Karl Tomm: Interventive Interviewing

Contents
Karl Tomm: Interventive Interviewing

Interventive Interviewing: Part I. Strategizing as a Fourth Guideline for the Therapist

Abstract

Introduction

The need for a Fourth Guideline

A minor reformulation of Hypothesizing, Circularity, and Neutrality

Strategizing about conceptual postures

References

Interventive Interviewing: Part II. Reflexive Questioning as a Means to Enable Self-Healing

Abstract

Introduction

A theoretical rationale

Types of Reflexive Questions

Future-Oriented Questions

Observer-Perspective Questions

Unexpected Context-Change Questions

Embedded-Suggestion Questions

Normative-Comparison Questions

Distinction-Clarifying Questions

Questions Introducing Hypotheses

Process-Interruption Questions

Concluding comments

References
Interventive Interviewing: Part III. Intending to Ask Lineal, Circular, Strategic, or Reflexive
Questions?

Abstract

Introduction

Therapist intentions and assumptions

Four major types of questions

Lineal Questions

Circular Questions

Strategic Questions

Reflexive Questions

The effects of different questions

Lineal Questions

Circular Questions

Strategic Questions

Reflexive Questions

Concluding comments

References
Fam Proc 26:3-13, 1987

Interventive Interviewing: Part I. Strategizing as a Fourth

Guideline for the Therapist
KARL TOMM, M.D.a
Family Therapy Program, Department of Psychiatry, University of Calgary, 3330 Hospital Drive, N.W., Calgary, Alberta,
Canada T2N 4N1.

Abstract
A clinical interview affords far more opportunities to act therapeutically than most therapists realize.
Because so many of these opportunities remain outside the therapist's conscious awareness, it is useful
to elaborate guidelines that orient his or her general activity in directions that are liable to be
therapeutic. The Milan associates defined three such basic guidelines: hypothesizing, circularity, and
neutrality. Hypothesizing is clear and easy to accept. The notions of circularity and neutrality have
aroused considerable interest but are not as readily understood. These guidelines may be clarified and
operationalized when reformulated as conceptual postures. This process is enhanced by
differentiating a fourth guideline, strategizing, which entails the therapist's decision making,
including decisions about how to employ these postures. This paper, the first in a series of three,
explores these four interviewing guidelines. The other papers will appear in a subsequent issue. Part
II will focus on reflexive questioning, a mode of inquiry oriented toward mobilizing the family's own
healing capacity. Part III will provide a scheme for analyzing and choosing among four major types
of questions: linear questions, circular questions, reflexive questions, and strategic questions.

Introduction
I have become fascinated with the variety of effects a therapist can have on individual
clients or families during the course of a clinical interview. In a conventional session, most
of the therapist's questions ostensibly are designed to help him or her formulate an
assessment. The questions themselves are not usually regarded as interventions to help
clients. Yet, many questions do have therapeutic effects on family members, (directly)
through the implications of the questions and/or (indirectly) through the verbal and nonverbal
responses of family members to them. At the same time, however, some of the therapist's
questions can be countertherapeutic.
The latter became painfully obvious to me a few years ago while reviewing a
videotape of a marital session. One of my "innocent" questions appeared to have stimulated
the re-emergence of serious marital conflict. It occurred during a follow-up session in which
the couple were talking about the fact that they had not had any arguments for several weeks.
In other words, there had been a major improvement in the marriage. After a lively and
enjoyable discussion about these changes, I asked, "What problems would you like to talk
about today?" Following this seemingly innocuous question, the couple gradually drifted into
a bitter argument about which of the two of them most needed further therapy. I (privately)
reconstrued the improvement as "transient and unstable" and resumed my treatment of their
chronic marital difficulties. I remained completely blind to the fact that I had inadvertently
triggered the deterioration until a colleague pointed it out to me on the videotape1. In
retrospect, the assumption behind the question, that problems needed to be identified and/or
clarified before I could act therapeutically, turned out to be limiting and pathogenic. It limited
the discussion to areas of dissatisfaction and served to bring forth pathological interactions.
Instead, I could have capitalized on the new developments and asked questions that were
designed to strengthen the recent changes. Unfortunately, I did not see that option clearly at
the time.
This blunder and other more positive learning experiences (reported in Part II) led me
to realize that a therapist is far more influential in what emerges during a session than I had
previously imagined. I began examining the interviewing process in greater depth and
eventually came to the conclusion that it would be more coherent and heuristic to regard the
whole interview as a series of continuous interventions. Thus, I began to think in terms of
"interventive interviewing," a perspective in which the range of therapeutic opportunities is
extended by considering everything a therapist does during an interview to be an intervention.
This perspective takes seriously the view that it is impossible for a therapist to interact
with a client without intervening in the client's autonomous activity2. The therapist assumes
that everything she or he says and does is potentially significant with respect to the eventual
therapeutic outcome. For instance, every question and every comment may be evaluated with
respect to whether it constitutes an affirmation or a challenge to one or more behavior patterns
of the client or family. As illustrated in the scenario described above, to ask about a problem
is to invite its emergence and to affirm its existence. In addition, to listen to and to accept the

