed in ice
water) indicative of minimal pain (Hilgard, 1979). Research on
excitation transfer shows that although participants gradually
lose awareness of physiologically detectable residual arousal
following vigorous exercise, they subsequently respond with
increased aggression or sexual excitement (depending on the
stimulus) when presented with anger-arousing or erotic stimuli
during this postawareness phase (Zillman, 1978), Thus, they
are acting on arousal of which they are unaware.
In a classic New Look study, participants more readily perceived neutral than taboo words presented briefly on a screen
and showed higher skin conductance responses (GSRs) for the
taboo words prior to conscious recognition of them (McGinnies,
1949). These findings, which not only replicated but proved
robust in the face of numerous rival hypotheses (see Broadbent,
1977; Dixon, 1971, 1981; Erdelyi, 1974, 1985), suggest a preconscious stage of processing in which information is evaluated
for its affective content, a position similarly arrived at by Bargh
(1997). Other researchers (as cited in Shevrin & Dickman,
1980) have discovered differences in amplitude of brain ERPs
in response to emotional versus neutral words presented subliminally. In one study, emotional words evoked more alpha waves
as assessed by electroencephalograph (EEG) than did neutral
words even when presented at luminance levels that rendered
them consciously imperceptible, which suggests that their emotional content was processed prior to their conscious recognition
(Heinemann & Emrich, 1971). More recently, Shevrin and his
colleagues (Shevrin, Bond, Brakel, Hertel, & Williams, 1996)

339

have been using brain wave activity to examine the impact

of subliminal presentation of words selected (on the basis of
extensive interview and projective data) by a team of clinicians
as relevant to patients' symptoms. Silverman and colleagues
(Silverman & Weinberger, 1985; J. Weinberger, in press) have
demonstrated that subliminal presentation of stimuli hypothesized to be psychodynamically meaningful can affect a wide
range of behavior. Despite the widely held perception that such
findings have only emerged in Silverman's laboratory, a metaanalysis of over 100 studies Conducted by numerous researchers
documented a modest but robust effect of subliminal stinmlation
of this sort; supraliminal stimulation often does not produce the
same effect (J. Weinberger & Hardaway, 1990).
Also of relevance is research by Wegner (as summarized
in Wegner, 1992) on thought suppression. Using such simple
manipulations as instructing participants not to think about a
white bear, Wegner has demonstrated that the act of suppression
instigates not only a conscious control process in which participants actively search for distractors but also an automatic, unconscious search process aimed at detection of the word to be
suppressed. When the automatic process detects the target word
(e.g., while participants are asked to think aloud or complete
a word-association task), it activates the conscious distractor
process. The automatic search process, however, has the unintended effect of keeping the to-be-suppressed thought activated
unconsciously. Blum (1954) demonstrated a similar phenomenon in which both the vigilance and suppression processes occur
outside of awareness. He exposed participants to threatening
and nonthreatening visual stimuli at speeds that rendered them
in one condition far below and in another condition near the
threshold of conscious perception. When asked which stimuli
"stood out" most clearly at extremely brief stirrmlus exposures,
participants indicated that the anxiety,provoking images were
more salient, although they could not report what they had seen.
At longer exposures, however, neutral stimuli were more easily
perceived; that is, anxiety-provoking stimuli took longer to recognize consciously.
Of particular relevance to the current discussion is the finding
that when people are instructed to suppress an exciting thought
about sex, they remain psychophysiologically aroused even
while the thought is outside awareness--as aroused, in fact, as
participants instructed to think about the sexual thought (Wegner, Shortt, Blake, & Page, 1990). Unlike the latter participants,
however, those instructed to suppress the thought do not habituate to it, so that when it returns to their consciousness, they
show physiological arousal again, unlike participants who have
kept it in mind the entire time. This suggests, much as Freud
did, that affect-laden thoughts kept from consciousness may
continue to have an affective press.
Unconscious affect and defensive processes. The studies
7 Curiously, Fazio and colleagues (1995), like many cognitive social
psychologistsworking in this area, did not conclude that they had elicited
or measured unconscious affect, although it seems highly unlikely that
participants with implicitly negativ~ but explicitly positive attitudes toward African Americans werd ~~d of activation of negative affect
while they were unwittingly sendi~ flegMiVesignals to the Black confederate. At the very least, however, the results clearly indicate that
implicit and explicit affective associations are macon-elated.

340

WESTEN

presented thus far on unconscious affect provide particularly

compelling evidence for a central hypothesis that has been propounded only by psychoanalytic theory and has guided psychoanalytic clinical work for a century: People can feel things without knowing they feel them, and they can act on feelings of
which they are unaware, as in subtly hostile, indifferent, or
defensive treatment of members of ethnic minority groups (such
as gratuitous, defensive smiles that members of these groups
immediately recognize as nongenuine) by people who believe
they harbor no ill feelings toward them. Psychoanalytic clinicians have attended to such signs of implicit affect for decades
by observing both nonverbal behavior (such as facial expressions, sighing, blocking, shifting posture, etc.) and verbal behavior (such as sudden changes in syntax, shifts in topic, disorganization in otherwise coherent narratives, and word use that seems
stilted or idiosyncratic). A discussion of unconscious affect
invariably leads to literatures pertaining to another distinctively
psychodynamic construct: psychological defense.
The most important research in this area has been conducted
by Vaillant and colleagues, who adapted Haan's (1977) defense
Q sort to study two longitudinal samples (Roston, Lee, & Vaillant, 1992; Vaillant, 1977, 1992). VaiUant has shown that reliably scored defenses, arranged according to four levels of adaptiveness (from grossly maladaptive defenses, such as blatant
denial, through more mature ways of dealing with discomfort,
such as humor), have a number of predictable correlates, both
longitudinally and cross-sectionally. For example, Vaillant and
Drake (1985) found strong associations between level of defense maturity and independently rated measures of psychological health-sickness and presence of a personality disorder. A
different methodology has produced similar results (Perry &
Cooper, 1989).
Other researchers have tried to operationalize defense mechanisms empirically as well (Bond, Paris, & Zweig-Frank, 1994;
Ihilevich & Glesel; 1986; Paulhus, Fridhandler, & Hayes, 1997;
Perry & Cooper, 1987). For example, Hauser and colleagues
(Hauser, Borman, Jacobson, Powers, & Noam, 1991) have found
links between coping and defense mechanisms in adolescents
and their parents' level of ego development as assessed by Loevinger's (1985) Sentence Completion Test. K. Davidson and
MacGregor (1995) found, as predicted, that overall degree of
defensiveness as assessed by observers during an interview moderates the discrepancy between self-reported and informant-reported neuroticism. The correlation between self- and informant
reports for low-defensiveness participants was moderately positive, whereas the correlation for high-defensiveness participants
was strongly negative.
Westen, Muderrisoglu, Fowler, Shedler, and Koren (1997)
assessed the reliability and validity of a Q-sort measure of affect
regulation that includes unconscious defensive and conscious
coping strategies. Factor analysis of clinician descriptions of 90
patients yielded three affect regulation factors. The first, labeled
adaptive regulation, is defined by the poles of active coping
versus acting impulsively. People high on this dimension tend
to use active coping strategies, such as anticipating problems
and seeking out information, whereas people low on this dimension use maladaptive defenses and behaviors, such as dissociating and using drugs and alcohol. The second factor, externalizing
defenses, is defined by the poles of blaming others versus blam-

ing the self. Individuals high on this factor deny responsibility

for their own problems and brag or dwell on their successes to
bolster flagging self-esteem, whereas those who are low on this
factor have difficulty expressing or acknowledging anger, tend
to feel bad or unworthy instead of feeling appropriately angry
at others, and remain passive in the face of distress. The third
factor, avoidant defenses, is defined by the poles of avoiding
consciousness of unpleasant affect versus containing affective
leakage. People who are high on this factor tend to avoid thinking about distressing experiences consciously and to think in
intellectualized or abstract terms about emotion-laden events,
whereas people low on this factor find themselves constantly
trying to contain breakthroughs of distress, vacillating between
clinging to others and pushing them away, expressing distress in
the form of physical symptoms, and ruminating when distressed.
Interrater reliability from clinical research interviews is relatively high (r = .78), and factor scores predict relatively objective data such as capacity to maintain employment, history of
psychiatric hospitalization, and history of suicidality, as well as
diagnosis and clinician ratings of adjustment.
A number of studies have added considerably to the literature
on defensive processes by using experimental, rather than solely
correlational, methods. For example, positivity of participants'
self-descriptions increases following a threat manipulation,
which suggests defensive alteration of conscious self-representations (Paulhus & Levitt, 1987). Cramer (1996) has developed
a measure of three defenses--projection, denial, and identificat i o n - f o r use with Thematic Apperception Test (TAT) responses. In one study, experimentally inducing anger led to increases in projection and identification (Cramer, 1991). A different projective measure of defense was used by Hedegard (1969),
who experimentally manipulated anxiety using hypnosis and
found that more intense anxiety predictably evokes less mature
defenses, as predicted by theories that view defenses as hierarchically organized (A. Freud, 1936; Vaillant, 1977). Other researchers have used cognitive or perceptual tasks to measure
defensive processes experimentally (see Erdelyi, 1990; G. J. W.
Smith & Westerlundh, 1980).
Numerous social psychological investigations have documented defensive processes as well (see Westen, 1994). Pyszczynski, Greenberg, and Solomon (1998) have demonstrated in
dozens of studies that making people more aware of their mortality (typically by having them complete a mortality salience
questionnaire) leads them to respond defensively in a number
of ways even though they have no awareness of either the affect
that triggered their responses or its effects on their conscious
decisions or behavior (see, e.g., Simon et al., 1997). Along
different lines, L. S. Newman, Duff, and Baumeister (1997)
found support in a series of studies f o r an ingenious theory
about the mechanisms by which one defense, projection, occurs.
Drawing on Wegner's (1992) work on thought suppression,
they found that people who avoid thinking about threatening
aspects of themselves and have a tendency to avoid awareness
of their feelings become chronically vigilant toward precisely
those traits they want to deny; this chronic accessibility leads
them to use those same trait constructs when making inferences
about other people. Although they used the descriptive language
of self-regulation rather than of defense, Bandura and colleagues
(Bandura, Barbaranelli, Caprara, & PastoreUi, 1996) showed

FREUD'S LEGACY
that children who use externalizing defenses--that is, who endorse items such as "Slapping and shoving someone is just a
way of joking" and "Kids cannot be blamed for using bad
words when all their friends do i t " - - a r e more likely to be
physically aggressive and less likely to behave prosocially than
children who do not reframe transgressions to avoid responsibility. Other research has found that members of couples defensively fail to recognize their partner's attraction to another person tO the extent that (a) the relationship is important to them,
(b) they are insecure about its stability, and (c) the potentially
threatening other is attractive (Simpson, Ickes, & Blackstone,
1995).
The classic psychodynamic defense hypothesis of prejud i c e - t h a t one function of stereotyping is to bolster self-esteem
by derogating outgroups--has recently received experimental
support in a series of studies in which manipulations that affirmed valued aspects of self or threatened self-esteem led to
predicted decreases or increases in prejudiced responding
(Fein & Spencer, 1997). A recent study with genital plethysmography has breathed new life into the classic psychoanalytic
theory of homophobia as a defense against threatening homosexual feelings (Adams, Wright, & Lohr, 1996). Participants identified by questionnaire as homophobic, unlike nonhomophobic
comparison participants, showed increases in penile circumference (indicating arousal) while exposed to sexually explicit
videotapes showing male homosexual activity. Although one
could attempt to explain these data in other w a y s - - f o r example,
suggesting that anxiety, novelty, or surprise could lead to penile
arousalJthe more parsimonious explanation of this and similar
electrophysiological findings reviewed below is that people can
be unaware of, and defend against, things they find threatening,
including their own feelings.
Of particular relevance is the literature on self-serving biases
and more recent empirical outgrowths of it on narcissism. For
years, researchers have documented numerous self-serving biases, such as the tendency to see oneself as above average on
positive characteristics such as intelligence (when, of course,
not everyone can be above average), to overrate one's role
in group projects with positive outcomes, and so forth (see
Dunning, Leuenberge~; & Sherman, 1995; Epstein, 1992;
Greenwald & Pratkanis, 1984). More recent research has begun
to track down individual differences in the tendency to see ourselves unlike others see us, and has found, as clinical observation
suggests, that narcissistic people tend to hold defensively grandiose views of themselves and their accomplishments and that this
is not without substantial social and personal cost. In one study,
MBA students ranked themselves and their peers on their contribution to a group problem-solving task, as did a group of psychologist observers (John & Robins, 1994). Participants' rankings of their peers were remarkably consistent with the psychologists' rankings, but their self-rankings correlated with peer and
psychologist rankings of them at only about r = .30; 60% overestimated their own performance, suggesting a self-serving bias.
Participants who showed the most substantial self-serving biases
were significantly more narcissistic by both observer report and
self-report. Indeed, the tendency of participants to self-enhance
in this study correlated around r = .50 with psychologists'
ratings of their narcissism.
A longitudinal study compared two groups of 1st-year college

341

students who were matched in level of ability on the basis of

high school grades and Scholastic Assessment Test (SAT) scores
(Robins & Beer, 1996). One group was strongly self-enhancing;
their perceived (self-reported) level of academic competence
was much greater than their actual past performance (assessed
from records in the admissions office). The other group was
relatively accurate in assessing their own abilities. At the end
of their sophomore year, the self-enhancers, who had expected
much greater academic success than they had achieved, reported
significantly higher subjective well-being. A behavioral measure, however, suggested otherwise: They were 32% more likely
to have dropped out of school.
Other research demonstrates that people whose views of
themselves are overly favorable, relative to reports by independent observers or peer informants, show numerous psychological
and social difficulties, both longitudinally and cross-sectionally
(Colvin, Block, &Funder, 1995). For example, men who were
highly self-enhancing at age 18 were described by observer Q
sorts at age 23 as having fluctuating moods, being deceitful and
distrustful, and having brittle defenses. Low self-enhancers, in
comparison, were described as cheerful, straightforward and
forthright, and dependable. (The data are not yet in on whether
some degree of self-enhancement, accompanied by "positive
illusions" [Taylor & Brown, 1988] that foster hope and effort,
is adaptive.) Studies of leadership (summarized in Hogan, Curphy, & Hogan, 1994) similarly have found that leaders who are
narcissistic (characterized by traits such as arrogance, vindictiveness, and selfishness) tend to fail despite the presence of
other traits typically associated with good leadership, such as
intelligence, perseverance, and ambition. Interestingly, these
leaders are difficult to detect by self-report, because they present
themselves as confident and high in self-esteem.
Defense against Unpleasant emotions and its physiological
correlates. Several other literatures bear on psychological defense and particularly on the hypothesis that people can prevent
themselves from consciously experiencing affect as a way of
trying to manage unpleasant feeling states. The literature on
adult attachment demonstrates that people inhibit conscious access to representations of the self, others, and relationships on
the basis of their affective qualities. Unlike adults with secure
attachment styles, who speak freely and openly about their relationships with their parents, avoidant adults dismiss the importance of attachment relationships or offer idealized generalizations about their parents yet are unable to back them up with
specific examples (Main, Kaplan, & Cassidy, 1985). Dozier
and Kobak (1992) have produced physiological evidence of
the discordance between what avoidant adults know and feel
consciously and unconsciously. The researchers monitored electrodermal response while participants were asked to recall memories involving separation, rejection, and threat from their parents. Avoidant adults, like avoidant infants, are hypothesized to
shut off or deactivate attachment-related feelings as a way of
coping with the distress associated with them. In fact, the more
individuals showed avoidant, deactivating strategies by disavowing feelings, the more physiological reactivity they manifested while answering affectively evocative questions about
separations, rejections, and parental threats.
Two entirely different lines of research have produced parallel
findings. Shedler, Mayman, and Manis (1993) studied partici-

342

WESTEN

pants in two studies with illusory mental health, who reported

themselves to be free of psychological distress and symptomatology but whose descriptions of their early memories (a projective measure) were rated as showing signs of psychological
disturbance. Participants underwent a mildly stressful procedure
that can be disturbing to someone who is highly defensive (reading aloud, performing a phrase association test, or providing
projective stories). Those participants who viewed themselves
as healthy but showed unconscious evidence of distress in their
early memories were significantly more reactive on a measure
of cardiac reactivity related to heart disease than participants
who were either low or high on both measures of distress. They
also showed more indirect signs of anxiety (such as stammering,
sighing, and avoiding the content of the stimulus) while simultaneously declaring themselves to be the least anxious during
these tasks. A subsequent study found that illusory mental health
predicted the number of visits students made to the health center
over the next year, as did the amount of negative affect present
in early memories, regardless of level of self-reported distress
(Cousineau, 1997). Strikingly, self-reported distress, stress, and
life events all predicted self-reported visits to the doctor, but
none predicted either actual visits or visits that led to a referral
from the nurse to a doctor.
Research on repressive coping styles (which refer not to the
use of repression but to a tendency to avoid feeling emotions
as a way of managing distress) documents a direct link between
defensive disavowal of affect, particularly anger, and physical
illness (Schwartz, 1990; D. Weinberger, 1992, 1995). Participants characterized by this coping style report low anxiety but
have high scores on the Marlowe-Crowne scale (Crowne &
Marlow, 1960), which appears to measure some combination
of defensiveness, social desirability, and overcontrol of affect.
Repressors, like avoidants in the Dozier and Kobak (1992) study
and like the participants with illusory mental health in the studies
by Shedler, Mayman, and Manis (1993), report low levels of
anxiety while demonstrating marked physiological reactivity.
They also have difficulty retrieving unpleasant memories
(Davis & Schwartz, 1987), as do people with avoidant attachment styles (Mikulincer and Orbach, 1995). Weinberger and
his colleagues (D. Weinberger, 1995) have demonstrated a link
between repressive coping (particularly suppression of aggressive or angry feelings) and both cholesterol levels and asthma.
Individuals with this coping style, like individuals who report
high levels of anxiety, also show higher salivary cortisol levels
than genuinely low-anxious individuals; this suggests that people
with a repressive coping style are hyperreactive to potentially
stressful encounters (Brown et al., 1996). Other research has
uncovered an association between repressive strategies and vulnerability to cancer (Jensen, 1987; D. Weinberger, 1992). Apparently, inhibiting conscious access to one's emotions places the
body, particularly the heart and the immune system, under considerable stress. Freud may be dead, but his theories are proving
increasingly useful in predicting who will be joining him sooner
rather than later.
These findings converge with a literature that has existed for
many years documenting a negative correlation between emotional expressiveness and physiological reactivity (see Gross &
Levenson, 1993). In a sample of several thousand medical patients, for example, Schwartz, Krupp, and Byrne (1971) found

that participants who minimized their conscious feelings of anxiety reported more organic problems, whereas those who focused on them reported more psychological problems. A recent
meta-analysis found that defensive constriction of emotional
experience was the best predictor of essential hypertension (elevated blood pressure of unknown origin) of any personality
variable yet studied (Jorgensen, Johnson, Kolodziej, & Schreer,
1996).
A program of research by Pennebaker ( 1989, 1997) has demonstrated that simply writing about or discussing painful experiences (such as job loss in unemployed professionals) produces
increases in immune functioning, physical health, and adaptive
behavior (such as getting a new job). Expressing unpleasant
emotion leads to a momentary increase in arousal but a decrease
in arousal in the long run (Hughes, Uhlmann, & Pennebaker,
1994), apparently because keeping oneself unaware of what one
feels prevents dissipation of the feeling, keeps related cognitions
active, and hence maintains the potential for unconscious priming of aversive arousal. (From a clinical perspective, one might
add a behavioral loop: Keeping oneself unaware of unpleasant
feelings prevents a person from exercising conscious control
over situations that might trigger the affect.) Of relevance to
these findings is a study in which group psychotherapy for
women with advanced breast cancer extended their lives by an
average of 18 months relative to control participants matched
for stage of the illness (Spiegel, Bloom, Kraemer, & Gottheil,
1989).
What is important about the studies described here is that
they demonstrate the existence not only of unconscious affective
and motivational processes (unconscious efforts to control aversive affective states) but of what Freud (1915/1957c) referred
to as dynamically unconscious processes; that is, those that are
kept unconscious for a reason. The existence of these processes
has long been a central assumption of psychoanalytic theory, an
assumption that continues to distinguish it from even contemporary cognitive approaches to unconscious processes (e.g., Kihlstrom, 1987, 1990). Although many researchers have difficulty
with the notion that such active, seemingly intentional processes
could occur outside of awareness, many of the mechanisms are
probably not unlike familiar conditioning processes (Dollard &
Miller, 1950; Wachtel, 1997; Westen, 1985, 1994). Just as people can learn to avoid a stimulus associated with pain through
negative reinforcement, they can learn to avoid focusing attention on particular cognitive or affective processes because doing
so is associated with shame, guilt, sadness, or anxiety. Nothing
about the architecture of memory requires that these procedures
or their triggers be accessible to consciousness; in fact, most
processes that guide attention are not accessible to introspection.
In the final section of this article I describe some possible mechanisms through which implicit affect-regulatory procedures of
this sort might occur.
At this point, I believe the data on" unconscious affective
processes are incontrovertible. Considerable affective processing occurs unconsciously in dally life whether or not a
person is responding defensively. Conscious affective experience, like conscious cognition, is likely assembled through the
action of multiple neural modules operating in parallel. The
capacity to respond with automatic affective responses, some
of which occur in the absence of conscious recognition of a

FREUD'S LEGACY
stimulus, is highly adaptive and likely guided our ancestors long
before they developed the kind of reflective self-awareness characteristic of contemporary humans. Blocking conscious emotional experience carries a cost, because people who chronically
keep themselves unaware of their feelings are more likely to
suffer from physical disorders such as heart disease. (Readers
who remain unconvinced should probably consult their
cardiologists.)

Unconscious Motivation
Consciousness, at least in its human form, appears to be a
relatively recent evolutionary development superimposed on an
information-processing system that worked relatively well for
millions of years (Reber, 1992). To assume that consciousness
is essential for goal-directed behavior suggests that our preconscious protohuman ancestors managed to escape extinction
for millennia through simple good luck, despite their lack of
motivation, until consciousness felicitously evolved. Much of
human behavior is in fact simultaneously motivated by multiple
goals, which would disrupt goal-directed behavior if they all
had to be represented in consciousness because they would
consume too much working memory. As I write these words,
for example, I am clearly motivated to make a coherent argument, write grammatically, present studies accurately, accommodate an imagined audience of whom I would expect many to
be initially skeptical, present myself as a competent scientist,
and act on a host of other less admirable motives that may have
expressed themselves despite m y best efforts to the contrary. At
the very least, motives must be capable of automatization and
hence unconscious activation, much as skills are. As shown
below, there is now considerable evidence for this last proposition (see Bargh, 1997).
Data on defensive processes simultaneously provide evidence
also for the existence of unconscious motivational processes,
because a defense is by definition-a motivated unconscious effort
to minimize painful or maximize pleasurable emotion. A growing body of evidence from other quarters, however, speaks even
more directly to the phenomenon of unconscious motivation,
suggesting that the distinction between implicit and explicit processes applies not only to cognition and affect but to motivation
as well. Some of the best evidence comes from research comparing the correlates of self-report and projective measures of mefives (see McClelland et al., 1989). The correlation between
these two types of measures typically hovers around zero; for
example, self-reported and projective assessments of the need
for power tend to show no relation to each other, leading adherents of each method to proclaim that the other is obviously
invalid. However, each type of measure predicts different kinds
of behavior. The correlates of each are highly predictable if one
views consciousness as an override mechanism or as a special
lens for examining information and choosing courses of action
when unconscious standard operating procedures have broken
down or require monitoring. For example, over the long run,
assessment of motives from TAT stories is much more predictive
of entrepreneurial or managerial success than are self-report
measures of need for achievement or power, which tend to have
little predictive validity. Tell people that they need to achieve on
a particular task, however, and their self-reported achievement

343

motivation will be a much better predictor of their effort and

performance than will motives assessed from TAT responses.
In both his own studies and a meta-analysis of objective and
projective studies of dependency, Bomstein (1995) has produced similar findings with respect to the construct of dependency: Self-report and projective measures have predictable but
different correlates. Peterson and Ulrey (1994) may have uncovered a similar phenomenon in research on explanatory style
(the tendency to attribute negative events globally, stably, and
internally). Both TAT-based and self-report measures of explanatory style correlated with self-reported depression, but the two
measures of explanatory style (one implicit and the other explicit) did not correlate with each other, suggesting that individu~
als may be pessimistic in their attributions consciously, unconsciously, or both. s
These and other data suggest that when conscious motives
are activated, they guide behavior. When they are not activated,
which is much of the time as people muddle through their lives,
unconscious motives guide behavior. This should come as no
surprise to people who have broken New Year's resolutions or
to cognitive psychologists who assume a similar phenomenon
with respect to cognition: Controlled information processing
operates when people are focusing attention on a problem or
task, whereas automatic information processing--including associative memory, activation of procedural skills, and activation
of schemas--guides thought, emotion, and behavior in the absence of consciously focused attention (see Schneider & Shiffrin, 1977).
Bargh (1997; Bargh & Barndollar, 1996) has produced another, quite different, line of research on unconscious motivation,
documenting the existence of Unconscious motivational processes experimentally. Bargh extends the literature on automatization of cognitive processes (Anderson, 1995) to motives, arguing that well-learned goals can be activated by environmental
stimuli--and attendant behavioral plans can run their c o u r s e - even without conscious awareness. Bargh takes issue with the
conception of unconscious processes prevalent in cognitive psychology (e.g., Greenwald, 1992) as simple and irrational. He
contends, instead, that the history of a person's learning in a
given situation is embodied in habitual and automatic motives.
These motives often provide a better guide to behavior than
does conscious and presumably rational analysis of a single
current instance, which may be ignorant of base rates and prior
automatic actions. This supposition accords with research by
T. D. Wilson (e.g., T. D. Wilson & Schooler, 1991; T. D. Wilson
et al., 1993), which demonstrates that people's "gut level"
feelings are often more effective guides to action and lead to
more subsequent satisfaction than do their reasoned reflections,
which can interfere with emotion-based judgments. Damasio
and colleagues (Bechara, Tranel, Damasio, & Damasio, 1996)
have obtained parallel findings with patients with ventromedial
prefrontal damage, whose reasoning processes are largely intact

s This finding also has potential implications for outcome measures

in studies of the efficacy of cognitive therapies for depression, which
may produce misleading findings because the treatment is, among other
things, a training procedure for altering or disavowing negative conscious cognitions about the self.

