Articles Written

Relational Paradox: Toward a Language of Interactional Sequences*

Brian L. Ackerman, M.D.**

The author discusses and clarifies the “Theoryof Logical Types” which forms the cornerstone of the “DoubleBind Hypothesis.”A clear distinction is made between single sentence double binds and double binds that evolve in an interaction over time. The crucial term “metacommunication”is redefined in a manner consistent with logical type theory. The locus of message-levelconfu- sionisshiftedfromthemessagesof thesendertothearenaof thecommunicant‘s interaction. Analysis of the interactional arena reveals a phenomenon which the author labels “Relational Paradox.” Several examples are given which illustrate how paradoxes can arise even if neither interactant is sending messages which are intrinsically bewildering.

On March 3rd and 4th, 1977,I attended an “historic”conferencein New York City entitled “Beyond the Double Bind.” The conference served the unique function not only of bringing together the authors of the “double bind theory” twenty years after its inception, but also of bringing together other original thinkers in the field of communications and family therapy who have been looking at similar phenomena from different perspectives.

The panel consisted of Gregory Bateson, Jay Haley, John Weakland, Murray Bowen, Albert Scheflen, Carl Whitaker, and Lyman Wynne. Bateson, Haley and Weakland (1956)werethreeoftheauthorsoftheseminalarticle“TowardaTheoryof Schizophrenia,” (The fourth author, Don Jackson, died in 1968).The article, which has served as a major impetus for the growth of family therapy, grew out of the work done in a research project under the direction of Bateson begun in 1954.This research group approached the phenomenon of schizophrenic communication from what was then a radically new viewpoint. Rather than characterize schizophrenia as a primarily intrapsychic disorder that secondarily affects interpersonal relationships, they won- dered what sequences of interpersonal experiences would induce schizophrenic behavior, and whether this behavior might actually be adaptive to its bewildering context.

The conference format included presentation of new papers, live family interviews, videotapes, and general discussion. Among the new papers were “Double Bind and Epistemology” (Bateson), “Levels of Double Binding” (Scheflen), “Meta-binding and Unbinding” (Wynne), and “Pursuing the Evident into Schizophrenia and Beyond” (Weakland). The conference was inspired by the presence of Margaret Mead, who contributed a number of poignant comments on family communication problems in

*This paper was presented as a plenary address at the Second International Congress of Family Therapy, Jerusalem, Israel, August 1977.

**Dr. Ackerman is Instructor in Psychiatry, University of Massachusetts Medical School, Directorof InserviceTraining,North Central MassachusettsMental Health Center, and a faculty member of the Cambridge Family Institute.


other cultures. The renewed effort to examine communicationdilemmas from a variety of perspectives-clinical practice, psychiatric research, anthropology, and epistemol- ogy-resulted in a n exhilarating family therapy conference.

Although the conference was successful in many respects, I felt the discussion often got bogged down in trying to “go beyond” the double bind theory. For the theory itself suffers from terminological confusion, which frequent reformulations have failed to resolve. In particular, the term “double b i n d has been used to mean several things indiscriminately:

(1) the situation in which a paradoxical injunction--“Be spontaneousY-cwurs; (2) the paradoxical injunction itself;
(3) the theory which postulates the occurrence of paradoxical injunctions and

their induction of schizophrenicbehavior;
(4) an interactional sequence where the relationship of the interactants itself is

part of the paradox.
The “double bind” cannot possibly be clear if it refeh to all these things

simultaneously. In order to unravel the terminologicalconfusion,Iwould like to return to the “Theory of Logical Types,” which forms the cornerstone of the original “Double Bind Hypothesis,” and which, I believe, offers the best hope of getting beyond it.

With reference to the four meanings of ‘double bind‘ listed above, a distinction should be drawn between paradoxical injunctions and what shall be referred to as relational paradoxes. I hope to show that while the original authors of the doublebind hypothesis succeeded in showing how paradoxical injunctions can be illuminated by the theory of logical types, they (and the revisionists) have failed to show how the theory of logical types can be used to analyze the manner in which members of a relationship actively participate in the binding process.