1
It would be easy to say that the couple had not yet "really" worked through their difficulties. To do so might
absolve me of any responsibility for the deterioration, but it would not help me become a more effective
clinician. I chose to construe my decision to ask that question as a mistake in order to restrain similar behaviors
in my future work.
2
Clients are, of course, continually intervening in the activities of the therapist as well. This important feature
of the therapeutic system is alluded to but not elaborated in these papers. For some insightful reflections on
this issue, see Deissler (3).
description of a problem is to concede power with respect to its definition (10). Within this
perspective, no statement or nonverbal behavior is assumed, a priori, to be inconsequential.
Nor is the absence of certain actions considered trivial. By not responding to particular events
the therapist may knowingly or unknowingly [pg 1] disappoint or fulfill certain expectations
of one or more family members. For instance, the failure to challenge explicitly a position
statement or a particular construal of a situation is often experienced by family members as
implicit agreement, support, and/or reinforcement. Thus, interventive interviewing refers to
an orientation in which everything an interviewer does and says, and does not do and does
not say, is thought of as an intervention that could be therapeutic, nontherapeutic, or counter-
therapeutic. While this perspective dilutes the conventional meaning of the term
"intervention", it opens the possibility of entertaining an enormous range of therapeutic
actions.
Interventive interviewing also takes seriously the view that the actual effect of any
particular intervention with a client is always determined by the client, not by the therapist.
The intentions and consequent actions of the therapist only trigger a response; they never
determine it. Although many deliberate therapeutic interventions do have their desired
effects, these effects can never be guaranteed. Listeners hear and experience only that which
they are capable of hearing and experiencing (by virtue of their history, emotional state,
presuppositions, preferences, and so on). Thus, a carefully prepared question that a therapist
intends as "a therapeutic intervention" may not turn out to have any therapeutic impact
whatsoever. Conversely, something that the therapist does not intend as a therapeutic
intervention could turn out to have a major therapeutic effect. For instance, an ordinary
exploratory question could pique the client's curiosity in a crucial area and precipitate a major
change in patterns of thought. Indeed, it is not uncommon for clients to report that they were
significantly influenced by a particular question that seemed relatively unimportant to the
therapist.
Adopting the perspective of interventive interviewing orients therapists to focus more
closely on their own behavior within the vicissitudes of the therapeutic system, not just on
the client system. When every action is regarded as an intervention, therapists are drawn
toward attending to the ongoing effects of their behaviors in order to distinguish those actions
that were, in fact, therapeutic from those that were not. In addition, when something
undesirable occurs among family members during the interview, therapists are more liable to
examine their own behavior as a possible trigger. With this increased scrutiny of the
interaction between therapist and client, the discrepancy between therapeutic in tent and
effect on the client becomes even more apparent. Consequently, therapists become more
inclined to reflect carefully on all their actions before acting, not just on those that they
previously might have chosen to define as "interventions”. However, it is impossible to
monitor every response and to reflect consciously on the details of every action before acting.
The complexities of this perspective could quickly become totally unmanageable unless the
therapist develops and implements some organizing priorities. One approach to this
complexity is to establish guidelines which, when mastered, can be adopted as nonconscious
therapeutic postures that facilitate desired actions and restrain undesired ones.

