DynaPsych Table of Contents

Schizophrenia and the Family:
Double Bind Theory Revisited

Mathijs Koopmans
York College/ City University of New York

Copyright Mathijs Koopmans, 1997
Parts of this paper were presented at the Annual Conventions of the International Congress of Psychology, Montreal, Canada, August, 1996, National Council on Family Relations, Portland, OR, November, 1995; and the American Psychological Association, New York, NY, August, 1995.


Double bind theory proposes that symptoms of schizophrenia are an expression of contradictory patterns of interaction in the family. Although considered practically useful, the double bind construct is typically seen as difficult to investigate empirically. The status of the theory therefore needs to be clarified. This paper addresses three aspects of this endeavor: (1) a review of the empirical evidence, (2) a clarification of the double bind construct and its operationalization, and (3) an assessment of the compatibility of double bind with other theories and findings, which are directly concerned with schizophrenia in the family, such as the behavioral-genetic interpretation of findings from twin and adoption studies.  


In the mid nineteen fifties, a number of scholars developed proposals that attributed schizophrenia to the exposure to, and participation in, dysfunctional communication patterns in the family. Some characterizations of such patterns are double bind interaction (Bateson, Jackson, Haley & Weakland, 1956), transactional disqualification (Sluzki, Beavin, Tarnopolsky, & Ver˘n, 1967), pseudomutuality (Wynne, Ryckoff & Day, 1958; marital schism (Lidz, 1968; Lidz, 1973; Lidz & Fleck, 1985) and "scapegoating" (Ackerman, 1958).

One of the best known models to arise out of this work has been Gregory Bateson's double bind theory (Bateson, Jackson, Haley & Weakland, 1956), which proposes that contradictions in the interaction between family members predisposes its members to schizophrenia. The double bind notion initially met interest in the academic and mental health communities (Schuham, 1967), and it has provided part of the standard vocabulary of many clinical practitioners, to conceptualize the relational complexities typically addressed in individual and family therapy. This vocabulary continues to be used in spite of the fact that the theory itself has been largely discarded for the lack of convincing empirical support (Abeles, 1976; Haller, 1989; Olson, 1972). Abeles (1976) concluded that although double bind may be a practically useful concept for therapists and counselors, it is an 'unresearchable' construct (Abeles, 1976, p. 113; see also Haller, 1989). For now, this conclusion has sealed the fate of double bind as an idea whose refutation lies outside of the purview of scientific inquiry, and the amount of empirical work undertaken after these reviews were published has indeed been scanty (Haller, 1989). This state of affairs creates the false impression that the double bind issue has been resolved, or that it can be dismissed as yet another 'grand theory' (Bebbington & Kuipers (1988), and it ignores some significant new work (Haller, 1989). There has also been a great deal of unclarity about how the insights of Bateson et al.'s theory are compatible with other theories and findings, such as behavioral-genetic notions (Gottesman, 1991), the expressed emotion literature (Brown, Birley & Wing, 1972; Kavanaugh, 1992; King & Dixon, 1995; Miklowitz, 1994; 1995; Vaughn & Leff, 1976), and other family process theories (e.g. Laing & Esterson, 1964; Lidz & Fleck, 1985; Sluzki, Beavin, Tarnopolsky & Ver˘n, 1967; Wynne, Ryckoff & Day, 1958).

This paper consists of two parts. The first section reviews the empirical evidence pertaining to Bateson et al.'s double bind theory. A second section discusses the compatibility of double bind with related theories concerning schizophrenia and the family. This paper will be specifically concerned with other family process theories, and with the compatibility of double bind with the findings from behavioral-genetic research (twin comparisons, adoption studies). A discussion of the relationship between Bateson et al.'s theory, and the expressed emotion literature is very important as well, but will be conducted elsewhere.

A number of earlier reviews of empirical studies were published (Abeles, 1976; Haller, 1989; Olson, 1972: Sluzki & Ransom, 1976; Schuham, 1967). Although these reviews have become increasingly pessimistic in tone about the ability of empirical research to address the status of double bind theory, recent work has been more rigorous than early attempts, and has also been able to produce more interpretable results (e.g. Blotchky, Tittler & Friedman, 1982; Haake, 1984). A need therefore exists to evaluate the bearing of this work on conclusions drawn earlier.

A new review of the double bind literature is also warranted because a significant portion of the empirical work consists of Doctoral Dissertations whose results have not been published in peer reviewed outlets, and some of the most recent publications have been in languages other than English (Aoki, 1993; Weakland, 1992). Moreover, repeated calls for further development and clarification on the theoretical level (Haller, 1989; Mishler & Waxler, 1965; 1968; Olson, 1972; Pine, 1984; Rotzien, 1989; Shuham, 1967; Watzlawick, 1974) have not met sufficient response, some significant attempts notwithstanding (Gundersen, 1988; Rotzien, 1989; Roy & Roy, 1987; Watzlawick, Beavin & Jackson, 1967).

It is also becoming increasingly clear from schizophrenia research taking place from other perspectives, such as the behavioral-genetic perspective, that the contribution of the rearing environment cannot be circumvented in our search for the origins of schizophrenia (Gottesman, 1991; Gottesman, McGuffin & Farmer, 1987; McGue & Gottesman, 1989; Kety & Ingraham, 1992; Tienari & Wynne, 1994; Tienari, Wynne, Moring, Lahti, Naarala, Sorri, Wahlberg, Saarento, Seitaama, Kaleva & L„ksy, 1994). We therefore need to re-evaluate those models which are specifically concerned with the nature of this environmental contribution.

We have also been unable to appreciate potentially relevant findings concerning schizophrenic patients and their families in terms of how compatible they are with Bateson et al.'s original propositions. For example, it has repeatedly been found that when patients are returned to their families after hospitalization, high levels of expressed emotion (criticism, involvement) in their families is a good predictor of a relapse (Kavanagh, 1992; Vaughn & Leff, 1976). To my knowledge, there have been no attempts to draw a meaningful connection between these findings and double bind theory.

There is the very persistent misconception in the literature that models connecting family processes to schizophrenic symptomatology in effect blame the parents for their children's ills, rather than recognizing the potential of such models to empower parents and caretakers to modify their interactive patterns to better accommodate the identified patient, and perhaps even prevent dysfunctional patterns of interaction from occurring in the first place.

There also have been few attempts thus far to make the connection between Bateson et al., and other dynamical models which connect schizophrenic symptomatology to family processes explicit. Lidz and Fleck (1985), for instance, conceptualize of dysfunctional interaction in the family in terms of a weakness in the relationship in the spousal subsystem, which often results in unduly involvement of the children in their parents' realm of concerns. N. W. Ackerman (1958) similarly considers the symptomatic behavior of children and adolescents as instrumental in the attempts of families to resolve their functional problems. Are these proposals compatible with double bind theory, and if so how?

The plan of this analysis is as follows. First, the double bind hypothesis will be summarized as originally formulated in the classical Bateson et al. paper, and empirical research done so far to determine its status will be reviewed. Next, some of the missing links in the theory will be clarified and better articulated in an attempt to make double bind more accessible to empirical scrutiny. An attempt will also be made to clarify some of the outstanding areas of confusion. Finally, the compatibility of double bind with other theories connecting schizophrenia to family interaction will be considered, and the compatibility of the findings of behavioral-genetic twin research with double bind will be assessed.      

Methodological Preliminaries


Part of the inconclusiveness of existing work results from terminological confusion concerning the double bind phenomenon. Double bind is interchangeably referred to as a hypothesis, and as a theory (see e.g. Abeles, 1976; Olson, 1972; Schuham, 1967, and many others). Since the requirements for empirical confirmation of a theory and a hypothesis are very different (Kerlinger, 1973), this terminological issue needs to be resolved. If double bind is seen as a theory, numerous different hypotheses can be legitimately derived, whose standing can be separately confirmed or disconfirmed in empirical research. If double bind is seen as a hypothesis, on the other hand, the accumulation of evidence either speaks in its favor or not in its favor, and existing research can be evaluated in terms of how well each operationalizes this one hypothesis. Existing reviews have taken the latter approach, evaluated the accumulated body of evidence as largely unsupportive, and attributed the lack of evidence to deficiencies in the operationalization of the relevant constructs. In this review, I will argue that Bateson et al's Double Bind needs to be understood as a theory, and that the evidence needs to be reviewed in light of the specific derivation of the theory which informs the research. This approach enables us to contemplate the possibility that some hypotheses derived from double bind theory enjoy stronger empirical support than others.  

Double Bind as a Theory


Bateson et al. (1956) proposed that schizophrenic symptoms are an expression of social interactions in which the individual is repeatedly exposed to conflicting injunctions, without having the opportunity to adequately respond to those injunctions, or to ignore them (i.e., to escape the field). For example, if a mother tells her son that she loves him, while at the same time turning her head away in disgust, the child receives two conflicting messages about their relationship on different communicative levels, one of affection on the verbal level, and one of animosity on the nonverbal level. It is argued that the child's ability to respond to the mother is incapacitated by such contradictions across communicative levels, because one message invalidates the other. Because of the child's vital dependence on the mother, Bateson et al. argue that the child is also not able to comment on the fact that a contradiction has occurred, i.e., the child is unable to metacommunicate (Bateson et al., 1956).

The symptomatology of schizophrenia, it is argued, reflects the accommodation of the individual to a prolonged exposure to such interactions. Once 'victims' have learned to perceive their universe in terms of contradictory environmental input, the inability to respond effectively to stimuli from the environment is no longer contingent on the extent to which stimuli from the environment are contradictory in specific interactive sequences. Instead, the individual will generally experience any input from the environment as conflicting information without being able to discriminate between different communicative levels. In the long run, this inability manifests itself as typically schizophrenic symptoms such as flattened affect, delusions and hallucinations, and incoherent thinking and speaking (Bateson et al., 1956).

It is further stipulated by Bateson et al. (1956) that double bind interactions have a pathogenic effect only if they occur in a context where the accurate discrimination of messages is of vital importance for the participants, and in a relational context which is characterized by intense levels of involvement between the participants. The interaction between parents and children within the nuclear family is a typical example of such a relational context.

At least the following twelve components can be distinguished in double bind theory with potential for operationalization as variables in empirical work: (1) schizophrenia, (2) conflicting injunctions, (3) communicative levels, (4) prolonged exposure, (5) ability to comment (6) escape from the field, (7) intense involvement, (8) discrimination of levels, (9) relational context, (10) learning to perceive, (11) double bind patterns, and (12) vital importance of the relationship. The interrelationship between these components is clearly stated in the theory, but too complex to be captured in a single hypothesis. We need to recognize that different hypotheses derived from this theory may fare differently upon empirical scrutiny.      