The original authors failed to distinguish between two important communication dilemmas. In each of these dilemmas, logical-type confusion plays a crucial role. In paradoxical injunctions, the logical-type error is buried in a message; in a relational paradox the logical-type error is buried in an interactional sequence.

Oncethe distinction between paradoxical injunctions and relational paradoxeshas been made, it ought to be clear, for clinical purposes, that the importance of paradoxical injunctions is secondary and that the power of the theory of logical types (and of the double bind theory derived from it) resides in its capacity to explicate relational paradoxes.

The Theory of Logical Types

The theory of logical types originates in the field of mathematics and logic. It was introduced by Bertrand Russell and Albert North Whitehead (1910)in a monumental work entitled Principia Mathematica. One essential axiom is that “whatever involves all of a collection (class) must not be one of the collection (member).” There is a discontinuity between a class and its members. A ‘class’is on a different level of abstraction, and, therefore, is said to be of a different logical type, than a ‘member.’

Prior logical theory was flawed by a number of embarrassing paradoxes; the theory of logical types was invented to establish a stable logic that avoided the paradoxes. One famous paradox that inspired the theory of logical types is known as Russell’sparadox. Since i t can be rather difficult to comprehend, it seems advisable to take it one step a t a time.

(1) Russell distinguishes two kinds of classes: classes which are members of themselves; and
classes which are not members of themselves.


The commonest type is classes which are not members of themselves. For example, the class of shoes is not itself a shoe, the class of dogs is not itself a dog, the class mankind is not itself a man. But there are also certain classes that are members of themselves. For example, the class of all conceptsis a concept; the class of all thinkable thoughts is a thinkable thought.

(2) Let N = the class of all classes that are not members of themselves.
(3) Ifwenowposethequestion:IsclassNamemberof itself,disasterstrikes.We run straight into Russell’s paradox. For if N is a member of itself, then it is not a member of itself. This startling paradox is not artificial; it is

straightforward logical deduction.
These kinds of paradoxes are generated if we talk about classes as if they were at

the same level of abstraction as their members, and if truth at one level entails falsehood at another.’ Russell suggested that to avoid these paradoxes, the logical system itself must be adjusted. Two levels of language must be distinguished, a n object language for the members and a metu-languagefor the classes. Then classes are defined in such a way that to be a member of a class, something must be of a lower level of abstraction than that of the class itself. Thus the possibilities of classes being members of themselves can be eliminated. (Watzlawick,Beavin, Jackson, 1967).

The Double Bind Hypothesis

The major insight of Bateson’s seminal article was the observation that in human communication the discontinuity between class and member is continually breached. The “Double Bind Hypothesis” attempted to illustrate the consequent paradoxical phenomena in human interaction. The article shows how peculiar impasses arise when messages structured precisely like classical paradoxes in formal logic (such as “Be spontaneous!”) are exchanged. The originators of the “Double Bind Theory” hypothe- sized that schizophrenic behavior might be an adaptive response to recurrent message- level confusion contained in parent-child communication. Their analysis suggested a causal sequence involving senders of hopelessly bewildering messages and their distraught victims.

Revisionists of the “Double Bind Theory” have attempted to “unbind the linear cause and effect model which views one person as a victim of another‘s communication. Revisionists have also underscored the importance of the relationship as a crucial feature of the double bind. However,the original theory and its subsequent revisions have suffered from three major problems.

1) Terminological inconsistency as to the meaning of the term ‘meta- communication’;

  1. 2)  Failure to show precisely how the relationship is involved in the binding process;

  2. 3)  Failure to explain paradoxical communication dilemmas other than those that arise when message-level confusion is present in single statements.

In what follows, various human communication dilemmas-paradoxical injunc- tions, contradictory messages, meta-communication-will be reexamined through the lens of logical-type theory in an attempt to develop a consistent terminology, out of which will emerge a new paradigm, “Relational Paradox.”