3~

WES~N

but who cannot use prior affective associative learning to guide

their responses adaptively.
Bargh suggests that priming procedures that have been shown
to influence the way people subsequently categorize or make
inferences should be equally capable of influencing their motives
and cites evidence from his own and others' laboratories in
support of this supposition. For example, researchers presented
participants with scenarios in which a protagonist was interested
in either accurately assessing a situation or making a good impression (Chaiken, Giner-Sorolla, &Chen, as cited in Bargh &
Barndollar, 1996). Then, in a "second study" that appeared to
be independent of the first, participants were asked to discuss
their opinions about a topic with a partner after being told what
the partner's attitudes toward it were. Participants who had been
primed in the first "study" with the impression management
story were more likely to align their attitude with that of their
partner, whereas those in the accuracy condition were unswayed
by their partner's alleged position. Thus, simply activating
thoughts about two different kinds of motives actually primed
the motives.
In an important series of studies paralleling the work of
McClelland et al. (1989), Bargh and his colleagues (Bargh,
1997) primed participants with words related to either achievement (e.g., strive) or affiliation (e.g., friend) in an allegedly
unrelated first task. Next, they placed participants in a situation
of motivational conflict: Participants worked with an incompetent partner on a puzzle task at which they could either succeed
and hence make their partner feel humiliated and stupid or not
be so successful but protect the confederate's self-esteem. As
predicted, participants who had been primed with achievement
words outperformed participants primed with affiliation words
(and control participants), even though upon debriefing they
were, not surprisingly, unaware of the potential influence of the
primes. In a second study, the same findings occurred, but the
manipulation wore off after the first few trials. One of the variables that then predicted success was participants' need for
achievement as assessed by the TAT. In other words, after recent
activation effects decayed, chronically active motives "kicked
in" to control behavior. Similar findings occurred in the study
by Chaiken and colleagues (as cited in Bargh & Barndollar,
1996).
What these studies essentially show is that many of the Same
processes found in research on unconscious cognitive processes,
such as the impact of both recently and chronically activated
mental processes on thought and behavior (Higgins, 1990),
ap~ly to motivation as well. Research on prospective m e m o r y - remembering to remember something in the future, such as
picking up a few items at the grocery store after w o r k - - h a s
shown that intending to carry out an action leads to heightened
activation of the information to be remembered below the threshold of consciousness even when the intention is no longer conscious, as evidenced in both explicit and implicit memory tasks
(such as reduced response latencies for recognizing to-be-remembered items from a list; Goschke & Kuhl, 1993). In other
words, active motives influence ongoing thought and action
whether or not they are conscious.
Further data on unconscious motives come from a variety of
other sources, including case studies of patients with neurological damage. An early demonstration occurred almost a century

ago when Claparede (described in Cowey, 1991) shook hands

with a patient suffering from Korsakoff's disorder. Claparede
had concealed a pin between his fingers, which pricked the
patient as their hands clasped. Upon meeting again, the patient
was unable to recognize Claparede (because of her amnesia for
recent events) but was nonetheless unwilling to shake his hand,
despite being unable to explain why. Gazzaniga (1985) has
described the similar behavior of split-brain patients who can
carry out instructions to produce some action presented to their
right hemisphere. When subsequently asked what they are doing,
they offer seemingly sensible but incorrect rationalizations for
their behavior, because the motivation is inaccessible to their
right hemisphere.
Similar hypnotic demonstrations, in which participants given
a suggestion for posthypnotic amnesia would carry out an action
and then try to explain it, were precisely the kind of data that
led Freud to his hypotheses about unconscious motivational processes. He was also struck by the fact that hysterical symptoms
could be created hypnotically, leaving their bearers as unable to
explain them as true hysterics, and that true hysterical symptoms
could sometimes be temporarily relieved by hypnotic suggestion
(see Erdelyi, 1985). The fact that such hypnotic cures tended
to fade may not be unrelated to the fading of priming effects and
the ultimate resurgence of unconscious influences on behavior.
Phenomena such as these are not limited to neurological patients, hypnotic subjects, and hysterical patients. Nisbett and
Wilson (1977) reviewed a wide array of studies demonstrating
that when people act on the basis o f motives or preferences for
which they cannot access reasons, they tend to "tell more than
they know," making up sensible explanations based on their
intuitive theories of themselves and of psychological causality
that are entirely post facto and often incorrect. In more recent
research, T. D. Wilson (e.g., T. D. Wilson & Schooler, 1991;
T. D. Wilson et al., 1993) has shown that when people subsequently act on the basis of such reasons or when they try to
think about why they might prefer one option over another
before expressing their preference, they often make worse
choices, and the original affective preference ultimately returns.
I have, no doubt, failed to cite a number of relevant studies,
some conducted by New Look researchers decades ago and
others by contemporary social psychologists, but I believe the
point is now sufficiently clear: Freud was fight in his most
central hypothesis, that much of mental life, including thought,
feeling, and motivation, is unconscious. Indeed, Freud and the
psychoanalytic theorists who have followed in his footsteps have
turned out to be right where competing psychological theories
were once completely wrong and where most contemporary
theories of consciousness and cognition continue to be either
wrong or incomplete (e.g., about unconscious defenses).
In retrospect, the heated arguments over the last century about
the proposition that much of mental life is unconscious are
somewhat perplexing in light of data showing that children in
Western cultures begin to recognize the nature and functions of
unconscious processes, notably defensive processes, by the time
they are 10 years old (Chandler, Paget, & Koch, 1978). Why
this debate has been so heated and why discourse on Freud,
unlike that on any other person in the history of psychology,
has often been so ad hominem and ignorant of subsequent theoretical and empirical developments related to the approach he

FREUD'S LEGACY
initiated (e.g., Crews, 1996), are questions for historians of
psychology. At this juncture, however, I believe the question
of whether unconscious cognitive, affective, and motivational
processes exist is no longer an interesting or informative one
and that we would do better to turn our attention to the implications of their existence and the mechanisms by which they influence information processing and behavior.

Conflict, Ambivalence, and the Modular Mind

A second characteristic that has distinguished psychodynamic
theory from other approaches to psychology since its inception
is its focus on mental conflict and the correlative assumption
that conscious experience and behavior reflect competition, collaboration, and conflict among quasi-independent psychological
events occurring below the threshold of consciousness (Brenner,
1982; Freud, 1900/1953b)i From the time Freud began observing patients, be noted the ubiquity of the conflict inherent in
wanting to be free of symptoms that are clearly volitional (such
as ritualized hand-washing or binge eating) but being unable to
eliminate them. From these observations, Freud made a simple
deduction: If the person has a conscious will to overcome a
symptom (such as bulimic bingeing) but cannot do so, and the
problem is not organic, some unconscious counterwill must be
blocking the conscious volition (see Erdelyi, 1985). To put it
another way, if a person is producing meaningful behavior involving voluntary musculature (such as bingeing) but cannot
explain why and wishes he or she could do otherwise, the motive
for producing the behavior not only must be stronger than the
conscious will but must also be unconscious.
Ultimately, this insight about intraPSYChic conflict led Freud
to the postulation, similar in certain respects to contemporary
connectionist or parallel distributed processing (PDP) models
in cognitive science (Rumelhart, McClelland, and the PDP Research Group, 1986), that multiple psychological processes can
proceed in parallel. Freud ( 1915/1957c, p. 170) noted that "the
different latent mental processes inferred by us enjoy a high
degree of mutual independence, as though they had no connection with one another" and argued against the theory that unconscious processes are really a dissociated "second" consciousness by suggesting that this would require postulation of an
infinite number of consciousnesses that conflict and cooperate
outside awareness.
Contemporary connectionist models view representations as
distributed across a network of interconnected units of information, analogous (or perhaps homologous) to a network of neurons, which, through repeated coactivation, become activated in
concert to represent a concept. This produces top-down, hypothesis-driven processing, so that, in a casino for example, a person
whose concept of gambling (rather than bakeries) is activated
uses this context to rule out the need to look under the feet of
the nearby gambler who says he is "on a roll." Activation of
some of the neural-cognitive units that represent a concept
facilitates categorization of a particular instance as a member
of a particular category (or, in the gambling example, facilitates
simple lexical interpretation), whereas activating others (in this
example, perhaps, if the gambler is at a bakery instead of a
casino) inhibits this categorization. The final "decision" the
brain makes when confronted with the task of categorizing a

345

novel stimulus (or completing any other cognitive or perceptual

task) reflects an unconscious process of constraint satisfaction,
in which the combination of " o n " and " o f f " signals from
thousands or millions of neurons or processing units produces
an imperfect but good-enough match between current input and
prior knowledge stored as patterns of firing.
Freud was as uninterested in the processes by which people
categorize stimuli as contemporary cognitive psychologists are
in the conflicts among wishes, fears, and values that attract the
attention of psychodynamic theorists and clinicians. However, a
fundamental similarity in their models is the postulation of multiple independent processing units that combine, conflict, and
collaborate to produce psychological decisions, some of them
conscious and others unconscious. As I suggest later, a point of
integration of the two models lies in the fact that the process
of constraint satisfaction described by artificial intelligence researchers leaves out a key set of Constraints that are central to
psychoanalysis: affective and motivational pulls from various
directions, which must be satisfied alongside cognitive constraints that maximize accuracy. When fans of two basketball
teams in a hotly contested championship game witness the same
act of physical contact between two players and those on one
side categorize it as a foul whereas the others do n o t - - e v e n
after watching a televised sl0w-motion r e p l a y ~ i t is clear that
even simple perceptual tasks can be influenced by affective and
motivational constraints. According to psychoanalytic theory,
this is the case with most cognitive "decisions," because humans tend to have feelings about most of the people and objects
with which they come in contact. Making matters more complicated, however, is the fact that their feelings and motives often
run in conflicting directions.
From a psychodynamic perspective, nothing guarantees that
an individual' s wishes, fears, values, or goals will be in harmony,
because these motives, or components of them, may be processed by relatively independent neural circuits. Thus, a person
can want and fear the same thing, which leads to conflict. From a
developmental perspective, conflict is virtually built into human
existence. Whom do children first learn to love? And who first
frustrates them and makes them enraged? For children in most
cultures and families, precisely the same people are associated
with security, comfort, distress, and disappointment: their primary caregivers. So conflict is built into the earliest social expectations, feelings, wishes, a n d fears simply by virtue of the
fact that memory is built on principles of association, a crucial
point of agreement and potential integration between psychodynamic and cognitive theories, to which I later return. When the
child of an alcoholic parent "finds" himself or herself with an
alcoholic spouse, is this accidental? Given the number of children of alcoholics who vow at some point in their adolescence
that they will never pick a mate like their parent or that they
will never expose their own children to the same kind of family
environment they experienced, surely some counterwill must be
expressing itself to override this conscious will. Such a child
may well have encoded some aspects of the parent's behavior
not only as something to be feared and avoided but also as part
of what it means to be an appropriate love object (for relevant
data, see McCord, 1988).
Or consider a more distinctly Freudian example of ambivalence. The anus is such a disgusting and smelly region that most

346

WESTEN

people will not even touch their own without the intervention
of a piece of paper. So why do people often say others have
"cute buns," and why is touching the buttocks or anus a relatively common aspect of people's sexual life or fantasies? Why
do many men enjoy looking at women in G-strings, thong underwear, or thong bikinis that accentuate and call attention to this
seemingly disgusting region of the body? Given that the vast
majority of children's first 10 years of associations to the genitals and contiguous regions involve excretion, it is a wonder
(and a tribute to the power of hormonal influences on thought
and motivation) that anyone can ever have enjoyable sex. Ambivalence is thus virtually built into human sexuality. Clinically,
this is most apparent in patients with sexual contamination issues, who are often most attracted to, and simultaneously repulsed by, the same sexual acts or body parts.
Once again, research supports this second fundamental psychodynamic supposition. That mental processes operate in parallel and that people have minimal access to the component processes that ultimately contribute to their conscious experience
and behavior are now taken for granted by most cognitive scientists, although they have tended not to consider affective and
motivational processes in this respect. The fact that positive and
negative affect are mediated by different neural circuits (see,
e.g., R. Davidson, 1992; Gray, 1990) lends credence to the view
that nothing about the mind requires that any given stimulus be
associated with only a single affective valence. Indeed, several
literatures are independently converging on the notion that positive and negative emotional responses and interpersonal interactions are only moderately correlated, are processed separately,
and have distinct correlates, including research on the impact of
supportive and harsh parenting (Pettit, Bates, & Dodge, 1997),
ingroup favoritism and outgroup derogation (Brewer & Brown,
1998), positivity and negativity in children's relationships
(Hartup, 1996), and acceptance and negative reciprocity in marital couples (Arkowitz-Westen, 1998).
The complexity of human cognition, particularly about the
significant others with whom individuals have repeated contact,
virtually assures that any given representation will be associated
with multiple affects. The research reviewed previously on implicit versus explicit emotional and motivational processes
clearly documents not only that people may have multiple feelings about the same object but that their conscious and unconscious evaluations can be in substantial conflict, as in conflicting
attitudes toward racial minorities or conflicting motives toward
power. If the affect associated with a representation motivates
approach or avoidance of it, if multiple affects are associated
with most important objects (certainly in the social realm, which
becomes readily apparent in any relationship that lasts as long
as 6 weeks), and if many of these evaluative reactions are
unconscious and can be unconsciously triggered--all of which
are now empirically supported propositions--then the view of
the mind as filled with conflict that once seemed so peculiarly
psychoanalytic now seems eminently sensible. Indeed, if part
of what the cognitive revolution did to prior research from a
behaviorist paradigm was to show how concepts such as stimulus generalization actually imply cognitive processes of schema
formation and categorization (i.e., stimuli are considered equivalent for the purposes of responding when sufficient overlap exists
between the neural networks that represent them), then what

contemporary psychodynamic thinking can do is to suggest that

approach-avoidance conflicts can similarly be structuralized
through repeated associations, which may or may not be
conscious.
A small body of research has examined conflict and ambivalence empirically (e.g., Glick & Fiske, 1996; Harter, 1986; Sincoff, 1992). In one study, students listed 15 of their personal
strivings or goals, defined as objectives " t h a t you are typically
trying to accomplish" (Emmons & King, 1988, p. 1042). Participants then rated the extent to which each striving conflicted
with every other striving, and the researchers used the matrix
of their ratings to construct a mean index of level of conflict for
each individual. Participants also reported the extent to which
success at the striving would lead to some unhappiness, a second
measure of ambivalencel Dependent variables included participants' daily mood reports taken twice a day over 21 consecutive
days; reports of somatic complaints such as headaches,
coughing, and acne; and visits to the health service. Conflict
and ambivalence scores showed modest but significant correlations with reported emotions and illness.
Periodically, attitude researchers have challenged the view
that attitudes can be represented by using a single bipolar evaluative dimension (negative to positive); this view is once again
ascendant, as researchers have begun focusing on attitudinal
ambivalence (Cacioppo, Gardner, & Berntson, 1997; Priester &
Petty, 1996; Thompson, Zanna, & Griffin, 1995). Drawing on
a range of social-psychological and psychophysiologicai studies,
Cacioppo et al. (1997) have concluded that attitudes include
two evaluative dimensions, positive and negative, which are relatively independent and can each be either relatively weak or
relatively strong. Low-positive/low-negative attitudes are likely
to have minimal impact on behavior and are quite distinct from
high-positive/high-negative (ambivalent) attitudes, which nonetheless yield similar moderate scores on traditional bipolar attitude measures. In one set of studies, for example, people who
were low on both positive and negative attitudes toward blood
donation were significantly less willing actually to donate blood
than those who were ambivalent, who were apparently able to
harness their social values to overcome their (probably classically conditioned) squeamishness (Gardner & Cacioppo, 1996,
as cited in Cacioppo et al., 1997).
Other investigators have examined conflict from a developmental perspective. As early as 12 months of age, infants consult
their primary caregivers when confronting a novel stimulus such
as a toy or person, and if they receive conflicting facial messages
from their two parents (a smile from one and a fearful response
from the other), they become distressed and confused (Hirshberg, 1990). Research with children whose attachment status is
defined as disorganized, that is, whose attachment figures are
highly unpredictable, shows how these children' s own confusing
responses (such as mixtures of approach and avoidance, often
both aborted and replaced with ritualized actions) reflect internalization of the confused messages they receive about the comfort and discomfort associated with seeking nurturance (see
Main, 1996; Main & Solomon, 1986). As cited previously, adults
with an avoidant or dismissive attachment style manifest signs of
covert emotional distress in attachment relationships alongside
efforts to disavow it (Dozier & Kobak, 1992). Along different
lines, a cross-sectional study asked children, adolescents,

FREUD'S LEGACY
and college students to rate the degree of support and conflict
they encountered in various relationships (Furman & Buhrmester, 1992). Of particular relevance was the finding that relationships with parents were at their most supportive at the same
time as they were at their most conflictual.
So how do people resolve conflicts when confronted, consciously or unconsciously, with multiple competing affective
and motivational pulls? One of the most important concepts in
contemporary psychodynamic theory is the notion of compromiseformation, which attempts to address that question (Brenner, 1982; Westen, 1985, chapter 2). A compromise formation
is a thought, feeling, or action that reflects a compromise among
multiple motives, such as the desires to maintain self-esteem,
to obtain some kind of gratification, to respond to moral imperatives, to escape unpleasant emotions, and to perceive reality
accurately. The televangelists who preached the evils of illicit
sex while engaging in it in one colorful variety or another in
the 1980s provide useful examples. Presuming that one gives
them the benefit of the doubt and does not assume that they were
simply conscious hypocrites who secretly enjoyed flaunting the
values they professed, one might hypothesize that their religious
preoccupation with sex allowed them to focus their attention on
sexual images and fantasies that their rigid sense o f right and
wrong precluded. At the same time, fighting this evil in other
people may have helped assuage their shame and guilt for their
thoughts (and, it turns out, actions). The result was a compromise that maximized feelings of both titillation and moral righteousness: They preached vociferously against sex acts that, on
occasion, they had difficulty containing.
This is of course speculation, because one cannot psychoanalyze at a distance, but the public details of these cases are well
known and hence provide a useful example. A simple clinical
example comes from the therapy of a man who was terribly
conflicted about whether to stay in a very unsatisfying marriage
or to end it. He was highly prone to guilt, which kept him in
the relationship for years beyond the point at which the balance
of pleasure to displeasure in the marriage had shifted overwhelmingly toward the debit column. When he finally left his
wife, he turned down opportunities to date two women who
really attracted him and rented a small, somewhat seedy apartment that seemed to be designed both to make him uncomfortable and to dissuade any woman from "coming back to his
place." When we explored his associations to these choices, the
compromise he had unconsciously crafted seemed apparent: He
could leave his wife, but to do penance for his guilt he could
not enjoy himself, particularly with other women.
Empirical research on motivated reasoning has indirectly documented the existence of similar processes. Extraverts who are
induced to believe that introversion is conducive to academic
success will come to see themselves as less extraverted, but
they will not completely deny their extraversion (see Kunda,
1990). That is, they will form an on-line compromise between
enduring self-representations and current motives. Similarly, in
a study described previously, the majority of MBA students
showed self-enhancing biases when asked to rank their contribution to a group decision-making task, but most of them ranked
themselves within one rank of the way their colleagues and
objective observers ranked them (John & Robins, 1994). In
other words, when faced with a conflict between accuracy and

347

self-enhancement, people compromise, creating a distorted selfportrait that allows them a balance of satisfaction of both mofives. The researchers did not ask, but I doubt that participants
were aware of doing this,
Similar findings emerged in a study of the accuracy of memory, in which college student participants recalled their math,
science, history, English, and foreign language grades from high
school (Bahrick, Hall, & Berger, 1996). When checked against
their actual high school grades, students recalled 71% of their
grades correctly. More interesting, however, was the pattern of
their errors: Participants correctly remembered 89% of their As
but only 29% of their Ds. Ds were as likely to be remembered
as Bs and Cs as they were as Ds, but they were virtually never
incorrectly remembered as As, which suggests a compromise
between veridical information processing and desired grades.
When the investigators created for each participant an index of
the direction of distortions, approximately 80% inflated their
remembered grades, whereas only 6% reported grades lower
than they had actually achieved. The remaining 14% remembered correctly.
One might similarly consider the social-psychological literature on self-enhancement versus self-verification in terms of
compromise formations. Decades of research have shown that
people tend to view themselves more positively than do neutral
observers, primarily because doing so protects their self-esteem.
On the other hand, Swann (1990; Swann, Stein-Seroussi, &
Geisler, 1992) has shown that people are motivated not only to
self-enhance but to confirm their preexisting views o f themselves (and presumably to see themselves accurately as well).
From a psychodynamic perspective, one would predict that conscious self-representations or self-schemas reflect a balance of
these often-conflicting motives. A recent series of studies has
documented the way people create compromise formations that
express precisely that balance (Morling & Epstein, 1997). In
three studies, self-enhancement and self-verification motives led
to compromise solutions in response to evaluative feedback,
leading participants to prefer feedback that was only moderately
self-enhancing. Participants high and low in self-esteem differed
substantially in the feedback they preferred, because each preferred feedback that enhanced them but not too far above what
they believed about themselves.

Continuity of Personality and the Role

of Childhood Experiences
A central set o f claims of psychodynamic theory from the
start has been that stable personality patterns begin in childhood
and that childhood experiences with caretakers play a crucial
role in shaping personality (especially in shaping later ways of
becoming attached to and intimate with others). For years, this
claim distinguished psychodynamic from other theories. Freud
argued that personality crystallizes in childhood, whereas learning theorists contended that behavior is (a) selected by its environmental consequences (Skinner, 1953), (b) shows minimal
consistency over time and situations (Mischel, 1968), and (c)
hence can be readily controlled and altered at any age (see
Bandura, 1977). The formative role of childhood has long been
a fulcrum of debate in psychology (see Kagan, 1980; Kagan &
Zentner, 1996).