Research Designs Addressing the Theory


There has been a tendency in the double bind literature to characterize inherent design limitations in empirical research as deficiencies in the operationalization of Double Bind as a construct (Abeles, 1976; Haller, 1989; Olson, 1972). For example, studies in which attempts are made to simulate the pathogenic effect of conflicting injunctions in a laboratory setting cannot be justifiably criticized for not considering actual family interactions (Olson, 1972). Similarly, designs comparing families with and without a schizophrenic member to determine whether conflicting injunctions occur more often in the families with a schizophrenic member (e.g. Blotchky, Tittler & Friedman, 1982), cannot be fairly criticized for not taking the longitudinal perspective it would require to assess whether victims have learned to perceive their universe in double bind patterns (Bateson et al., 1956). Except for Sluzki and Ransom (1976), reviews have not distinguished different types of research design as addressing different aspects of the double bind theory, and have not sufficiently acknowledged the inherent limitations of the different research designs used.

To recognize the strengths and limitations of the research conducted so far, it is necessary to make a distinction between three types of study: (1) simulation studies, in which double bind contingencies are simulated in the laboratory context in the relationship between the experimenter and the subject (e.g. contradictory task instructions), and some outcome is measured, (2) family observation studies in which it is determined to what degree families with and without a schizophrenic member display the hypothesized patterns of interaction, and (3) double bind tasks, in which the differential ability of schizophrenic versus nonschizophrenic subjects to recognize and respond to double bind task content is assessed using paper-and-pencil. While each of these designs have their limitations in addressing double bind theory, each design arguably addresses different aspects of the double bind concept. It is argued in this review that recognition of the inherent limitations of each design results in a different appreciation of the theory.  

Criteria for selection of studies


In this review, only studies are considered which explicitly address the status of the double bind hypothesis empirically. Research which uses a double bind framework to interpret its findings without being explicitly concerned with the hypothesis itself (e.g. Gooblar, 1978; Summers & Walsh, 1977; Wooley & Vigilante, 1984) were eliminated from this review, as were studies whose findings could be construed as evidence for double bind, but were not designed to test it (e.g. Koopmans, 1994; Laing & Esterson, 1964; Mishler & Waxler, 1968; Shoham, Weissbrot, Markowsky & Stein, 1977). Also excluded were articles which discuss the application of the double bind construct in relevant clinical settings without investigating its empirical status (Ferris, 1975; Klessman & Klessman, 1983; Koopmans, 1995; Krefting, 1990). Also excluded were studies which addressed the status of double bind theory on the basis of clinical observation rather than systematic collection of evidence (Watzlawick, 1963). Table 1 lists the studies thus included for this review, and itemizes which aspects of the theory are addressed.

Table 1.  Characteristics of empirical studies concerned with the
status of Double Bind Theory

                                   Design    Features

Abeles (1975)                      DBT       1, 5, 7, 9, 11, 12
Berger (1965)                      DBT       1, 9
Beakel & Mehrabian (1969)          OBS       3, 9
Beavers et al. (1965)              OBS       1, 6, 9
Blotchky et al. (1982)             OBS       2, 3, 7, 9
Bugenthal et al. (1971)            OBS       2, 3, 9
Ciotola (1961)                     DBT       1, 2
Dush & Brodsky (1981)              SIM       4, 5, 9
Guindon (1971)                     DBT       1, 3
Haake (1984)                       DBT       2, 5, 9,11
Hall & Levin (1980)                OBS       1, 3
Heilveil (1980)                    DBT       1, 3, 9
Karr (1983)                        DBT       4, 9, 11
Kellogg (1980)                     DBT       7, 9
Kingsley (1970)                    SIM       1, 4, 5, 8, 9
Kuiken & Hill (1985)               SIM       4, 5, 9
Loeff (1965)                       DBT       1, 3, 8
McDonough (1975)                   SIM       9
Melowsksi (1975)                   SIM       2, 9
Mehrabian & Wiener (1967)          DBT       3
Newman (1977)                      DBT       1, 3, 8
Nida (1976)                        DBT       5, 7
Phillips (1970)                    DBT       9
Potash (1965)                      DBT       1, 6
Rierdan & Brooks (1977)            DBT       1, 2, 3, 4, 9, 10
Ringuette & Kennedy (1965)         OBS       1, 9
Rotzien (1989)                     OBS       9
Schaeffer (1972)                   DBT       1, 5, 8
Schreiber (1970)                   SIM       2, 9, 12
Smith (1973; 1977)                 SIM       2, 4, 5, 9
Sojit (1969; 1971)                 SIM       1, 5, 9, 12

Key to designs:

SIM: Double Bind Simulation
DBT: Double Bind Task
OBS: Family Observation
Key to features:
1.   Comparison schizophrenic vs. nonschizophrenic subjects,  i.e., 
     schizophrenic subjects in particular
2.   Conflicting injunctions aimed at subjects
3.   Measurement distinct communicative levels,  i.e., metacommunication
4.   Operationalization of prolonged exposure to double bind
5.   Ability to comment on contradiction considered
6.   Escape of the field operationalized
7.   Intensity relationship considered
8.   Ability to discriminate levels, or recognize double bind
9.   Relational context of double bind considered
10.  'Learning to perceive' operationalized
11.  Paradox measured
12.  Vital importance of discimination messages 

Current empirical standing of the Double Bind Theory


Family Interaction Studies


Some researchers have argued that it would be a confirmation of the hypothesis if it can be demonstrated that double bind interactions occur more often in families with a schizophrenic member, than in families without a schizophrenic member (Beakel & Mehrabian, 1969; Blotchky, Tittler & Friedman, 1982; Bugenthal, Love, Kaswan & April, 1971). Berger (1965) investigated this possibility retroactively by administering a questionnaire consisting of thirty double bind statements, which subjects were asked to rate on a four point scale in terms of how frequently they recalled their mothers using such statements. There were four groups, one consisting of schizophrenic subjects, and three consisting of non-schizophrenic subjects. The schizophrenic group consistently reported a higher incidence of such statements than one of the comparison groups (college students), but not significantly higher than the other comparison groups which also had medical or psychiatric conditions. As with any retroactive study, interpretability of findings is limited by the fact that recall may be unequally biased in the schizophrenic group.

Blotchky et al., observed 15 families who had a child enrolled in a short term residential treatment program. They found that mothers communicated a larger proportion of messages, in which verbal content was rated as being conflicting with nonverbal expression, to the symptom bearing child (71%) than to other family members, whereas the father did not send more conflicting messages to the symptom-bearing child than to the other family members. Mothers also contributed more to the total of conflicting messages exchanged in these interactional episodes than other family members.

Bugenthal et al. found that a greater number of mothers in families with disturbed members (referred by schools because of chronic behavior and emotional problems) produced messages which were conflicting along the verbal - nonverbal spectrum than mothers in families without disturbed members during a 5 minute interactional sequence in which families were asked to discuss family related issues. Beakel and Mehrabian (1969) found no differences between those parents ranked by clinical observers as having higher levels of psychopathology, and those rated as having lower levels, with respect to the level of incongruity between verbal content and accompanying nonverbal postures in the interaction with their adolescent children.

These findings, appear to indicate that within the family, the symptom bearer is more often engaged in contradictory interactional sequences than other family members, whereas no differences are discernible if interaction within the family as a whole is considered. Furthermore, the mother appears to be engaged more often in the exchange of contradictory messages involving children than other adults, a finding that would be consistent with Bateson et al.'s stipulation that discrimination of messages needs to be of vital importance for double bind interactions to have a pathogenic effect.

Results of these studies should be seen as tentative, however, as the comparisons in all three studies are not without problems. Beakel and Mehrabian's study has no information of differences between the groups, Blotchky et al. only included problem families, and Bugenthal et al.'s findings do not rule out differences between the two groups due to family structure: nine out of twenty of the disturbed families in this latter study were single parent families, whereas all non-disturbed families were dual parent families; it is not clear to what degree this structural difference may affect the likelihood of double binds occurring.

In a reanalysis of samples of data from Mishler and Waxler (1968), from Wild and Shapiro (1977), and from Wild, Shapiro and Goldenberg, 1975), Hall and Levin (1980) found no difference between families with and without a schizophrenic member in the degree to which verbal and non-verbal communication were in agreement. To the extent that inconsistencies in communication occur more often in schizophrenic families, then, it is not necessary to assume that they are played out across the verbal versus nonverbal one channel.

Beavers, Blumberg, Timken and Weiner (1965) investigated the metacommunicative ability of schizophrenic families. They found that mothers of schizophrenic children more often respond evasively to interview questions about their children than mothers of nonschizophrenic members, suggesting a difference in interactive style with respect to the degree of evasiveness. It is not clear from their study, however, whether this evasiveness is associated with contradictory patterns of interaction. Blotchky Tittler & Friedman (1982), who addressed this latter question, found that children indeed respond more often to conflicting messages with avoidance rather than metacommunicative statements, a finding which supports, albeit tentatively, Bateson et al.'s contention that conflicting communication patterns result in an inability to metacommunicate.

Bateson et al.'s claim that double binds typically occur in instances of intense mutual involvement between participants is supported in Blotchky et al.'s finding that conflicting messages tend be associated with extreme interpersonal closeness or distance, as measured by the Felt figures task, a figure placement test. The possibility that conflicting interactions affect the regulation of interpersonal distances in the family deserves closer empirical scrutiny, because finding such a relationship may enable us to perhaps better understand other aspects of family interaction traditionally not associated with family process models, such as symbiosis (Minuchin & Fishman 1981), and lack of cohesiveness in problem families (Pratt & Hansen, 1987).

The empirical work on family interaction is not without its problems. By designing those studies in such a manner that confirmation of the theory depends on the ability of the double bind construct to distinguish between 'normal' and 'schizophrenic' families, it is assumed that schizophrenics and family members continuously talk in double binds, will talk in double binds when third parties are present, and that outsiders are able to detect double binds without fully appreciating the context in which the interaction occurs (Bateson, 1969/1972). Mishler and Waxler (1965) rightly note that those messages that an observer may characterize as incongruent may carry implicit meanings for family members which produce congruence, or vice versa. The inability of outsiders to understand exactly all facets of what is being said, what is being implied, and what is being assumed in the interaction between two people who know each other well challenges the validity of research designs which rely on the evaluation of those communicational patterns by third parties (Abeles, 1976).

A related concern is that all family interaction studies on double bind interactions conducted so far rely on randomly selected chunks of family interaction, rather than using complete interactional episodes, thereby limiting the ability to investigator to detect contradictory patterns of interaction. It is especially important to consider this latter point because it does not necessarily follow from double bind theory that double binds occur very often (Bateson, 1969). The use of larger segments of family interaction in empirical source material should be one of the priorities of double bind research in future work.