Paradoxical Injunctions

Paradoxical injunctions served as the prototype for the “double bind.” One famous example is the injunction to “Be spontaneous!,”which is demanding a behavior which by its very nature cannot be spontaneous because it is commanded. (Thestatement can


be paraphrased as “I command you to be spontaneous.”)Analysis shows that a logical- type confusion is buried in the message. Specifically, the imperative classificatory assertion “Icommand ...”functions at a logical level different from that of the actual request made.’ By blurring the logical-typedistinction, the speaker is able to impose, with seeming innocence, the dreadful situation in which rule compliance would entail rule violation. (Watzlawick, Weakland, Fisch, 1974).

Another example of a paradoxical injunction is the sign that reads, “Don’t read this sign.” Again there is a logical-typeconfusion buried in the message. On one logical level is the implied message of the sign qua sign: ‘Here is something to be read.’ On another level is the negative message itself, (‘Don’t read. . .’) which subverts the sign’s normal function. To put the matter in Russellian class-member terminology, the message of this particular sign paradoxically enjoins the message of the sign as a member of the class of objects that are usually read. In the trap of self-reference,the boundary between class and member gets blurred, generating the paradox. The two levels of the injunction are pragmatically incompatible with each other. Once again, compliance at one level entails violation at another.

Contradictory Messages

In order to understand paradoxical injunctions more clearly, it is important to distinguish them from contradictory messages. In her review of double bind theory, Abeles (1975,p. 209)emphasizes the importance of this distinction.

It is essential to distinguish between paradox and other kinds of contradictions and incongruencies since the double bind is so often interpreted as meaning inconsistent communicationor contradictory messages and the like. Unless such definitions further specify that the contradiction occurs between levels of abstraction or different logical types, the definition is one of simple contradiction rather than paradox.

To avoid confusion, I suggest that ‘Contradictory messages’ be thought of as two messages, mutually inconsistent, but at the same level of abstraction. For example, a driver who comes to an intersection and encounters a stop sign and a green light is faced with instructions that are on the same level of abstraction, but that directly contradict each other.

Partial contradictions may involve inconsistencies between the verbal content of a message and the various qualifying messages conveyed by voice tone, inflection, facial expression, gesture, and context. These qualifying messages do not necessarily function on a different logical level than that of the verbal messages which they accompany; they may be simply consistent or inconsistent. For example, a husband would be sending inconsistent messages to his wife if he were to say to her in a half- hearted tone, “I’d really like to go to the movies.” Of course, qualifying messages and verbal messages can mesh perfectly well. For example, there are no contradictory messages when a mother makes a punitive statement in an angry tone, with a gesture that suggests that she means what she says in a situation where her statement is appropriate.

In response to contradictory messages, a person may choose to heed one of the messages and ignore the other. In response to partially contradictory messages, a person may give more or less weight to the verbal message or the qualifying message, or he may ask for clarification (“You said you were not mad, but you sound angry”). In response to paradoxical injunctions, however, the recipient is faced with two conflicting levels of message; no choice is possible any more than one can choose between a ‘fried egg’and‘truth!’(Sluzki&Ransom,1976).Thereader ofthesign,‘Don’treadthissign’is caught the moment the sign is read; clarification does not avoid the paradox. Paradoxical injunctions, thus, bewilder the recipient by paralyzing effective response.



Lack of clarity as to what constitutes meta-communication is central to the terminological inconsistency of the double bind theory. The term “meta- communication” generally refers to “communication about communication.” In the original hypothesis it was suggested that the “victim”is ‘hund“because he is unable to ffmeta-communicate”or comment on the intrinsically bewildering statements to which he is subjeded. A double bind is said to occur if no such comment (meta- communication) is pragmatically possible.

In a later article Haley (1959,p. 362)used the term meta-communicate as follows: “Whenever any piece of communication is about or qualifies another piece of communication they can be said to be of different (logical)levels.” The problem with Haley’s usage is that it is too general. The fact that one message (statement) comments upon, “qualifies,” or is “about” another does not entail that the two messages are of different logical types.

Comments which raise logical type issues need to be distinguished from comments which clearly proceed at the same logical level. Below is an interchange which specifically exploits logical type distinctions.

Person 1: “You’re angry.” Person 2: (angrily) “No, I’m not!” Person1: “Yousee!”