The need for a Fourth Guideline

In their original paper (13) on how to conduct a systemic interview, the Milan team
outlined three principles to guide the therapist. These principles or guidelines are now fairly
well known, and "circular interviewing" is the term often used to refer to the style of inquiry
associated with their application. Several authors have begun describing and elaborating
various aspects of this method of enquiry (3, 4, 6, 8, 11, 12, 14, 15, 17). At the end of the
original paper, the Milan team raised an intriguing question: "Can family therapy produce
change solely through the negentropic effect of our present method of conducting the
interview without the necessity of making a final intervention?" (p. 12)3. I would like to
propose an affirmative answer: "Yes, circular interviewing alone can, and does, trigger
therapeutic change". The basis for this affirmative response is clarified if one distinguishes a
fourth interviewing guideline, namely, "strategizing," and recognizes circular questioning as
a type of interventive interviewing.
Anyone who has observed the members of the Milan team conduct therapy will know that they plan
each and every move with great care. The process of generating plans of action, evaluating
them, and deciding on which course to follow is not limited to the intersession discussion
when they prepare the final intervention. It occurs throughout the session. Indeed, the
interviewers are continually making decisions on a moment-to-moment basis as the interview

3
The term "negentropic" as used by the Milan team implies "ordering" or "organizing”. See the original paper
(13) for an elucidation of this concept.
unfolds. In effect, they are posing questions to themselves and are answering them, either
consciously or nonconsciously. Some of these questions might be: "Which hypothesis should
I explore now?"; "Is the family ready to talk openly about that subject?"; "What would it
mean not to explore that area just yet?"; "Which question should I ask?"; "What effect do I
want?"; "How should the question be formulated?"; "To whom should I address it?"; "Should
I pursue this issue further or explore another?"; "Should I pick up on the child's sadness now
or ignore it?"; "Should I lean forward and offer Kleenex or should I ask a question that might
trigger other family members to respond?", and so on. The answers to these questions arise
from the therapist's history of socialization as a human being in general and of his or her
specific development as a therapist. The team behind the mirror is also actively evaluating
the therapist's activity, and if they have suggestions for a significant shift in the course of the
interview they interrupt the session and call the therapist out to confer briefly. Most observers
would readily agree that the whole therapeutic endeavor revolves around judgments about
what a therapist should and should not do when interacting with the client or family.
This decision-making process is implied but not adequately accounted for in the three
interviewing guidelines that the Milan associates originally described. Hence the
appropriateness of delineating a fourth to guide therapists in making these choices.
Strategizing may be defined as the therapist's (or team's) cognitive activity in evaluating the
effects of past [pg 2] actions, constructing new plans of action, anticipating the possible
consequences of various alternatives, and deciding how to proceed at any particular moment
in order to maximize therapeutic utility. As an interviewing guideline, it entails the therapists'
intentional choices about what they should do or should not do in order to guide the evolving
therapeutic system toward the goal of therapeutic change. In labeling this guideline, I chose
the root term "strategy" to emphasize that therapists adopt a stance with a definitive
commitment toward achieving some therapeutic goal. The gerund form, -ing, was chosen to
emphasize its active nature; that is, it is an active process of maintaining a network of
cognitive operations that result in decisions for action4.

4
The notion of "strategizing" has much in common with, but is not equivalent to, that of "strategic therapy”.
The latter refers to a commitment to a specific school of therapy, just as systemic therapy and structural therapy
imply commitments to alternative schools. Strategizing implies a commitment to purposive therapeutic change
in general, and as a guideline may be applied to all therapies. The intentionality inherent in strategizing will be
discussed in Part III.
 It is possible to distinguish several levels of strategizing. In these papers, I will focus
mainly on two: strategizing about general conceptual postures for a therapist to adopt, and
strategizing about specific verbal actions for a therapist to enact. The four interviewing
guidelines will be presented as conceptual postures (in Part I) while the questions asked in
the session will exemplify actions (see Parts II and III). These levels are, of course,
intertwined in that certain actions are easier to perform when a therapist has assumed one
posture as opposed to another. For instance, it is easier to ask a genuinely exploratory
question from a posture of neutrality, and it is easier to ask a confrontative question from a
posture of strategizing. Having chosen to adopt a particular posture, the therapist can focus
attention on other details and remain assured that the posture itself will guide his or her
ongoing actions.