In an attempt to analyze family interaction segments more systematically, Rotzien (1989) applied the Structural Analysis of Social Behavior (SASB, Humphrey & Benjamin, 1986), a methodological strategy specifically developed to measure elusive family processes, by rating interpersonal as well as intrapsychic components in the observation of interactional sequences. However, rather than specifically operationalizing double bind theory, SASB proposes a model of interpersonal behavior, and its hypothesized intrapsychic correlates, in which freedom vs. control, and sexuality vs. aggression are the critical dimensions. Since SASB is an entirely different theoretical formulation, examining this model lies outside of the purview of double bind research. Moreover, Rotzien's analysis indicates that double bind interactions can only be indirectly inferred from response patterns of subjects, rather than being directly observed. Given these complications, a more direct measurement of double bind interactions remains a more desirable course.  

Tasks with double bind content


A number of studies have used paper-and-pencil tasks with double bind content to determine whether the responses of schizophrenic subjects were different from those of nonschizophrenic subjects, as Bateson et al would predict. Berger (1965) simply asked schizophrenic and non-schizophrenic subjects whether they remembered having been exposed to double bind statements, and he found that schizophrenic subjects indeed report a higher incidence of such statements by their mothers. However, it is difficult to interpret these findings, as recall may by affected by the pathological condition. Potash (1965) found that when faced with a choice between high risk - high yield, and low risk - low yield response categories (i.e., the prisoner's dilemma, see e.g. Hargreaves Heap & Varoufakis, 1995), schizophrenic did not withdraw more often than nonschizophrenic subjects would from the situation. Whether having to choose between these two options can be called a double bind analog, or paradox, is open to debate.

Phillips' (1970) developed and administered a double bind test ("douBT") in which subjects would rate parent to child communications. Response styles were coded as being cynical, neutral, or euphemistic. No association was found between response style on the double bind test and personality characteristics of the subject such as ego strength, super-ego strength, and guilt of the subjects, or parenting style, as measured in the mothers of the most markedly cynical, or most markedly euphemistic respondents (35 and 26 mothers respectively). The only exception was a significant negative correlation between cynical response mode, and super-ego strength. The results were interpreted as challenging double bind theory.

Loeff (1965), Mehrabian and Wiener (1967), Newman (1977), and Guindon (1971) had subjects rate tape recorded statements on the degree of consistency between verbally and non-verbally expressed messages. In the comparison between schizophrenic versus nonschizophrenic subjects' rating, results are inconsistent. Loeff (1965) found that schizophrenic, delinquent, and 'normal' subjects were equally able to identify conflicting messages as such, but that patterns of evaluation were different: schizophrenic and delinquent subjects tend to evaluate on tone and content simultaneously, whereas the other subjects tend to evaluate on content alone. The groups were equally able, however, to discriminate double bind contingencies from non-double bind contingencies.

Mehrabian and Wiener (1967) found that variations in tone rather than content affected variations in response in subjects without psychiatric classification, and that in the event of contradictions between these two channels the nonverbal input prevails. This finding underlines the importance of distinguishing different levels of communication but does not directly confirm or disconfirm the hypothesis.

Guindon (1971), on the other hand, found that schizophrenic as well as non-schizophrenic subjects tend to resolve conflicts toward the verbal rather than the nonverbal channel. Although his method of distinguishing pathological groups is more rigorous than that in the work previously mentioned (MMPI classifications, rather than whether or not referred for psychiatric help or seeking treatment), the number of subjects in each group is rather small (4 groups of 5 subjects).

In sum, differences in response to verbal vs. nonverbal discrepancies in the tape recorded segments do not reliably discriminate schizophrenic and non-schizophrenic responses, although schizophrenic subjects may have greater difficulty with double bind items than nonschizophrenic subjects (Guindon, 1971; Newman, 1977; Schaeffer, 1972, see Abeles, 1976).

Considering different types of items that may fall under the broad definition of double bind, Guindon (1971) made a distinction between items in which the contradiction was purely substantive (e.g. 'You can always talk to me, but don't bother me with unimportant problems' p. 104), and items in which verbal content and affect were conflicting ('Oh, my darling how perfect you are', in an angry tone, p. 104). Injunctions conflicting across channels were rated as less sincere than those which were nonconflicting across channels. Subjects found items to be more difficult if their nature was personal ('You are disgusting' p. 103), than if their nature was nonpersonal ('There are two people, dancing and clapping their hands together in time to the music' p. 103).

Abeles (1975) compared subjects' response styles of schizophrenic, alcoholic, and normal subjects to paradoxical items. Example of a paradoxical item would be: 'A man tells you that all men are liars. Since he is a man, he must be a liar too. Therefore, when he says that all men are liars, he is (a) telling the truth, (b) lying.' (p. 169). No differences were found in response styles between the different groups in terms of how much they were 'bound' by the items, i.e., reasoned about it in a vicious circle. There were differences, however, in the degree to which subjects response styles were affected by the anticipation of being evaluated. Schizophrenic and alcoholic subjects' performance was more affected than that of normal subjects if they were led to expect that a good performance was of critical importance to them.

Recognizing the importance of 'relational context', Abeles included items which were concerned with relationships, such as the following item: "A man says to his wife 'You are too dependent on me and I don't like it. I insist that you develop a mind of your own' If she does as he asks, she is being (a) dependent, (b) independent" (Abeles, 1975; p. v). However, Abeles' items are concerned with a verbal description of relationships, rather than that they operationalize the embeddedness of double bind in its actual relational context, i.e., between experimenter and subject.

Karr (1983) similarly placed paradoxical statements in a parent-child context and asked college students to place themselves in the position of the source or recipient of the conflicting injunctions. Her results indicate no difference in the ratings of 'boundedness' depending on the point of view taken by the rater (source or recipient). These results emphasize the importance of the relational context of double binds, and, more specifically, the reciprocal nature of interactive exchanges. Consistent findings were obtained by Kellogg (1980) who had subjects read a double bind scenario, and then rate interactions on a semantic differential scale.

Nida (1976) used self report measures of double bind in the parent-child context to assess the relationship between double bind experiences and personality variables such as communicative apprehension, anxiety and introversion. The association between reported double bind experiences and the personality variables turned out significant.

Rather than comparing schizophrenic and nonschizophrenic subjects, Ringuette and Kennedy (1965) used scholars well versed in double bind theory as subjects to assess the reliability of their judgement when rating letters, received by schizophrenic patients written by their parents, versus letters received by non-schizophrenic patients from their parents. It would be a confirmation of the theory, it was argued, if those investigators would be able to reliably rate the extent to which double binds occurred in the letters received by the schizophrenic patients, and do a better job at it than clinicians and lay-persons. It turned out that the correlation of interrater agreement was .16 for the double bind experts, .26 for trained clinicians, and .39 for naive judges (i.e. with no training in medical of social sciences). If investigators highly conversant in double bind theory are unable to agree on what constitutes double bind interaction in the communication between schizophrenic children and their parents, it was argued, the construct is in all likelihood without a clear empirical referent (Shuham, 1967). It should be noted, however, that it is inconsistent with a theory, which places messages in their relational context, to assume that outsiders are in a position to fully familiar with the context in which the interaction in written correspondence originally occurred. The lack of consideration of the relational context, which is an inherent limitation in all studies relying on the evaluation of double bind statements out of context, clearly places these findings in perspective: regardless of whether subjects are double bind experts, or - victims, it is difficult for third parties to assess which communicative episodes constitute double interactions.  

Simulation double bind process in laboratory settings


In simulation studies, double bind interactions are deliberately created in the context of the relationship between the experimenter and his or her subjects. Examples of the kind of double bind contingencies that can be created in the simulation studies are contradictory instructions, rewards, or task input (Ciotola, 1961; Dush & Brodsky, 1981; Kingsley, 1970; Kuiken & Hill, 1985; Melowsky, 1975; McDonough, 1975; Rierdan & Brooks, 1977; Schreiber, 1970; Smith, 1973; 1976; Sojit, 1969; 1971).

Ciotola presented psychotic and nonpsychotic psychiatric patients with an auditory discrimination task; Subjects were instructed to discriminate between tones; while discrimination was actually impossible (the double bind analog). In a subsequent auditory discrimination tasks, there were no differences between the psychotic and the non-psychotic subjects in response times, level of tension during the task performance, and level of concreteness of responses.

Kingsley (1969) found that there was more anxiety and confusion among those subjects exposed to double bind situations, if they were also less able to comment on the task conditions they were exposed to (i.e., they were unable to metacommunicate). Kingsley hypothesized a decline in performance on concept formation, and visual discrimination tasks, that would be greater after subjects had been exposed to double bind situations, than after ordinary stress situations, and that a greater decline would be observed for poor premorbid, than good premorbid schizophrenic subjects. Results only partially confirmed these predictions. Performance was poorer in less adjusted than well adjusted group, but the level of sophistication of responses was not different depending on whether subjects were or were not exposed to double bind. However, after having been exposed to double bind, subjects gave more concrete answers on the concept formation task (explaining proverbs) than in ordinary stress situations. Performance on this task also tended to be more concrete among poor premorbid schizophrenics, whereas among good premorbid schizophrenics tended to choose more abstract responses. On a visual discrimination task, input depicting a mother figure was more unfavorably affected by previous double bind conditions than task input depicting a father figure, but there was no difference in level of concreteness of response.

Rierdan and Brooks (1977) examined the effects of reinforcement incongruous with respect to the verbal vs. nonverbal channel on the learning performance of schizophrenic patients vs. nonschizophrenic subjects. There turned out to be a difference in favor of the nonschizophrenic subjects in learning when reinforcement was incongruous but not when it was congruous, suggesting a better ability to deal with double binds in nonschizophrenic subjects.

Smith (1973; 1976) found that when subjects were to respond to sets of questions about fictional letters from 'mothers' to 'daughters' with or without conflicting content, and with or without a punishing response by the experimenter (white noise), a combination of high ambiguity and a punishing response by the experimenter appeared to be more anxiety arousing to the subjects than ambiguity or punishment alone. It was made difficult for subjects to escape the field by a threatened reduction in course credit if subjects who initially committed themselves to the task did not complete it.

Dush and Brodsky (1981) replicated this study but also explicitly blocked subjects' option to metacommunicate by not allowing requests for clarification of task requirements. In addition to more anxiety, it was found that subjects produced digit recall patterns similar to those typically found in schizophrenic patients, i.e., better performance in backward recall.