348

WESTEN

Continuity Over lime: A More Recent Interpretation

Research in behavior genetics (Loehlin, 1992; Plomin,
DeFries, McClearn, & Rutter, 1997) has demonstrated the importance of distinguishing the two subcomponents of this third
psychodynamic proposition: (a) that adult personality characteristics begin to crystallize in childhood a n d ( b ) that childhood
experiences are formative for later personality and social functioning. Freud was a strong believer in temperament, even suggesting (1908/1953a) that the tendency to become fixated at
one stage or another might have biological influences. However,
neither he nor anyone else in the field ever imagined the kinds
of heritability estimates (most in the range of .30 to .50) that
twin and adoption studies have produced.
With respect to the crystallization of personality in childhood, considerable continuity between child and adult personality has been well established since Block's ( 1971 ) landmark
Lives Through Time. More recent research has supported the
proposition that, although the child is not the sole contributor
to the man, he is at least the father, as intimated by poets such
as Wordsworth and as developed as a psychological principle
a century ago by Freud. The data are particularly strong for
a link between two personality styles or types--inhibited and
undercontrolled--and subsequent personality characteristics
and psychopathology. For example, behavioral inhibition to
the unfamiliar, which includes anxiety when faced with novelty and shyness, shows substantial continuity from infancy
through early childhood (Kagan & Snidman, 1991) and from
middle childhood through adulthood (Gest, 1997), with correlations across several years as high as .50. In a longitudinal
study, children described as inhibited at age 3 were more
likely than comparison participants to be depressed at age 21,
whereas children described at age 3 as undercontrolled and
impulsive were more likely to meet diagnostic criteria for
antisocial personality disorder 18 years later (Caspi, Moffitt,
Newman, & Silva, 1996). Another longitudinal study found
that 8- to 10-year-old boys characterized as ill tempered were
distinguished at age 30 by maladaptive personality traits, poor
occupational performance, and disrupted marriages (Caspi,
Elder, & Herbener, 1990). Ill-tempered boys were rated 20
years later as undercontrolled, irritable, moody, unethical, and
undependable. A Swedish study found that children (particularly boys) rated as aggressive at ages 10 and 13 by their
teachers, were disproportionately represented among criminals, especially perpetrators of violent crime, at age 26 (Stattin & Magnusson, 1989). Several other studies have shown
impressive correlations between preschool variables, such as
impulsivity, irritability, and passivity, and personality patterns
and behavior problems in adolescence and young adulthood
(Block, 1993; Caspi, Henry, McGee, Moffitt, & Silva, 1995;
D. L. Newman, Caspi, Moffitt, & Silva, 1997; Tremblay, Pihl,
Vitaro, & Dobkin, 1994).
Mischel (1968, 1973) was once one of the most persistent
critics of psychodynamic ideas, particularly the notion of consistency of personality over time and across situations. More recently, however, he has presented compelling longitudinal evidence to the contrary and has placed his own work in the context
of theorists who view the mind in terms of interacting dynamic
processes, of whom he considers Freud a prime example (Mis-

chel & Shoda, 1995). In a major study (Shoda, Mischel, &

Peake, 1990), he and his colleagues found that 4-year-old children who were able to delay gratification (resisting grabbing a
desired object while looking directly at it) were substantially
different from their peers as adolescents: They were rated by
their parents as more likely to exhibit self-control when frustrated, as more likely to resist temptations, as more intelligent,
and as less distractible than children who were less able to delay
in preschool.
The first half of the proposition, then, that adult personality
characteristics begin to be evident in childhood and that many
show considerable consistency over time, has received substantial empirical support. Data from research in behavioral genetics
have demonstrated, however, that a substantial part of the variance in individual differences in personality, and bY extension
a substantial part of the variance in documented consistency
of personality over time, reflects genotypic variance (Plomin,
DeFries, McClearn, & Rutter, 1997). The repeated finding that
environmental effects, which typically account for at least half
the variance in individual differences in personality, tend to
reflect nonshared rather than shared environment among siblings, has led many to conclude that psychoanalytic ideas about
the importance of childhood have finally been put to rest. In
other words, if similarities between siblings who grow up in
the same household are attributable to their genes and not to
similar experiences at home, then surely childhood experiences
and processes such as identification cannot have rnach of an
impact on personality development.
Research in behavioral genetics is clearly revolutionizing
our understanding of personality and development and rewriting every account of personality development, including
Freud' s. Nevertheless, several caveats are essential to understand the implications of research into the behavioral genetics
of personality, many of which have been repeatedly pointed
out by prominent behavior geneticists but are too easily forgotten. Three are of particular importance from the present
perspective.
First, heritability coefficients are estimates of the influence of
genes on individual differences within a population~ They are not
estimates of the relative importance of genes and e n v i r o ~ t (see
Plomin et al., 1997). Height was highly heritable in 1865 just as
it was in 1965--that is, genes accounted for the lion's share of
differences among individuals in height in the population at each
of those times--but the average soldiex's uniform during the Vietnam War was almost a foot longer than the average uniform of
soldiers in the U.S. Civil War because of changes in nutrition that
shifted the entire population mean upward. For many physical and
psychOlogical traits, genetic factors make environmental factors
relevant, and environmental factors activate genes, so that parsing
the relative roles of nature and nurture is the wrong question. The
only reason lengthy separations in childhood can lead to later
effects on personality, social behavioI; and psychopathology is that
humans are genetically prepared to need other people, and the
absence of appropriate caregivers at particular junctures leads to
alterations in brain physiology and gene expression.
This underscores a second caveat, that heritability estimates
are highly dependent on the sample. If a typical study of
heritability included a substantial minority of inner-city children, heritability coefficients would likely drop considerably

FREUD'S LEGACY
because the "average expectable e n v i r o n m e n t " in middleclass homes and neighborhoods is quite different from that in
inner cities (for a relevant discussion, see Scarr, 1993).9 Decades of research on culture and personality have revealed
substantial cultural differences in the way people express
anger, form intimate relationships, take care of their children,
pursue goals such as achievement, regulate emotion, and so
forth (see LeVine, 1982). Including a random sample of individuals from dozens of cultures would lead to radical revisions
in estimates of the impact of shared environment on personality (see Mandler, 1997).
Third, genetic and environmental events may influence phenotypic behavior in fundamentally different ways that may
substantially bias the way heritability estimates are interpreted. Individuals whose genotype predisposes them to schizophrenia are more likely than other individuals to appear phenotypically schizophrenic or to show subsyndromal signs of
thought disorder or other signs of schizotypy, such as emotional flatness or interpersonal peculiarities. Environmental
events, in contrast, may not lead to such straightforward links
between underlying causal and resulting phenotypic behavior.
The highly variable ways in which humans can respond to
the same event can lead to artificially deflated estimates of the
influence of shared environment. To take a c o m m o n clinical
example, two siblings exposed to an alcoholic parent may
respond very differently, with one becoming impulsive and
alcoholic and the other constricted and workaholic. In this
case, the same environmental event has led to radically different styles of impulse and affect regulation, even though both
styles are clearly adaptations to the same shared environment.
What makes this scenario particularly problematic in research
attempting to draw aggregate conclusions about the impact
of childhood experiences on adult characteristics is what happens when averaging together the characteristics of these two
individuals, both of whose personalities have clearly been
impacted by the same environmental pathogen: The environmental effect will average to zero if relatively equal numbers
of individuals with each of these two opposite patterns are
included in the sample. 1
In fact, the concept of nonshared environment is actually a
misnomer because it includes shared events to which different
children respond differently. Perhaps most important, it includes
gene-environment interactions that lead two children confronted with the same situation to behave differently. Thus, early
separation from a parent may have a substantially greater effect
on a child who is temperamentally higher in negative affect than
one with an easier temperament, even though the environmental
event is identical and hence shared. To put it another way, it
includes the kinds of nonshared effects of shared environments
described above, in which different children construe a situation
differently, cope differently, have different emotional responses,
and so forth, and hence move in divergent directions despite
shared experiences.
I do not introduce these considerations to minimize the impact
of genetic influences on personality development. Research is
just beginning to document the hidden impact of genotypic variance on life events that are often treated as environmental stressors or assumed to be of psychosocial origins (see Jockin,
McGue, & Lykken, 1996; Kendler, 1995). The point is simply

349

that a comprehensive picture of the relative roles of heredity,

shared environment, and nonshared environment in creating personality and maintaining personality continuity from childhood
through adulthood has not yet been painted.

9 The dependence of heritability estimates on the sample can be seen

in a study that found the lack of autonomic responsivity hypothesized
to be genetically related to psychopathy to be characteristic of delinquent
boys from intact homes but not boys from broken homes (Lytton, !990).
Environmental risk factors may be so substantial (or, alternatively, genetic variance so minimal) in disrupted families as to overshadow genetic effects.
l01 know of few empirical studies directly addressing this issue, but
I suspect that the reason such data are not available is that the problem
has not been recognized and that reanalysis of a number of data sets
might shed light on it. From a methodological standpoint, one should
not routinely begin with the assumption that the individuals in samples
are homogeneous with respect to patterns of covariation; where samples
are large enough, this assumption should first be assessed with techniques such as cluster analysis, which allows discrimination among subgroups that may differ in type rather than in degree. For example, Jennifer Harnden and I (Westen & Harnden, 1998) have recently collected a
sample of patients with anorexia and bulimia and found that the two
groups do not differ in central tendency with regard to a number of
variables, including developmental history variables such as child sexual
abuse, number of psychiatric hospitalizations, number of suicide attempts, and other measures of level of functioning.~However; when we
began instead by clustering participants empirically using Q-factor analysis (inverted factor analysis, in which cases rather than items are factored) and then tested for differences, we found huge group differences
on nearly every variable we examined. The reason, as we discovered,
is that anorexia and bulimia are not homogeneous diagnoses. The Qfactor analysis pointed instead to the existence of three groups whose
membership was only partially related to symptoms: (a) a high-functioning, perfectionistic, self-critical group that includes patients with both
anorexia and bulimia; (b) a low-functioning,constricted, avoidant, overcontrolled group that consists almost exclusively of patients with anorexia; and (c) another low-functioning group, consisting almost exclusively of patients with bulimia, that is undercontrolled, impulsive, and
hyperemotional. Averaging high- and low-functioning patients with anorexia or bulimia led to the apparent conclusion that bulimics and anorexics are similar, just as averaging overcontrolled and undercontrolled
children of alcoholics might lead to the conclusion that parental alcoholism has minimal psychological effect as an environmental variable on
dimensions such as impulsivity, when in fact it can produce two different
types of outcomes. Roberts and Chapman (1997) have discovered a
similar phenomenon in examining self-reported well-being in a 30-year
longitudinal study of women. When they simply looked at mean differences over time, well-being seemed relatively stable across three decades. When they applied more sophisticated statistical procedures to
distinguish subgroups, however, they found that 30% of the sample
increased by more than one standard deviation and another decreased
by the same amount--leading to a net effect across the sample of
zero--but that the subgroups could be distinguished based on other
theoretically meaningful variables. Similarly, Harrist, Zaia, Bates,
Dodge, and Pettit (1997) used cluster analysis to distinguish four types
of children usually called socially withdrawn. Whereas one group was
relatively uninterested in relationships and another was withdrawn because they were depressed, the others were interested in having friends
but either anxious and hence avoidant or actively isolated by their peers.
These four groups differed on a number of variables, including ways of
processing social information.

350

WESTEN

The Role of Childhood Experiences

When environmental events are in fact measured rather than
calculated by subtracting out genetic effects, they are much
more likely to show a substantial impact on personality (see
Caspi, 1998). The formative role of childhood experiences has
been most clearly demonstrated in longitudinal studies and other
research linking these experiences with the way adults and adolescents form and experience relationships.
One of the complexities in this research is that one cannot
assume a simple model, such as "Experience x leads to trait y
in childhood, which persists into adulthood," which does not
capture the complexity of human development, for multiple reasons. First, childhood experiences can have delayed effects. Promiscuity is a common sequela of childhood sexual abuse that
typically does not manifest until adolescence (Cavaiola &
Schiff, 1988). Second, trait y may have different behavioral
manifestations at different points in the life span. Aggressiveness
in a child may be indexed by physical aggression toward peers,
but given the low incidence of such behavior in adulthood, how
does one measure thesame trait in adults? Third, as individuals
develop and assume an ever-widening set of roles, the ways in
which they can express trait y may become much more differentiated and varied. An aggressive toddler may become an aggressive adult who displays his aggression verbally or physically,
with peers or with his own children, passively or actively, etc.
Fourth, many traits are better understood in the broader context
of the individual's personality structure or organization. Certain
parenting styles, for example, may lead to anger at age 6 but
depression at age 21 as the person develops a self-critical personality style. Similarly, what may look like delayed effects
of a childhood pathogen such as sexual abuse may be better
understood in terms of the way organized personality patterns
change and manifest behaviorally at different life stages.
Several studies from the longitudinal project conducted by
Block and Block (1980) demonstrate the importance of taking
a more complex view of the ways in which childhood experiences can affect adult personality. These studies have demonstrated impressive correlations among childhood characteristics,
parenting styles, and later attributes, such as drug use and depression. One study found that adolescent drug use and abuse,
as well as personality characteristics such as rigidity or alienation, can be predicted from mothers' interactions with their
preschool children (Shedler & Block, 1990). Another study
demonstrated the complexities that may be involved in predicting adult traits from their childhood antecedents (Block,
Gjerde, & Block, 1991). Male participants who were undercontrolled and impulse-ridden in early childhood were more likely
to be depressed as young adults, whereas female participants
who were overcontrolled as children were more likely to be
depressed 15 years later. These findings are more readily explained in terms of culture, socialization, and personality organization than temperament, unless one could develop a coherent
genetic explanation (sex-linked ~genetic transmission) for the
gender-specific correlation between over- and undercontrol in
childhood and depression in adulthood. Studies by other research groups have documented the influence of parent-child
interactions and parental socialization on subsequent personality
and social behavior (e.g., Baumrind, 1967, 1991; Franz,

McCleUand, & Weinberger, 1991), and research by Patterson

and colleagues has shown that experimentally manipulating parental behavior through skills training can produce decreases in
delinquent behavior over time (Bank, Marlowe, Reid, & Patterson, 1991).
A body of data in developmental psychopathology has documented links between experiences such as abuse, neglect, and
disruptions in caretaking in childhood and later interpersonal
pathology and personality disorders (see Herman, Perry, & van
der Kolk, 1989; Rutter, 1986; Tizard & Hodges, 1978; Westen,
Ludolph, Block, Wixom, & Wiss, 1990). Sexual abuse in childhood leaves individuals vulnerable to a number of untoward
psychological experiences in adulthood such as anxiety, depression, chronic self-destruc.tiveness, and suicidality (e.g., Boudewyn & Liem, 1995; Briere & Runtz, 1993). A longitudinal
study of boys at risk for delinquency found that paternal alcoholism was a strong risk predictor of alcoholism in the son decades
later (McCord, 1988). In support of models of childhood identification or modeling, the study found that a variable assessed
prospectively--the extent to which the child's mother appeared
to esteem the father--doubled the son's chance of being diagnosed with alcoholism 40 years later (from 35% to 70%). Maternal esteem of the father was not similarly predictive of alcoholism in at-risk boys whose fathers were not alcoholic, suggesting a clear interaction of paternal alcoholism and materual
esteem.
Nonclinical samples have similarly begun to bolster the case
for an association between at least some childhood experiences
and later clinical or subclinical signs of psychopathology. In a
stratified random sample of over 5,000 adults, a history of parental separation or loss in childhood was associated with both
lower attachment security and higher ratings of insecure attachment (Mickelson, Kessler, & Shaver, 1997). All nine childhood
interpersonal traumas assessed in the study (e.g., assault, serious
neglect, sexual molestation) were .associated with anxious attachment in adulthood. In a study of adult female twins, a history
of parental loss through separation (but not through death) was
associated with later vulnerability to depression and alcoholism,
even after stringent procedures to remove direct and indirect
genetic effects (Kendler; Neale, Kessler, Heath, & Eaves, 1992;
Kendler et al., 1996; Tennant, 1988). The data are not yet conclusive, but the association between parental loss through separation and later psychiatric problems appears likely to stem in
part from the broader constellation of variables in the family
environment (such as parental conflict, abuse, etc.) that precipitate family dissolution (see Kessler, Davis, & Kendler, 1997).
Although some of this research is subject to potential biases of
retrospective assessment, longitudinal research similarly shows
that individuals who experience more negative events in childhood tend to have lower occupational success and satisfaction
than their peers decades later (McClelland & Franz, 1992).
Traumatic events later in life can similarly produce lasting scars,
although age of trauma tends to be negatively correlated with
severity of later problems (e.g., Bunce, Larsen, & Peterson,
1995).
Of particular interest are results from a recent study showing
that although individuals with a history of childhood parental
loss do not consciously self-report more negative mood than
comparison subjects, they show less intense positive mood and

FREUD'S LEGACY
less positive feelings regarding intimacy in day-to-day life when
completing a mood diary over 28 days (Bunce, Bernat, &
Shevrin, 1997). They also respond differently than comparison
subjects on implicit measures (facial EMG and brain-evoked
potentials) when exposed subliminally to words related to intimacy and excited positive mood; interestingly, these differences
do not appear with supraliminal presentation. These findings, if
replicated, are important, because they suggest that the impact
of some aversive childhood events may be expressed behaviorally and manifest in implicit emotional responses in adulthood
but not be amenable to the typical self-report methods. These
are just the kind of responses addressed in psychoanalytic forms
of psychotherapy. The findings also make sense in light of the
research reported by McClelland and colleagues (1989), previously described, which found substantial correlations (r
.40) bet~veen childhood experiences and implicit motives assessed by TAT decades later. In related research, Bunce and
colleagues (Bunce, Snodgrass, & Larsen, 1996) found, in a
nonclinical sample, that participants reporting a history of traumatic events showed greater baseline and imagery-generated
sympathetic reactivity as assessed by skin conductance but less
reactivity as assessed by facial EMG; this pattern suggests defensive efforts at emotional suppression, because several studies
have shown this same asynchronous pattern to occur when participants attempt to inhibit affective expression or suppress
thoughts. Interestingly, this pattern holds in combat veterans
without posttraumatic stress disorder (PTSD) who are exposed
experimentally to combat sounds but not in combat veterans
with PTSD, whose efforts at affect regulation have proven ineffective and who instead tend to experience an alternation of
emotional constriction and lability (Bunce, Casada, Amdur, &
Liberzon, 1997).
A considerable body of research, also much of which is longitudinal, has established links between parenting styles and the
subsequent tendency of adult offspring to be depressed or dysphoric. In one study (Lefkowitz & Tesiny, 1984), 19-year-old
participants whose parents reported more rejecting attitudes toward them 10 years earlier showed elevated Depression scores
on the Minnesota Multiphasic Personality Inventory (MMPI;
Butcher, Dahlstrom, Graham, Tellegen, & Kaemmer, 1989). In
another, parental coldness rated at age 5 was highly predictive of
Depression scores 36 years later in midlife (Franz, Weinberger,
Kremen, & Jacobs, 1996). The data from this study also supported a number of theoretical predictions about gender differences between specific childhood variables (such as maternal
tolerance of the child's dependence) and adult depression. As
with the work by Block, Gjerde, and Block ( 1991 ), the pattern
of findings cannot be readily explained by a genetic hypothesis.
Multiple regression with childhood variables as predictors accounted for approximately 40% of the variance in adult dysphoria, with the strongest predictors being lack of parental warmth
and a composite index of difficult childhood experiences (such
as loss, divorce, frequent moves, etc.). This latter finding is
particularly impressive, although, as with most longitudinal
studies, parsing out genetic influences on life events and parenting styles is difficult.
Other research shows that poor parental care in childhood
leaves individuals vulnerable to depression in response to stressful events in adulthood (Harris, Brown, & Bifulco, 1986). In

351

research involving retrospective reports of parents by adults, a

pattern of low perceived care and overprotection by parents
is highly correlated with depression and predicts subsequent
vulnerability to depressive episodes (Blatt & Homann, 1992;
Burbach & Borduin, 1986; Gotlib, Whiffen, Wallace, & Mount,
1991; Parker, 1983). Kaplan & Gotlib (1996) demonstrated
hypothesized three-way interactions between quality of perceived parenting (low care or overcontrol), quality of depression
(interpersonally focused or autonomy focused), and stressful
life events. For example, a person who is interpersonally sensitive and perceives her parents as low in caring is more likely to
become depressed by life events that involve loss or disruption
of social support. Again, these data are difficult to explain by
reference to a temperamental tendency toward neuroticism, because very specific diatheses based on the meaning of particular
types of events interacted with very specific stresses to produce
depression.
Studies of attachment have now established the role of expectations shaped by interactions between children and their caregivers, as early as the first 12 months of life, in shaping later
social adjustment (Bretherton, 1990; Lyons-Ruth, Connell, Grunebaum, & Botein, 1990; Ricks, 1985; Sroufe & Fleeson, 1986).
Maternal responsiveness has proven the most robust predictor
of infants' patterns of attachment (see Pederson et al., 1990),
which in turn predict social adjustment and relationship patterns
at least into adolescence (the current general age of the first
participants being followed in longitudinal attachment research
since infancy). Other research has shown that parents' attachment patterns vis-h-vis their own parents are highly predictive
of their unborn infants' later patterns of attachment to them,
documenting the impact of parental attachment style on the
child's developing experience of relationships (Fonagy,
Steele, & Steele, 1991; H. Steele, Steele, & Fonagy, 1996).
The almost uniformly higher predictive power of maternal over
paternal attachment style in these studies (see, e.g., Main et al.,
1985) again militates against a primarily genetic explanation.
None of this is to suggest that childhood experiences inevitably leave indelible marks that later experiences cannot correct
or obviate. Many longitudinal studies have shown only modest
associations between early childhood experiences and temperament on the one hand and adult personality and psychopathology
on the other (see Kagan & Zentner, 1996). Not all children who
are securely or insecurely attached at 12 months will show
similar patterns of attachment at 6 years (Lamb, 1987), and
the study of resilient children certainly makes clear that the
determinism assumed by many psychoanalysts through the
1940s was too extreme (see Masten, Best, & Garmezy, 1990).
Today, psychologists are more likely to recognize the transactional nature of development and the interaction of thought,
feeling, and behavior patterns established in childhood with
subsequent experiences that may reinforce them. Thus, a child
whose primary caretaker finds holding, touching, and tenderness
uncomfortable is likely to discover early that such experiences
are sources of as much anxiety as comfort and is likely to try
to shut off needs for them (Main et al., 1985; Sullivan, 1953).
In so doing, the child learns to behave and process information
in a way that may distance others, who in turn respond with
behavior that confirms the expectation. Interrupting a cycle of

352

WESTEN

this sort would likely require substantially changed life

circumstances.
Psychoanalytic theory since Freud has undoubtedly underestimated the importance of postchildhood events in shaping personality as well. Adult traumas such as rape and torture can
permanently affect a range of cognitive, affective, and behavioral
processes, as can events that are less obviously catastrophic but
nevertheless tremendously powerful (such as divorce or enduring occupational derailments) depending on their meaning to
the person. More generally, with the exception of Erikson
(1963) and some of the neo-Freudians, psychoanalytic theorists
have tended to underestimate the importance of schemas about
situations that arise only after early childhood, such as repeated
frustrations and feelings of inadequacy in learning-disabled
school-age children or repeated romantic rejections in the teenage years, which can have a wide impact on the self-concept,
expectancies regarding the self and others, and so forth.
.Nevertheless, what psychoanalytic theory has stressed from
the start, and what seems both theoretically sensible and empirically supported, is that most important human experiences-namely, those involving feelings of intimacy and adequacy-have precursors in childhood that leave their precipitates not
only in working models of the self, others, and the world but
also in motives, emotions, and ways of regulating emotions in
later situations that activate these prototypes. To the extent that
these processes reflect repeated experiences or experiences that
are painful and conflictual and hence have engendered automatized affect-regulatory procedures to cope with them, they are
more likely to be resistant to change and to require more extraordinary experiences to alter them.