A minor reformulation of Hypothesizing, Circularity, and Neutrality

In describing these three guidelines as conceptual postures, I am trying to bring them
out of the abstract realm of transcendent principles into the concreteness of clinical activity,
and to encourage therapists to accept more personal responsibility for adopting them. A
conceptual posture may be defined as an enduring constellation of cognitive operations that
maintain a stable point of reference which supports a particular pattern of thoughts and
actions and implicitly inhibits or precludes others. Like a physical posture, it may be adopted
without conscious awareness during the spontaneous flow of activity during an interview.
Alternatively, it could be taken up deliberately as a way of preparing for certain actions or
avoiding others. Conscious awareness in assuming a specific posture may be helpful when a
therapist is learning to develop new patterns of behavior, but once mastered, the posture tends
to become part of the therapist's nonconscious flow of activity (much like the physical posture
of an actor, a musician, or an athlete).
To choose to adopt a posture of hypothesizing is to apply deliberately one's cognitive
resources in order to create explanations. One activates those cognitive operations that seek
connections among observations, reported data, personal experience, and prior knowledge,
in order to formulate a generative mechanism that might explain the phenomenon one wishes
to understand. The Milan team's statement about hypothesizing includes an excellent
description of the issues involved. I encourage any reader not already familiar with their
paper (13) to study it carefully. The only issue I wish to emphasize here is the difference
between circular hypothesizing and lineal hypothesizing. If our conceptual posture is oriented
to create circular and systemic explanations, we will tend to ask circular questions. If our
posture is oriented to create lineal explanations, we will ask lineal questions. At the same
time, however, circular and lineal questions as interventions are liable to have quite different
effects in the interview. Circular questions usually have far more therapeutic potential than
lineal ones (see Part III). Thus, to optimize our therapeutic impact during the process of
interventive interviewing, it is useful to develop expertise in adopting a posture of circular
hypothesizing.
To describe circularity as a conceptual posture requires some preliminary comments.
This principle, as originally described by the Milan team, has led to considerable confusion,
with varying interpretations about what is entailed. The confusion appears to have resulted
from a failure to draw a clear distinction between the circular aspects of the observed system
(the family) and the circularity of the observing system (the therapist-family unity). This
distinction separates first-order cybernetics (the cybernetics of observed systems) from
second-order cybernetics5 (the cybernetics of observing systems) and delineates very
different domains (even though the first is incorporated as a component in the second). In
this discussion, I will limit the notion of circularity as an interviewing guideline to the second
domain and apply it to the recursive feedback in the therapeutic (observing) system. Other
aspects of the original definition are allocated to other guidelines. For instance, first-order
cybernetic descriptions regarding reciprocity in "differences" and circular patterns among
family members are regarded as part of circular hypothesizing. Decisions about what kinds
of questions to ask, such as triadic questions to reveal a family's circular patterns, are
accounted for in the new guideline of strategizing.
Given this reformulation, circularity refers to the dynamic structural coupling between
the therapist and the family that makes it possible for the therapist to draw distinctions about
the family. As a conceptual posture, it entails an acute sensitivity on the part of therapists to
nuances in their own sensory responses during their recursive interaction with clients. It
includes an acknowledgement of the discontinuity between intent and effect (as described in
the introduction), and orients therapists to attend to what they perceive as the ongoing

5
Heinz von Foerster (5) has been a central figure in the elaboration of this distinction. For an excellent historical
account see Keeney (7).
behavior of clients in the evolving therapeutic system. The more astute the observing, the
more the therapeutic responses can be refined to fit the family's responses, and the more
closely therapist and family become coupled. Therapists are by no means passive in this
observing process. Just as the eye, in order [pg 3] to see, needs to move back and forth with
a continual micro-nystagmus in order to distinguish "differences" in the patterns of light
falling upon the retina, so therapists must continually probe family members by asking
questions, paraphrasing their answers, and noting their verbal and nonverbal responses in
order to draw distinctions about their experiences. Indeed, this activity on the part of the
therapists is the major reason this guideline is labelled "circularity" rather than simply
"observing”. Continual movement on the part of the therapist in relation to the movements
of the client or family is essential if therapists are to become more refined in structural
coupling with them in the therapeutic system. It is the nature of this coupling that provides
the foundation for all of the other cognitive operations in the course of therapy6.
As with hypothesizing, there are variations in the posture of circularity. Two
contrasting modes may be referred to as "caring circularity" and "obligatory circularity”. The
first is based on natural human love, the second on coercion. They lead to different ways of
being structurally coupled in the therapeutic system. To adopt a loving posture is to attend
selectively to those differences in the responses of the client(s) that offer space for the
therapist to be genuinely supportive of the autonomous growth and development of the
client(s). On the other hand, to adopt a posture of circularity by "necessity," perhaps because
the therapist realizes that she or he must do so in order to be a "good" clinician, is to attend
selectively to those responses of the client(s) that provide openings for the therapist to be
therapeutically efficacious. While these variations in circularity may not always be mutually
exclusive, whichever posture the therapist adopts as the priority will significantly influence
the direction and tone of the interview. On the one hand, the client(s) may experience warm
and sensitive understanding and, on the other hand, insensitive and penetrating scrutiny.
Neutrality as an interviewing principle is a difficult notion to understand because,
strictly speaking, it is physically and logically impossible to remain absolutely neutral. At
the moment one acts, one is not being neutral with respect to that specific action; the behavior