Schreiber (1970) used tasks with instruction intended to heighten performance anxiety, accompanied by a denial of the appropriateness of anxiety reactions. Communicational disruption (as measured by variations in type/token ratio and responses in a cloze test) occurred only if the task was related to the course for whom the experimenter was the instructor, and subjects were students, while disruption did not occur when tasks were unrelated to that context. This result again underlines the importance of Bateson et al.'s stipulation that the relational context in which communicative events take place, is a critical factor in determining whether or not double bind interactions have a pathogenic effect. It should be added, however, that measurements were very indirect in this study. 'Reduced communicative levels' was operationalized as a smaller number of different words in the first 100 words of the writing subjects produced (type token ratio), and on the basis of the results of a cloze test in which subjects writings were assessed on their coherence after each 10th word had been deleted from their text. At best, these measurements can be seen as a very indirect assessment of communicativeness.

McDonough (1975) found that responses to a scanning task were affected to a significant extent by whether subjects received contradictory instructions (the double bind condition), than when they received unambiguous instructions.

Sojit (1969; 1971) found that in the context of a doctor- patient relationship, contradictory task input gave rise to less metacommunication and more incongruent responses in the parents of schizophrenics than in the parents of non-schizophrenics. However, responses of parents of juvenile delinquents and youngsters referred for ulcerative colitis were similar to the parents of schizophrenic rather than to those of non- schizophrenic subjects. This result again stresses the importance of the relational context in which the effects of contradictions in communication are observed. In Sojit's study, the ambiguity has consequences for the subjects in this study for whom the ambiguity occurs in the context of a doctor-patient relationship. This factor appears to be more critical than whether or not the reason for referral is schizophrenia, or another type of disorder.

The absence of a differential ability of schizophrenic versus nonschizophrenic parents to respond effectively to the task requirements speaks against another important aspect of the double bind theory, namely that responses of schizophrenic families to double bind input will be different from those of nonschizophrenic families.

Melowski (1978) had psychology students rate a psychologist in a taped therapy segment. Instructions for the subjects were conflicting on the verbal versus the nonverbal level with respect to whether there were 'right' and 'wrong' answers. The verbal injunction for all subjects was that there were no 'right' or 'wrong' answers, and that not giving honest answers would ruin the experiment. In the double bind condition, additional nonverbal feedback to subjects suggested that subjects were giving the wrong ratings. Differences between treatment and control group in response time were indeed different, but there were no differences in self-reported anxiety.

It appears, then, that double binds produced by the experimenter create anxiety in subjects only if their inability to respond to it jeopardizes task success, and this lack of task success has tangible consequences for the subjects. Double bind situations also appear to create anxiety if escape options and metacommunicative options are blocked for the subjects (Dush and Brodsky, 1981; Smith, 1973; 1976). These findings support one important aspect of Bateson et al.'s model, namely that conflicting input indeed has an effect if discrimination of messages really matters.

Contrary to Bateson et al.'s model, the results of simulation studies are unspecific as to which type of pathological behavior is triggered by contradictory injunctions. The finding that contradictory instructions in task environments where successful performance matters produces heightened anxiety does not demonstrate a pathogenic effect, let alone the etiology of schizophrenia. In fact, the possibility that anxiety in response to ambiguous experimental conditions may not demonstrate abnormality receives indirect support in Kingsley's (1969) finding that in good premorbid schizophrenic patients' responses, double bind conditions produced more anxiety than in the responses of poor premorbid schizophrenic patients.

Double Bind Theory emphasizes the relational context in which the exchange of messages occurs, because it is precisely this context that is being contradicted in double bind communication (Abeles, 1976). In empirical research conducted in artificial settings the relational context of the subject with the experiment is not necessarily contradicted by contingencies such as cryptic instructions, contradictory rewards or confusing advice.

The scope of simulation studies is also inherently limited in its ability to assess the pathogenic effect of 'prolonged exposure' to double bind, as the theory would require (Olson, 1972). Similarly, it is difficult to capture the 'intense involvement' between participants in simulation designs although there have been some attempts to vary the level of mutual involvement between experimenter and subjects (Kuiken & Hill, 1985; Schaeffer, 1972).

In an attempt to address the operational difficulties with the double bind construct, Kuiken and Hill (1985) limited the definition of double bind to communications which satisfy the following two criteria (1) "two or more persons are involved in a relationship in which their conceptions of appropriate behavior in the relationship are ambiguous or unclear" (p. 85), and (2) "a communication is given which includes two mutually incompatible self-presentations or a metacommunicatively [sic] disqualified self-presentation and, hence, includes contradictory definitions of appropriate behavior by the recipient" (p. 86). In their operationalization, experimenters gave contradictory self- disclosures, and they invited subjects to do likewise. Although this operationalization does not require us to deal with the more problematic aspects of the original concept, such as the vitality of the relationships, and inability to escape the field, concepts like 'self-presentation' and 'appropriate behavior' are difficult to capture in experiments as well. Kuiken and Hill's reconceptualization, then, does not make double binds less intangible, and it is therefore difficult to accept the omission of the more critical elements of the original definition.

It has been argued in the double bind literature that double bind should be primarily seen as a language, a way of thinking and speaking, which, although practically useful, may not necessarily be accessible to systematic empirical investigation (Abeles, 1975; Bateson, 1969/1972; Watzlawick, 1965) given the limitations of empirical methods we have available. The necessity to develop constructs with can be operationalized is thereby eliminated, while double bind remains useful as heuristic for family therapy practice (Haller, 1989).

However, the tentative nature of many of the research findings clearly has at least in part to do with a lack of replications rather than a lack of encouraging preliminary evidence. Longer segments of family interaction would enhance our ability to detect double binds, systematically varying escape options of length of exposure could be suggestive of the pathogenic effects of double binds in experimental settings. Moreover, in simulated settings, outcomes such as arousal and anxiety can be measured more precisely. Empirically establishing a link between double bind contingencies and anxiety measures would also enable us to further determine the compatibility of these outcomes with the hypothesis of schizophrenia as a long term effect of such contingencies. In designs which compare schizophrenic to nonschizophrenic subjects, a greater homogeneity of groups, and larger sample sizes would enable to more clearly detect double bind effects, as would our ability to control for the intake of medication (Guindon, 1971). In addition, psychopathology could be better operationalized by using MMPI scores rather than psychiatric diagnoses (Guindon, 1971).

Addressing these questions would clearly be within reach of what empirical research is capable of doing. However, the lack of determination on the part of researchers to pursue these issues is understandable considering the vagueness of the theory itself (Olson, 1972), and its inability to connect with other research traditions which are concerned with schizophrenia. For example, researchers interested in the communicational aspects of schizophrenia have not connected double bind to research findings on expressed emotion (Brown, Birley & Wing, 1972; Vaughn & Leff, 1976), or to the findings from behavioral genetic twin and adoption studies, and those investigators who did comment on these findings from a family dynamics perspective were not specifically concerned with the double bind theory (Jackson, 1960; Lidz, Blatt and Cook, 1981). An assessment of the significance of this work from a double bind point of view, is long overdue. The objective of the second part of this paper is to clarify the more critical concepts of Bateson et al.'s theory, and to make an attempt to connect double bind theory to other family process models, and to other research which is concerned with schizophrenia in families.  

Theoretical Extensions

There have been repeated and long neglected calls for a further delineation and specification of the double bind phenomenon (Abeles, 1976: Olson, 1972; Schuham, 1967; Weakland, 1974). The question where double bind interactions originate, or how they should be defined, is addressed only in a perfunctory manner by the authors of the double bind paper. Having described a situation where, in the context of a mother - child relationship, approaching behavior is accompanied by anxiety and avoidance on the part of the mother, Bateson et al. observe: 'Since this is a formal description we are not specifically concerned with why the mother feels this way about the child' (p. 213 of steps). The assumption that mother-child relationships are primarily of interest is not explicitly justified Bateson et al.'s work, nor is it indicated which particular feature of this relationship produces a pathogenic effect in case of ambiguity in the mother child alliance.

One of the sources of doubt about the viability of double bind theory as an explanatory framework for schizophrenia lies in the fact that in the course of daily social exchanges, conflicting injunctions are not uncommon, either within, or outside of the family, as role conflicts are an inextricable aspect of self-regulatory behavior in any social system, families in particular (see e.g. Broderick, 1993; Biddle, 1979; Minuchin & Fishman, 1981). From the fact that discriminating messages might be vital in some relational contexts, it does not necessarily follow that duplicity of meaning produces pathological results. For example, when an adolescent son complains to a parent about being treated like a child, and the parent responds by saying 'You are my child' (Sluzki, Beavin, Tarnopolsky, & Ver˘n, 1967; p. 498), there is a measure of duplicity in the parent's response in the sense that biological status is confounded with relative age status. However, it is unclear why such a response would reflect a 'pathogenic' pattern of self regulation. The traditional line of argument in double bind theory is that such interactions involve a confusion of communicative levels, or logical types, and that the participant(s) in the double bind interaction gradually internalize this confusion. In this example, one could perhaps also argue that the parent expresses reluctance about yet treating the son as an adult, while trying not to offend him by explicitly reinforcing a generational boundary, thereby expressing what could be seen as a very ordinary parental dilemma (Steinberg, 1990).

One possibility which deserves closer scrutiny is that there is a connection between Bateson et al.'s theory, and other theories concerning the relationship between schizophrenia, and social dynamical processes, such as N. W. Ackerman (1958) and Lidz and Fleck (1985). The possibility needs to be considered that by displaying problematic behavior, children conform to the role of a 'referral-seeker' (N. W. Ackerman, 1958): features of the behavior of the child deemed problematic provide an occasion for families to seek help without having to come to face the fact that they have functional problems as a family. Any remark or action from family members which supports dysfunctional behavior on the part of the identified patient would constitute a double bind contingency: a child who gets a shower of adult attention when throwing a tantrum, while at the same time being reprimanded, receives conflicting injunctions concerning the tantrum behavior (reinforcement yet discouragement). The functional significance of interactions of this sort is that adults deflect from complications at the family level by strengthening problem behavior in their child. From the child's point of view, participating in this scenario involves a critical ambiguity concerning his or her family membership. On the one hand, the child plays his or her role, assigned by the family, and by doing so (s)he is the 'odd person out'. The communicative process through which this role is assigned to the child, then, can be seen as double bind communication.

Role conflicts in the family are not necessarily associated with the referral process, however. They can occur in any situation in which there are marked violations of the boundaries between generations (Lidz & Fleck, 1985), such as when a parent is behaves more as a child than as a spouse with respect to the other parent, when a parent acts as a rival to one of the children, or when a child is used as an emotional substitute for a spouse (Lidz & Fleck, 1985). When generational boundaries are breached in the family, the communicative flow between parents and children contains conflicting injunctions with respect to the nature of the relationship between them. Since children critically rely on those relations for their survival, their ability to discuss such contradictions is limited by their fear of losing the relationship (Olson, 1972).