The first two statements function on the same logical level. One person makes an assertion and the other denies it. The comment ‘‘You see!” functions differently. It implies that person 2’s remark is an instance (lower logical type) of the previous generalization about his feeling state (higher logical type).

It appears that Haley’s very general definition of meta-communication blurs the distinction between comments in general and that subset of comments in which questions are raised about the logical status of previous comments. Given that logical- type theory is the cornerstone of the double bind hypothesis, it would appear more economical and consistent to choose a definition of meta-communication that confines the term to comments that raise logical-typeissues. The followingdefinition of meta- communication is suggested:

Meta-communication:Message A is “meta”to message B only if it classifiesmessage B, thereby relegating the message it classifies to the status of a member.

The comment “You see!” in the example above reclassifies the statement “No, I’m not!”, by treating it as an instance of person 2’s alleged anger. The message “You see!” is “meta” to the message “No, I’m not!” More generally, the proposed definition of meta-communication clarifies the possibility of a hierarchy in the logical relationships among ~taternents.~

Toward a Language of Interactional Sequences

Translated into an interactional language, use ofthe term meta-communicationas suggested illuminates the manner in which a relationship may be involved in a “double bind.” The two levels of language, classificatory-level assertions and member-level utterances may be in constant oscillation during an interactional sequence. In general, comments within a relationship may be said to be bivalent, for example, when a husband asks his wife, “Did you go shopping?’, regardless of his intention, the remark may be interpreted as a straightforward question or as some sort of indirect comment on the definition of the wife’s role. The communication may not intrinsically specify one level or the other; in most interactions, there is a tacit agreement about the intended level.


Interactional sequences often become problematic when the intended message level and interpreted message level are out of harmony. For example,if a husband says to his wife, “The lights were left on again!”, a quarrel might ensue, not over the correctness of the information conveyed,or the manner in which it was expressed,but over the implication that she is being classified as a careless and irresponsible wife (and that leaving the lights on again is an example of behavior meriting this classification). The metamessages, “View me as responsible,”and “View my telling you about the lights” as an example of why “You should regard me as being responsible,” have the dimension of the relationship as a referent, not the lights. The meta-message at the relationship level is, “This is how you are to see me in relationship to you-I’m responsible-you’re irresponsible.” Suppose the wife responded by saying, “Damn it, would you stop criticizing me!” By this statement she would reclassify the husband’s alleged ‘responsible behavior’-pointing out the burning lights--as an example of his irresponsible recurrent criticisms of her, thereby defining the relationship as one where he is a critic and she is a victim. As Jackson (1968)suggests, an interaction can be defined as two or more communicants in the process of, or at the level of defining the nature of their relationship. The communications of both members are crucial because they have a reciprocally influencing effect on the outcome of interactional sequences.

In addition, a relationship usually operates within certain limits or parameters, in that “rules for the relationship” and an “internal homeostatic set point” become established over time (Jackson, 1968).If one member’s communicationbegins to offset the established balance, the other usually acts to bring it back in line. When the attempts at re-equilibration backfire, the interactional sequencebecomes bogged down even further. It is in the entanglement of reciprocally influencing claims and dysfunctional attempts at reequilibration that the ‘binding’process of a ‘relational paradox’ evolves. The notion that the binding process is circular, distinguishes this paradigm from and goes ’beyond‘a linear cause and effect model. Rather than locate the logical-type confusion in the simultaneous incongruent message levels of the sender, as in paradoxical injunctions, in a ‘relational paradox’ the message-level confusioncanbediscernedinthejuxtaposition ofincongruentmessagelevelsburiedin the interaction.