6
For a theoretical grounding regarding the nature of cognition upon which this view of circularity has been
elaborated, see Maturana and Varela (9).
affirms itself. Thus, the clearest behavioral manifestation of neutrality might be "not to act”.
However, in situations where action is expected, not to act may be construed as a definitive
action; furthermore, it is antithetical to the need for action in circularity. In actual practice,
the therapist does act (as guided by the other guidelines) but strives to balance the movements
so that the net result is to maintain an overall posture of neutrality. Thus, time is an important
component of this posture. The therapist participates in an ongoing "dance" with the client
or family and maintains a careful balance in relation to the various desires of family members
(much like the continual movements of a tightrope walker to maintain balance in relation to
gravity). The logical difficulty applies at the level of meanings and values where a therapist
either takes a certain position on an issue or does not. To not take a position is to take the
position of not taking one, that is, to be noncommittal, to decide not to decide, or to be
deliberately evasive. Nor does the synthesis of "both/and" escape the dilemma. The synthesis
is the beginning of a new dichotomy; both/and versus either/or. The problem in drawing
distinctions is inherent in the language, which we cannot escape. With respect to meanings
and values, neutrality is closest to taking the position of remaining noncommittal.
Despite these difficulties, neutrality is an extremely important guideline in systemic
therapy. To be neutral in an interview is to adopt a posture in which the therapist accepts
everything as it is taking place in the present, and avoids any attraction to, or repulsion from,
anything that the client(s) says or does. The therapist remains open to whatever happens, and
flows with the stream of spontaneous activity, not against it. At the same time, however, the
therapist avoids being drawn into taking a position for or against any person or issue. In
addition, the therapist remains open to reconsider whatever she or he has construed to be
happening. By releasing any attachments to his or her own perceptions and intentions, the
therapist's neutrality insures more flexibility in overall interventive behavior. There is more
space for the intuitive, nonconscious aspects of cognition to emerge and become active in the
therapeutic process. In neutrality, the therapist does not claim to know what is accurate or
true, what is useful or not useful, but instead places "objectivity in parenthesis"7. For instance,