If double binds are seen as an expression of conflicting family constellations operating simultaneously (Koopmans, 1995; Wooley & Vigilante, 1984), it becomes more feasible to find concrete examples. Laing and Esterson (1964) discuss a family, where the member being diagnosed with schizophrenia had been taught to address her family members as follows:  

BIOLOGICAL STATUS                                       TITLES TAUGHT TO THE
father                                                  uncle
mother                                                  mummy
aunt (mother's sister)                                  mother
uncle (mother's sister's husband)                       daddy, later uncle
cousin                                                  brother

This constellation in effect hides the identity of the biological parents, and makes the boundaries between nuclear and extended family systems incomprehensible to the youngster. The family systems are not consistent in the biological vs. communicative sense. From the child's point of view, the double bind lies in the fact that two persons are identified as the mother, thereby simultaneously affirming and denying that relationship for the child with each of those two persons, and thereby raising doubts about the identity of the biological father as well. The ongoing interaction between this child and her caretakers takes place against the background of uncertainty with respect to the relational premises on which the communication is based within this system.

To bring Bateson et al.'s theory in line with other sociodynamical models of schizophrenia (N. W. Ackerman, 1958; Laing & Esterson, 1964; Lidz & Fleck, 1985), pathogenic double bind injunctions in the family needs to be confined to those instances in which they express a confounding of basic family roles in the communicative process. Such a specification would eliminate an important source of confusion in double bind theory, namely that double bind interactions involve a confusion of logical types, or communicative levels (Kingsley, 1969; Mishler & Waxler, 1965). Such communicative levels have turned out to be extremely difficult to define (B. L. Ackerman, 1979; Relph, 1991). As a result, there is a chronic lack of concrete examples of double bind situations in the literature. This state of affairs has long been an obstacle to the development reliable assessment procedures (Ringuette & Kennedy, 1965; Olson, 1972; Haller, 1989; Rotzien, 1989). To the extent that double bind interaction involves a confusion of levels of communication, the proposed specification would clarify the nature of the confusion: Rather than confusing the levels of communicative expression themselves, e.g. verbal vs. nonverbal communication, double binding reflects an error or contradiction in the classification of the relationships presumed by the participants in the course of the interaction between them (Gundersen, 1988; Roy & Roy, 1987). Because of its vital functions (reproduction, caretaking), the family offers a particularly compelling case to consider the consequences of such classification errors.

It should be emphasized, however, that double bind theory does not require us to consider the family as the only context in which a discrimination of messages can be of vital importance. Other relational contexts in which the discrimination of messages is typically of vital importance are the following: infirmity, material dependence, captivity, friendships, love, loyalty to creed, cause, or ideology, contexts informed by norms and traditions, and therapeutic relations (Rotzien, 1989; Watzlawick, Beavin & Jackson, 1967). The pathogenic potential of double bind interactions in these contexts deserves further contemplation as well.  

Behavioral-genetic versus double bind interpretations of concordance findings within families


A discussion of the entire corpus of relevant data on schizophrenia to determine compatibility is beyond the scope of this paper. Two strands of research, whose relevance to double bind theory is most immediately apparent, will be considered from a double bind perspective, namely (1) the behavioral-genetic interpretation of findings on schizophrenia from twin comparisons (Gottesman, 1991; Kendler, 1983; 1993) and from adoption studies (Kety, Rosenthal, Wender, & Schulsinger, 1968; Kety, 1988; Kringlen, 1991; Rosenthal, Wender, Kety, Schulsinger, Welner & þstergaard, 1968; Tienari & Wynne, 1994; Tienari, Wynne, Moring, Lahti, Naarala, Sorri, Wahlberg, Saarento, Seitaama, Kaleva & L„ksy, 1994), and (2) the compatibility of double bind theory with the findings on expressed emotion and communicative deviance in the families to which the schizophrenic patients has returned after hospitalization (Brown, Birley & Wing, 1972; Kavanagh, 1992; King & Dixon, 1995; Miklowitz, 1995; Vaughn & Leff, 1976). Since these research traditions are directly concerned with the role of the family in etiology and course of schizophrenic disorders, the compatibility of double bind theory with findings from this work needs to be examined.  

Pedigree Studies

The purpose of pedigree studies is to determine the rate of coincidence of schizophrenia within the family (Gottesman, McGuffin & Farmer, 1987; McGue & Gottesman, 1989). These studies have established a high rate of coincidence, or concordance, within the family, with a higher concordance for members of closer genetic relatedness (see Table 2). It has traditionally been argued that the statistical association between concordance for the disorder and genetic closeness in families is an indication of the hereditary nature of the disorder. It has also been recognized, however, that one cannot separate hereditary from environmental influences when the subjects being compared share both. The aggregation of the disorder within family systems may therefore reflect a shared environment (Pardes, Kaufman, Pincus & West, 1989). More importantly, the concordance rates themselves do not reveal the nature of the transmission process, and therefore do not exclude the possibility that this process is of a double bind nature. The interaction process in families with multiple instances of the disorder therefore needs to be investigated in greater depth to examine how, and to what extent, dynamical social interaction processes contribute to these concordance figures.

Table 2.  Percentage of relatives of schizophrenic patients
affected by the disorder

                                               N         %
Familial Relationship

Monozygotic Twins                             106      44.3
Offspring two schizophrenic parents           134      36.6
Dizygotic Twins                               149      12.1
Siblings                                     7523       7.3
Offspring one schizophrenic parent           1678       9.4
Half-siblings                                 442       2.9
Nieces or nephews                            3965       2.7
Grandchildren                                 739       2.8
First cousins                                1600       1.6
Spouses                                       399       1.0

Adapted from Rowe (1994)

Twin Comparisons

The intention of twin comparisons has been to resolve the complications in the interpretation of concordance figures within the family. Since identical twins are genetically the same, while fraternal twins are not, the observed differences in how often the disorder is shared by identical versus fraternal twins has been used as an estimate of the degree to which the disorder is hereditary (see e.g. Rosenthal, 1970; Rowe, 1994; Vandenberg, 1976). It has indeed been found that identical twins more often share schizophrenia than fraternal twins do (see Table 2, and Kendler, 1983; 1993; McGue & Gottesman, 1989; Onstad, Skre, Torgersen & Kringlen, 1991). These findings clearly pose a challenge for a purely environmental reading of double bind theory.

To interpret the greater co-occurrence of schizophrenia in identical twins as demonstrating a genetic basis for the disorder, it needs to be assumed that rearing environments are equally similar or different for twins regardless of their zygosity (Kendler, 1983). If identical twins are treated more similarly by their caregivers than fraternal twins, greater environmental and genetic similarity may both account for the greater co-occurrence of schizophrenia in identical twins. If it cannot be assumed that environments are shared to the same degree by identical and fraternal twins, it cannot be known to what degree higher twin concordance is attributable to genetic factors, and to what degree to environmental factors (Lewontin, 1975; Lewontin, Rose & Kamin, 1984).

A greater similarity in treatment of identical twins in the course of interaction with caregivers has indeed been found (Lytton, 1977). Although this similar treatment may occur in response to traits and behavioral features in the child that are of a genetic origin, these traits and features are not necessarily related to the schizophrenic disorder itself, only to the fact that concordance has occurred.

It is conceivable that Lytton's findings would apply to double bind interactions: caregivers may be more inclined to simultaneously engage identical twins in double bind interactions than they are to engage fraternal twins in such interactions, and such a difference between identical and fraternal twins could occur both in the presence or in the absence of a genetic vulnerability to the disorder.

To rule out an effect of zygosity status on similarity in treatment, it has been examined to what extent twins reared apart share the disorder. However, because of the rarity of the phenomenon (Kendler, 1993), data are too scarce at this point to be conclusive (Farber, 1981; Ratner, 1982).    

Adoption Designs

To address the problems with associated with the establishment of a genetic basis for schizophrenia on the basis of twin comparisons, adoption designs have also been used, because the genetic and child-rearing aspects of parenthood can be separated (Kringlen, 1991; Tienari & Wynne, 1994). It is argued that any concordance between adoptees and biological parents points to a hereditary transmission, because the biological parents do not rear the adoptee; any concordance between adoptees and adoptive parents, on the other hand, points to a social transmission, because these parents do rear the adoptee without there being a genetic relationship of descent (Kety, 1988; Kringlen, 1991; Rosenthal, 1970; Tienari & Wynne, 1994; Tienari, et al., 1994).

Three types of adoption designs have been distinguished, namely (1) those studies in which it is determined to what extent the adopted away off-spring of biological parents, known to be affected by the disorder, are affected as well, (2) studies in which it is determined to what extent biological relatives or adoptees known to be affected by the disorder, are also affected, and (3) studies which determine whether children are affected whose biological parents are not affected, but whose adoptive parents are affected (Kringlen, 1991).

It appears that more children adopted from schizophrenic biological parents have severe psychiatric diagnoses at some point (30%) than children adopted from nonschizophrenic biological parents (15%, Tienari, 1992; Tienari et al., 1994), but 'clear differences between these two groups of off-spring emerged only in [adoptive] families who were rated as disturbed (Tienari, 1992, p.56; italics added).' These results suggest that both innate vulnerability and a disturbed rearing environment are necessary to produce the condition.

In the most rigorous study to date examining the relatives of adoptees with the disorder (Kety & Ingraham, 1992) found that 24 out of 46 adoptees with schizophrenia had one or more instances of the disorder among relatives, while the remaining 22 adoptees with the disorder had no instances of the disorder among relatives.

It also appears that a greater number first than second degree biological relatives of adoptees with schizophrenia share the disorder, and that adoptees with schizophrenia have more relatives who also have the disorder than adoptees in the comparison group (i.e., adoptees without the disorder (Kendler & Gruenberg, 1984; Tienari & Wynne, 1994). It needs to be emphasized, however, that the numbers are small in this sample.