Relational Paradox: Toward a Language of Interactional Sequences
Definition: An interactional sequence in a relationship, in which the exchange of messages a t incongruent levels, generates a spiral of reciprocal perspec- tives and responses binding the interactants. The bind itself resides neither in the sender, the receiver, or the message,but in the interplay of

message levels in the interaction. The ingredients are:




2) 3)


An interactional sequence.
Two or more persons involved in a significant relationship.
In such a context messages are exchanged in which:
a) One interactant asserts something about the relationship
b) “he other interactant makes a meta-assertion which explicitly or impli-

citly classifies the previous assertion, redefining the relationship.
c) “he assertion and the meta-assertion are incongruent.
The nature of the interactants’relationship is a part of the paradox in that:
a) The interactants are bound by their respective and reciprocally influ-

encing claims about the relationship.
b) Either interactant has the power to classify any remark as a particular

utterance or as a claim on the definition of the relationship.
c) The relationship is bound to correct the incongruity (deviance) by a further exchange of messages that may paradoxically entangle the

sequence further. To illustrate:

I recall that during a n escalating marital squabble years ago, I attempted to leave the ‘field’by walking away. But when I turned to leave, my wife snapped, “I knew you would walk away!” I stopped in my tracks, momentarily paralyzed from moving in either direction.

Let us take a closer look a t this interactional sequencein light of the above criteria for a relational paradox. First, this is an interactional sequence in which two persons are involved in a significant relationship. In the messages exchanged there are two incongruous assertions. I assert that my walking away is no longer participating in the argument; however, my wife classifies that same behavior as part of the argument, if not a further escalation of it. In the interactional sequence, three levels can be distinguished:

1) A message, that given by the behavior ‘walkingaway.’
2) Meta-messages, indicating how the act of walking away is classified by both

interactants, as arguing or not arguing.
3) Meta-messages, indicating the implication of the exchange of messages for the

relationship itself.
In locating the relational paradox then, we must look for the incongruent levels in

the interaction, and not for the internal inconsistencies in the messages of either interactant. Returning to the example, I can certainly walk away in a manner that is congruent with my view that I am no longer arguing and that I have a right to make such a claim. But this may indeed be incongruent with my wife’s view or with her right to deny my claim. Furthermore, because we are in a relationship, my behavior, how I see my behavior, how I see my wife seeing my behavior, is influenced by her behavior, how she sees her behavior, and how she sees me seeing her behavior. My ‘walking away’ is incongruent a t t h e relationship level, because i t is usually understood t h a t a n argument will continue until both parties decide to stop. My wife adroitly reframes and reclassifies my rule violation (attempting to end the argument unilaterally) as rule compliance; (“I knew you would). As suggested before, deviance in a relationship elicits responses that usually bring the deviance back into balance. Thus, I respond to the arguing, which I feel is throwing the relationship off balance, by walking away, itself deviant; my wife responds to my deviance by reframing it as non-deviance. We are bound in this sequence by incongruent logical levels of communication, by conflicting relationship claims, and by the escalating entanglement, resulting from attempts at reequilibration.

Between a parent and child the stakes are often higher. Consider a mother who says to her daughter during a family therapy session, “Look,Jean, we are here to be honest, you can say anything you want.” The mother’s tone is quite convincing, although the expression on Jean’s face indicates that she finds the mother’s statement inconsistent with her usual attitude. Jean then cautiously begins to tell her mother that she is too domineering, to which the mother responds,“Howdare you, after all I’ve done for you . . .” (and begins to weep). Obviously, the mother’s initial statement and subsequent behavior are contradictory, but, more significantly, the incongruity of message levels lies in the mother’s assertion that her daughter is free to speak her

mind,juxtaposed with the daughter’s assertion that she is notfree toclaim whetheror not she isfiee. Furthermore, in attempting to CORect the incongruency,the communication knot tightens; the daughter’s quietness is viewed by the mother as sickness, while the


daughter views it as a response to her view of her mother as domineering. The daughter’s caution and view of her mother as domineering are both a response to and serve to elicit the mother’s domineering behavior. Indeed, the daughter is prevented from saying anything she wants, not primarily because the mother has sent the paradoxical injunction, ‘speak freely,’but because mother and daughter are locked into a way of relating that elicits a repetitive behavioral pattern.

