J Psychother Pract Res 8:213-224, July 1999
© 1999 American Psychiatric Association
Mardi Horowitz, M.D. and Hansjoerg Znoj, Ph.D.
Received September 10, 1998; revised December 23, 1998; accepted January 6, 1999. From the Department of Psychiatry, University of California-San Francisco, and the Department of Psychology, University of Bern, Switzerland. Address correspondence to Dr. Horowitz, Department of Psychiatry, University of California-San Francisco, 401 Parnassus Avenue, San Francisco, CA 94143.
Defensiveness in intrapsychic and interpersonal activities is a generally accepted concept among psychodynamic theorists,but a theoretically grounded classification of emotional controlprocesses is needed. As a result of intensive case-by-case clinicaland empirical studies, such a system was assembled. The systemis organized by three major categories of processes that canregulate emotions. These are sets of mental operations thatcontrol 1) content of thought and communications, 2) form ofthought and communications, and 3) person schemas that organizebeliefs and interpersonal expressions. Each category of defensivecontrol processes is linked to observable outcomes at intrapsychicand interpersonal levels. This classification system can beused to formulate how patterns of avoidance and distortion areformed.(The Journal of Psychotherapy Practice and Research 1999;8:213–224)
Defensive control of emotion is a classic psychodynamic concept and an element of case formulations. For decades, medical studentsand psychiatric residents have had to pass tests showing anability to define traditional psychoanalytic defense mechanisms. Yet the concept is not fully accepted in nondynamic approaches,and defensive styles were not accepted, after controversy, as a part of the multiaxial diagnostic system in DSM-III1 and DSM-IV.2
One problem has been the theory of classification of defenses. A related problem has been the degree of reliability and validity of empirical variables related to the traditional definitions of defense mechanisms. A third problem has been divergence in schools of psychotherapy. The time is at hand for an integration and a revision of the theory of classification of defenses based on empirical evidence. This paper summarizes such a revision, and some of the evidence. It is intended as a teaching document.
Sigmund Freud and other early psychoanalysts developed names and definitions for defense mechanisms. Anna Freud3 offered an organization of her father’s observations, yet inconsistencies of categorization remained.4–10 Some defense mechanismswere defined as simple units of mental processing and mental outcomes, as in repression of a single, specific wish. Otherdefense mechanisms were complex combinations of mental and social operations, as in identification with the aggressor. What isneeded now is a better theory of classification, one that clearly relates processes to observable outcomes.
The earlier defense classification system evolved in the context of a psychoanalytic motivational theory that emphasized drives:the ego formed defenses to mediate between id urges and superego injunctions.11 Subsequent theory from cognitive sciences revised this view, replacing it with an information-processing approach.This modern view allows a shared theoretical language acrossdisciplines and a redefinition of emotional control processes.12–14
The context for control of emotion is usually the processing of a difficult, emotionally challenging topic. An example is processing a piece of bad news from a stressor event, one that requires revision of prior knowledge, cognitive map, or schema. When a person gets bad news, it may be discordant with existing mental models. Such a mismatch between event interpretations and cognitive maps will activate negative emotions. These emotions in part function as motives to focus attention, fixing it on the problem of how to reconcile the incongruities between thebad news and existing schemas. With reconciliation, the person can reduce emotional alarms and shift attention to another topic.
Information processing helps the person shift from a state of mismatch to one of match, as illustrated in Figure 1. Figure 2 adds the element of emotion to the cognitive model of Figure 1. In addition to current emotional evocation, information processing evokes anticipation of where emotion might lead. One possibilityis anticipation of entry into a dreaded state of mind with intense, out-of-control negative emotions. To avoid current or anticipatedexcessive emotional arousal, the person then increases control processes to regulate the flow of information by selective inhibitionsand facilitations. This is a contemporary model of signal affects, like Freud’s early formulation of signal anxiety.3, 11 The following case example provides an illustration.