A behavioral-genetic interpretation of findings from adoption studies is contingent upon our ability to address the following concerns, not addressed in previous work: (1) Timing and circumstances of adoption process itself should be controlled, because of the potential of early exposure to possible double binding influences of the biological family system on the adoptive system. Tienari (1992) specifies residence of four years or less as a criterion for being included in the sample, allowing still for the possibility of considerable variation between those adoptees who do versus those who do not have the disorder. (2) Evidence using medical records and interviews with relatives needs to be supplemented with direct observations of family interaction in adoptive families in their natural setting, to assess the degree to which double binds are present or absent to the same degree in adoptive families of adoptees with versus without schizophrenia in their biological families. Tienari et al.'s (1994) findings indicate that disturbance rating of the adoptive family is a critical variable in the comparison between the two groups. It therefore needs to be determined whether the interactive patterns in adoptive families is qualitatively different in those families whose adoptees have versus have no schizophrenia in their biological families; specifically, it needs to be determined to what extent a simultaneous affirmation and denial of vital relationships characterizes the interaction between adoptees and other members in the adoptive family system. (3) degree of exposure to interaction with the biological family after the adoption needs to be controlled, because of its potentially double binding nature: the ties to the biological and adoptive family could potentially contradict each other with respect to the adoptee's claim to family membership. (4) It is also important that the question of selective placement be addressed empirically, because any similarity between biological family and adoptees may be an expression of the effects of characteristics of the biological families on placement decisions (Baumrind, 1992; Lewontin, Rose & Kamin, 1984; Rowe, 1994).

Adoption is a potentially double binding situation in the narrow sense defined above (i.e., injunctions indicating that the adoptee belongs, yet does not belong to the adoptive family, Ward & Koopmans, 1996; 1997), and contact with the biological family makes it even more so. The extent to which this duplicity manifests itself in the interaction between adoptive family members is a question that needs to be resolved empirically. It also needs to be determined empirically to what degree there are differences in this regard between families who adopt someone from a schizophrenic vs. a nonschizophrenic family, and whether there was knowledge in the adoptive family about the mental health condition of the biological parents.

Double bind and behavioral-genetic theories differ in their interpretation of the finding that biological relatives share the disorder. From a double bind perspective, the fact that parent and child are engaged in double bind interactions does not require the assumption that they will both display other symptoms of the disorder.

Moreover, from the fact that biological parents and their adopted away offspring share the disorder, it does not necessarily follow that there has been a genetic transmission of the disorder (Lidz, 1994). It is possible that adoptee and adoptive parent(s) are engaged in double bind interactions to the detriment of the adoptee, while the biological parent will have schizophrenia as well. A behavioral genetic would argue for a genetic transmission process, while double bind theory would allow for the possibility that parental schizophrenia affects the likelihood of adoption, which is a potentially double binding situation for anyone involved, regardless of eventual vulnerability factors.

It is possible that an innate vulnerability to the disorder will be expressed provided that certain environmental conditions, such as, perhaps, double bind contingencies in family interaction, are met. Double bind theory does not require assuming such a vulnerability to be able to account for the findings of behavioral-genetic research, but does allow for it.  


Since both genetic and social transmission occur in families, family members who are in contact can share schizophrenia on account of either one, or both, types of transmission. Double bind theory specifies an environmental transmission process which would be able to account for high rates of schizophrenia within the family. It is important to remember, however, that double bind theory does not require that those who are engaged in double bind interactions will all share the disorder in any other sense than that their interaction is duplicitous. 'Healthy' parents can, for instance, under the pressure of circumstances, interact in a pathogenic fashion with their children.

Double bind theory would suggest that identical twins may share the disorder more often than fraternal twins because the former are double bound together more often than the latter. The reasons for possible more frequent double binding may be the greater genetic similarity of identical twins, but these more similar traits and behaviors are not necessarily related to the disorder, only to the concordance thereof. Adopted away offspring are in a potentially double binding environment, and may be even more so if contact with the biological family remains. Patterns of exposure to such contingencies may be different for those adoptees who have versus those who do not have the disorder. Medical records typically do not contain enough information to decide on such issues, and the interviews with relatives conducted in the Finnish and Danish adoption studies primarily served to diagnose them, rather than to obtain information about the adoption process itself.

Detailed observations of the interactive behavior in adoptive families, and families with identical versus fraternal twins is therefore needed to determine whether the assumptions we need to make for a behavioral-genetic interpretation of these findings are warranted. As they stand, the findings from pedigree, twin, and adoption studies, are not incompatible with double bind theory. The viability of a behavioral-genetic interpretation of these findings therefore critically depends on our ability to rule out such compatibility.

A detailed treatment of the methodological issues associated with a behavioral-genetic interpretation of findings from twin and adoption studies is beyond the scope of this paper, and can be found in Baumrind (1992), Cavalli-Sforza & Freeman (1973); Lewontin (1975), Lewontin, Rose, and Kamin (1984), Lidz, Blatt & Cook (1981), Plomin and Daniels (1987), and Plomin and Bergeman (1991). However, in light of the extensiveness of the ramifications of adoption and twinhood for the dynamical aspects of interactions within the systems of interest (Baumrind, 1992), the determination of genetic vulnerability to schizophrenia is better left to molecular - genetic research (National Advisory Mental Health Council, 1995; Prescott & Gottesman, 1993).  

Summary and Conclusion

Given the complexity of Bateson et al.'s proposal, it needs to be defined as a theory, which consists of at least twelve distinct components with a complex web of interrelationships between those components. Hypotheses derived from this theory typically address the relationship between some, but not all, of the components of the theory. It is important to raise this issue because it sheds a different light on the existing evidence concerning double bind: the accumulation of evidence which would lead one to disconfirm double bind as a hypothesis, addresses only one specific aspect of double bind as a theory, namely that schizophrenic and nonschizophrenic subjects will differ in their ability to detect double binds in paper and pencil tasks. Other double bind hypotheses, which arguably better operationalize double bind constructs, have fared somewhat better. In contexts in which a proper understanding of a message is essential, exposure to conflicting injunctions indeed produces high levels of discomfort and self-reported anxiety.

A connection still needs to be made, however, between these outcomes and schizophrenic symptomatology. Such a connection requires a more detailed analysis of subjects' responses in simulated settings, and longitudinal studies in natural settings, in which the ramifications of repeated exposure to conflicting injunctions can be assessed.

Evidence from family observation studies reveals greater interactional ambiguity in families, which include a schizophrenic member than in families which do not include a schizophrenic member, specifically between identified patients and their mothers. To determine how strong this result is, family observation studies need to work with larger segments of interaction than has been done so far, because the theory does not predict that double bind interactions to occur very often, only that they are functionally significant (Bateson, 1969/1972).  

Research on double bind theory has largely concentrated thus far on discrepancies between information verbally and non- verbally conveyed. There are many other communicative levels that can be considered as potentially conflicting, such as gestures, facial expression, ritual behaviors, fantasy and play, or literal vs. metaphoric meanings. The frequency of occurrence of conflicting injunctions across these levels is accessible to empirical investigation as well. More important, perhaps, is the exploration of the relational context of these conflicting injunctions (Abeles, 1976; Olson, 1972). Specifically, the possibility needs to be systematically investigated that a conflicting classification of relationships, as reflected in interactive episodes between family members, produces symptomatology under certain circumstances (Lidz, 1968; Lidz, 1984; Lidz & Fleck, 1985). When investigating double bind in a family context, Bateson et al.'s conceptualization of the outcome as perceiving the universe in double bind patterns could be fruitfully narrowed to the perception specifically of family relationships in double bind patterns. According to this definition, then, we can define double bind literally as a situation in which two conflicting ties are asserted between participants in communicative exchanges. To investigate how, and to what extent, double binds operate in families, there is a great need to study the behavior of families in their natural environment, where the hypothesized pathogenic processes actually take place.

The double bind literature has not articulated a clear position about where double bind interactions in the family originate. Since the theory makes an etiological claim with respect to schizophrenia, answering this question is an important aspect of its credibility. It is possible that incidental variations in interactive patterns, which would normally have little effect, may, in periods of family disruption, result in the establishment of new, and possibly double binding, patterns of interaction (Koopmans, 1996; Ward & Koopmans, 1996; 1997). Such a scenario connect pathogenic patterns of family interaction with the life events which typically create disruption in families such as the death of a family member, loss of a job, moving to a new neighborhood, and so on (King & Dixon, 1995). Weak spousal alliances, large power discrepancies between spouses, single parenthood, or lack of clearly defined parental roles could exacerbate the effects of external pressure and turbulence, in that such circumstances further affect the likelihood that adults turn to children to address their adult needs (Lidz & Fleck, 1985), thereby creating the double bind contingencies to which an adaptive response is required from the child or youngster.

There is currently agreement in the literature that the findings of behavioral-genetic research do not warrant ruling out environmental models of the etiology and course of schizophrenia (Kety & Ingraham, 1992; McGue & Gottesman, 1989; Tienari et al., 1994; National Advisory Mental Health Council, 1995). In fact, degree of disturbance in adoptive families appears to be a critical predictor of disturbance in the adoptees (Tienari et al., 1994), and it is estimated that somewhat less than half of the adoptees with schizophrenia do not appear to have any relatives who share the disorder (Kety & Ingraham, 1992).

The findings from adoption studies cannot be used to establish the extent of a genetic or social transmission, unless the assumption can be empirically confirmed that variations in adoption procedures, and in the adoption situation, can be held constant in the comparison between adoptees with and without schizophrenia in their biological families. Particularly in light of the potentially double binding nature of the adoption situation and of the adoption process, we need to know more about the circumstances under which the adoptions have taken place in the samples that have been used in these studies. Moreover, to better understand why level of disturbance in adoptive families is a predictor of schizophrenia in adoptees, patterns of interaction need to be observed in adoptive families in general, those whose adopted children have a history of mental illness in their biological families in particular.

In light of the apparent consensus about the genetic basis of schizophrenia (Bailey & Pillard, 1993), the ability of double bind theory to challenge those very findings on which this consensus is based should inspire researchers to address the extent to which double binding occurs in adoptive families and families with twins vigorously. Such research should be part of a more comprehensive assessment of double binding in the nuclear family, and its ramifications for individual members, children and adolescents in particular.