Paradoxes can be generated when a statement functions on two logical levels and truthononelevelentailsfalsehoodattheother.Giventhebivalenceof languagewithin relationships, ‘beingright‘ononelevelmaybetantamountto‘beingwrongonanother. In the present example, the mother’s remark (“How dare you ...”) is “right” at the utterance level asan expression of indignation overbeing criticized,but “wrong”at the relationship level because it reinforces the daughter’s belief that she is not free to speak freely.

Let me add one final example of relational paradox. John Spiegel(1957) described the problems of the Bonnelli family. Joanne, the 11-yearold and middle of three, feared rejection by her father, and repeatedly made demands upon him for gifts. At first the father satisfied her demands intermittently, the daughter gradually defined his intermittent withholding as confirmation of her fear of rejection, and subsequently intensified her demands. The father defined this as pestering, felt victimized by it, and responded with increased withholding and disapproval. This interactional sequence reveals:

1) The manner in which discrepant labeling can precipitate an impasse in a relationship, and

2) Thestrugglethatensueswheneachinsistsonthevalidityofhisorherlabel.

In analyzing this sequence from an interactional point of view, emphasis would be placed on the relationship of behaviors to one another. For example, ‘rejecting’and ‘demanding’ can be seen as reciprocally stimulating and reinforcing. A logical-type perspective would draw attention to the interplay of message levels. In the Bonnelli sequence, messages of different levels are interchanged. One of the daughter’s messages is an instruction, uiz., ‘give me,’ the other is an instruction at the relationship level,-‘Don’t see me as being demanding.’ The fact that the daughter is giving somewhat inconsistent messages is fairly obvious. But more significant is the process whereby additional incongruity arises when the various message levels are juxtaposed in the interchange. Thus, it is Joanne’s instruction ‘Giveme’ in the context ofherfather’smeta-instruction, ‘Youaretoseemeasavictimofyourdemandingness,’ that is incongruous, or the father’s instruction, ‘Don’t demand‘ in the context of Joanne’s meta-instruction, “Youare to see me as a victim of your deprivation,’that is incongruous.

Now that the incongruity in the interchange has been demarcated, let us look at how the interactants become bound. In any ongoing interaction, the a d s of one are both the cause and effect of the act of the other. Here, adopting the role of victim, each implicitly assigns the other the role of victimizer. The father’s behavior can be seen as a rejection of both the label ‘withholding’and of the role of victimizer assigned at the relationship level. Likewise, Joanne’s behavior can be seen as both a rejection of the label ‘demanding’and of the role of victimizer assigned at the relationship level. Her ‘demanding’behavior is not only a response to her father’s ‘withholding’behavior, but a stimulus for more ‘withholding.’The father’s ‘withholding‘is both a response to and serves to elicit more ‘demanding.’Each interadant holds a perspective, and acts in a way consistent with it. Ironically, these actions may serve only to confirm the other’s perspective.

The central importance from the point of view of family and couple therapy of the kind of examples that have been discussed is: that a ‘relational paradox’ and not a


‘paradoxical injunction’ is prototypical of “The Double Bind,” that such relationally locked interchanges embody the original intuition of the double bind theorists; and that usage of ‘doublebind’ in the manner suggested will be illuminating clinically.

The focus in the examples given has been on diadic interactions. The extension of these ideas to three or more persons is beyond the scope of this paper. The work of David Kantor (1975) has underscored the crucial role “witness” in the interactional field. The responses of the witness can play a pivotal role in determining the outcome of an interactional sequence. Kantor’s framework might be extended even further to include people outside the interactional field, but whose presence is felt in the interactional field. The work of Sullivan on therapist-client interaction demonstrates how a therapist is continually “shadow-boxing” the “other people” in the room (Havens, 1976). In the example of the Bonnelli family, for instance, no mention was made of Mrs.Bonnelli’sresponses during the entangled interaction of her husband and daughter. Also, there was no mention made of the history of the father’s giving and receiving transactions with his own father. All these elements and more are a part of the social matrix in which the messages exchanged are embedded.