Steve, a first-year surgical resident, was at a crucial level of training that would dictate his future career, and his goalwas to become a great surgeon. The attending surgeon who had observed him told Steve that he had not met the required standards of judgment and skill for advancement in the residency training program. The new information—that he had only mediocre skills in a key area—was sharply discordant with Steve’s goal. This mismatch threatened to produce intense shame. He anticipated entering a state of excessive humiliation and suicidal preoccupation.
Steve had had states of shame and despair in the past. He anticipated a repetition of such dreaded moods, and to avoid them he activated defensive control processes that inhibited his conscious contemplationof the recent deflating news, as illustrated in Figure 3.
Steve’s defensive avoidance could be variously categorized. One way is to use the classic psychoanalytic terminology of defense mechanisms.3,6,15–17 If his inattention to thetopic of his insufficient surgical skills was a conscious choice, then his acts would be called suppression. If it was an unconsciously chosen inhibition of an important topic, then it would be called repression. The defensive control processes classification tobe presented16–20 uses one simple term, topic inhibition, to identify the basic process without getting into the difficult inference of the role of reflective consciousness in information processing.
Through topic inhibition, Steve affected the content of his attention. He could also have affected the form of his attentionas well as its mental content. Instead of having a distressing visual fantasy of failing to receive his certificate of completion when other residents received theirs at the year-end ceremony, he could inhibit imagery. By being aware only of words predicting such a future, and not imagining the scene visually, he might reduce shame. Instead of an undermodulated state, Steve might maintain a well-modulated state while anticipating a tolerable degree of embarrassment.
By a third type of control, Steve could alter person schemas. He could change the relative activity of his various possible roles of self and his relationship models. Steve could inhibit conceptualizing himself as a degraded man and instead facilitate concepts of himself as a competent man performing well in an alternative career. Figure 4 illustrates how deciding to be an anesthesiologist shifted his state from a degraded to a competent self schema. Steve could also alter role schemas of his attending surgeon, by shifting from viewing him as an admired and competent mentor to viewing him as an incompetent judge.
Person Schemas Theory
Contemporary cognitive theory employs the concept of parallel processing. That is, multiple appraisals and revisions of information on the same topic can occur at the same time in different information-constructionchannels. Schemas (cognitive maps) can organize the constructionof informational sets in each parallel processor. Supraordinate schemas help in fitting together the output of several processing channels. The fittings and choices may tend to accentuate one schema from the medley, leading to the working model that organizesa current state of mind.21–24
Organizing schemas influence the changing strengths of associational linkages among these networks of patterned meanings. The most complex of these organizing schemas are probably the schemas of self and others (person schemas). Like other cognitive maps, person schemas revise, smooth, integrate, and package information.25–27Current motives are guided to suitable targets by plans andscripts for actions. Person schemas include transactional sequences that lead to anticipated consequences, as well as roles andvalues used in assigning blame and praise. Control processesare guided by anticipation of such outcomes as success, failure,esteem, or degradation; that is, anticipation involves both desired and dreaded states of mind. In this way, defensive control processes are placed within a motivational matrix.18–21
|A CLASSIFICATION OF DEFENSIVE CONTROLS: PROCESSES AND OUTCOMES|
Three larger categories organize this classification. Some controls shift the content. Other control processes alter the form. Stillother defensive control processes can shift person schemas.These three sets will be discussed in that order. For each set, control process outcomes will be described at two levels of observation: reflective awareness (which can be assessed by self-reports) and communication (which can be judged by clinical raters).
Contents of expressions can be selected by control processes that can alter 1) topics, 2) concepts, 3) the designation of importance of concepts, or 4) the threshold for disengaging attention from a topic.
1. Altering Topics:
Potential topics may be inhibited or facilitated relative toone another, leading to a shift in attention from one topicto another.28 Adaptively, shifting focus away from an unresolved topic can reduce emotion and ward off entry into a dreaded stateof mind. Maladaptively, extended forgetting, disavowal, or denial can lead to a failure in coping.
2. Altering Concepts:
Concepts are potential ideas and feelings to be used in contemplating a topic. These elements are activated from memory because of the strength of connections in complex networks of associational linkages. Priming can determine priorities for what conceptis likely to be “next” in the expression or representation of a chain of concepts. Altering the next concept can change emotion.