The Role of Mental Representations of the Self, Others,

and Relationships
That mental representations of the self, others, and relationships play a crucial role in the way people establish intimate
relationships and in the forms of psychopathology they suffer
from was once, like the emphasis on childhood experience,
a peculiarly psychoanalytic supposition. Freud (1914/1957b,
1917 / 1957a) first began developing hypotheses along these lines
about 80 years ago, as he examined the representations of people
suffering from mania and melancholic depression. From these
beginnings emerged object relations theories in the 1930s and
1940s, which focused on representations of the self, others, and
relationships (called, unfelicitously, object representations for
reasons not worth describing here) and on the interpersonal
behavior believed to reflect the interaction between these mental
representations and people's wishes, fears, and emotions (see
Greenberg & Mitchell, 1983; Westen, 1991). Today, most psychodynamic theorists and therapists rely heavily on object relations theories (or, to use the more human-sounding term, relational theories; Aron, 1996; Mitchell, 1988). 11
Object relations theories have led to the most productive area
of psychodynamically inspired research, in part because of their
accessibility to measurement from self-reports, projective tests,
and interview methods (see Blatt & Lerner, 1991; Levy, Blatt, &
Shaver, 1998; Stricker & Healey, 1990; Westen, 1991). For example, one study assessed children's representations of their
relationships with their mothers by asking 4- and 5-year-olds to

complete story stems of emotionally charged situations (Oppenheim, Emde, & Warren, 1997). Themes of positive interaction
and nonabusive discipline were negatively correlated with maternal reports of the child's internalizing behavior (depression
and distress), whereas presence of negative themes (such as
physical or verbal abuse) strongly predicted maternal reports of
externalizing behavior (aggression and misbehavior). The same
relations held between thematic content and mothers' reports of
their own distress, which suggests that children's representations, assessed projectively, bore some relation to actual parental
behavior" or feeling states. A stfidy just completed in my laboratory (Calabrese, 1998) found that measures of dimensions of
object relations (such as the capacity to invest emotionally in
relationships) coded from the narratives of adults predicted both
their own relationship status and that of their parents (i.e.,
whether their parents' marriage was still intact), whereas selfreport measures of attachment, self-esteem, and fear of intimacy
predicted only other self-report measures.
Again, one of the major sources of data pertaining to this
fourth basic psychodynamic tenet is attachment research, which
is not surprising, because Bowlby (1969) was himself an object
relations theorist. Mary Main (1990) and her colleagues provided a key tool for assessing representational processes by
developing the Adult Attachment Interview (AAI; Main et al.,
1985). A large body of literature now documents that adult
attachment styles as assessed by the AAI and other measures
(Brennan, Clark, & Shaver, 1998) are highly predictive of variables such as the way people respond interpersonally to threatening events (either by seeking comfort or by withdrawing;Simpson, Rholes, & Nelligan, 1992), the extent to which they acknowledge distress (Dozier & Kobak, 1992), and their styles of
coping (Mikulincer & Florian, in press; Mikulincer, Florian, &
Weller, 1993). Research on the intergenerational transmission
of attachment documents that the internal working models or
representations of relationships assessed in adults are highly
predictive of similar models assessed from the behavior of their
infants and young children (Main et al., 1985; van IJzendoorn,
1995).
By the 1970s, similar constructs were being developed by
researchers and clinical theorists of very different theoretical
orientations. Mischel (1973, 1979) and later Cantor and Kihlstrom (!987) assimilated notions such as Kelly's (1955) personal constructs (views of people idiosyncratic to individuals
that guide the way they think about themselves and others) to
cognitive-social approaches to personality, eventually integrating them with information-proceSsing theory and research. Higgins (1990) has shown that people interpret social information
on the basis of chronically accessible as well as recently activated constructs and that their emotions reflect discrepancies
H In the survey of psychodynamic clinicians described above, participants were asked to rank the extent to which they subscribe to the
following five theoretical perspectives: classical psychoanalytic, object
relations, self psychology, neo-Freudian, and "other." Of the 86 respondents, 48.1% ranked object relations theory first and 38.0% ranked it
second, for a total of 86.1%. In contrast, the perspective that received
the next highest rankings was classical psychoanalysis, which 22.8%
endorsed as their preferred perspective and 21.5% ranked second (for a
total of 44.3%).

FREUD'S LEGACY
between various types of self-representations or self-schemas.
Strauman and his colleagues (Strauman, 1992; Strauman,
Lemieux, & Coe, 1993) have demonstrated that such discrepancies not only distinguish different kinds of patients (anxious vs.
depressed) but even predict immune response when confronted
with different kinds of self-relevant information.
The psychoanalytic concept of transference is predicated on
the view that people's wishes, fears, feelings, and perceptions
of significant others will resemble prototypes from the past (a
term Freud himself used) to the extent that aspects of the person
or situation activate those mental prototypes (see Singer, 1985;
Wachtel, 1981; Westen, 1988). Andersen and Cole (1990) documented the cognitive side of transference experimentally by asking participants to describe significant others and then embedding descriptions from their responses in descriptions of fictional characters. When subsequently asked to describe the
characters, participants mistakenly attributed traits to the characters that were part of their schemas of the significant other but
were not originally included in the character's description. More
recently, Andersen and her colleagues have extended this ingenious experimental paradigm to demonstrate the transference of
affect from significant others to descriptions of unknown others
allegedly seated next door (Andersen & Baum, 1994). In their
latest work, Andersen, Reznik, and Manzella (1996) showed
that these affective evaluations lead to differential tendencies
to avoid or approach the person next door, demonstrating that
transferential processes can influence cognition, affect, and motivation--a central claim of psychoanalytic theory since Freud.
In a crucial demonstration, they have now also shown that the
effects they have produced in this program of research can
be obtained unconsciously through subliminal presentation of
significant-other descriptors (Glassman & Andersen, 1997).
Other researchers, notably Luborsky and colleagues, have examined transference processes empirically from narrative data, particularly psychotherapy transcripts (Luborsky & CritsChristoph, 1990).
Researchers are currently examining the impact of disturbances in social cognition and object relations on the understanding and treatment of various forms of psychopathology
(Blatt, 1994; Higgins, 1990; Mineka & Sutton, 1992; Strauman,
1992; Westen, 1991). Research on depression provides a good
example in that distinctions between representational triggers
were first explored by psychodynamic researchers (Blatt & Zuroff, 1992; Mongrain, Vettese, Shuster, & Kendal, 1998) and are
now being actively explored by cognitively oriented researchers.
Investigations of clinically depressed patients have found that
roughly half of these patients have concomitant personality disorders (Shea, Glass, Pilkonis, Watkins, & Docherty, 1987; Zimmerman, Pfohl, Coryell, Stangl, & Corenthal, 1988). Given that
the diagnostic criteria for all the DSM-1V (Diagnostic and
Statistical Manual of Mental Disorders [4th ed.]; American
Psychiatric Association, 1994) personality disorders are heavily
interpersonal, this suggests that many depressed patients have
substantial object-relational pathology. An important question
is whether depressed patients with pathological object relations
differ in some important ways from their depressed peers.
The data suggest that they do. For example, two initial studies
compared borderline patients, with and without concomitant
depression, to other depressed patients using Blatt, D'Affliti,

353

and Quinlan's (1976) Depressive Experiences Questionnaire

(Westen et al., 1992; Wixom, Ludolph, & Westen, 1993). Both
studies found that the quality or phenomenology of depression
differs substantially depending on whether or not participants
have borderline pathology. Borderline participants manifested
a type of depression characterized by feelings of emptiness,
loneliness, desperation vis-h-vis attachment figures, and labile
and diffuse negative affeetivity. Subsequent studies have replicated these findings (e.g., Rogers, Widiger, & Krupp, 1995).
One study examined the relationship between intensity of depression (as assessed by both interview and self-report) and the
extent to which participants endorsed items on a borderline
depression factor that tapped this particular quality of depression (Westen et al., 1992). For borderline patients, the correlation was .70; within the limits of reliability of measurement,
this suggests that depression in borderline patients is essentially
the same thing as loneliness, emptiness, etc. For nonborderline
major depressives, however, the correlation was -.50: The more
they endorsed borderline depression items, the less depressed
they were. Research that does not distinguish between these two
different types of depression may thus be mixing groups that
are psychologically highly dissimilar. Indeed, Peselow, Robins,
Sanfilipo, Block, and Fieve (1992) documented differential response to tricyclic antidepressant medication depending on
whether depressed patients' depression centered on interpersonal concerns such as loss or on concerns about failure and
threats to autonomy. 12
One might wonder whether object-relational variables may
be mediating or moderating a number of relations between depression and other variables, such as the association between
maternal depression and childhood psychopathology. Studies of
risk factors for psychopathology, such as those examining the
relation between maternal depression and psychopathology (see,
e.g., Rutter, 1986), have seldom explored precise mechanisms
through which parental psychopathology leads to pathology in
the child. Given that at least half of the population of patients

12Although no one has yet explored the relation between objectrelational variables and the dimensions of the five-factor model of personality (McCrae & Costa, 1990), notably neuroticism, the data from
these studies cast doubt on the rival hypothesis that these findings are
all reducible to level of neuroticism. In the studies reported above as
well as in several others (for a review, see Westen, 1991), substantial
differences emerged on dozens of variables between two groups high on
neuroticism: patients with major depressive disorder and with borderline
personality disorder (BPD). The data showed that a lonely, rejectionsensitive depression (along with expectations of interpersonal malevolence, peculiarities in making attributions about interpersonal events,
need for self-soothing through polysubstance abuse, etc.) distinguished
patients with BPD from depressed patients without borderline pathology.
Note that these differences are probably not readily understood by use
of the five-factor model (FFM) even at the level of facets (subfactors)
because the FFM includes no process variables, such as fears of abandonment or ways of coping, that tap many of the differences between
major depressives with and without BPD. In recent research from our
laboratory, we have found that factor analysis of sophisticated personality data from expert observers does not yield a simple Neuroticism factor
in a large personality-disordered sample (N = 530), because there are
simply too many ways to be neurotic (Westen & Shedler, 1998, in
press).

354

WESTEN

with major depression is comorbid for personality disorders, the

possibility cannot be ruled out that part of the link between
maternal depression and. psychopathology in offspring is really
a result o f long-standing maternal disturbances of object relat i o n s - t h a t is, of personality, not of episodic mood state. The
differences in parenting may well be profound between a caregiver without a personality disorder who becomes periodically
depressed and the parenting of a caregiver whose depression
occurs in the context of grossly unstable relationships and a
tendency to see the child in contradictory or malevolent ways.
Similarly, my colleagues and I are currently examining
whether the well-documented association between social support and health (e.g., House, Landis, & Umberson, 1988; Martikainen & Valkonen, 1996) may reflect in part a link between
object relations and health. Social support is generally operationalized as the number of people participants have in their
social networks, the number of people they feel close to and
confide in, the extent to which they perceive others as helpful,
supportive, empathic, etc., or some combination of these (see
Pierce, Lakey, Sarason, & Sarason, 1997). Less empirical attention, however, has been paid to the variables that explain why
a person has or does not have close relationships or why the
person experiences others as supportive or nonsupportive. From
a psychodynamic perspective, object-relational processes, such
as the nature of a person's mental representations of the self,
others, and relationships, should play a substantial role in accounting for the capacity to form and maintain supportive social
relationships. A preliminary study from the author's research
group found self-reported object relations to be a better predictor of self-reported health and visits to the doctor over the
p r i o r 12-month period than were measures of social support
(Costanzo, 1996).
A recent study by Russek and Schwartz (1997) may shed
some light on the links between representations and health as
well as on the role of representations formed in childhood on
subsequent health status. Using a longitudinal design, the investigators found that ratings of perceived parental caring made by
participants in college were highly predictive of their health
status in midlife. Participants who perceived their parents as
less caring in college were significantly more likely to suffer at
midlife with illnesses such as coronary artery disease, hypertension, duodenal ulcer, and alcoholism, even holding constant family history of illness and adult risk factors such as smoking.

Developmental Trajectories
A final central proposition of contemporary psychodynamic
theory is that personality development entails not only learning
to manage sexual and aggressive impulses but moving from an
immature, dependent state to a mature, interdependent one.
Freud focused on the development of his two drives, sex and
aggression. Thus, he emphasized both psychosexual development and the ways children gradually learn to handle their aggressive impulses. Since Freud's time, psychoanalytic theorists
have shifted from a primary focus on the development of sexuality to a focus on the evolving capacity for forming and maintalning intimate relationships.

Freud's Psychosexual Theory

From the start, Freud focused most on psychosexual development, arguing that personality development reflects the child's
quest for gratification and the process of learning to deal with
impediments to it, many of them socially imposed. Freud and
his followers no doubt overemphasized the role of sexuality
in personality and offered some wild speculations that have
contributed to the low regard in which many in academic psychology hold psychoanalysis. Nevertheless, several of Freud's
psychosexual hypotheses actually make considerable developmental sense when understood from the point of view of the
child. Consider, for example, the experience of the child undergoing toilet training. What does a toddler experience when her
parents are suddenly demanding something from her, namely
that she control her own body and excrete some of its contents
when and where they see fit? How could she not wrestle with
the issues of dirtiness and cleanliness and of compliance and
defiance that Freud attributed to children at this age? "Why,"
she might ask, if only she had the language, "don't you poop
when I tell you to poop?" She is learning an important lesson
during this period--on the potty and off--about who is the
boss. As previously cited research on transferential processes
suggests, the way she copes with it will influence (though certainly not completely determine) the way she subsequently processes information about and copes with similar circumstances
in the future.
Although the most compelling data regarding Freud's psychosexual theory do not come from the laboratory, decades
of research (reviewed in Fisher & Greenberg, 1985, 1996;
Kline, 1981) have attempted to test specific hypotheses derived from it. The data are mixed on many hypotheses; some,
however, have received Substantial support. For example, several researchers have produced evidence for the construct
validity of an oral character style, characterized by features
that could only be predicted by Freudian theory. Such features
include the combination of an orientation toward seeking
nurturance and contact with others (measured by self-report,
by projective responses, and by behavior, such as help seeking, volunteering, and touching, in experimental settings) and
indexes of orality in the literal sense, such as a tendency to
produce oral responses on the Rorschach (see Bornstein &
Masling, 1985; Fisher & Greenberg, 1996; Juni, Masling, &
Brannon, 1979; Masling, 1986; Masling, Price, Goldband, &
Katkin, 1981). Factor analytic questionnaire studies have
similarly supported the concept of an anal character style,
characterized by traits such as orderliness, stinginess, stubbornness, rigidity, concerns around control, and issues surroun.ding spending versus retaining money (Fisher &
Greenberg, 1996; Kline, 1981; Torgersen, 1980). A study using a multimethod assessment of anality (including such measures as questionnaires, a word-stem completion task, and
frequency of references to dirt in spontaneous thought) found
correlations between enjoyment of anal humor (as opposed to
oral or other sexual humor) and measures of anality (O'Neill,
Greenberg, & Fisher, 1992). This finding is particularly important because a link between a personality style and enjoyment of anal humor is neither intuitive nor predicted by other

FREUD'S LEGACY
theories. Evidence linking these character styles to particular
developmental phases, however, is much more tenuous.
Studies of the Oedipus complex have also produced some
striking findings (again, for a review, see Fisher & Greenberg,
1985, 1996). For example, Friedman ( 1952) scored the extent
to which children produced positive or negative endings when
completing two stories: (a) one in which they were engaging
in a pleasurable activity alone with their same-sex parent and
were then met by the other parent and (b) one in which they
were engaging in a pleasurable activity with their oppositesex parent and were then interrupted. As predicted, the children produced more negative endings to oedipal stories in
which they were initially alone with their opposite-sex parent.
Interestingly, similar findings emerged with a second projective measure but not with measures assessing conscious feeli n g s - c o n s i s t e n t with the previously described data on implicit versus explicit cognitive, affective, and motivational
processes.
More recently, Watson and Getz (1990) asked parents of
children ages 3 - 6 to record over a 7-day period the number of
affectionate and aggressive acts they displayed toward their
same- and opposite-sex parents. Supportive of Freud's theory,
affection toward the opposite-sex parent and aggression toward
the same-sex parent were significantly more common than the
reverse. This oedipal pattern was strongest at age 4 and began
to decline by age 5. These findings are particularly important
because they are neither intuitively obvious nor predictable from
other theories; in fact, on the basis of socialization, social learning, and reinforcement history, one might expect more aggression toward fathers from children of both sexes. Other research
has found that subliminal presentation of the stimulus Beating
Dad is OK produced increases in dart-throwing performance
among male participants, whereas Beating Dad is Wrong produced decreases; similar stimuli (such as Beating Morn is Wrong
or Beating Him is Wrong) did not have the same effect (see
Palumbo & Gillman, 1984).
More broadly, data from the Sears, Maccoby, and Levin
(1957) longitudinal data set have documented the influence
of socialization for sex and aggression on later personality (J.
Weinberger & McClelland, 1990), as have numerous crosscultural studies (see Whiting & Child, 1953). Longitudinal
studies conducted by both Block (1971) and Vaillant (1977)
have demonstrated correlations between parenting styles
in early life and later sexual comfort or inhibition in adulthood.
Although these examples are impressive, most psychosexual hypotheses are obviously difficult to test in the laboratory,
and many are, no doubt, too sweeping or simply wrong. One
should not, however, ignore the myriad instances in which
Freudian theory can provide a compelling explanation, especially where other theories can offer no rival explanations.
For example, I recently received an anxious call from a man
whose 3-year-old son had suddenly, despite their previously
close and loving relationship, become very rejecting. The last
straw was when the little Oedipus told him, " M o m m y says
she doesn't want you to be my daddy anymore. So you have
to go away." When I told this to a colleague, he said he could
top that story: His 4-year-old son had recently said, "Mommy,
daddy has to leave. I don't like him anymore." When the

355

child's mother asked why, the boy baldly acknowledged, " H e

has a bigger penis than I do." I would be very interested in
an interpretation of these data from the perspective of the
five-factor model. Or consider an occasion I witnessed myself, when two young boys in the dressing room of a clothing
store were distraught because every time they pulled the curtain closed in one direction, it opened in the other. "You have
to close it," one of them admonished, " s o no one can come
in and steal your 'ding.' " Presumably the boy had not received instruction from his parents to protect his ding in J.C.
Penney's.~3

An Expanded Concept of Development

Erikson (1963) broadened Freud's model of development by
drawing out an implication that was sometimes explicit and
sometimes implicit: Freud's theory of psychosexual development was to be understood not only literally as a progression
of sensual pleasures arising from different bodily zones but also
more metaphorically as the evolution of the child's experience
of the self and the social world. Thus, the oral stage is not only
a time in which children will stick anything in their mouths but
a period of profound dependence, wherein they learn what it
means to rely on others. The anal stage is not only a period in
which children learn about a particular orifice but an era in
which they begin to develop attitudes toward control, cleanliness, and the demands of authority. Investigations of Eriksonian
concepts of identity (Marcia, 1994), intimacy (Orlofsky, 1993),
and generativity (Bradley, 1997; McAdams, 1992) have been
some of the most methodologically sound studies inspired by
psychoanalytic theories of development.
Contemporary psychodynamic theorists do not discount the
importance of psychosexual development as Freud conceived it
or of the impact of socialization for sex and aggression on
personality. Like Erikson, however, they tend to focus more on
the development of the capacity for intimacy and relatedness to
others. Nearly a half century ago, Fairbairn (1952) took issue
with Freud's notion that people seek others for their orifices,
proposing instead that people need people. He thus posited that
development entails a movement from an immature, dependent
state in infancy to a mature, interdependent state in adulthood.
His conception of development foreshadowed many contempo~3Personally, I did not find any of Freud's psychosexual notions compelling until I began doing clinical work and found myself confronting
phenomena that simply could not be easily explained by alternative
theories. Some patients, who were highly competitive with members of
the same sex, reported the combination of a highly flirtatious relationship
with their opposite-sex parent and the devalued view of their same-sex
parent that Freud described. Others (a) repeatedly chose inappropriate
partners who resembled their parents in some critical way, (b) were
only attracted to much older lovers, (c) could readily pinpoint key sexual
fantasies in childhood, or (d) recalled distinctly Freudian childhood
thoughts and feelings about their genitals that persisted into adulthood
and continued to influence their sexuality (such as feeling deeply
ashamed about the relative size of their penis or feeling that their vagina
was damaged, deformed, or disgusting). These were not simply instances
of schema-driven assimilation, because I was actually quite resistant to
several of these notions and initially found these observations more
discrepant than confirming.

356

WESTEN

rary feminist views of relational development (e.g., Miller,

1994).
Object relations theorists have offered a variety of developmental stage theories. Although they have tended to be ignorant
of debates in psychology about the stagelike or continuous nature of development and of empirical literatures that bear on
their theories (see Westen, 1989, 1990c), in general these theorists have argued that with development children become increasingly capable (a) of seeing others in more complex, less
egocentric ways, (b) of investing emotionally in others for who
they are and not primarily for the gratifications they can provide,
and (c) of putting their moral values at times above their immediate needs and wishes, Once again, research from a number of
domains, .such as the development of person perception (Livesl e y & Bromley, 1973), perspective taking (Selman, 1980), conceptions of friendship (Damon, 1977), the self-concept (Damon & Hart, 1988; Harter, 1998), moral reasoning (Kohlberg &
Kramer, 1969; Rest, 1983; Turiel, 1998), and ego development
(Loevinger, 1976), supports this view. So too does research
specifically designed to examine development from 1st through
12th grades on various dimensions of object relations, as assessed from interview and TAT responses (Westen et al., 1991).
This research has documented a progressive ability to represent
the self and others in complex and integrated ways and a capacity for caring about others' needs (rather than just using others
for pleasure or self-soothing) that grows throughout childhood
and adolescence (and probably beyond). As we have seen, other
research has shown what happens to later patterns of intimacy
when childhood experiences such as sexual abuse, nonnurmrant
parenting, or extended parental separations disrupt normal developmental processes.

Implications o f a Revised Understanding o f

P s y c h o d y n a m i c s for Psychological Science
The five propositions examined here were all originally associated exclusively with psychoanalysis and were hotly contested.
The empirical data described above challenge the commonly
held view, voiced clearly by Crews (1996) in the quotation that
began this article, that psychodynamic theory has no empirical
basis and belongs, along with other Mesozoic curiosities, in a
museum of natural history. Much of the research cited here was
not conducted by researchers with any interest in psychodynamics, but taken together, these studies from diverse literatures
suggest the need for substantial modification of current views
of the approach initiated by Sigmund Freud.
So are we all Freudians now? Or could one equally contend
that psychology has independently come around to some of the
basic propositions of contemporary psychodynamic theory and
has done so in an empirically sound way so that contemporary
psychodynamic thought remains superfluous? In this final section, I hope to show the fallacy of the latter sentiment, arguing
that as psychologists we can no longer afford to mistake Oedipus
Rex for q'~'rannosaurus Rex. 14 Constraints of space (and intellect) prevent me from a thorough description of the ways in
which a reconsideration of psychodynamic theory could enrich
multiple areas of psychological research and practice. I briefly
explore three examples: implications for (a) social psychology,

(b) cognitive science, and (c) our understanding of what constitutes data in empirical psychology.

Psychodynamics and Social Psychology

Many areas of social psychology might benefit substantially
from an integration of psychodynamic concepts into mainstream
psychology, such as research on social cognition and attitudes.
In describing attitudes or schemas in social information processing, researchers might do well to specify several of their
qualities emphasized by psychodynamic clinicians in clinical
work, such as the multiple affective associations linked to various aspects of them, the strength of these affective associations,
and the denseness of connections among aspects of the representation with differing affective valence.

Incongruence of Implicit and Explicit Processes

For example, one of the most striking anomalies in the literature on couples and marital interactions is that the mean level
of marital satisfaction reported by men and women on a selfreport measure such as the Kansas Marital Satisfaction Scale
(Schumm, Paff-Bergen, Hatch, & Obiorah, 1986) is roughly 18
points out of a possible 21. Yet the divorce rate in the United
States is over 50%. So what does this mean about the on-line
construction of conscious, explicit beliefs about one's marriage?
Presumably over the course of millions of interactions, spouses
are going to associate aspects of each other with the entire gamut
of feelings; however, when a researcher asks spouses about their
partner, about their marriage, or about marriage in general,
somehow a selection process is activated that allows most people
to ignore most or all of the negative instances or associative
links--that is, unless the marriage is ending or over, in which
case the explicit attitude or schema shifts in affective valence.
The study of social cognition has relied heavily on hypothetical people as targets, so that such basic processes as affective
representational shifts, and the mechanisms that underlie them,
have not been studied. Yet in reality, the norm in human life is
that people have conflicting feelings toward most people about
whom they process information, and we know little about the
way in which conflicting social information, and particularly
information associated with contradictory affect states, is processed. Recent research on attitudinal ambivalence is an exciting
development in this regard (e.g., Priester & Petty, 1996). Many

141 do not explore here the clinical implications of taking empirically

validated psychodynamic propositions seriously--other than to suggest
that advocates of alternative treatment strategies need to reconsider some
of their fundamental assumptions in light of these new data, particularly
on unconscious processes (Westen & Arkowitz-Westen, in press). If
patients can act on the basis of unconscious networks of association,
motives, and defenses, then any approach to treatment that fails to address such ubiquitous processes is likely to need substantial supplementation. Indeed, if behavior can reflect unconscious structures of meaning,
learning to interpret these structures must be a central aspect of Clinical
training. Psychoanalysts will equally need to reconsider lingering archaic
views of "the unconscious" as a primitive, instinctual cauldron of untamed motives, in light of research on the manifold varieties of unconscious processes, many of which are quite adaptive (see Bargh & Barndollar, 1996; Epstein, 1994; Westen, in press-b).