Abeles, G. (1975) The double bind: Paradox in relationships.
       Doctoral Dissertation, Boston University.
Abeles, G. (1976) Researching the unresearchable: Experimentation
       on the double bind. In C. E. Sluzki & D. C. Ransom (Eds.)                           
       Double bind: The foundation of the communicational  approach
       to the family. New York: Grune & Stratton. Pp. 113-149.
Ackerman, B. L. (1979) Relational paradox: Toward a language of                            
       interactional sequences. Journal of Marital and Family                              
       Therapy, 5, 29-38.
Ackerman, N. W. (1958) The psychodynamics of family life: Diagnosis
       and treatment of family relationships. New York: Basic.
Aoki, M. (1993) Influences of double bind communication on human
       information processing system and emotion. Japanese Journal of
       Educational Psychology, 41, 31-39.
Bailey, J. M. .& Pillard, R. C. (1993) In reply to Lidz. Re: A
       genetic study of male sexual orientation. Archives of General
       Psychiatry, 50, 240-241.
Bateson, G. (1969/1972) Double bind 1969. In G. Bateson Steps  to
       an ecology of the mind. New York: Ballantine. Pp. 271-278.
Bateson, G., Jackson, D. D., Haley, J. & Weakland, J. (1956) Toward
       a theory of schizophrenia. Behavioral Science, 1, 251-264.
Baumrind, D. (1993) The average expectable environment is not good
       enough: Response to Scarr. Child Development, 64, 1299-1317.
Beakel, N. G. & Mehrabian, A. (1969) Inconsistent communication in
       psychopathology. Journal of Abnormal Psychology, 74, 126-130.
Beavers, W. R., Blumberg, S., Timken, K. R. & Weiner, M. F. (1965)
       Communication patterns of mothers of schizophrenics.                         Family
       Process, 4, 95-104.
Bebbington, P. & Kuiper, L. (1988) Social influences on
       schizophrenia. In P. Bebbington & P. McGuffin (Eds.)
       Schizophrenia: The major issues. Oxford: Heinemann Medical
Berger, A. (1965) A test of the double-bind hypothesis of                                  
       schizophrenia. Family Process,  4, 198-205.
Biddle, B. J. (1979) Role Theory: Expectations, Identities, and
       Behaviors. New York: Academic Press.
Blotchky, A. D., Tittler, B. I. & Friedman, S. (1982) The double-
       bind situation in families of disturbed families. Journal of
       Genetic Psychology, 141, 129-142.
Broderick, C. B. (1993) Understanding family process: Basics of
       family systems theory. Newbury Park, CA: Sage.
Brown, G. W., Birley, J. L. T. & Wing, J. K. (1972) Influence of
       family life on the course of schizophrenic disorders: A                             
       replication. British Journal of Psychiatry, 121, 241-258.
Bugenthal, D. E., Love, L. R., Kaswan, J. W.  & April, C. (1971) 
       Verbal-nonverbal conflict in parental messages to normal and
       disturbed children. Journal of Abnormal Psychology, 77, 6-10.
Cavalli-Sforza, L. L. & Feldman, M. W. (1973) Cultural versus 
       biological inheritance: Phenotypic transmission from parents
       to children. American Journal of Human Genetics, 25, 618-637.
Ciotola, P. V. (1961) The effects of two contradictory levels of
       reward and censure on schizophrenics. Doctoral Dissertation,
       University of Missouri. University Microfilms No. 61-2278.
Dush, D. M. & Brodsky, M. (1981) Effects and implications of                               
       the experimental double bind. Psychological Reports, 48, 895-
Farber, S. L. (1981) Identical twins reared apart: A  reanalysis.
       New York: Basic Books.
Ferris, W. P. (1975) The incidence and prevention of double binding
       communication patterns in the secondary school. Doctoral
       Dissertation, Rensselaer Polytechnic Institute. Dissertation
       Abstracts 76-3694.
Gelinas, D. J. (1978) Conceptual frameworks in psychology: A
       modified Kuhnian analysis of the emergence of double bind
       family therapy. Doctoral Dissertation, University of
       Massachusetts, Dissertation Abstracts 78-16,261.
Gooblar, H. M. (1978) Double bind: Aspects of the communicational
       style of the father-absent family. Doctoral Dissertation,
       California School of Professional Psychology, Dissertation
       Abstracts 79-01,804.
Gottesman, I. I. (1991) Schizophrenia genesis: The origins of                              
       madness. New York: Freeman.
Gottesman, I. I., McGuffin P. & Farmer, A. E. (1987) Clinical
       genetics as clues to the "Real" genetics of schizophrenia (A
       decade of modest gains while playing for time). Schizophrenia
       Bulletin, 13, 23-47.
Guindon, J. E. (1971), Paradox, schizophrenia and the double bind
       hypothesis: An exploratory study. Doctoral Dissertation,
       University of Washington.
Gundersen, J. H. (1988) The theory of logical types: A cul-de-sac
       in communication theory. Nordisk Psykiatrisk Tidsskrift, 42,
Haake, M. L. (1984) Paradoxical communication reconsidered.
       Doctoral Dissertation, University of Massachusetts,
       Dissertation Abstracts 84-10259.
Hall, J. A. & Levin, S. (1980) Affect and verbal-nonverbal                                 
       discrepancy in schizophrenic and on-schizophrenic family                            
       communication. British Journal of Psychiatry, 137, 78-92.
Haller, K. A. (1989) The double bind: Paradox in Human Behavior.
       Doctoral Dissertation, Teachers College/Columbia University,
       Univ. Microfilms 90-02502.
Hargreaves Heap, S. P. & Varoufakis, Y. (1995) Game theory: A
       critical introduction. London: Routledge.
Heilveil, I. P (1980) Verbal and nonverbal channel congruency in
       schizophrenia. Doctoral Dissertation, California School of
       Professional Psychology, Univ. Microfilms 81-14759
Humphrey, L. L. & Benjamin, L. S. (1986) Using structural analysis
       of social behavior to assess critical but elusive family
       processes: A new solution to an old problem. American
       Psychologist, 41, 979-989.
Jackson, D. D. (1960) A critique of the literature on the genetics
       of schizophrenia. In  Jackson, D. D. (Ed.) The Etiology of                          
       schizophrenia. New York: Basic Books.
Karr, K. B. (1984) Source and recipient perceptions of double bind
       patterns. Doctoral Dissertation, Arizona State University,
       Dissertation Abstracts 84-05155.
Kavanagh, D. J. (1992) Recent developments in expressed emotion and
       schizophrenia. British Journal of Psychiatry, 160, 601-620.
Kellogg, L. C. (1980) The effect of double bind conflict on
       interaction perceptions of college students. Doctoral
       Dissertation, University of Cincinnati, Univ. Microfilms 81-
Kendler, K. S. (1983) Overview: A current perspective on twin
       studies of schizophrenia. American Journal of Psychiatry, 140,
Kendler, K. S. (1993) Twin studies of psychiatric illness: Current
       status and future directions. Archives of General Psychiatry
       50, 905-915.
Kendler, K. S. & Gruenberg, A. M. (1984) An independent analysis
       of the Copenhagen sample of the Danish Adoption study of
       Schizophrenia: The pattern of psychiatric illness as defined
       by DSM-III-R in adoptees and relatives. Archives of General
       Psychiatry, 41, 555-564.
Kerlinger, F. N. (1973) Foundations of Behavioral Research 2nd
       Edition. London: Holt, Rinehart & Winston. 
Kety, S. S. (1988) Schizophrenic illness in the families of
       schizophrenic adoptees: Findings from the Danish National
       Sample. Schizophrenia Bulletin, 14, 217-222.
Kety, S. S. & Ingraham, L. J. (1992) Genetic transmission and
       improved diagnosis of schizophrenia from pedigrees and
       adoptees. Journal of Psychiatric Research., 26, 247-255.
Kety, S. S., Rosenthal, D., Wender, P. H. & Schulsinger, F. (1968)
       The types and prevalence of mental illness in the biological
       and adoptive families of adopted schizophrenics. In D.
       Rosenthal & S. S. Kety (Eds.) The transmission of
       schizophrenia. Oxford: Pergamon. Pp. 345-362.
King, S. & Dixon, M. J. (1995) Expressed emotion, family dynamics,
       and symptom severity in a predictive model of social
       adjustment for schizophrenic young adults. Schizophrenia
       Research, 14, 121-132.
Kingsley, V. C. (1969) The effects of the double bind conflict and
       sex of the experimenter on the conceptual functioning and  
       visual discrimination  of male good and poor premorbid
       schizophrenics. Doctoral Dissertation, New York Univ., Univ.
       Microfilms 70-16,078.
Klessman, E. & Klessman, H. A. (1983) Anorexia Nervosa - Eine                              
       therapeutische Beziehungsfalle? Ein Fazit nach 13 Jahren                            
       ambulanter Therapie. Praxis der Kinderpsychologie und                               
       Kinderpsychiatrie, 32, 257-261.
Koopmans, M. (1994) Self reports of sexual abuse, mental health                            
       status, and perceived confusion of family roles. Psychological
       Reports, 75, 339-347.
Koopmans, M. (1995) Applicability of a family systems paradigm to
       teenage suicidal behavior.  Adolescence, 30, 87-94.
Koopmans, M. (in press) Chaos theory and the problem of change in
       family systems. To appear in Nonlinear Dynamics, Psychology,
       and Life Sciences.
Krefting, L. (1990) Double bind and disability: The case of
       traumatic head injury. Social Science Medicine, 30, 859-865.
Kringlen, E. (1991) Adoption studies in functional psychosis.
       European Archives of Psychiatry and Clinical Neuroscience,
       240, 307-313.
Kuiken, D. & Hill, K. (1985) Double-bind communications and                                
       respondents' reluctance to affirm the validity of their
       self-disclosures. Perceptual and Motor Skills, 60, 83-95.
Laing, R. D. (1961) The Self and others: Further studies in
       sanity and madness. London: Tavistock.
Laing, R. D. & Esterson, A. (1964) Sanity, madness, and the
       family. Harmondsworth: Pelican.
Lewontin, R. C. (1975) Genetic aspect of intelligence. Annual
       Review of Genetics, 9, 387-405.
Lewontin, R. C., Rose, S. & Kamin, L. J. (1984) Not in our genes:
       Biology, ideology, and human nature. New York: Pantheon
Lidz, Th. (1968) The family, language, and the transmission of
       schizophrenia. In D. Rosenthal & S. S. Kety (Eds.) The
       transmission of schizophrenia. Oxford: Pergamon. Pp. 175-
Lidz, Th. (1973) The origin and treatment of schizophrenia. New
       York: Basic.
Lidz, Th. (1993) "A genetic study of male sexual orientation":
       Reply. Archives of General Psychiatry, 50, 240.
Lidz, Th. (1994) Nature vs. nurture in schizophrenia: The
       struggle continues. Integrative Psychiatry, 2. 144-145.
Lidz, Th., Blatt, S. & Cook, B. (1981) Critique of the Danish-
       American studies of the adopted away offspring of
       schizophrenic parents. American Journal of Psychiatry, 138,
Lidz, T. & Fleck, S. (1985)  Schizophrenia and the family Second