A tightly linked chain of ideas is likely to be clear. Leaps and flights in linking concepts together can lead either to confusing chains or to new combinations that provide stunning and creative insights. Altering concepts can lead to both adaptive and maladaptive consequences.
3. Altering the Importance of a Chain of Concepts to the Self:
A chain of ideas may appear to lead to a solution to a problem. Another chain can be formed that also appears to solve the problem. Reflective conscious awareness can compare the alternatives,and each chain is then weighed for relative probabilities ofreal success.
The significance of a chain can be exaggerated or minimizedrelative to an alternative. This shift in weights can lead torational evaluations or irrational sliding of meanings. Distortedattitudes may be formed or important possibilities can be ignored.
4. Altering Threshold for Disengagement:
Some patients in psychotherapy declare a topic concluded whenthe therapist believes it is not. The problematic topic, withits emotional conflicts, is interrupted. The habitual tendencynoted is one of moving away from hard decisions. Personal dilemmas remain unresolved because of these short circuits.
Outcomes of Control Processes:
Either by self-reports (from introspection) or by direct observation (e.g., of recorded communications), clinicians study the outcomes of defensive controls and infer the underlying processes. Most outcomes result from the combination of several control processes, but it is clarifying to give an example of a prime outcome for each category of process….
In our research in developing this classification, we examined how independent judges rated psychotherapy transcripts for different kinds of elaboration and dyselaboration of emotional ideas. The subcategories of dyselaboration in the manual used by judges followed the “Maladaptive Outcome” column of Table 2.
The judgments of the raters were reliable, and the averaged category frequencies showed increased defensive-level outcomes during discourse on the patient’s most conflictual and unresolved topics.29,30 In single-case studies, we found the results from this classification to be clearer than analyses of some portions of the data thatwere based on classical psychoanalytic defense mechanism categories.31
Controlling the Form of Conscious Thought
Emotion can be controlled by altering forms of thought. Several such processes include altering 1) mode of representation, 2)time span, 3) logic level, 4) level of action planning, and 5) arousal level.
1. Altering Mode of Representation:
People can shift the relative dominance of words, images, and somatic enactions in thought. People can have different states in which there is high or low translation of meaning betweenmodes. Isolation of meaning in one mode, especially in the lexical mode, can reduce emotion, whereas increasing imagery may increase emotion. If vivid images occur without translation of the images into lexical meanings, the focal conscious experiences may seemlike perceiving rather than recollecting. This can result in a sense of reliving a past experience, rather than having a memory or fantasy contemplation.
2. Altering Time Span:
Time spans for contemplation can be set by intentions to limit an associational search for knowledge. Time spans can be set to encompass long or short periods or toward recollecting and imagining past or future. The remote or recent past, as well as the near or distant future, can be specifically focused on in setting an attentional frame. Emotional arousal can be limited by focusing attention tightly away from a time when bad things happened and might happen again.
Focusing consciousness on the very recent past and the very near future can help the person plan immediate action without being swamped by emotions about dire but distant futures. Forexample, when receiving news of a laboratory test that shows the presence of an unexpected cancer, a person can think of whom to tell right now, rather than think about what it will be like to have the disease exacerbate in the future. People can also reduce emotion by focusing on the remote past or by reliving a memory or fantasy in order to ignore current threats. At other times, threatened by the possibility of recalling traumatic memories, they keep the time frame of attention set for only right now and for what detailed actions they should perform next to keep themselves busy.
3. Altering Logic Level:
People can influence the rules of thought; they can shift fromhaving tight, narrow rules for logically linking associations to making broad, creative, and even illogical associations. Some people blunt emotion caused by broad awareness by emphasis on plans that involve careful logic. Others who are distressedby a focus on planning may instead have a broad and glorious fantasy.
4. Altering Level of Action Planning:
In restful repose, a person may have a mental set for contemplation without bodily action. On a basketball court, that person’s mental set may intend swift reactive movements without the distraction of contemplation. In playing ball, action is at lightning speed; in playing chess, action is restrained until a move is fully thought through. Control processes can alter such setpoints for degree of thought and degree of concomitant motor activity. Defensively, a person can act too impulsively instead of thinking or, like Hamlet, ruminate to avoid taking action.