FREUD'S LEGACY
studies, however, have used fictional targets about whom participants know little and have no emotional investment or have
assessed attitudinal ambivalence consciously by self-report. Future studies shouM consider the potential role of implicit affective responses associated with an attitudinal object. Furthermore, clinical experience and research suggest that individuals
differ tremendously in the extent to which they can integrate
affectively mixed qualities into their explicit schemas of the self
and others (Baker, Silk, Westen, Nigg, & Lohr, 1992). Integrating social-cognitive and psychodynamic explanations might offer a rich way of thinking about the conditions and individualdifference variables that foster the capacity to construct multivalent explicit schemas of the self and others.
The word consciousness almost never appears in discussions
of schemas, yet diametrically opposed explicit and implicit representations of the self or others can be active at any given time,
leading to behavior that betrays conflicting feelings (see Westen,
1991). We should never speak of representations, expectancies,
or attributions without specifying their level of consciousness.
Gilbert ( 1991 ), for example, has shown that the automatic attributions people make under heavy cognitive load are quite different from the consciously corrected attributions they make when
given the chance. Recent research predicting drug and alcohol
use from explicit outcome expectancies and from implicit associations between drugs and their affective consequences has
found implicit expectancies more predictive (Stacy, 1997). Research on defensive pessimism (Cantor & Harlow, 1994; Spencer & Norem, 1996) shows that conscious, self-reported expectancies (e.g., " I ' m going to fail" ) may be quite discrepant from
the unconscious expectancies that can be inferred from the fact
that such individuals persist and succeed. The implications of
potentially conflicting conscious and unconscious attitudes
would be a ripe area for research on cognitive dissonance as
well. For example, does dissonance reduction banish discordant
information or simply render it unconscious? If the latter is
the case, can feelings inconsistent with self-reported attitudes
nevertheless find implicit behavioral expression? And does inhibition of dissonant cognitions lead to their chronic activation,
as in Wegner's (1992) research on thought suppression?
Indeed, we would do well to avoid words like activation
and accessibility without specifying whether we are referring to
conscious or unconscious activation or accessibility. The most
well-intentioned of conscious liberals may decry racial stereotypes and deny his acceptance of them despite their repeated
implicit expression when he is choosing a seat on the subway.
As suggested by some dynamically informed approaches to stereotyping (e.g., Katz, 1981), he may even hold or construct
"on-line" conscious beliefs that are particularly liberal precisely because he is defending against a negative stereotype that
is part of his unconscious associational networks but conflicts
with his conscious values (e.g., consciously believing that, even
if O. J. Simpson was guilty, his acquittal is justified because of
the history of oppression of other Black men, while unconsciously sharing a cultural representation of the dangerous Black
man). In this case, which schema is active? The schema of the
oppressed Black man or the schema of the violent, dangerous
Black man whom one avoids sitting next to on the subway?
Which schema is chronically accessible?
Or consider the recent extraordinary series of experiments

357

by C. Steele (1997; C. Steele & Aronson, 1995) in which a

manipulation as simple as asking African American students to
fill in their race before taking an exam described as difficult
primed negative stereotypes about Blacks and achievement (and
hence about themselves) and led to substantial decrements in
performance. In cases such as this, the cognitive and affective
aspects of a schema that chronically reach consciousness may
be very different from those that do not. An African American
student may express unambivalent racial pride and consciously
disavow any of the negative beliefs about her race that a r e
pervasive in her culture. Yet the same student may show the
effects of these beliefs in her test-taking behavior. What, then,
are this student's " r e a l " attitude and " r e a l " self-schema? is
Three Axes o f Activation

These arguments converge with neurological data to suggest

that we should distinguish three orthogonal axes involved in
"activation" that influence thought and memory. The first is
the distinction between conscious and unconscious activation,
which maps loosely onto anterior (prefrontal) versus posterior
cortical activity. When people perceive or represent information
in their minds without activation of working memory, posterior
regions of the brain in the occipital, temporal, and parietal cortex
are activated, depending on the nature of the representation.
When they bring that information into consciousness or hold it
in mind briefly in the absence of the stimulus, areas of the
dorsolateral prefrontal cortex, among others, are activated (see,
e.g., Frith & Dolan, 1996). As argued below, representations
selected for consciousness are not necessarily those with the
highest level of unconscious cognitive activation.
A second axis of activation is excitation versus inhibition.
Most contemporary models of associative memory assume that
when activation spreads from one representation to another, it
can either facilitate or inhibit it. Thus, when one first tries to
classify a cat as Siamese or Burmese, both representations are
likely to be active below the threshold of consciousness. As
soon as one of these representations "wins out," the other is
inhibited. Multiple areas of cognitive research support this distinction, such as experimenlts on directed forgetting (Bjork,
Bjork, & Anderson, 1998). Instructing people to forget the first
part of a word list reduces recall and proactive interference with
the next part of the list, demonstrating that this procedure does
in fact inhibit retrieval. However, both implicit memory and
recognition memory for the to-be-forgotten words are unimpaired, which suggests that the representations remain active. In
many respects, this is a nonaffective analog of a psychodynamic

15C. Steele (1997) argued that these results reflect knowledge of

stereotypes but not internalization of them and cited consistent findings
over decades that show that Black and White students do not differ in
self-reported self-esteem. However, in his own study he measured what
is essentially implicit self-esteem by telling participants that the test was
diagnostic of their ability (and hence stereotype relevant ) and then giving
them a word-stem completion task with stems like lo--- and du--. Black
students were more likely to complete the word stems with words reflecting low implicit self-esteem (such as loser and dumb). They were
also more likely to complete word stems lending themselves to negative
racial stereotypes with those words (such as lazy from la--).

358

WESTEN

process of motivated inhibition of memory, in which efforts at

conscious inhibition do not similarly eliminate unconscious or
implicit activation; the major difference lies in the motivation
for inhibition.
A third axis of activation is cognitive versus affect-regulatory.
Affect can influence excitation in at least two ways. First, with
some important qualifications (see Reisberg & Heuer, 1995),
humans and other mammals tend to have better memory for
emotionally significant events. From an evolutionary standpoint,
memory for emotionally significant events should be stronger
because these events are more likely to carry adaptive significance. In this sense, affect is simply one more source of excitatory cognitive activation; that is, it increases excitatory activation of implicit or explicit representations.
Affect, however, serves a second and distinct activational role:
It activates approach or avoidance toward stimuli as well as
inhibition or facilitation of conscious activation of thoughts,
feelings, or impulses, independently of its impact on level of
activation in the more strictly cognitive sense. The importance
of the distinction becomes clear when one considers phenomena
that involve negative affect such as punishment and avoidance
learning. From a behaviorist perspective, repeated trials in which
an animal receives an electric shock each time it presses a bar
previously associated with reward will weaken the response
tendency, inhibiting a previously reinforced response. From a
cognitive perspective, however, the mechanism for this inhibition
is that repeated trials strengthen the cognitive connection between the bar, shock, and fear. Aversive stimuli such as electric
shocks are motivationally inhibitory, leading to inhibition of
behavior, but cognitively excitatory, reinforcing a neural connection and heightening activation of learned associations (see Westen, 1985, chapter 2). To put it another way, feelings (such as
pain or fear) not only provide information but trigger affectregulation strategies, which may produce behavioral and cognitive effects directly opposite to the effect of the same emotion on
cognitive activation. (Relevant research on avoidance learning in
rats has shown that the effects of amphetamines on memory
enhancement and reinforcement are neurally independent; see
White, 1995.)
For example, people are motivated to avoid memories of painful events, but these are nonetheless more vividly remembered.
In this case, negative affect produces countervailing cognitive
effects (excitatory activation) and affect-regulatory effects (activation of inhibitory mechanisms). This may help explain the
oscillation between memory intrusion and memory failure for
traumatic memories in individuals with PTSD (see Horowitz,
1987). Freud used the concept of repetition compulsion (a motivational explanation) to try to explain intrusive memories, and
many cognitive theorists have proposed cognitive explanations.
An alternative account combines motivational and cognitive explanations: Unresolved traumatic memories remain cognitively
activated precisely because they have been warded off. Their
cognitive activation triggers feelings that require regulation, and
one way people try to regulate them is through avoidance. As
Wegner's (1992) research suggests, however, this strategy actually maintains a heightened level of activation. Keeping representations of traumatic events chronically active necessitates
chronic activation of affect-regulation strategies, creating a cy-

cle that cannot easily be broken without conscious exposure to

the feared representations.
What this analysis suggests, then, is that conscious activation
is not simply a function of level of cognitive activation, in which
crossing a threshold brings representations into consciousness.
Rather, conscious activation is probably a joint function of (a)
cognitive activation (priming), (b) affective attention to emotionally significant events, (c) affective inhibition of unpleasant
information (which can run counter to attention to emotionally
significant events), and (d) affective excitation of pleasurable
information. It is probably no accident that regions of the prefrontal cortex contiguous to those involved in directing action
and attention receive projections from limbic structures that
guide approach and avoidance to both environmental events
(through motor behavior) and internal events (through attention
and consciousness). Individuals differ substantially in the extent
to which they are comfortable with or seek cognitive ambiguity,
unpleasant feelings, and positive affect, which means that people
will differ in the computations that lead to conscious activation.

Connectionism and the Psychodynamics of Parallel

Constraint Satisfaction
These considerations suggest what is probably the most important area of potential integration of empirically validated
psychodynamic principles into mainstream psychology, the integration with cognitive science, which is the closest thing to a
paradigm in psychology in the Kuhnian sense (see Robins &
Craik, 1994). I n this section, I highlight the importance of psychodynamics for understanding affective and motivational processes that could be readily assimilated into connectionist models in cognitive science.

Associations and Equilibration in Psychodynamic and

Connectionist Models
Psychoanalysis and connectionism share a common ancestor:
associationism. Although these two descendants may at first
glance seem like very divergent branches on the family tree, the
family resemblance may have some important implications.
When Freud developed the method of free association as a
therapeutic tool, a primary aim was to bypass the conscious
theories that patients held about their minds and symptoms and
to try to map the structure of their associational networks. In
contemporary practice, if a patient presents with some form of
avoidance--difficulty completing projects he actually wants to
complete, sexual anxiety and attendant avoidance of situations
that might lead to sexual encounters, anxiety about spending
money on herself, and so f o r t h - - a psychodynamic therapist is
likely to explore specific instances and, for each, to follow the
trail of associations until patterns begin to emerge that may
provide insight into the cognitive-affective associations underlying the agoidance. The assumption is that if a person does not
know why he is avoiding something he consciously wants to
do, then at some level the action must be associated with a fear
(or with something else associatively connected to it that itself
is fear-inducing). Because associational networks are not available to introspection, the only way to try to map them is to see
what thoughts, feelings, memories, images, and so forth emerge

FREUD'S LEGACY
when the person relaxes consciousness and simply says whatever
comes to mind. Thoughts or memories that themselves elicit
avoidance during this process--that come to mind but that the
patient does not want to divulge--are likely to be particularly
relevant because they are not only associatively connected but
also linked to some kind of unpleasant feeling.
Although cognitive models have not traditionally been attentive to these more motivational processes, the link between
Freud's free-associational method and many cognitive procedures for trying to map the implicit organization of associative
networks should be clear to any cognitive scientist. Cognitive
psychologists use measures such as reaction time following
priming to assess the degree of association between two nodes
on a network (e.g., bird and feathers), and there is no reason
that one would not expect meaningful idiographic data similarly
to emerge by ease of activation of specific associations to a
prime (a memory, dream element, symptom, feeling, etc.).
An important point of complementarity between psychodynamic and connectionist models is that the former have elaborated on the role of affect and motivation in associative processes, whereas the latter have focused on the role of more
cognitive and perceptual constraints. Despite their very different
foci, however; the two approaches share certain features that
make integration not only possible but potentially very useful.
First, both assume that psychological processes occur simultaneously, in parallel, but also presume a serial processing capacity (consciousness)superimposed on a parallel architecture.
Second, both assume that psychological processes active below
the threshold of consciousness at any given time can move the
system in competing directions and that the result is an equilibrated solution to multiple and potentially competing forces. In
connectionist models, this means that a process such as categorization occurs through the simultaneous activation of units of
information (nodes) that can excite or inhibit each other. In
psychoanalytic theory, a person can attach multiple evaluative
responses, and hence have competing motivational tendencies,
toward the same object and is likely to construct a compromise
solution to this affective-motivational "problem" outside of
awareness. For example, a person can channel competitive or
aggressive impulses into socially acceptable practices, such as
sports or academic achievement.
A third premise shared by both approaches is that the meaning
of an object, concept, or behavior is not contained in any single
unit or dynamic but can be understood only in terms of a broader
configuration. In PDP models, the meaning of an object is distributed throughout a network of processing units that, through
experience, have become activated in tandem. Each of these
units attends to some small aspect of the representation, and
none alone "stands for" the entire concept. From a psychodynamic perspective, the meaning of any conscious belief, feeling,
behavior, or symptom can be understood only in the context of
the dynamic interplay of unconscious processes that create it.
Although these similarities may seem more like analogies
than shared features, in fact they reflect an underlying holistic,
equilibratory, dynamic, and field-systemic approach that Read,
Vanman, and Miller (1997) have recently argued is shared by
connectionist and Gestalt models in perception and social psychology. In each case, a unified " d e c i s i o n " - - w h e t h e r an approaching insect can be categorized as a moth or whether a

359

symptom or conscious belief represents a stable if not optimal

solution to a conflict--is an equilibrated solution emerging from
a configuration of dynamic processes active outside of
awareness.

Cognitive and Affective Constraint Satisfaction

From the present perspective, of particular importance with
respect to the potential utility of psychodynamic constructs is
the notion of parallel constraint satisfaction central to most
connectionist models. Constraint satisfaction refers to the tendency to equilibrate to a solution that satisfies as many constraints as possible in order to achieve the best fit to the data.
The nodes in a PDP network essentially represent hypotheses
as to the presence or absence of a given feature (Read et al.,
1997). The hypothesis can be as simple as whether a line has
a horizontal orientation or as complex as an inference about
whether a particular tone of voice was intended as an insult.
The links among nodes are weightod, and these ever-changing
weights represent the extent to which the information in associated nodes is mutually supportive or conflicting. These weighted
links reflect the extent to which the two nodes have been coactivated, and they increase the more times two nodes are activated
in tandem. Uncorrelated units of information will have weights
of zero connecting them, whereas units of information that tend
to covary will have strong positive weights, so that activating
one will spread activation to the other. Nodes that represent
incompatible (negatively correlated) information (such as a
hostile tone of voice and a friendly facial expression) will have
negative weights connecting them and thus inhibit each other as
activation spreads from one to the other. If, however, a person's
experience changes so that two objects or representations no
longer tend to covary or contradict one another (as when a
conditioned stimulus is no longer paired with the unconditioned
stimulus and hence no longer predicts its presence), the weight
will gradually move toward zero.
Psychodynamic theory can augment a connectionist model in
proposing that affects and affectively charged motives provide
a second set of constraints, distinct from strictly cognitive or
informational ones, that influence the outcomes of parallel constraint-satisfaction processes. The decisions people implicitly
make as they are confronted with tasks of categorization, inference, problem solving, and so forth are constrained not only by
the data but by the hedonic significance of different decisions.
According to this integrated view, the mind is not only cognitively self-regulating but affectively self-regulating.
For example, Ali Feit and I (Westen & Feit, 1998) recently
completed a study in which we attempted to predict participants'
beliefs about what happened between President Clinton and
Kathleen Willey, who accused him of sexual harassment, in the
week after the scandal broke. We hypothesized that the extent
to which people believed Willey's version of the events would
reflect some combination of cognitive constraints (what they
know about Clinton and the scandal) and affective constraints
imposed by their feelings toward the Democratic and Republican parties, Clinton, infidelity, and feminism. We assessed cognitive constraints by asking participants 10 factual questions
about Clinton' s life that would indicate longstanding knowledge
about his political and personal history and 10 factual questions

360

WESTEN

about the scandal (such as how Willey's husband died, how

much she had sought as a book advance, etc.). We assessed
affective constraints by asking multiple questions about participants' feelings toward the Democrats, Republicans, infidelity,
feminism, etc. and by factor analyzing their scores on these
variables.
As predicted, beliefs about the Willey affair reflected a compromise formation, or equilibration process, that yielded a conscious solution that minimized negative affect but did not completely disregard the facts. Three variables predicted beliefs and
certainty about what happened: feelings toward Democrats and
Republicans (/3 = .46), feelings toward high-status men who
philander (/3 = - . 3 8 ) , and knowledge about Clinton'S life (/3
= .28), with a multiple R = .63 (R 2 = .39). In other words,
the less participants liked Republicans, disliked philandering
men, and knew about Clinton, the less likely they were to construct the belief that he had molested Willey in the Oval Office.
In this case, cognitive constraints were actually less influential
than affective constraints in shaping conscious judgment.
Similar ideas and data are beginning to emerge in the cognitive literature as well as in research in social and personality
psychology (see Kunda & Thagard, 1996; Mischel & Shoda,
1995) as researchers are increasingly recognizing the role of
affect and motivation in constraint-satisfaction processes. Perhaps the best example is Holyoak and Thagard's (1989) model
of analogical reasoning--the process by which people understand a novel situation in terms of a familiar analog--through
implicit selection of the best fitting of multiple relevant analogs
stored in memory. For example, during the Gulf War, U.S. President George Bush compared Saddam Hussein to Hitler; if one
accepts this premise, then Iraq's invasion of Kuwait was like
Germany's invasion of its neighbors at the start of World War II,
which implied that Saddam had to be stopped before becoming a
danger to the world (Spellman & Holyoak, 1992).
According to Holyoak and Thagard ( 1989, 1997), selecting
the best fitting analog is a constraint-satisfaction process that
involves three sets of constraints. Two are cognitive: The current
situation must be similar enough to the analog stored in memory
to activate it, and the elements of the novel situation and the
analog mast be readily mapped onto each other. Thus, Iraq's
occupation of Kuwait did not activate the analog of one team
defeating another in a soccer match, because Kuwait was not
a willing participant in the "game." The third constraint is
motivational: the reasoner's goals. Had President Bush not
wanted to intervene after the Iraqi takeover of Kuwait, he could
easily have chosen a different analog, such as the Vietnam analog that blocked U.S. intervention in many similar international
confrontations for two decades. In a paragraph that makes clear
the resemblance between this approach and Freud's concept of
compromise formations, Holyoak and Thagard wrote that these
three kinds of constraint satisfaction function "like the various
pressures that guide an architect engaged in creative design,
with some forces in convergence, others in opposition, and their
constant interplay pressing toward some satisfying compromise
that is internally coherent" (1997, p. 36). That Bush chose the
analog of World War II instead of Vietnam is probably no accident, because he had fought in World War I / a n d its memory
was likely associated with much more affect. Experimental research in fact has begun to document that manipulating partici-

pants' goals changes their analogical mappings when confronted

with ambiguous situations that could be mapped in more than
one way (Spellman & Holyoak, 1996).
A psychodynamically informed connectionist model can be
applied to a range of phenomena described in the empirical
literature, some of which was reviewed previously, such as selfserving biases. With self-serving biases, one of the constraints
that enters into the equation is that the conclusion should maximize positive feelings about the self and minimize negative
feelings. A solution that is too far afield from reality as the
person has experienced it will likely elicit a connectionist version of cognitive dissonance (see Shultz & Lepper, 1996; E. R.
Smith, 1996), whereas a solution that threatens self-esteem will
elicit affective dissonance. As a result, a person is likely to be
drawn to a conclusion that compromises affective considerations
with accuracy. As Morling and Epstein (1997) have shown, this
is precisely the kind of compromise that" tends to occur when
people make judgments about themselves.
This kind of model is also capable of handling individual
differences, such as the tendency of narcissistic individuals to
rate themselves much more positively than objective observers
rate them on a variety of dimensions and the tendency of people
with a depressive bent of mind to underrate themselves (John &
Robins, 1994). For narcissists, the threat to self-esteem of realistic appraisals is strong enough that affective considerations cim
override accuracy considerations more easily than they do for
most people, who only moderately self-enhance. This suggests,
as long argued by psychoanalytic clinicians, that beneath the
conscious grandiosity of the narcissist is usually a fragile baseline of implicit or unconscious self-esteem or a pervasive deficit
in the capacity for self-esteem regulation (Kemberg, 1975;
Reich, 1960; Westen, 1990b). Clinically, affect appears to have
particular power to alter the configurations of equilibrated solutions, and the extent to which this is the case depends on the
person, the affect, the intensity of the affect, and the chronically
activated issues that trigger strong aversive affective reactions.
For some depressive individuals, whose affective "pulls" often run in the opposite direction of the narcissist's, affective
constraint-satisfaction processes are likely to override accuracy
considerations in the opposite direction. Cognitive models of
schematic processing have never adequately explained this phenomenon because thinking bad things about oneself is so unpleasant that such schematic processes would surely be extinguished over time through operant conditioning mechanisms,
giving way to the moderate self-enhancement characteristic of
most people. From a psychodynamic perspective, motivated
avoidance of positive thoughts and feelings about the self is
likely to reflect identification with a critical parent, a learning
history in which pride and self-esteem were associated with
feeling fraudulent, being disappointed at the lack of response
from (or criticism by) parental figures, guilt at outdoing a significant other, activation of wishes for more esteem from significant others that the person has learned to expect not to be
forthcoming, fear of subsequent humiliation, and so forth (Westen, 1985, chapter 2; Westen, in press-a). These motivational
factors, like those that lead to distortions in processing information about the self in narcissistic individuals (who similarly have
problems regulating self-esteem but have different preferred
solutions), can be readily integrated into a connectionist model

FREUD'S LEGACY
that views data-fitting and hedonic value as two sets of constraints that influence information processing, often in opposing
directions.
As suggested above in the discussion of the multiple meanings
of activation, a psychodynamically informed connectionist
model would also distinguish more carefully between implicit
and explicit products of equilibrated solutions. Just as narcissistic individuals can respond to perceived failures with defensively
grandiose conscious self-representations, people can hold positive explicit beliefs about negatively stereotyped groups that are
incompatible with their implicit stereotypes. In each of these
cases, the system has settled on two solutions, one at odds with
the other. (Alternatively, one might consider the combination of
the two incongruent representations to be a single equilibrated
solution.)