The major insight of the double bind hypothesis was that classmember disconti- nuity is continually being breached in human communication. The original Bateson article provided the first link between communication dilemmas and logical type theory. However, the scope of this linkage was limited to the message-levelconfusion that may be contained in single statements. By recognizingthe crucial role of social context on behavior, the double bind theory took a significant step beyond the theories of its day. But it was also restricted by the social context of its day, in particular, the prevailing notions of psychological trauma and linear causality. In this context, the theorists hypothesized that schizophrenicbehavior might be a responseto or an effectof recurrent double binds. Attempting to unravel the ‘how’of schizophrenicbehavior they got stuck in the ‘why’ofthat behavior. Subsequentrevisions have attempted to ‘unbind‘ the theory from such linear cause and effect thinking. However, these reformulations have failed to show precisely how both interactants actively participate in the binding process.Thispaper suggeststheparadigm of‘relationalparadox’inordertoextendthe concept of message-level confusion to the interactional arena and to unravel how “double binds” occur in interactional sequences over time. In addition, the following points have been highlighted:

1) The term ‘double bind’ has referred to too many things indiscriminately to allow one to ‘go beyond.’

2) Meta-communication,acrucialtermindoublebindterminology,isdefinedina manner fully consistent with logical type theory.

  1. 3)  Paradoxical injunctions contain message levels of different logical types, but are not the same as relational paradoxes.

  2. 4)  In a relational paradox, the incongruity of message levels is located in the juxtaposition of messages in an interchange, not in the interactant, or in the message.

5) The theory of logical types, the cornerstone of double bind theory, helps to explicate the conflicting message levels in interactional sequences.

6) All participants reciprocally influence the binding process of an interaction.

In presenting the paradigm of ‘relational paradoxes,’this paper has attempted to unbind the ‘double bind’ from previous terminological confusion. A language of


interactional sequences that incorporates the circular causality of behavioral influence may, indeed, offer hope of going ’beyond the double bind.’4


Abeles, G. The double bind: Parador in relationship. Boston University, University Microfilms, 1975,209.

Bateson, G., Jackson, D., Haley, J., & Weakland, J. “Toward a theory of schizophrenia.” Behavioral Science, 1956,1,251-264.

Bateson, G. & Ruesch, J. Communication: The social matrix of society. New York W.W. Norton, 1951.

Haley, J. “The family of a schizophrenic: A model system.” Journal ofNeruous and Mental Disorders, 1959,362.

Havens, L. Participunt observation. New York Jason Aronson Inc. 1976,2332.
Jackson, D. “The question of family homeostasis.” In Communicution,Family, and Marriage, Palo

Alto: Science and Behavior Books, 1968.
Kantor, D. & Lehr, W. Znside the family. San Francisco: Jossey-Bass, 1975.
Sluzki, C. & Ransom, D. (Eds.), Double bind. New York: Grune and Stratton, 1976. Spiegel,J.“Theresolution ofroleconflictwithinthefamily.”Psychiatry,1957,20,1-16. Watzlawick, P.,Beavin,J., &Jackson,D.,Prugmuticsof humancommunication.NewYork:W.W.

Norton, 1967.
Watzlawick, P., Weakland, J., & Fisch, R. Change. New York: W.W. Norton, 1974.
Weakland, J. T h e double bind theory by self-reflexive hindsight.” Family Process, 1974,13. Whitehead, A. & Russell, B. Principia Mathematica, 3. Mass.: Cambridge University Press, 1910.

‘The classical paradoxical statement (uttered by a Cretan) that “All Cretans are liars” is paradoxical because if it is a true classification, then it is a false utterance, and vice versa. Two

logical levels can mesh perfectly well; there would be nothing paradoxical in a Cretan saying “All

Cretans’ statements contain words,” as the statement is both a true classification and utterance. *Bateson(1951)noted that messages can be said to have two distinct aspects or functions, report and command. Translated into logical-type theory, the report is at one logical level, the command, which instructs how the report should be received or how the members of the relationship should

view each other is at a meta-communicative level.
31nhuman communication there are not only verbal messages which are meta-communicative, but

posture, facial expression, and tone of voice which give rules for interpreting what is said.
‘I would like to express my gratitude to David Ackerman for his editorial and substantive contributions to this paper, and to Dr.Henry Grunebaum and Iva Gardner for their interest and