5. Altering Arousal Level:
People seek stimuli, take drugs, meditate, or choose calming or arousing activities. Giddiness, avoidant sleeping, or sexual promiscuity are sometimes used to avoid serious issues. Both thrills and lethargies can reduce emotional threat.
Outcomes of Control Processes:
The introspective outcomes of these defensive processes that control form are summarized in Table 3. The outcomes observablein interpersonal communications, such as records of evaluationor therapy sessions, are summarized in Table 4.
Controlling Person Schemas
Emotions can be changed by shifts in how self and others are viewed. People can change an internal working model of an actual social transaction by altering 1) self schema, 2) the schema for the other person, 3) role relationship models, 4) value schemas, and 5) executive-agency schemas.
1. Altering Schemas of Self:
Control of schemas of self can lead to shifts in roles and procedures for role-related behaviors. It can directly and indirectly alter the emotions that color a state of mind. By shifting roles within a repertoire, people can reduce the intensity of an unwanted emotion. One example is shifting to the role of indignant accuser from the role of weak victim. Getting angry provides escape from a dreaded state of fear.
2. Altering the Schema for the Other Person:
In a similar manner, a person can shift the roles of others. Altering the role of the other person leads to different emotional interpretations, expectations, and plans. An individual may at first view the other person as a highly desirable companion and shudder at an anticipated rejection. A defensive shift in role for the other may change this emotional arousal. The other is viewed as ordinary, unworthy of intense pursuit. The anxious affects over feared rejection are quieted.
3. Altering Role Relationship Models:
Role relationship models are an internal map for social transactions. In an important affiliation, a series of role relationship modelsdevelops. Of these different role relationship models, somemay be desired, some feared, and others used as a compromise to avert dreaded states. When a wish/fear dilemma links desiredto dreaded views of a relationship, the person can shift to a state organized by such compromise role relationship models.32
The compromise avoids the threat of the dreaded consequences of the desire. For example, a person may have an urge to flirt with a new acquaintance but fears the other would find the self deplorable. Anxious states of mind could occur: will the self be rejected? A compromise state of mind can be organized by a different role relationship model, one in which the self is self-sufficient and the other is viewed as only slightly interesting.The dilemma between excitement of intimacy and feared pain of rejection is reduced by a shift into the compromise view. The defensive shift says, in effect, “Oh, I was just conversing to pass time with you until I go on to something more interesting.”Unfortunately, the compromise seldom promotes real satisfaction.
4. Altering Value Schemas:
An important conscious activity is judging actions. People use alternative sets of values to make these critical analyses. Different value schemas may be given different priorities indifferent states.
When they are in danger of blame, when emotions of shame and guilt threaten to demolish self-esteem, people can control distressby shifting values. For example, a patient lied to his bestfriend, saying he was not flirting with the friend’s girlfriend. This lie made him feel guilty because telling the truth is of value. He then shifted to asserting that “love is war” and that in war deceit is part of “good strategy.” Similar rationalizations occurred in his business. As business agent, he sold a product containing pollutants because he was “just following orders”and was “loyal to the corporation.” In other states, he was furious about firms that committed such practices.
5. Altering Executive-Agency Schemas:
The executive agent is the person believed to be in charge of forming plans and instigating action. In different states of mind, an individual may vary in how a sense of executive agency influences decisions. In everyday life, the executive agent may be the “I” in a sense of identity. This executive agent may be shifted to something other than “I.” There are other leaders, the “we” of a family, friendship, work group.
Sometimes seeing “I” as an executive agent feels too impoverished, and one’s mood deflates. The schema for who is executive maybe shifted to an idealized other or larger group. De Gaullesaid, “I am France.” Self-abnegation or a sense of merger with cosmic powers may occur. That can have adaptive or maladaptive consequences; morale may be restored as grand agents are assumed to be in charge, or grandiose delusions may be formed.