Methods and Data

Elsewhere I have described some of the methodological implications of taking psychodynamic constructs more seriously, particularly the notion that emotions and motives can influence
behavior outside of awareness and that people can regulate their
affects unconsciously (e.g., Westen, 1995, 1996, 1997; Westen
et al., 1997). The overreliance on self-reports in personality,
clinical, and attitude research can no longer be maintained in
the face of mounting evidence that much of what we do, feel,
and think is inaccessible to consciousness; indeed, on the basis
of neuropsyehological data on the generation of fear responses,
LeDoux (1997) has made precisely the same point. Here, however, I briefly highlight a very different implication for psychological methods and data of integrating psychodynamic thinking
into mainstream psychology, more at the level of philosophy of
science.
Psychoanalytic theorists have been describing, categorizing,
and exploring the workings of unconscious mental processes
for a century, whereas the existence of these processes has just
been acknowledged in mainstream research in psychology in
the last decade. This should give us pause to wonder whether
the impllcit, if not the explicit, philosophy of science that guides
our work requires some revision. We explicitly acknowledge in
introductory psychology and methods texts that case studies are
a legitimate form of psychological inquiry, albeit one open to
bias. However, our implicit attitude--expressed, for example,
in the absence of descriptions of anything other than neurological (i.e., "hard science") cases in APA journals--bespeaks an
attitude toward data that can only be understood in historical
context.
In the late 19th century, when psychology emerged from philosophy as a discipline, psychologists were determined to turn
psychology into an objective science that could answer questions
empirically rather than through the seemingly endless speculations of one philosopher after another about the nature of mind,
knowledge, free will, innate ideas, and the like. Psychophysics
provided a model of how this could be done, and early psychologists such as Titchener were determined to route subjectivity
out of psychological science. Behaviorists took this a step further, and although cognitive psychologists are now gradually
letting emotions and motives back into what began as a cold,
cognitive mind, there remains in academic psychology a perva-

361

sive antipathy toward data from the clinic, clinical inference,

and especially psychoanalytic modes of inference.
Anyone who reads the psychoanalytic literature can readily
see the perils of developing theory and making inferences without the constraints imposed by a more systematic empirical
attitude. Operationalizing variables not only allows one to test
hypotheses but forces one to clarify the meaning of one's constructs. Everything Titchener, Watson, Skinner, and others appropriately wanted to avoid can be seen routinely in the pages of
even the best psychoanalytic journals. On the other hand, it is
no accident that psychoanalysts have been making some very
sophisticated observations about unconscious processes for decades, whereas experimental psychologists have only recently
acknowledged their existence. Researchers are now discovering,
for example, that word associations can provide a way of assessing individual differences in the linkages among concepts
in memory (Stacy, Leigh, & Weingardt, 1997).
Clinical observation provides a remarkably important, even
if at times foggy, window to the operations of the mind. The
reason is that it allows observation of individuals who are motivated to disclose their most personally significant and affectladen experiences longitudinally for months or years as they
describe their experiences; offer associations to thoughts or feelings they cannot consciously explain; and express patterns of
thought, feeling, motivation, and behavior in their interactions
with a significant other with whom they have considerable rapport and who has considerable experience with the way they
behave and impose meaning on events. None of this can be
obtained through other commonly used methods.
The philosophical and methodological issue is, I believe, essentially one of signal detection: At what point does one report
a h i t - - t h a t is, take an observation as credible enough to include
in formulating a theory or m o d e l - - o r dismiss a potentially significant piece of data as noise? For reasons that can only be
understood in the context of their respective histories--consistent with the signal detection a n a l o g y - - I would argue that psychoanalysis has chronically set the threshold far too low,
whereas experimental psychology has set it too high. A more
pragmatic philosophy of science might consider clinical observations as potentially relevant data, particularly when they appear to disconfirm o r suggest modifications in current thinking
(because disconfirming instances need not be experimental),
while not according these data the evidentiary weight of more
systematic scientific observations.
Psychodynamic theory and observation is, I believe, extremely valuable in one other respect. Philosophers of science
often distinguish between (a) the context of scientific discovery,
in which initial observations are made and hypotheses are
framed, and (b) the context of justification, in which hypotheses
are tested by use of the scientific method. I have just suggested,
essentially, that we not ignore clinical data in the context of
discovery. Clinical data may, however, be useful in another respect suggested by a century of psychodynamic thinking and
observation. In the human sciences, one could argue that alongside the more familiar context of scientific justification is an
interpretive context of justification, in which a theory demonstrates its relative superiority to lay theories or competing scientific theories if it can confer intelligibility to otherwise less
intelligible phenomena by offering principles for drawing infer-

362

WESTEN

ences about the meanings that guide behavior. In this view,

prediction and control and interpretive understanding are two
ways of knowing that one knows something about people, even
if their evidentiary weights are not equal.

Conclusion
So what conclusions can one draw about the legacy of Sigmund Freud a century after he began his inquiries into human
nature? The answer is complex. Freud advanced several fundamental propositions, once highly controversial and unique to
psychoanalysis, that have stood the test of time, and many of
the implicit and explicit assumptions today at the heart of the
school of thought he initiated have a considerably stronger empirical basis than has been widely supposed. This is probably
the best any thinker could hope for in a rapidly developing
discipline like ours 60 years after his death. No doubt, many
psychoanalytic writings are obscure, muddleheaded, and ignorant of relevant empirical work. But clinical investigation has
led to a set of propositions about the mind, now well verified
empirically, that can inform the experimental investigation of
human mental life and behavior and the theories that drive and
derive from it. In particular, a wedding of psychoanalysis and
cognitive science might produce some important, if unlikely,
offspring.
Was Freud wrong in some of his fundamental ideas about
human nature? Without doubt. His theory of drives was highly
problematic, his view of aggression was far too mechanistic,
his hypothesized death instinct was evolutionarily untenable, his
theory of development was too exclusively focused on sexuality,
his theory of female development was simplistic, and so on.
Fisher and Greenberg ( 1996, pp. 19-21 ) have recently distilled
and laid out in propositional form the general theory of psychopathology that Freud proposed, and I must admit that as a practicing dynamically oriented clinician I can scarcely find any
proposition on their three-page list with which I agree even
remotely.
Grand theorists like Freud, Piaget, and Skinner are routinely
the grandest purveyors of falsehood in the business. This reflects
simple mathematics: The more propositions one advances (and
the bolder those hypotheses are), the higher the probability that
several, will be wrong. 16 But on some of the central postulates
of psychodynamic theory, such as the view that much of mental
life is unconscious, Freud has left an i m p o r t a n t - - a n d I believe
i n d e l i b l e - - m a r k on human self-understanding. As psychology
moves into its second century, we would do well to attend to
and integrate some of these disavowed psychodynamic ideas,
which need not remain, like classic psychodynamic symptoms,
outside the consciousness of the scientific community.
16Popper (1957) was right in prescribing that scientists make risky,
falsifiable predictions, but he neglected to consider the relative contributions of broad versus circumscribed theories. The philosophy of science
he proposed, and that most psychologists tacitly accept, is biased toward
microtheories devoid of a broader paradigmatic context because these
theories are the most likely to withstand efforts at falsification.

References
Adams, H.E., Wright, L.W., & Lohr, B.A. (1996). Is homophobia
associated with homosexual arousal? Journal of Abnormal Psychology, 105, 440-445.
American Psychiatric Association. (1994). Diagnostic and statistical
manual of mental disorders (4th ed.). Washington, DC: Author.
Andersen, S., & Baum, A. (1994). Transference in interpersonal relations: Inferences and affect based on significant-other representations.
Journal of Personality, 62, 460-497.
Andersen, S., & Cole, S.W. (1990). "Do I know you?" The role of
significant others in social perception. Journal of Personality and
Social Psychology, 59, 384-399.
Andersen, S., Reznik, I., & Manzella, L. (1996). Eliciting facial affect,
motivation, and expectancies in transference: Significant-other representations in social relations. Journal of Personality and Social Psychology, 71, 1108-1129.
Anderson, J. (1995). Learning and memory: An integrated approach.
New York: Wiley.
Arkowitz-Westen, L. (1998). The roles of negative reciprocity and acceptance in couples: Predicting relationship quality. Unpublisheddoctoral dissertation, University of Nevada, Reno.
Arlow, J. ( 1991 ). Psychoanalysis: Clinical theory and practice. Madison, C~ International Universities Press.
Akron, L. (1996). A meeting of minds: Mutuality in psychoanalysis.
Hillsdale, NJ: Analytic Press.
Bahrick, H., Hall, L. K., & Berger, S. A. (1996). Accuracy and distortion
in memory for high school grades. Psychological Science, 7, 265271.
Baker, L., Silk, K. R., Westen, D., Nigg, J. T., & Lohr, N. E. (1992).
Malevolence, splitting, and parental ratings by borderlines. Journal
of Nervous and Mental Disease, 180, 258-264.
Banaji, M., & Hardin, C. (1996). Automatic stereotyping. Psychological
Science, 7, 136-141.
Bandura, A. (1977). Social learning theory. Englewood Cliffs, NJ: Prentice Hall.
Bandura, A., Barbaranelli, C., Caprara, G., & Pastorelli, C. (1996).
Mechanisms of moral disengagement in the exercise of moral agency.
Journal of Personality and Social Psychology, 71, 364-374.
Bank, L., Marlowe, J.H., Reid, J.B., & Patterson, G.R. (1991). A
comparative evaluation of parent-training interventions for chronic
delinquents. Journal of Abnormal Child Psychology, 19, 15-33.
Bargh, J. A. (1997). The automaticity of everyday life. In RI S. Wyer,
Jr. (Ed.), The automaticity of everyday life: Advances in social cognition (Vol. 10, pp. 1-61). Mahwah, NJ: Erlbaum.
Bargh, J., & Barndollar, K. (1996). Automaticity in action: The unconscious as repository of chronic goals and motives. In P.M. Gollwitzer & J. Bargh (Eds.), The psychology of action: Linking cognition
and motivation to behavior (pp. 457-481 ). New York: Guilford Press.
Barron, J.W., Eagle, M.N., & Wolitzky, D.L. (Eds.) (1992). The
interface of psychoanalysis and psychology. Washington, DC: American Psychological Association.
Baumrind, D. (1967). Child care practices anteceding three patterns of
preschool behavior. Genetic Psychology Monographs, 75, 43-88.
Baumrind, D. (1991). The influence of parenting style on adolescent
competence and substance use. Journal of Early Adolescence, 11,
56-95.
Bechara, A., Tranel, D., Damasio, H., Adolphs, R., Rockland, C., &
Damasio, A. ( 1995, August ). Double dissociation of conditioning and
declarative knowledge relative to the amygdala and hippocampus in
humans. Science, 29, 1115-1118.
Bechara, A., Tranel, D., Damasio, H., & Damasio, A. (1996). Failure
to respond autonomically to anticipated future outcomes following
damage to prefrontal cortex. Cerebral Cortex, 6, 215-225.
Biederman, J., Milberger, S., Faraone, S., Kiely, K., Guite, J., Mick, E.,

FREUD'S LEGACY
Warburton, R., & Reed, E. (1995). Family-environment risk factors
for attention-deficit hyperactivity disorder. Archives of General Psychiatry, 52, 464-470.
Bjork, E., Bjork, R., & Anderson, M.C. (1998). Varieties of goaldirected forgetting. In J. M. Golding & C. M. MacLeod (Eds.), Intentional forgetting: Interdisciplinary approaches (pp. 103-137). Mahwall, NJ: Erlbanm.
Blatt, S. (1994). Therapeutic change: An object relations approach.
New York: Plenum.
Blatt, S., D'Affliti, J., & Quinlan, D. (1976). Experiences of depression
in normal young adults. Journal of Abnormal Psychology, 85, 385389.
Blatt, S., & Homann, E. (1992). Parent-child interactions in the etiology
of dependent and self-critical depression. Clinical Psychology Review,
12, 47-91.
Blatt, S., & Lernet; H. ( 1991 ). Psychodynamic perspectives on personality theory. In M. Hersen, A. Kazdin, & A. S. Bellack (Eds.), The
clinical psychology handbook (2nd ed., pp. 147-169). New York:
Pergamon Press.
Blatt, S., & Zuroff, D.C. (1992). Interpersonal relatedness and selfdefinition: Two prototypes for depression. Clinical Psychology Review, 12, 527-562.
Block, J. (1971). Lives through time. Berkeley, CA: Bancroft.
Block, J. (1978). The Q-sort method in personality assessment and
psychiatric research. Palo Alto, CA: Consulting Psychologists Press.
Block, J: (1993). Studying personality the long way. In D. Fundel; R. D.
Parke, C. Tomlinson-Keasey, & K. Widaman (Eds.), Studying lives
through time: Personality and development (pp. 9-41 ). Washington,
DC: American Psychological Association.
Block, J., & Block, J. (1980). The role of ego-control and ego-resiliency
in the organization of behavior. In W. A. Collins (Ed.), Minnesota
Symposium on Child Development: Vol. 13. Hillsdale, NJ: Erlbaum.
Block, J. H., Gjerde, P. F., & Block, J. H. (1991). Personality antecedents of depressive tendencies in 18-year-olds: A prospective study.
Journal of Personality and Social Psychology, 60, 726-738.
Blum, G. S. (1954). An experimental reunion of psychoanalytic theory
with perceptual vigilance and defense. Journal of Abnormal and Social Psychology, 49, 94-98.
Bond, M., Paris, J., & Zweig-Frank, H. (1994). Defense styles and
borderline personality disorder. Journal of Personality Disorders, 8,
28-31.
Bornstein, R . E (1995). Sex differences in objective and projective
dependency tests: A meta-analytic review. Assessment, 2, 319-331.
Bornstein, R., & Masling, J. (1985). Orality and latency of volunteering
to serve as experimental subjects: A replication. Journal of Personality Assessment, 49, 306-310.
Boudewyn, A., & Liem, J. (1995). Childhood sexual abuse as a precursor to depression and self-destructive behavior in adulthood. Journal
of Traumatic Stress, 8, 445-459.
Bowlby, J. (1969). Attachment and loss: Vol. 1. Attachment. New York:
Basic Books.
Bradley, C. (1997). Generativity-stagnation: Development of a status
model. Developmental Review, 17, 262-290.
Brennan, K. A., Clark, C. L., & Shaver, P. (1998). Self-report measurernent of adult romantic attachment: An overview. In J. A. Simpson &
W. S. Rholes (Eds.), Attachment theory and close relationships (pp.
46-76). New York: Guilford Press.
Brenner, C. (1982). The mind in conflict. Madison, C'12. International
Universities Press.
Breslau, N., Davis, G. C., Andreski, P., Peterson, E. L., & Schultz, L. R.
(1997). Sex differences in posttraumatic stress disorder. Archives of
General Psychiatry, 54, 1044-1048.
Bretherton, I. (1990). Communication patterns, internal working mod-

363

els, and the intergenerational transmission of attachment relationships.

Infant Mental Health Journal, I1, 237-252.
Brewer, M., & Brown, R. J. (1998). Intergroup relations. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), Handbook ofsocialpsychology
(Vol. 2, pp. 554-594). New York: McGraw-Hill.
Briere, J., & Runtz, M. (1993). Childhood sexual abuse: Long-term
sequelae and implications for psychological assessment. Journal of
Interpersonal Violence,& 312-330.
Broadbent, D. E. (1977). The hidden preattentive processes. American
Psychologist, 32, 109-118.
Brown, L. L., Tomarken, A., Orth, D., Loosen, P., Kalin, N., & Davidson,
R. (1996). Individual differences in repressive-defensiveness predict
basal salivary cortisol levels. Journal of Personality and Social Psychology, 70, 362-371.
Bruyer, R. ( 1991 ). Covert face recognition in prosopagnosisa: A review.
Brain and Cognition, 15, 223-235.
Bunce, S., Bernat, E., Wong, P., & Shevrin, H. (1995, August). Eventrelated potential and facial EMG indicators of emotion-relevant unconscious learning. Paper presented at the 103rd Annual Convention
of the American Psychological Association, New York.
Bunce, S., Casada, J., Amdur, R., & Liberzon, L (1997). Combat sounds
elicit asynchronous physiological response pattern among combat
controls. Unpublished manuscript, University of Michigan.
Bunce, S. C., Bernat, E., & Shevrin, H. (1997). Complex unconscious
cognitions, affective systems and life experiences: The unconscious
influence of early loss on positive and intimate emotions. Paper presented at the 9th Annual Convention of the American Psychological
Society.
Bunce, S., Larsen, R., & Peterson, C. (1995). Life after trauma: Personality and daily life experiences of traumatized people. Journal of
Personality, 63, 165-188.
Bunce, S., Snodgrass, M., & Larsen, R. J. ( 1996, August). Psychophysiological responses to mental imagery in a subclinical trauma population. Paper presented at the 104th Annual Convention of the American
Psychological Association, Toronto, Ontario, Canada.
Burbach, D., & Borduin, C. (1986). Parent-child relations and the
etiology of depression: A review of methods and findings. Clinical
Psychology Review, 6, 133-153.
Butcher, J. N., Dahlstrom, W. G., Graham, J. R., Tellegen, A., & Kaemmet; B. (1989). Minnesota Multiphasic Personality Inventory-2
(MMPI-2 ): Mahual for administration and scoring. Minneapolis:
University of Minnesota Press.
Cacioppo, J.T., Gardner; W.L., & Berntson, G.G. (1997). Beyond
bipolar conceptualizations and measures: The case of attitudes and
evaluative space. Personality and Social Psychology Review, 1, 325.
Cadoret, R. J., Yates, W. R., Troughton, E., Woodworth, G,, & Stewart,
M. A. (1995). Adoption study demonstrating two genetic pathways
to drug abuse. Archives of General Psychiatry, 52, 42-52.
Calabrese, M. (1998). Object representations as they relate to adult
attachment, intimacy and self-esteem. Unpublished doctoral dissertation, Columbia University.
Cantor, N., & Harlow, R. (1994). Personality, strategic behavior, and
daily-life problem solving. Current Directions in Psychological Science, 3, 169-172.
Cantor, N., & Kihlstrom, J. E (1987). Personality and social intelligence. Englewood Cliffs, NJ: Prentice Hall.
Caspi, A. (1998). Personality development across the lifespan. In W.
Damon (Series Ed.) & N. Eisenberg (Vol. Ed.), Handbook of child
psychology: Vol. 3. Social, emotional, and personality development
(2nd ed., pp. 311-388). New York: Wiley.
Caspi, A., Elder, G. E., & Herbener, E. (1990). Childhood personality
and the prediction of life-course patterns. In L. N. Robins & M. Rutter

364

WESTEN

(Eds.), Straight and devious pathways from childhood to adulthood

(pp. 13-35). New York: Cambridge University Press.
Caspi, A., Henry, B., McGee, R. O., Moffitt, T., & Silva, P. (1995).
Temperamental origins of child and adolescent behavior problems:
From age 3 to age 15. Child Development, 66, 55-68.
Caspi, A., Moffltt, T., Newman, D. L., & Silva, P. (1996). Behavioral
observations at age 3 years predict adult psychiatric disorders. Archives of General Psychiatry, 53, 1033-1039.
Cavaiola, A., & Schiff, M. (1988). Behavioral sequelae of physical and/
or sexual abuse in adolescents. Child Abuse & Neglect, 12, 181 - 188.
Ceci, S. J., & Bronfenbrenn~ U. (1991). On the demise of everyday
memory: "The rumors of my death are much exaggerated" (Mark
Twain). American Psychologist, 46, 27-31.
Chaiken, S. (1980). Heuristic versus systematic information processing
and the use of source versus message cues in persuasion. Journal of
Personality and Social Psychology, 39, 752-766.
Chandler, M. J., Paget, K. E, & Koch, D. A. (1978). The child's demystification of psychological defense mechanisms: A structural and developmental analysis. Developmental Psychology, 14, 197-205.
Colvin, C. R., Block, J., & Funder, D. C. (1995). Overly positive selfevaluations and personality: Negative implications for mental health.
Journal of Personality and Social Psychology, 68, 1152-1162.
Coplan, J., Trost, R.C., Owens, M.J., Cooper, Z B., Gorman, J.M.,
Nemeroff, C.B., & Rosenblum, L.A. (1998). Cerebrospinal fluid
concentrations of somatostatin and biogenic amines in grown primates
reared by mothers exposed to manipulated foraging conditions. Archives of General Psychiatry, 55, 473-477.
Cork, R. C. (1996). Implicit memory during anesthesia. In S. R. Hameroff, A. Kaszniak, & A. C. Scott (Eds.), Toward a science of consciousness: The first Tucson discussions and debates (pp. 295-302). Cambridge, MA: MIT Press.
Costanzo, K. (1996). The connection between object relations, social
support, social adjustment, and health. Unpublished master's thesis,
Harvard University.
Cousineau, T.M. (1997). Psychological predictors of health service
utilization in college students (Doctoral dissertation, A_delphi University, 1997). Dissertation Abstracts International, 57, 4700.
Cowey, A. (1991). Grasping the essentials. Nature, 349, 102-103.
Cramer, P. (1991). Anger and the use of defense mechanisms in college
students. Journal of Personality, 59, 39-55.
Cramer, P. (1996). Story-telling, narrative, and the Thematic Apperception Test. New York: Guilford Press.
Crews, E ( 1993, November 18). The unknown Freud. New York Review
of Books.
Crews, F. (1996). The verdict on Freud. Psychological Science, 7, 6367.
Crowne, D. P., & Marlow, D. (1960). A new scale of social desirability
independent of psychopathology. Journal of Consulting Psychology,
24, 349-354.
Damon, W. (1977). The social world of the child. San Francisco: JosseyBass.
Damon, W., & Hart, W. (1988). Self-understanding in childhood and
adolescence. New York: Cambridge University Press.
Davidson, K., & MacGregor, B. A. ( 1995, August). Clinically-judged
defensiveness as a moderator of self-friend neuroticism consensus.
Paper presented at the 103rd Annual Convention of the American
Psychological Association, New York.
Davidson, R. (1992). Emotion and affective style: Hemispheric substrates. Psychological Science, 3, 39-43.
Davis, P., & Schwartz, G. (1987). Repression and the inaccessibility
of affective memories. Journal of Personality and Social Psychology,
52, 155-162.
Devine, P. (1989). Stereotypes and prejudice: Their automatic and con-

trolled components. Journal of Personality and Social Psychology,

56, 5-18.
Dixon, N. E ( 1971 ). Subliminal perception: The nature of a controversy.
New York: McGraw-Hill.
Dixon, N. E ( 1981 ). Preconscious processing. New York: Wiley.
Dollard, J., & Mille~ N. (1950). Personality and psychotherapy: An
analysis in terms of learning, thinking, and culture. New York:
McGraw-Hill.
Dovidio, J., & Gaertner, S. (1993). Stereotypes and evaluative intergroup
bias. In D. Mackie & D. Hamilton (Eds.), Affect, cognition, and
stereotyping: Interactive processes in group perception (pp. 167193). San Diego, CA: Academic Press.
Dozier, M., & Kobak, R. (1992). Psychophysiology in attachment interviews: Converging evidence for deactivating strategies. ChiM Development, 63, 1473-1480.
Dunning, D., Leuenberger, A., & Sherman, D. (1995). A new look at
motivated inference: Are self-serving theories of success a product of
motivational forces? Journal of Personality and Social Psychology,
69, 58-68.
Eagle, M. (1959). The effects of subliminal stimuli of aggressive content
upon conscious cognition. Journal of Personality, 27, 678-688.
Ellenberger, H. (1970). The discovery of the unconscious: The history
and evolution of dynamic psychiatry. New York: Basic Books.
Emmons, R., & King, L. A. (1988). Conflict among personal strivings:
Immediate and long-term implications for psychological and physical
well-being. Journal of Personality and Social Psychology, 54, 10401048.
Epstein, S. (1992). Coping ability, negative self-evaluation, and overgeneralization: Experiment and theory. Journal of Personality and Social
Psychology, 62, 826-836.
Epstein, S. (1994). Integration of the cognitive and the psychodynamic
unconscious. American Psychologist, 49, 709-724.
Erdelyi, M. (1974). A new look at the new look: Perceptual defense
and vigilance. Psychological Review, 81, 1-25.
Erdelyi, M. (1985). Psychoanalysis: Freud's cognitive psychology. San
Francisco: W. H. Freeman.
Erdelyi, M. (1990). Repression, reconstruction, and defense: History
and integration of the psychoanalytic and experimental frameworks.
In J. Singer (Ed,), Repression and dissociation: Defense mechanisms
and personality styles: Current theory and research. Chicago: University of Chicago Press.
Erikson, E. (1963). Childhood and society. New York: Norton.
Esteves, E, Dimberg, U., & Ohman, A. (1994). Automatically elicited
fear~ Conditioned skin conductance responses to masked facial expressions. Cognition and Emotion, 8, 393-413.
Fairbairn, W. R. D. (1952). Psychoanalytic studies of the personality.
London: Routledge & Kegan Paul.
Fazio, R. (1990). Multiple processes by which attitudes guide behavior:
The MODE model as an integrative framework. Advances in Experimental Social Psychology, 23, 75-109.
Fazio, R., Jackson, J. R., Dunton, B., & Williams, C. J. (1995). Variability in automatic activation as an unobtrusive measure of racial attitudes: A bona fide pipeline? Journal of Personality and Social Psychology, 69, 1013-1027.
Fein, S., & Spencer, S. J. (1997). Prejudice as self-image maintenance:
Affirming the self through derogating others. Journal of Personality
and Social Psychology, 73, 31-44.
Fisher, S., & Greenberg, R. (1985). The scientific credibility of Freud's
theories and therapy. New York: Columbia University Press.
Fisher, S., & Greenberg, R. P. (1996). Freud scientifically reappraised:
Testing the theories and therapy. New York: Wiley.
Fonagy, P., Steele, H., & Steele, M. (1991). Maternal representations
of attachment during pregnancy predict the organization of infant-

FREUD'S LEGACY
mother attachment at one year of age. Child Development, 62, 891905.
Franz, C., McClelland, D., & Weinberger, J. ( 1991 ). Childhood antecedents of conventional social accomplishment in midlife adults: A 36year prospective study. Journal of Personality and Social Psychology,
60, 586-595.
Franz, C., Weinberger, J., Kremen, W., & Jacobs, R. (1996). Childhood
antecedents of dysphoria in adults: A 36-year longitudinal study.
Unpublished manuscript, Williams College.
Freud, A. (1936). The ego and the mechanisms of defense. Madison,
C'12. International Universities Press.
Freud, S. (1953a). Character and anal eroticism. In J. Strachey (Ed.
and Trans.), The standard edition of the complete psychological works
of Sigmund Freud (Vol. 9, pp. 169-175). London: Hogarth Press.
(Original work published 1908)
Freud, S. (1953b). The interpretation of dreams. In J. Strachey (Ed.
and Trans.), The standard edition of the completepsychological works
of Sigmund Freud (Vol. 4, pp. 1-338). London: Hogarth Press. (Original work published 1900)
Freud, S. (1953c). Symptoms, inhibitions, and anxiety. In J. Strachey
(Ed. and Trans.), The standard edition of the complete psychological
works of Sigmund Freud. London: Hogarth Press. (Original work
published 1926)
Freud, S. (1957a). Mourning and melancholia. In J. Strachey (Ed. and
Trans.), The standard edition of the completepsychological works of
Sigmund Freud (Vol. 14, pp. 237-260). London: Hogarth Press.
(Original work published 1917)
Freud, S. (1957b). On narcissism. In J. Strachey (Ed. and Trans.), The
standard edition of the complete psychological works of Sigmund
Freud (Vol. 14, pp. 67-102). London: Hogarth Press. (Original work
published 1914)
Freud, S. (1957c). The unconscious. In J. Strachey (Ed. and Trans.),
The standard edition of the completepsychological works of Sigmund
Freud (Vol. 14, pp. 159-215). London: Hogarth Press. (Original
work published 1915)
Friedman, S. M. (1952). An empirical study of the castration and Oedipus complexes. Genetic Psychology Monographs, 46, 61-130.
Frith, C., & Dolan, R. (1996). The role of the prefrontal cortex in
higher cognitive functiOns. Cognitive Brain Research, 5, 175-181.
Furman, W., & Buhrmester, D. (1992). Age and sex differences in
perceptions of networks of personal relationships. Child Development,
63, 103-115.
Gabbard, G. (1994). Psychodynamic psychiatry in clinical practice:
The DSM-IV edition. Washington, DC: banerican Psychiatric Press.
Garcia, J., & Rusiniak, K.W. (1980). What the nose learns from the
mouth. In D. Muller-Schwarz & R. M. Silverstein (Eds.), Chemical
signals. New York: Plenum.
Gazzaniga, M. S. (1985). The social brain: Discovering the networks
of the mind. New York: Basic Books, Inc.
Gest, S.D. (1997). Behavioral inhibition: Stability and associations
with adaptation from childhood to early adulthood. Journal of Personality and Social Psychology, 72, 467-475.
Gilbert, D. T. ( 1991 ). How mental systems believe. American Psychologist, 46, 107-119.
Glassman, N., & Andersen, S. (1997). Activating transference without
consciousness: Using significant-other representations to go beyond
the subliminally given information. Unpublished manuscript, New
York University.
Glick, P., & Fiske, S. (1996). The ambivalent sexism inventory: Differentiating hostile and benevolent sexism. Journal of Personality and
Social Psychology, 70, 491-512.
Goschke, T., & Kuhl, J. ( 1993 ). Representations of intentions: Persisting
activation in memory. Journal of Experimental Psychology: Learning,
Memory, and Cognition, 19, 1211-1226.