       Edition. New York: International Universities.
Loeff, R. G. (1965) Differential discrimination of conflicting
       emotional messages by normal, delinquent, and schizophrenic
       adolescents. Doctoral Dissertation, Indiana University,
       University Microfilms No. 66-1470.
Lytton, H. (1977) Do parents create, or respond to, differences
       in twins? Developmental Psychology, 13, 456-459.
McDonough, J. S. (1975) An investigation of the effect of
       double bind instructions on perceptual functioning as a
       function of maternal child-rearing behaviors. Doctoral
       Dissertation, University of Arkansas, Dissertation Abstracts
       75-27, 661.
McGue, M. & Gottesman, I. I. (1989) Genetic linkage in                                     
       schizophrenia: Perspectives from genetic epidemiology                               
       Schizophrenia Bulletin, 15, 453-464.
Mehrabian, A. & Wiener, M. (1967) Decoding of inconsistent                                 
       communications. Journal of Personality and Social
       Psychology, 6, 109-114.
Melowski, F. (1978) Escaping double bind: Effects of locus of
       control and double bind information. Doctoral Dissertation,
       Texas Tech University, Dissertation Abstracts 78-19,895.
Miklowitz, D. J. (1994) Family risk indicators in
       schizophrenia. Schizophrenia Bulletin, 20, 137-149.
Miklowitz, D. J. (1995) The evolution of family based 
       psychopathology. In R. H. Mikesell, D. D. Lusterman & S. H.
       McDaniel (Eds.) Integrating family therapy: Handbook of
       family psychology and systems theory. Washington, DC:
       American Psychological Association (Pp. 183-197)
Minuchin, S. & Fishman, H. C. (1981) Family therapy techniques.
       Cambridge, MA: Harvard.
Mishler, E. G. & Waxler, N. E. (1965) Family interaction and                               
       schizophrenia: A review of current theories. Merril-Palmer                          
       Quarterly, 14, 269-315.
Mishler, E. G. & Waxler, N. E. (1968) Interaction in families: An
       experimental study of family processes and schizophrenia.
       New York: Wiley.
National Advisory Mental Health Council (1995) Basic behavioral
       science research for mental health: a national investment.
       Bethesda, MD: U.S Department of Health and Human Services.
Newman, E. H. (1977) Resolution of inconsistent attitude
       communications in normal and schizophrenic subjects. Journal
       of Abnormal Psychology, 86, 41-46.
Nida, R. A. (1976) The double bind theory of communication as it
       relates to communication apprehension and associated
       personality development. Doctoral Dissertation, Ohio
       University, Dissertation Abstracts 77-3489.
Olson, D. H. (1972) Empirically unbinding the double bind: A
       review of research and conceptual formulations. Family
       Process, 11, 69-94.
Onstad, S., Skre, I., Torgersen, S. & Kringlen, E. (1991) Twin
       concordance for DSM III-R schizophrenia. Acta Psychiatrica
       Scandinavica, 83, 395-401.
Pardes, H., Kaufman, C. A., Pincus, H. A. & West, A. (1989).                               
       Genetics and psychiatry: Past discoveries, current dilemmas,
       and future directions. American Journal of Psychiatry,
       146, 435-443.
Philips, M. (1970) Response to "Double Bind" messages in relation
       to four dimensions of personality and to two maternal                        child
       rearing attitudes: A study of interpretational and feeling
       preferences of late adolescents in response to "Double                              
       Binding"  type communications. Doctoral Dissertation, New
       York Univ., Univ. Microfilms 70-26,439.
Pine, J. J. (1984) The double bind and some of its implications.
       Doctoral Dissertation, University of Illinois at Urbana-
       Champaign, Dissertation Abstracts 84-09,982.
Plomin, R. & Bergeman, C. S. (1991) The nature of nurture:
       Genetic influence on "environmental" measures. Behavioral
       and Brain Sciences, 14, 373-427.
Plomin, R. & Daniels, D. (1987) Why are children in the same
       family so different from one another? Behavioral and Brain
       Sciences, 10, 1-60.
Potash, H. M. (1965) Schizophrenic interaction and the concept of
       the double bind. Doctoral Dissertation, Michigan State
       Univ., Univ. Microfilms 65-2052.
Pratt, D. M. & Hansen, J. (1987) A test of the curvilinear
       hypothesis with FACES II and FACES III. Journal of Marital
       and Family Therapy, 13, 387-392.
Prescott, C. A. & Gottesman, I. I. (1993) Genetically mediated
       vulnerability to schizophrenia. Psychiatric Clinics of North
       America, 16, 245-267.
Ratner, C. (1982) Do studies on identical twins prove that
       schizophrenia is genetically inherited? International
       Journal of Social Psychiatry, 28, 175-178.
Reiss, D., Plomin, R. & Hetherington, E. M. (1991) Genetics and
       psychiatry: An unheralded window on the environment.
       American Journal of Psychiatry, 148, 283-291.
Relph, A. (1991) Family therapy and the theory of logical types.
       Australian and New Zealand Journal of Family Therapy, 12, 1-
Rierdan, J. & Brooks, R. (1977) Verbal conditioning of middle and
       lower socioeconomic class schizophrenics. Journal of
       Abnormal Psychology, 86, 369-378.
Ringuette, E. L. & Kennedy, T. (1966) An experimental study of
       the double bind hypothesis. Journal of Abnormal Psychology,
       71, 136-141.
Rosenthal, D. (1970) Genetic theory and abnormal behavior. New
       York, NY: McGraw-Hill.
Rosenthal, D., Wender, P. H., Kety, S. S., Schulsinger, F.,
       Welner. J. & þstergaard (1968) Schizophrenics' offspring
       reared in adoptive homes. In D. Rosenthal & S. S. Kety
       (Eds.) The transmission of schizophrenia. Oxford: Pergamon.
       Pp. 377-392.
Rotzien, M. K. (1989) A critique of the double bind system:
       Theory, methodology, and research. Doctoral Dissertation,
       Fuller Theological Seminary, School of Psychology, Univ.
       Microfilms 89-24359.
Rowe, D. (1994) The limits of family influence: Genes,
       experience, and behavior. New York: Guilford.
Roy, L. & Roy, S. (1987) Does the theory of logical types inform
       a theory of communication? Journal of Genetic Psychology,
       148, 519-526.
Schreiber, A. (1970). An experimental double bind and
       communicativeness. Doctoral Dissertation, City University of
       New York.
Schuham, A. I. (1967) The double bind hypothesis a decade later.
       Psychological Bulletin, 68, 409-416.
Shoham, S. S., Weissbrod, L., Markowsky, R. & Stein, Y. (1977)
       The differential pressures toward schizophrenia and
       delinquency. Genetic Psychology Monographs, 96, 165-210.
Sluzki, C. E., Beavin, J., Tarnopolsky, A. & Ver˘n, E. (1967).                             
       Transactional disqualification  research on the double bind.
       Archives of General Psychiatry, 16, 494-504.
Sluzki, C. E. & Ransom, D. C. (1976)  Comment on Abeles' review.
       In C. E. Sluzki & D. C. Ransom (Eds.) Double bind: The
       foundation of the communicational  approach to the family.
       New York: Grune & Stratton. Pp. 151-164.
Smith, E. K. (1973) The effect of double bind communication upon
       the state anxiety of normals. Doctoral Dissertation,
       University of New Mexico.
Smith, E. K. (1976) Effect of double-bind communication on the                             
       anxiety level of normals. Journal of Abnormal Psychology,
       85, 356-363.
Sojit, C. M. (1969) Dyadic interaction in a Double Bind
       situation. Family Process, 8, 235-259.
Sojit, C. M. (1971) The Double Bind hypothesis and the parents 
       of schizophrenics. Family Process, 10, 53-74.
Steinberg, L. (1990) Autonomy, conflict, and harmony in the
       family relationship. In  S. S. Feldman & G. R. Elliot (Eds.)
       At the threshold: The developing adolescent. Cambridge, MA:
       Harvard University Press. (Pp. 255-276) 
Summers, F. & Walsh, F. (1977) The nature of the symbiotic bond
       between mother and schizophrenic. American Journal of
       Orthopsychiatry, 47, 484-494.
Tienari, P. (1992) Implications of adoption studies on                                     
       schizophrenia. British Journal of Psychiatry, 161 (Suppl.
       18), 52-58.
Tienari, P. & Wynne, L. C. (1994) Adoption studies of
       schizophrenia. Annals of Medicine, 26, 233-237.
Tienari, P., Wynne, L., Moring, J., Lahti, I., Naarala, M.,
       Sorri, A., Wahlberg, K. E., Saarento, O., Seitaama, M.
       Kaleva, M. & L„ksy, K. (1994) The Finnish adoptive family
       study of schizophrenia: Implications for family research.
       British Journal of Psychiatry, 164 (suppl. 23), 20-26.
Vandenberg, S. G. (1976) Twin studies. In A. R. Kaplan (Ed.)
       Human behavior genetics. Springfield, Ill: C. C. Thomas.
Vaughn, C. E. & Leff, J. P. (1976) The influences of family and
       social factors on the course of psychiatric illness: A
       comparison of schizophrenic and depressed neurotic patients.
       British Journal of Psychiatry, 129, 125-137.
Vetter, H. (1969) Language, behavior and psychopathology.
       Chicago: Rand McNally. Pp. 167-180.
Ward, M. & Koopmans, M. (1996, November) Chaos and nonlinear
       perspectives on family processes: the case of older child
       adoption. Paper presented at the meeting of the National
       Council on Family Relations, Kansas City, MO.
Ward, M. & Koopmans, M. (1997, August) Dynamic pattern formation
       in families: II. Nonlinear and chaotic processes in Older-
       Child adoption. Paper presented at the annual meeting of the
       Society for Chaos Theory in Psychology and Life Sciences,
       Milwaukee, WI.
Watzlawick, P. A. (1963) A  review of the double bind theory.
       Family Process, 2, 132-153.
Watzlawick, P. (1965) Paradoxical predictions. Psychiatry, 28,
Watzlawick, P., Beavin, J. H. & Jackson, D. D. (1967) Pragmatics
       of Human Communication. New York: Norton.
Weakland, J. H. (1974) The double bind theory in self reflective
       hindsight. Family Process, 13, 269-277.
Weakland, J. H. (1992) The development and significance of the
       double bind theory.  Japanese Journal of Family Psychology,
       6, 25-38.
Whitehead, A. N. & Russell, B. (1962) Principia Mathematica                                
       Abridged version. Cambridge: Cambridge University Press.
Wild, C. M. & Shapiro, L. N. (1977) Mechanisms of change from
       individual to family performance in male schizophrenics and
       their parents. Journal of Nervous and Mental Disease, 165,
Wild, C. M., Shapiro, L. N. & Goldenberg, L. (1975) Transactional 
       communication disturbances in families of male
       schizophrenics. Family Process, 14, 131-160.
Wooley, M. J. & Vigilanti, M. A. (1984) Psychological separation
       and the sexual abuse victim. Psychotherapy, 21, 347-352.
Wynne, L. C., Ryckoff, I. M. & Day, J. (1958) Pseudo-mutuality in
       the family relations of schizophrenics. Psychiatry, 21, 205-