Outcomes of Control Processes:
|RESEARCH FINDINGS AND CLINICAL APPLICATIONS|
The classification presented in Tables 1–6 evolved throughresearch efforts and discussions to develop a shared cognitive-psychodynamic language. These efforts began with careful case studies of patients in psychotherapy for stress response syndromes.18,21,33–35Later, quantitative investigations of reliability and validity of the categories occurred.36,37 Because multiple observerscan study and rate recordings, both quantitative and qualitative studies of interpersonal manifestations are likely to allow for tests of consensus and predictions. Judges did reliably score signs of defensive control, and these did locate the most conflictual topics in psychotherapy.30 Defensive organizations inferred as configurations of schemas were reliable as scoredby independent raters,38 and independent formulation teams, blind to each other’s inferences, arrived at similar configurations for the same case material.39,40 Additional empirical checks on this classification included separate studies of observerand self-report measures based on this classification system. Subjects reported on their own habitual control processes with highly significant levels of test-retest reliability over time,and observers agreed to a significant degree in independent rating of videotapes21(Horowitz and Znoj, unpublished data).
Tables 1–6 presented the details of this classification of emotional control processes and are useful for detailed theory and research. A simplification is desirable for clinical teaching and practice, and the focus there should be on what is important to help a patient change. Accordingly, a list of common maladaptive avoidances and distortions appears as Table 7. From such a list, the clinician can develop techniques for modification of themaladaptive defensiveness.
The levels of observation listed in the tables lead to inferences about how ideational and interpersonal avoidances and distortions occur. Psychotherapists can then show a patient how to confront emotional and conflictual topics more directly.
This research was supported by grants from the John D. and CatherineT. MacArthur Foundation and the Swiss National Foundation.
- American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 4th edition. Washington, DC, American Psychiatric Association, 1994
- Freud A: The Ego and Mechanisms of Defense. New York, International Universities Press, 1936
- Cramer P: The Development of Defense Mechanisms: Theory, Research, and Assessment. New York, Springer-Verlag, 1991
- Cooper S: The empirical study of defensive processes: a review, in Interface of Psychoanalysis and Psychology, edited by Barron JW, Eagle MN, Wolitsky DL. Washington, DC, American Psychological Association, 1992
- Hentschel U, Smith G, Ehlers W, et al (eds): The Concept of Defense Mechanisms in Contemporary Psychology. New York, Springer-Verlag, 1993
- Vaillant GE: Ego Mechanisms of Defense: A Guide for Clinicians and Researchers. Washington, DC, American Psychiatric Press, 1992
- Vaillant GE: Ego mechanisms of defense and personality psychopathology. J Abnorm Psychol 1994; 102:44–50
- Conte H, Plutchik R: Ego Defenses: Theory and Measurement. New York, Wiley, 1995
- Horowitz M: Cognitive Psychodynamics: From Conflict to Character. New York, Wiley, 1998
- Freud S: The ego and id (1923), in The Standard Edition of the Complete Psychological Works of Sigmund Freud, vol 17, edited by Strachey J. London, Hogarth, 1955, pp 1–66
- Lazarus RS: Psychological Stress and the Coping Process. New York, McGraw-Hill, 1966
- Lang PJ: The motivational organization of emotion: affect-reflex connections, in Emotions: Essays on Emotion Theory, edited by van Goozen SHM, Van de Poll NE, Sergeant JA. Hillsdale, NJ, Lawrence Erlbaum, 1994, pp 61–93
- Ledoux JE: Emotion: clues from the brain. Annu Rev Psychol 1995; 46:209–235[Medline]