365

Gotlib, I., Whiffen, V., Wallace, P. M., & Mount, J. H. ( 1991 ). A prospective investigation of postpartum depression: Factors involved in
onset and recovery. Journal of Abnormal Psychology, 100, 122-132.
Gray, J. A. (1990). Brain systems that mediate both emotion and cognition. Cognition and Emotion, 4, 269-288.
Greenberg, J. R., & Mitchell, S. (1983). Object relations in psychoanalytic theory. Cambridge, MA: Harvard University Press.
Greenwald, A. (1992). New look 3: Unconscious cognition reclaimed.
American Psychologist, 47, 766-779.
Greenwald, A. G., & Banaji, M. R. (1995). Implicit social cognition:
Attitudes, self-esteem, and~ stereotypes. Psychological Review, 102,
4-27.
Greenwald, A., & Pratkanis, A. (1984). The self. In R. Wyer & T.
Srull (Eds.), Handbook of social cognition (Vol. 3, pp. 129-178).
Hillsdale, NJ: Erlbaum.
Gross, J. J., & Levenson, R. (1993). Emotional suppression:Physiology,
self-report, and expressive behavior. Journal of Personality and Social
Psychology, 64, 970-986.
Haan, N. (1977). Coping and defending: Processes of self-environment
organization. New York: Academic Press.
Halligan, P. W., & Marshall, J. C. (1991). Left neglect for near but not
far space in man. Nature, 350, 498-500.
Harris, T., Brown, G., & Bifulco, A. (1986). Loss of parent in childhood
and adult psychiatric disorder: The role of lack of adequate parental
care. Psychological Medicine, 16, 644-659.
Harrist, A., Zaia, A., Bates, J.E., Dodge, K., & Pettit, G.S. (1997).
Subtypes of social withdrawal in early childhood: Sociometric status
and social-cognitive differences across four years. ChiMDevelopment,
68, 278-294.
Harter, S. (1986). Cognitive-developmental processes in the integration
of concepts about emotions and the self. Social Cognition, 4, 119151.
Harter, S. (1998). The development of self-representations. In W. Damon
(Ed.) & N. Eisenberg (Vol. Ed.), Handbook of child psychology: Vol.
3. Social emotional and personality development (pp. 553-617).
New York: Wiley.
Hartup, W. (1996). The company they keep: Friendships and their developmental significance. Child Development, 67, 1-13.
Hauser, S., Borman, E., Jacobson, A.M., Powers, S., & Noam, G.
( 1991 ). Understanding family contexts of adolescent coping: A study
of parental ego development and adolescent coping strategies. Journal
of Early Adolescence, 11, 96-124.
Hedegard, S. (1969). A molecular analysis of psychological defenses.
Unpublished doctoral dissertation, University of Michigan.
Heinemann, L., & Emrich, H. (1971). Alpha activity during inhibitory
brain processes. Psychophysiology, 7, 442-450.
Helmholtz, H. v o n ( 1971 ). Selected writings of Hermann von Helmholtz. R. Kahl (Ed. and Trans.). Middletown, C'12.Wesleyan University
Press. (Original work published 1863 )
Herman, J., Perry, J.C., & van der Kolk, B.A. (1989). Childhood
trauma in borderline personality disorder. American Journal of Psychiatry, 146, 490-495.
Higgins, E. T. (1990). Personality, social psychology, and person-situation relations: Standards and knowledge activation as a common language. In L. Pervin (Ed.), Handbook of personality: Theory and
research (pp. 301-338). New York: Guilford Press.
Hilgard, E. (1979). Divided consciousness in hypnosis: The implications of the hidden observer. In E. Fromm & R. E. Shor (Eds.), Hypnosis: Developments in research and new perspectives (2nd ed.). Chicago: Aldine.
Hirshberg, L. M. (1990). When infants took to their parents: II. Twelvemonth-olds' response to conflicting parental emotional signals. Child
Development, 61, 1187-1191.
Hogan, R., Curphy, G., and Hogan, J. (1994). What we know about

366

WESTEN

leadership: Effectiveness and personality. American Psychologist, 49,

493-504.
Holyoak, K., & Spellman, B. (1993). Thinking. Annual Review of Psychology, 44, 265-315.
Holyoak, K. J., & Thagard, P. (1989). Analogical mapping by constraint
satisfaction. Cognitive Science, 13, 295-355.
Holyoak, K., & Thagard, P. (1997). The analogical mind. American
Psychologist, 52, 35-44.
Horowitz, M.J. (1987). States of mind: Configurational analysis of
individual psychology (2rid ed.). New York: Plenum.
Horowitz, M. (1988). Introduction to psychodynamics: A synthesis.
New York: Basic Books.
House, J. S., Landis, K. R., & Umberson, D. (1988, July). Social relationships and health. Science, 241, 540-545.
Hughes, C., Uhlmann, C., & Pennebaker, J. (1994). The body's response
to processing emotional trauma: Linking verbal text with autonomic
activity. Journal of Personality, 62, 565-585.
Ihilevich, D., & Gleser, G. (1986). Defense mechanisms: Their classification, correlates, and masurement with the Defense Mechanisms
Inventory. Owosso, MI: DMI Associates.
Jacoby, L., & Kelly, C. M. (1992). A process-dissociation framework
for investigating unconscious influences: Freudian slips, projective
tests, subliminal perception, and signal detection theory. Current Directions in Psychological Science, 1, 174-179.
Jensen, M. R. (1987). Psychobiological factors predicting the course of
breast cancer. Journal of Personality, 55, 317-342.
Jockin, V., McGue, M., & Lykken, D. (1996). Personality and divorce:
A genetic analysis. Journal of Personality and Social Psychology,
71, 288-299.
John, O., & Robins, R. W. (1994). Accuracy and bias in self-perception:
Individual differences in self-enhancement and the role of narcissism.
Journal of Personality and Social Psychology, 66, 206-219.
Johnson, M. K., Kim, J. K., & Risse, G. (1985). Do alcoholic Korsakoff's syndrome patients acquire affective reactions? Journal of Experimental Psychology: Learning, Memory, & Cognition, 11, 22-36.
Jorgensen, R.S., Johnson, B.T., Kolodziej, M.E., & Schreer, G.E.
(1996). Elevated blood pressure and personality: A meta-analytic
review. Psychological Bulletin, 120, 293-320.
Juni, S., Masling, J., & Brannon, R. (1979). Interpersonal touching and
orality. Journal of Personality Assessment, 43, 235-237.
Kagan, J. (1980). Perspectives on continuity. In O. Brim, Jr., & J. Kagan
(Eds.), Constancy and change in human development (pp. 26-74).
Cambridge, MA: Harvard University Press.
Kagan, J., & Snidman, N. (1991). Temperamental factors in human
development. American Psychologist, 46, 856-862.
Kagan, J., & Zentner, M. (1996). Early childhood predictors of adult
psychopathology. Harvard Review of Psychiatry, 3, 341-350.
Kaplan, B. L., & Gotlib, I. H. (1996). Predicting change in depressive
symptoms from parenting quality, personality style, and life events:
An integrative approach. Unpublished manuscript, Northwestern
University.
Katz, I. (1981). Stigma: A socialpsychological analysis. Hillsdale, NJ:
Erlbaum.
Katz, I., & Hass, R. G. (1988). Racial ambivalence and American value
conflict: Correlational and priming studies of dual cognitive structures. Journal of Personality and Social Psychology, 55, 893-905.
Kelly, G.A. (1955). Psychology of personal constructs. New York:
Norton.
Kendler, K. S. (1995). Adversity, stress and psychopathology: A psychiatric genetic perspective. International Journal of Method in Psychiatric Research, 5, 163-170.
Kendler, K. S., Neale, M. C., Kessler, R. C., Heath, A. C., & Eaves, L.
(1992). A population based twin study of major depression in women:

The impact of varying definitions of illness. Archives of General

Psychiatry, 49, 257-266.
Kendler, K. S., Neale, M. C., Prescott, C. A., Kesslea; R. C., Heath, A. C.,
Corey, L.A., & Eaves, L.J. (1996). Childhood parental loss and
alcoholism in women: A causal analysis using a twin-family design.
Psychological Medicine, 26, 79-95.
Kernberg, O. (1975). Borderline conditions and pathological narcissism. New York: Jason Aronson.
Kessler, R. C., Davis, C. G., & Kendler, K. S. (1997). Childhood adversity and adult psychiatric disorder in the U.S. National Comorbidity
Survey. Psychological Medicine, 27, 1101 - 1119.
Kihlstrom, J. (1987, September). The cognitive unconscious. Science,
237, 1445-1452.
Kihlstrom, J. (1990). The psychological unconscious. In L. Pervin
(Ed.), Handbook of personality: Theory and research (pp. 445-464).
New York: Guilford Press.
Kline, P. (1981). Fact and fantasy in Freudian theory (2rid ed.). London: Methuen.
Kohlberg, L., & Kramer, R. (1969). Continuities and discontinuities in
childhood and adult moral development. Human Development, 12,
93-120.
Kohut, H. (1977). The restoration of the self. Madison, C'12.International
Universities Press.
Kolb, J., Cooper, S., & Fishman, G. (1995). Recent developments in
psychoanalytic technique: A review. Harvard Review of Psychiatry,
3, 65-74.
Kunda, Z. (1990). The case for motivated reasoning. Psychological
Bulletin, 108, 480-498.
Kunda, Z., & Thagard, P. (1996). Forming impressions from stereotypes, traits, and behaviors: A parallel-constraint-satisfaction theory.
Psychological Review, 103, 284-308.
Lamb, M. E. (1987). Predictive implications of individual differences
in attachment. Journal of Consulting and Clinical Psychology, 55,
817-824.
Lazarus, R. S., & McCleary, R. A. (1951 ). Autonomic discrimination
without awareness: A study of subeeption. Psychological Review, 58,
113-122.
LeDoux, J.E. (1989). Cognitive-emotional interactions in the brain.
Cognition and Emotion, 3, 267-289.
LeDoux, J. (1995). Emotion: Clues from the brain. Annual Review of
Psychology, 46, 209-235.
LeDoux, J. (1997). Emotion, memory, and pain. Pain Forum, 6, 3637.
Lefkowitz, M., & Tesiny, E. (1984). Rejection and depression: Prospective and contemporaneous analyses. Developmental Psychology, 20,
776-785.
Leventhal, H., & Everhart, D. (1979). Emotion, pain, and physical
illness. In C. E. Izard (Ed.), Emotion in personality andpsychopathology (pp. 263-299). New York: Plenum.
LeVine, R. (1982). Culture, behavior, and personality (2nd ed.). Chicago: Aldine.
Levy, K. N., Blatt, S.J., & Shaver, P. (1998). Attachment styles and
parental representations. Journal of Personality and Social Psychology, 74, 407-419.
Lewicki, P. (1986). Nonconscious social information processing. New
York: Academic Press.
Livesley, W. J., & Bromley, D. B. (1973). Person perception in childhood and adolescence. New York: Wiley.
Loehlin, J. (1992). Genes and environment in personality development.
New York: Guilford Press.
Loevinger, J. (1976). Ego development: Conceptions and theories. San
Francisco: Jossey-Bass.
Loevinger, J. (1985). Revision of the sentence completion test for ego

FREUD'S LEGACY
development. Journal of Personality and Social Psychology, 48, 420427.
Luborsky, L., & Crits-Christoph, P. (1990). Understanding transference: The core conflictual relationship theme method. New York:
Basic Books.
Lyons-Ruth, K., Connell, D., Grunebaum, H., & Botein, S. (1990).
Infants at social risk: Maternal depression and family support services
as mediators of infant development and security of attachment. ChiM
Development, 61, 85-98.
Lytton, H. (1990). Child and parent effects in boys' conduct disorder:
A reinterpretation. Developmental Psychology, 26, 683-697.
Main, M. (1990). Cross-cultural studies of attachment organization:
Recent studies, changing methodologies, and the concept of conditional strategies. Human Development, 33, 48-61.
Main, M. (1996). Disorganization and disorientation in infant strange
situation behavior: Phenotypic resemblance to dissociative states. In
L. K. Michelson & W. J. Ray (Eds.), Handbook of dissociation (pp.
107-138). New York: Plenum Press.
Main, M., Kaplan, N., & Cassidy, J. (1985). Security in infancy, childhood, and adulthood: A move to the level of representation. In I.
Bretherton & E. Waters (Eds.), Growing points of attachment theory
and research. Monographs of the Society for Research in Child Development, 50(1-2), 67-104.
Main, M., & Solomon, J. (1986). Discovery of a new, insecure-disorganized/disoriented attachment pattern. In T. Brazelton & M. Yogman
(Eds.), Affective development in infancy (pp. 95-124). Norwood,
NJ: Ablex.
Mandler, G. (1997). Human nature explored. New York: Oxford University Press.
Marcia, J.E. (1994). The empirical study of ego identity. In H.A.
Bosma, T. L. G. Graafsma, H. D. Grotevant, & D. J. de Levita (Eds.),
Identity and development: An interdisciplinary approach (pp. 6780). Thousand Oaks, CA: Sage.
Martikainen, P., & Valkonen, T. (1996). Mortality after the death of a
spouse: Rates and causes of death in a large Finnish cohort. American
Journal of Public Health, 86, 1087-1093.
Masling, J. (1986). Orality, pathology, and interpersonal behavior. In J.
Masling (Ed.), Empirical studies of psychoanalytic theories (Vol. 2,
pp. 73-106). New York: Analytic Press.
Masling, J., Price, J., Goldband, S., & Katkin, E. S. (1981). Oral imagery and autonomic arousal in social isolation. Journal of Personality
and Social Psychology, 40, 395-400.
Masten, A. S., Best, K., & Garmezy, N. (1990). Resilience and development: Contributions from the study of children who overcome adversity. Development & Psychopathology, 2, 425-444.
McAdams, D.P. (1992). The intimacy motive. In C.P. Smith, J.W.
Atkinson, D. C. McClelland, & J. Veroff (Eds.), Motivation andpersonality: Handbook of thematic content analysis (pp. 224-228).
Cambridge, England: Cambridge University Press.
McClelland, D.C., & Franz, C.E. (1992). Motivational and other
sources of work accomplishments in midlife: A longitudinal study.
Journal of Personality, 60, 679-707.
McC!elland, D. C., Koestner, R., & Weinberger, J. (1989). How do selfattributed and implicit motives differ? Psychological Review, 96, 690702.
McCrae, R., & Costa, P. (1990). Personality in adulthood. New York:
Guilford Press.
McConaghy, N. (1976). Is a homosexual orientation reversible? British
Journal of Psychiatry, 129, 556-563.
McCord, J. (1988). Identifying developmental paradigms leading to
alcoholism. Journal of Studies on Alcohol, 49, 357-362.
MeGinnies, E. (1949). Emotionality and perceptual defense. Psychological Review, 56, 244-251.
Mickelson, K. D., Kessler, R. C., & Shaver, P. R. (1997). Adult attach-

367

ment in a nationally representative sample. Journal of Personality

and Social Psychology, 73, 1092-1106.
Mikulincer, M., & Florian, V. (in press). Maternal-fetal bonding, coping
strategies, and mental health during pregnancy: The contribution of
attachment style. Journal of Social and Clinical Psychology.
Mikulincer, M., Florian, V., & Wellel; A. (1993). Attachment styles,
coping strategies, and posttraumatic psychological distress: The impact of the Gulf War in Israel. Journal of Personality and Social
Psychology, 64, 817-826.
Mikulincer, M., & orbach, I. (1995). Attachment styles and repressive
defensiveness: The accessibility and architecture of affective memories. Journal of Personality and Social Psychology, 68, 917-925.
Miller, J. B. (1994). Women's psychological development: Connections,
disconnections, and violations. In M. M. Better (Ed.), Women beyond
Freud: New concepts offeminine psychology (pp. 79- 97 ). New York:
Brunner/Mazel.
Milner, B., Corkin, S., & Teubez, H. L. (1968). Further analysis of the
hippocampal amnesic syndrome: Fourteen year follow-up study of
H. M. Neuropsychologia, 6, 215-234.
Mineka, S., & Sutton, S. K. (1992). Cognitive biases and the emotional
disorders. Psychological Science, 3, 65-69.
Mischel, W. (1968). Personality and assessment. New York: Wiley.
Mischel, W. (1973). Toward a cognitive social learning reconceptualization of personality. Psychological Review, 80, 252-283.
Mischel, W. (1979). On the interface of cognition and personality:
Beyond the person-situation debate. American Psychologist, 34,
740-754.
Mischel, W., & Shoda, Y. (1995). A cognitive-affective system theory
of personality: Reconceptualizing situations, dispositions, dynamics,
and invariance in personality structure. Psychological Review, 102,
246-268.
Mitchell, S. (1988). Relational concepts in psychoanalysis. Cambridge,
MA: Harvard University Press.
Mongrain, M., Vettese, L. C., Shustex, B., & Kendal, N. (1998). Perceptual biases, affect, and behavior in the relationships of dependents and
self-critics. Journal of Personality and Social Psychology, 75, 230241.
Morling, B., & Epstein, S. (1997). Compromises produced by the dialectic between self-verification and self-enhancement. Journal of Personality and Social Psychology, 73, 1268-1283.
Morokoff, P. J. (1985). Effects of sex guilt, repression, sexual "arousability," and sexual experience on female sexual arousal during erotica
and fantasy. Journal of Personality and Social Psychology, 49, 177187.
Murphy, S. T., Monahan, J. L., & Zajonc, R. (1995). Additivity of nonconscious affect: Combined effects of priming and exposure. Journal
of Personality and Social Psychology, 69, 589-602.
Murphy, S. T., & Zajonc, R. (1993). Affect, cognition, and awareness:
Affective priming with optimal and suboptimal stimulus exposures.
Journal of Personality and Social Psychology, 64, 723-739.
Neisser, U. ( 1991 ). A case of misplaced nostalgia. American Psychologist, 46, 34-36.
Newman, D. L., Caspi, A., Moffitt, T., & Silva, P. (1997). Antecedents
of adult interpersonal functioning: Effects of individual differences
in age 3 temperament. Child Development, 33, 206-217.
Newman, L. S., Duff, K., & Baumeister, R. (1997). A new look at
defensive projection: Thought suppression, accessibility, and biased
person perception. Journal of Personality and Social Psychology, 72,
980-1001.
Niedenthal, P., & Kitayama, S. (1994). The heart's eye: Emotional
influences in perception and attention. San Diego, CA: Academic
Press.
Nietzsche, E (1956). The birth of tragedy and the geneaology of morals.
Garden City, NY: Doubleday.

368

WESTEN

Nisbett, R., & Wilson, T. (1977). Telling more than we can know: Verbal
reports on mental processes. Psychological Review, 84, 231-259.
Ohrnan, A. (1994). "Unconscious anxiety": Phobic responses to
masked stimuli. Journal of Abnormal Psychology, 103, 231-240.
O'Neill, R. M., Greenberg, R. P., & Fisher, S. (1992). Humor and anality.
Humor, 5, 283-291.
Oppenheim, D., Emde, R., & Warren, S. (1997). Children's narrative
representations of mothers: Their development and associations with
child and mother adaptation. ChiM Development, 68, 127-138.
Orlofsky, J. (1993). Intimacy status: Theory and research. In J. E. Marcia, A. S. Waterman, D. R. Matteson, S. L. Archer, & J. L. Orlofsky
(Eds.), Ego identity: A handbook for psychosocial research. New
York: Springer-Verlag.
Palumbo, R., & Gillman, I. (1984). Effects of subliminal activation of
Oedipal fantasies on competitive performance. Journal of Nervous
and Mental Disease, 172, 737-741.
Parker, G. (1983). "Affectionless control" as an antecedent to adult
depression. Archives of General Psychiatry, 40, 956-960.
Paulhus, D., Fridhandler, B., & Hayes, S. (1997). Psychological defense:
Contemporary theory and research. In R. Hogan, J. Johnson, & S. R.
Briggs (Eds.), Handbook of personality psychology (pp. 543-579).
San Diego, CA: Academic Press.
Paulhus, D., & Levitt, K. (1987). Desirable responding triggered by
affect: Automatic egotism? Journal of Personality and Social Psychology, 52, 245-259.
Pederson, D. R., Moran, G., Sitko, C., Campbell, K., Ghesquire, K., &
Acton, H. (1990). Maternal sensitivity and the security of infantmother attachment: A Q-sort study. ChiM Development, 61, 19741983.
Pennebaker, J. (1989). Stream of consciousness and stress: Levels of
thinking. In J. S. Uleman & J. A. Bargh (Eds.), Unintended thought
(pp. 327-350). New York: Guilford Press.
Pennebaker, J. (1997). Opening up: The healing power of expressing
emotions (Rev. ed.). New York: Guilford Press.
Perry, J. C., & Cooper, S. H. (1987). Empirical studies of psychological
defense mechanisms. In R. Michels & J. O. Cavenar, Jr. (Eds.), Psychiatry. Philadelphia: J. B. Lippincott.
Perry, J. C., & Cooper, S. (1989). An empirical study of defense mechanisms: I. Clinical interview and life vignette ratingg. Archives of General Psychiatry, 46, 444-460.
Peselow, E., Robins, C., Sanfilipo, M., Block, E, & Fieve, R. (1992).
Sociotropy and autonomy: Relationship to antidepressant treatments
response and the endogeonous-nonendogenous dichotomy. Journal
of Abnormal Psychology, 101, 479-486.
Peterson, C., & Ulrey, L. (1994). Can explanatory style be scored from
TAT protocols? Personality and Social Psychology Bulletin, 20, 102106.
Pettigrew, T., & Meertens, R. (1995). Subtle and blatant prejudice in
Western Europe. European Journal of Social Psychology, 25, 57-75.
Pettit, G., Bates, J., & Dodge, K. (1997). Supportive parenting, ecological context, and children's adjustment: A seven-year longitudinal
study. Child Development, 68, 908-923.
Petty, R. E., & Cacioppo, J. T. (1986). The elaboration likelihood model
of persuasion. In L. Berkowitz (Ed.), Advances in experimental social
psychology (Vol. 19, pp. 123-205). San Diego, CA: Academic Press.
Pierce, G.R., Lakey, B., Sarasori, I., & Sarason, B. (Eds.). (1997).
Sourcebook of social support and personality. New York: Plenum
Press.
Plomin, R., DeFries, J. C., McClearn, G. E., & Rutter, R. (1997). Behavioral genetics (3rd ed.). New York: Freeman.
Pope, K., Tabachnick, B., & Keith-Spiegel, P. (1987). Ethics of practice:
The beliefs and behaviors of psychologists as therapists. American
Psychologist, 42, 993-1006.

Popper, K. (1957). The poverty of historicism. London: Routledge &

Kegan Paul.
Priester, J.R., & Petty, R.E. (1996). The gradual negative threshold
model of ambivalence: Relating positive and negative bases of attitudes
to subjective ambivalence. Journal of Personality and Social Psychology, 71, 431-449.
Pyszczynski, T., Greenberg, J., & Solomon, S. (1998). A terror management perspective on the psychology of control: Controlling the uncontrollable. In M. Kofta & G. Weary (Eds.), Personal control in action:
Cognitive and motivational mechanisms (pp. 85-108). New York:
Plenum Press.
Rachman, S. J. (1978). Fear and courage. San Francisco: Freeman.
Raine, A., Brennan, P., & Mednick, S. (1994). Birth complications
combined with early maternal rejection at age 1 year predispose to
violent crime at age 18 years. Archives of General Psychiatry, 51,
984-988.
Read, S. J., Vanman, E. J., & Miller, L. C. (1997). Connectionism, parallel constraint satisfaction processes, and Gestalt principles: (Re)introducing cognitive dynamics to social psychology. Personality and Social Psychology Review, 1, 26-53.
Rebel; A. (1992). The cognitive unconscious: An evolutionary perspecfive. Consciousness and Cognition, I, 93-133.
Reich, A. (1960). Pathologic forms of self-esteem regulation. Psychoanalytic Study of the Child, 15, 215-232.
Reisberg, D., & Heuer, E (t995). Emotion's multiple effects on memory.
In J.L. McGangh, N. Weinberger, & G. Lynch (Eds.), Brain and
memory: Modulation and mediation of neuroplasticity (pp. 84-92).
New York: Oxford University Press.
Rest, J. R. (1983). Morality. In J. H. Flavell & E. M. Markman (Eds.),
Handbook of chiM psychology: Vol. 3. Cognitive development. New
York: Wiley.
Ricks, M. (1985). The social transmission of parental behavior: Attachment across generations. In I. Bretherton & E. Waters (Eds.), Growing
points of attachment theory and research. Monographs of the Society
for Research in Child Development, 50( 1-2, Serial No. 209), 211227.
Roberts, B. W., & Chapman, C. N. ( 1997, August). A 30-year longitudinal study linking role quality with changes in dispositional well-being.
Paper presented at the 105th Annual Convention of the American
Psychological Association, Chicago.
Robins, R. W., & Beer, J. S. (1996). A longitudinal study of the adaptive
and maladaptive consequences of positive illusions about the self.
Unpublished manuscript, University of California, Berkeley.
Robins, R.W., & Craik, K. (1994). A more appropriate test of the
Kuhnian displacement thesis. American Psychologist, 49, 815-816.
Roediger, H. L. (1990). Implicit memory: Retention without remembering. American Psychologist, 45, 1043-1056.
Rogers, J. H., Widiger, T., & Krupp, A. (1995). Aspects of depression
associated with borderline personality disorder. American Journal of
Psychiatry, 152, 268-270.
Roston, D., Lee, K., & Vaillant, G. (1992). A Q-sort approach to identifying defenses. In G. Valllant (Ed.), Ego mechanisms of defense: A
guide for clinicians and researchers (pp.. 217-233). Washington,
DC: American Psychiatric Association.
Rubin, D., Wallace, W., & Houston, B. (1993). The beginnings of
expertise for ballads. Cognitive Science, 17, 435-462.
Rumelhart, D. E., McClelland, J. L., & PDP Research Group (1986).
Parallel distributed processing: Explorations in the microstructure
of cognition: Vol. 1. Foundations. Cambridge, MA: MIT Press.
Russek, L. G., & Schwartz, G. E. (1997). Perceptions of parental caring
predict health status in midlife: A 35-year follow-up of the Harvard
Mastery of Stress Study. Psychosomatic Medicine, 59, 144-149.
Rutter, M. (1986). Meyerian psychobiology, personality development

FREUD'S LEGACY
and the role of life experiences. American Journal of Psychiatry, 143,
1077-1087.
Sandier, J. (1987). Toward a reconsideration of the psychoanalytic theory of motivation. In J. Sandier (Ed.), From safety to the superego:
Selected papers of Joseph Sandier. New York: Guilford Press.
Sandier, J., & Rosenblatt, B. (1962). The concept of the representational
world. Psychoanalytic Study of the Child, 1Z 128-145.
Scarr, S. (1993). Biological and cultural diversity: The legacy of Darwin
for development. Child Development, 64, 1333-1353.
Schacter, D.L. (1992). Understanding implicit memory: A cognitive
neuroscience approach. American Psychologist, 47, 559-569.
Schacter, D. (1995). Implicit memory: A new frontier for cognitive
neuroscience. In M. Gazzaniga (Ed.), The cognitive neurosciences
(pp. 815-824). Cambridge, MA: MIT Press.
Sehacter, D., & Buckner, R. L. (1998). Priming and the brain. Neuron,
20, 185-195.
Schneider, W., & Shiffrin, R. (1977). Controlled and automatic human
information processing: I. Detection, search, and attention. Psychological Review, 84, 402-411.
Schopenhauer, A. (1958). The world as will and representation (E. E J.
Payne, Trans.). Indian Hills, CO: Falcon's Wing Press. (Original work
published 1883)
Schumm, W., Paff-Bergen, L., Hatch, R., & Obiorah, E (1986). Concurrent and discriminant validity of the Kansas Marital Satisfaction Scale.
Journal of Marriage & the Family, 48, 381-387.
Schwartz, G.E. (1990). Psychobiology of repression and health: A
systems perspective. In J. L. Singer (Ed.), Repression and dissociation: Defense mechanisms and personality styles: Current theory and
research. Chicago: University of Chicago Press.
Schwartz, G., Krupp, N., & Byrne, D. (1971). Repression-sensitization
and medical diagnosis. Journal of Abnormal Psychology, 78, 286291.
Sears, R., Maccoby, E., & Levin, H. (1957). Patterns of child rearing.
Evanston, IL: Row, Peterson.
Seger, C. A. (1994). Implicit learning. Psychological Bulletin, 115, 163196.
Selman, R. (1980). The growth of interpersonal understanding. San
Diego, CA: Academic Press.
Shea, M., Glass, D., Pilkonis, P., Watkins, J., & Docherty, J. (1987).
Frequency and implications of personality disorders in a sample of
depressed outpatients. Journal of Personality Disorders, 1, 27-42.
Shedier, J., & Block, J. (1990). Adolescent drug use and psychological
health: A longitudinal inquiry. American Psychologist, 45, 612-630.
Shedler, J., Mayman, M., & Manis, M. (1993). The illusion of mental
health. American Psychologist, 48, 1117-1131.
Shedler, J., & Westen, D. (1998). Refining the measurement of Axis II:
A Q-sort procedure for accessing personality pathology. Assessment,
5, 335-355.
Shevrin, H., Bond, J., Brakel, L., Hertel, R., & Williams, W. J. (1996).
Conscious and unconscious processes: Psychodynamic, cognitive,
and neurophysiological convergences. New York: Guilford Press.
Shevrin, H., & Dickman, S. (1980). The psychological unconscious: A
necessary assumption for all psychological theory? American Psychologist, 35, 421-434.
Shoda, Y., Mischel, W., & Peake, P. (1990). Predicting adolescent cognitive and self-regulatory competencies from preschool delay of gratification: Identifying diagnostic conditions. Developmental Psychology,
26, 978-986.
Shultz, T. R., & Lepper, M. R. (1996). Cognitive dissonance reduction
as constraint satisfaction. Psychological Review, 103, 219-240.
Silverman, L., & Weinberger, J. (1985). Mommy and I are one: Implications for psychotherapy. American Psychologist, 40, 1296-1308.
Simon, L., Greenberg, J., Harmon-Jones, E., Solomon, S., Pyszczynski,
T., Arndt, J., & Abend, T. (1997). Terror management and cognitive-

369

experiential self-theory: Evidence that terror management occurs in

the experiential system. Journal of Personality and Social Psychology, 72, 1132-1146.
Simpson, J.A., Ickes, W., & Blackstone, T. (1995). When the head
protects the heart: Empathic accuracy in dating relationships. Journal
of Personality and Social Psychology, 69, 629-641.
Simpson, J. A., Rholes, W. S., & Nelligan, J. S. (1992). Support seeking
and support giving within couples in an anxiety-provoking situation:
The role of attachment styles. Journal of Personality and Social Psychology, 62, 434-446.
Sincoff, J. (1992). Ambivalence and defense: Effects of a repressive
style of normal adolescents' and young adults' mixed feelings. Journal of Abnormal Psychology, 101, 251-256.
Singer, J. (1985). Transference and the human condition: A cognitiveaffective perspective. Psychoanalytic Psychology, 2, 189-219.
Skinner, B.E (1953). Science and human behavior. New York:
Macmillan.
Smith, E.R. (1996). What do connectionism and social psychology
offer each other? Journal of Personality and Social Psychology, 70,
893-912.
Smith, G. J. W., & Westerlundh, B. (1980). Perceptogenesis: A process
perspective on perception-personality. In L. Wheeler (Ed.), Review of
personality and social psychology (pp. 94-124). London: Sage.
Spellman, B. A., & Holyoak, K. (1992). If Saddam is Hider then who
is George Bush? Analogical mapping between systems of social roles.
Journal of Personality and Social Psychology, 62, 913-933.
Spellman, B. A., & Holyoak, K. (1996). Pragmatics in analogical mapping. Cognitive Psychology, 31, 307-346.
Spencer, S. M., & Norem, J. (1996). Reflection and distraction: Defensive pessimism, strategic optimism, and performance. Personality &
Social Psychology Bulletin, 22, 354-365.
Spiegel, D., Bloom, J. H., Kraemer, H. C., & Gottheil, E. (1989). Effect
of psychosocial treatment on survival of patients with metastatic breast
cancer. Lancet, 888-891.
Squire, L. (1987). Memory and brain. New York: Oxford University
Press.
Squire, L., & Zola-Morgan, S. ( 1991 ). The medial temporal lobe memory system. Science, 232, 1612-1619.
Sroufe, L. A., & Fleeson, W. (1986). Attachment and the construction
of relationships. In W. W. Hartup & Z. Rubin (Eds.), Relationships
and development. HiUsdale, NJ: Erlbaum.
Stacy, A. W. (1997). Memory activation and expectancy as prospective
predictors of alcohol and marijuana use. Journal of Abnormal Psychology, 106, 61-73.
Stacy, A.W., Leigh, B.C., & Weingardt, K. (1997). An individualdifference perspective applied to word association. Personality & Social Psychology Bulletin, 23, 229-237.
Stattin, H., & Magnusson, D. (1989). The role of early aggressive
behavior in the frequency, seriousness, and types of later crime. Journal of Consulting and Clinical Psychology, 57, 710-718.
Steele, C. (1997). A threat in the air: How stereotypes shape intellectual
identity and performance. American Psychologist, 52, 613-629.
Steele, C., & Aronson, J. (1995). Stereotype threat and the intellectual
test performance of African Americans. Journal of Personality and
Social Psychology, 69, 797-811.
Steele, H., Steele, M., & Fonagy, P. (1996). Associations among attachment classifications of mothers, fathers, and their infants. Child Development, 67, 541-555.
Stein, N. (Ed.). (1997). Memory for everyday and emotional events.
Mahwah, NJ: Erlbanm.
Strauman, T. (1992). Self-guides, autobiographical memory, and anxiety
and dysphoria: Toward a cognitive model of vulnerability to emotional
distress. Journal of Abnormal Psychology, 101, 87-95.
Strauman, T., Lemieux, A., & Coe, C. (1993). Self-discrepancy and

370

WESTEN

natural killer cell activity: Immunological consequences of negative

self-evaluation. Journal of Personality and Social Psychology, 64,
1042 - 1052.
Stricker, G., & Healey, B. J. (1990). Projective assessment of object
relations: A review of the empirical literature. Psychological Assessment, 2, 219-230.
Sullivan, H.S. (1953). The interpersonal theory of psychiatry. New
York: Norton.
Swann, W. (1990). To be adored or to be known: The interplay of
self-enhancement and self-verification. In R.M. Sorrentino & E. T.
I-Iiggins (Eds.), Handbook of motivation and cognition (Vol. 2, pp.
408-448). New York: Guilford Press.
Swann, W., Stein-Seroussi, A., & Geisler, R. (1992). Why people selfverify. Journal of Personality and Social Psychology, 62, 392-401.
Taylor, S. E., & Brown, J. D. (1988). Illusion and well-being: A social
psychological perspective on mental health. Psychological Bulletin,
103, 193-210.
Tennant, C. (1988). Parental loss in childhood: Its effect in adult life.
Archives of General Psychiatry, 45, 1045-1050.
Thompson, M., Zanna, M., & Griffin, D. W. (1995). Let's not be indifferent about (attitudinal) ambivalence. In R. E. Petty & J. A. Krosnick
(Eds.), Attitude strength: Antecedents and consequences (pp. 361386). Mahwah, NJ: Erlbaum.
Tizard, B., & Hodges, J. ( t 978). The effects of early institutional rearing
on the development of eight-year-old children. Journal of Child Psychology and Psychotherapy, 19, 99-108.
Torgersen, S. (1980). The oral, obsessive, and hysterical personality
syndromes. Archives of General Psychiatry, 37, 1272-1277.
Tremblay, R., Pihl, R., Vitaro, F., & Dobkin, P. (1994)! Predicting early
onset of male antisocial behavior from preschool behavior. Archives
of General Psychiatry, 51, 732-739.
Turiel, E. (1998). The development of morality. In W. Damon (Ed.) &
N. Eisenberg (Vol. Ed.), Handbook of child psychology: Vol. 3. Social
emotional and personality development (pp. 863-932). New York:
Wiley.
Underwood, G. (Ed.). (1996). Implicit cognition. New York: Oxford
University Press.
Vaillant, G. (t977). Adaptation to life. Boston: Little, Brown.
Vaillant, G. (Ed.). (1992). Ego mechanisms of defense: A guide for
clinicians and researchers. Washington, DC: American Psychiatric
Association.
Vaillant, G., & Drake, R. (1985). Maturity of defenses in relation to
DSM-III Axis II personality disorders. Archives of General Psychiatry, 42, 597-601.
van IJzendoorn, M. (1995). Adult attachment representations, parental
responsiveness, and infant attachment: A meta-analysis on the predictive validity of the Adult Attachment Interview. Psychological Bulletin, 117, 387-403.
Wachtel, P. (1981). Transference, schema, and assimilation: The relationship of Piaget to the psychoanalytic theory of transference. Annual
of Psychoanalysis, 8, 59-76.
Wachtel, P. (1993). Therapeutic communication. New York: Guilford.
Watson, M. W., & Getz, K. (1990). The relationship between Oedipal
behaviors and children's family role concepts. Merrill-Palmer Quarterly, 36, 487-505.
Wegner, D. (1992). You can't always think what you want: Problems
in the suppression of unwanted thoughts. Advances in Experimental
Social Psychology, 25, 193-225.
Wegner, D., Shortt, J., Blake, A. W., & Page, M. S. (1990). The suppression of exciting thoughts. Journal of Personality and Social Psychology, 58, 409-418.
Weinberg, N.M., Gold, P.E., & Sternberg, D.B. (1984, February).
Epinephrine enables Pavlovian fear conditioning under anesthesia.
Science, 223, 605-607.

Weinberger, D. (1992, August). Not worrying yourself sick: The health

consequences of repressive coping. Paper presented at the 100th Annual Convention of the American Psychological Association, Washington, DC.
Weinberger, D. (1995). The construct validity of the repressive coping
style. In J. Singer (Ed.), Repression and dissociation (pp. 337-386).
Chicago: University of Chicago Press.
Weinberger, J. (in press). The unconscious. New York: Guilford Press.
Weinberger, J., & Hardaway, R. (1990). Separating science from myth
in subliminal psychodynamic activation. Clinical Psychology Review,
10, 727-756.
Weinberger, J., & McClelland, D. (1990). Cognitive versus traditional
motivational models: Irreconcilable or complementary? In E. T. Higgins & R. Sorrentino (Eds.), Handbook of motivation and cognition:
Foundations of social behavior (Vol. 2, pp. 562-597). New York:
Guilford Press.
Westen, D. (1985). Self and society: Narcissism, collectivism, and the
development of morals. New York: Cambridge University Press.
Westen, D. (1988). Transference and information processing. Clinical
Psychology Review, 8, 161-179.
Westen, D. (1989). Are "primitive" object relations really preoedipal?
American Journal of Orthopsychiatry, 59, 331-345.
Westen, D. (1990a). Psychoanalytic approaches to personality. In L.
Pervin (Ed.), Handbook of personality: Theory and research (pp.
21-65). New York: Guilford Press.
Westen, D. (1990b). The relations among narcissism, egocentrism, selfconcept, and self-esteem. Psychoanalysis and Contemporary Thought,
13, 185-241.
Westen, D. (1990c). Toward a revised theory of borderline object relations: Implications of empirical research. International Journal of
Psycho-Analysis, 71, 661-693.
Westen, D. ( 1991 ). Social cognition and object relations. Psychological
Bulletin, 109, 429-455.
Westen, D. (1994). Toward an integrative model of affect regulation:
Applications to social-psychological research. Journal of Personality,
62, 641-647.
Westen, D. (1995). A clinical-empirical model of personality: Life after
the Mischelian ice age and the NEO-lithic era. Journal of Personality,
63, 495-524.
Westen, D. (1996). A model and a method for uncovering the nomothetic
from the idiographic: A comparison with the five factor model. Journal of Research in Personality, 30, 400-413.
Westen, D. (1997). Divergences between clinical and research methods
for assessing personality disorders: Implications for research and the
evolution of Axis II. American Journal of Psychiatry, 154, 895-903.
Westen, D. (1998). Implicit cognition, affect, and motivation: The end of
a century-long debate. In R. Bornstein & J. Masling (Eds.), Empirical
studies of unconscious processes. Washington, DC: American Psychological Association.
Westen, D. (in press-a ). Psychodynamic theory and technique in relation
to research on cognition and emotion: Mutual implications. In T.
Dalgleish & M. Power (Eds.), Handbook of cognition and emotion.
New York: Wiley.
Westen, D. (in press-b). The scientific status of unconscious processes:
Is Freud really dead? Journal of the American Psychoanalytic
Association.
Westen, D., & Arkowitz-Westen, L. (in press). Psychotherapy integration: Integrating psychodynamic and cognitive-behavioral approaches. In C. R. Snyder & R. Ingram (Eds.), Handbook of psychological change: Psychotherapy processes and practices for the 21st
century. New York: Wiley.
Westen, D., & Feit, W. (1998). [All the President's women: Affective
constraint satisfaction in ambiguous social cognition]. Unpublished
raw data.

FREUD'S LEGACY
Westen, D., & Harnden, J. (1998). Categorizing eating disorders by
personality profiles: Bridging the chasm between Axis I and Axis IL
Unpublished manuscript, Harvard Medical School.
Westen, D., Klepser, J., Ruffins, S.A., Silverrnan, M., Lifton, N., &
Boekamp, J. (1991). Object relations in childhood and adolescence:
The development of working representations. Journal of Consulting
and Clinical Psychology, 59, 400-409.
Westen, D., Ludolph, P., Block, M. J., Wixom, J., Wiss, E C. (1990).
Developmental history and object relations in psychiatrically disturbed
adolescent girls. American Journal of Psychiatry, 147, 1061-1068.
Westen, D., Moses, M. J., Silk, K. R., Lohr, N. E., Cohen, R., & Segal,
H. (1992). Quality of depressive experience in borderline personality
disorder and major depression: When depression is not just depression.
Journal of Personality Disorders, 6, 382-393.
Westen, D., Muderrisoglu, S., Fowler, C., Shedier, J., & Koren, D.
(1997). Affect regulation and affective experience: Individual differences, group differences, and measurement using a Q-sort procedure.
Journal of Consulting and Clinical Psychology, 65, 429-439.
Westen, D., & Shedler, J. (1998). Bridging the gap between clinical
and research perspectives on personality disorders: A nonreplication
of the Five Factor model using a clinically sensitive Q-sort procedure.
Unpublished manuscript, Harvard Medical School.
Westen, D., & Shedler, J. (in press). Revising and assessing Axis II: II.
Toward an empirically and clinically sensible taxonomy of personality
disorders. American Journal of Psychiatry.
White, N. M. (1995). Emotional memory: Conceptual and methodological issues. In J.L. McGaugh, N. Weinberger, & G. Lynch (Eds.),
Brain and memory: Modulation and mediation of neuroplasticity (pp.
93-100). New York: Oxford University Press.
Whiting, J. W. M., & Child, I. L. ( 1953 ). Child training and personality:
A cross-cultural study. New Haven, C'12.Yale University Press.
Wilson, T.D., Lisle, D., Schooler, J., Hedges, S.D., Klaaren, K., &

371

LaFleur, S. (1993). Introspecting about reasons can reduce postchoice satisfaciJon. Personality and Social Psychology Bulletin, 19,
331-339.
Wilson, I". D., & Schoolel; J. W. (1991). Thinking too much: Introspection can reduce the quality of preferences and decisions. Journal of
Personality and Social Psychology, 60, 181-192.
Wilson, W. R. (1975). Unobtrusive induction of positive attitudes. Unpublished doctoral dissertation, University of Michigan.
Wixom, J., Ludolph, P., & Westen, D. (1993). Quality of depression in
borderline adolescents. Journal of the American Academy of Child &
Adolescent Psychiatry, 32, 1172 - 1177.
Wong, P., Bernat, E., Bunce, S., & Shewin, H. (in press). Brain indices
of nonconscious associative learning. Consciousness and Cognition:
An International Journal, 6, 519-544.
Wong, P., Shevrin, H., & Williams, W. J.(1994). Conscious and nonconscious processes: An ERP index of an anticipatory response in a
conditioning paradigm using visually masked stimuli. Psychophysiology, 31, 87-101.
Zajonc, R. B. (1968). The attitudinal effects of mere exposure. Journal
of Personality and Social Psychology, 9(2, Pt. 2), 1-27.
Zajonc, R. (1980). Feeling and thinking: Preferences need no inferences.
American Psychologist, 35, 151 - 175.
Zillman, D. (1978). Attribution and misattribution of excitatory reactions. In J. H. Harvey, W. Ickes, & R. E Kidd (Eds.), New directions
in attribution research (Vol. 2). Hillsdale, NJ: Erlbaum.
Zimmerman, M., Pfohl, B., Coryell, W., Stangl, D., & Corenthal, C.
(1988). Diagnosing personality disorder in depressed patients. Archives of General Psychiatry, 45, 733-737.
Received December 31, 1996
Revision received June 3, 1998
Accepted June 4, 1998