- Freud S: Beyond the pleasure principle (1920), in The Standard Edition of the Complete Psychological Works of Sigmund Freud, vol 18, edited and translated by Strachey J. London, Hogarth Press, 1955, pp 3–66
- Haan N: Coping and Defending: Processes of Self-environment Organization. New York, Academic Press, 1977
- Perry JC, Kardos ME: A review of the Defense Mechanism Rating Scales, in Ego Defenses: Theory and Measurement, edited by Conte H, Plutchik R. New York, Wiley, 1995
- Horowitz MJ: Introduction to Psychodynamics: A New Synthesis. New York, Basic Books, 1988
- Horowitz MJ, Markman HC, Stinson CH, et al: A classification theory of defense, in Repression and Dissociation: Implications for Personality Theory, Psychopathology and Health, edited by Singer J. Chicago, University of Chicago Press, 1990, pp 61–84
- Horowitz MJ, Stinson CH: Consciousness and processes of control. J Psychother Pract Res 1995; 4:123–139
- Horowitz MJ, Znoj H, Stinson C: Defensive control processes: use of theory in research, formulation, and therapy of stress response syndromes, in Handbook of Coping, edited by Zeidner M, Endler N. New York, Wiley, 1996, pp 532–553
- Kihlstrom JF: The cognitive unconscious. Science 1987; 237:1445–1452
- Kosslyn S: Mental imagery, in Visual Cognition and Action: An Invitation to Cognitive Science, vol 2, edited by Osherson S, Kosslyn S, Hollervach J. Cambridge, MA, MIT Press, 1990, pp 2–40
- Neisser U: Learning from the children, in Knowing and Remembering in Young Children, edited by Fivush R, Hudson J. New York, Cambridge University Press, 1990
- Anderson CA: Abstract and concrete data in the perseverance of social theories: when weak data leads to unshakable beliefs. J Exp Soc Psychol 1983; 19:93–108
- D’Andrade R: The identification of schemas in naturalistic data, in Person Schemas and Maladaptive Interpersonal Patterns, edited by Horowitz MJ. Chicago, University of Chicago Press, 1991, pp 279–302
- Rumelhart DE, McClelland JL, PDP Research Group: Parallel Distributed Processing: Exploration in the Microstructure of Cognition, vol I: Foundations. Cambridge, MA, MIT Press, 1986
- Posner MI: Hierarchical distributed networks in the neuropsychology and neurolinguistics, in Cognitive Neuropsychology and Neurolinguistics: Advances in Models of Cognitive Function and Impairment, edited by Caramazza A. Hillsdale, NJ, Lawrence Erlbaum, 1990, pp 187–210
- Horowitz MJ, Milbrath C, Reidbord S, et al: Elaboration and dyselaboration: measures of expression and defense in discourse. Psychotherapy Research 1993; 3:278–293
- Horowitz MJ, Milbrath C, Stinson C: Signs of defensive control locate conflicted topics in discourse. Arch Gen Psychiatry 1995; 52:1040–1057
- Horowitz MJ, Cooper S, Fridhandler B, et al: Control processes and defense mechanisms. J Psychother Pract Res 1992; 1:324–336[Abstract/Free Full Text]
- Horowitz MJ: Formulation as a Basis for Planning Psychotherapy Treatment. Washington, DC, American Psychiatric Press, 1997
- Horowitz M: Stress Response Syndromes. Northvale, NJ, Jason Aronson, 1986
- Horowitz MJ: A model of mourning: change in schemas of self and other. J Am Psychoanal Assoc 1990; 38:297–324[Medline]
- Horowitz MJ, Fridhandler B, Stinson C: Person schemas and emotion. J Am Psychoanal Assoc 1991; 39(suppl: “Affect: Psychoanalytic Perspectives”):173–208
- Horowitz MJ, Stinson C, Curtis D, et al: Topics and signs: defensive control of emotional expression. J Consult Clin Psychol 1993; 61:421–430[Medline]
- Horowitz MJ, Milbrath C, Jordan DS, et al: Expressive and defensive behavior during discourse on unresolved topics: a single case study. J Pers 1994; 62:527–563[Medline]
- Horowitz MJ, Eells TD: Case formulations using role-relationship model configurations: a reliability study. Psychotherapy Research 1993; 3:57–68
- Eells TD, Horowitz MJ, Singer J, et al: The role relationship models method: a comparison of independently derived case formulations. Psychotherapy Research 1995; 5:161–175
- Horowitz MJ, Eells T, Singer J, et al: Role relationship models for case formulation. Arch Gen Psychiatry 1995; 53:627–632