7
In his theoretical work on cognition, Maturana draws the important distinction between objectivity and
objectivity in parenthesis. The latter entails the recognition that an object, event, idea, belief, and so on, is a
distinction made by an observer. There can be as many different, yet valid, distinctions as there are observers
to make them; and any individual observer can distinguish as many objects or phenomena as the coherences in
his or her cognitive operations allow.
when a husband complains that his wife is being unreasonable with a child, the therapist
listens and accepts the husband's complaint as his action in the present, then listens to and
accepts what the wife has to say. The therapist does not agree or disagree with the husband's
or the wife's views, that is, avoids aligning with either of them. Nor does the therapist insist
that the husband's statement was, indeed, "a complaint”. By releasing any attachment to such
a perception, the possibility that other intuitive perceptions might emerge are enhanced. For
example, the husband's statement could constitute "a plea" to the wife that she be more
accepting of him. But, if the therapist was committed to the complaint construal, the
possibility that it was a plea would not be entertained. During the course of the interview the
therapist may or may not choose to indicate a lack of agreement or disagreement (for
example, with the content or intent of the husband's statement) in the form of a question or
comment, but this choice has to do with strategizing. Neutrality itself is limited to a
conceptual posture in which the therapist is immersed in experiencing the present as fully as
possible and accepting everything that occurs as necessary and inevitable, including the
family's and his or her own construals.
It is possible to distinguish several variations in this posture. Indifferent neutrality, the
purest form, implies a posture in which a therapist attends to and accepts everything with
equal interest. In so doing, however, this may convey a relative lack of concern for clients as
unique human beings. Affirming neutrality is more differentiated. It orients a therapist to
attend to individuals as persons and to accept them as being human, in whatever way they
happen to be. It tends to support therapist behavior that confirms the other and, hence, is very
engaging. In this respect, caring circularity and affirming neutrality constitute mutually
supportive and synergistic postures. Aloof neutrality emerges when a therapist has difficulty
[pg 4] accepting others without agreeing with them. Consequently, the therapist takes a
metaposition and remains somewhat distant. Strategic neutrality implies a slippage toward
strategizing, to using neutrality as a strategic change technique rather than as a posture of
acceptance. For instance, to remain deliberately neutral with respect to persons by equalizing
talk-time, when the therapist perceives the family to be organized as having a single
spokesperson, reflects a choice that arises out of strategizing.
In essence, however, neutrality contrasts sharply with strategizing. While neutrality is
founded on an acceptance of "what is," strategizing is based on a commitment to "what ought
to be”. Leaning too far in either direction can impede a therapist's potential. If a therapist
adopts too much neutrality, and only accepts things as they are, eventually he or she stops
doing therapy. Thus, this danger is self-limiting. On the other hand, if a therapist adopts too
much strategizing, becomes too purposive, she or he may become blind and violent. In his
writings on the mind, Bateson (1, 2) warns us about the inherent blindness and lack of
wisdom in too much purpose. Unless therapists are able to adopt some degree of neutrality
they will not be able to see and experience "the other side" of an issue. Furthermore, therapists
who are too strongly committed to their own ideas and values about "correct" solutions can
easily become "violent" in imposing them on a "resistant" client or family. When this
happens, the strategic means defeat the therapeutic end, and more neutrality is clearly in
order. Fortunately, a strategic commitment to neutrality as a posture in its own right, that is,
not to be so purposive, can help reduce the blindness and potential violence of excessive
purposiveness.
A brief clinical vignette may help illustrate the therapeutic consequences of neutrality.
While interviewing a man who had been incestuously involved with his stepdaughter, I found
myself becoming increasingly frustrated by his unwillingness to acknowledge responsibility
for what he had done. I was intent on getting him to accept personal responsibility as a first
step toward a commitment to change his patterns of behavior. I realized that I was not being
sufficiently neutral but, being repulsed by his behavior, found myself unable to change my
posture. When my frustration almost reached the point of anger, I excused myself and left
the therapy room. Once I was in the hallway, I could concentrate on trying to regain a neutral
stance. By developing some circular hypotheses about how certain activities of his wife and
stepdaughter (as well as some memories from his childhood) participated in a systemic
pattern that included his incestuous behavior, I was able to regain a conceptual and emotional
posture of neutrality. When I returned and resumed the interview, he began responding to my
change (in manner and tone) by becoming progressively more open. Only then could I begin
to see that he was far more frustrated with himself than I was. Indeed, he was furious with
himself to the point of being suicidal for what he had done. I then proceeded to work with
these feelings and to help him modify some of his inappropriate ideas and behavior. Thus,
giving the posture of neutrality priority proved to be very therapeutic in this case.
 Is it reasonable to wonder whether it is possible to adopt postures of strategizing and
neutrality at the same time? After all, they are contradictory positions in many respects.
Fortunately, the human nervous system is sufficiently complex so that we can operate at
multiple conceptual levels and within different domains simultaneously. Thus, we can be
strategizing about the need to maintain neutrality at one level, adopt the latter relational
stance at another, and at the same time be asking questions out of circular hypothesizing and
adjusting ourselves to the client's sensitivities in circularity in other domains. Indeed, we are
probably employing some aspects of each conceptual posture nonconsciously most of the
time while conducting therapy.

Strategizing about conceptual postures

As noted earlier, the guideline of strategizing may be applied at several levels. Indeed,
it could orchestrate the entire spectrum of the therapist's perceptual, conceptual, and
executive activities. In this way, the inherent commitment to therapeutic change could
permeate the whole interviewing process, even down to the level of nonverbal and paraverbal
behaviors like hand and leg movements, body orientation, direction of gaze, tone of voice,
cadence of speech, and so on. What would be extremely important to include, however, is
strategizing about our own strategizing. This has already been alluded to above and requires
hypothesizing about ongoing developments in the therapeutic system. We need to keep
noticing whether or not our decisions to act therapeutically are, indeed, being therapeutic at
any particular moment. For example, I needed to recognize that my earlier choice to
encourage, cajole, push, and even "force" the incest-prone father to explicitly acknowledge
his responsibility was constraining my therapeutic potential, otherwise I may not have
abandoned that course of action and could have lost the case altogether. On other occasions
I have found it useful to try helping by not helping (16). Therapists are more liable to develop
this capacity to strategize about strategizing if they choose to opt for a posture of personal
strategizing, by which I mean that they decide to take full personal responsibility for their
choices and actions. This stance may be contrasted with projective strategizing in which
decisions are made because the therapist "was forced to" or "had no choice" as a result of
external factors (for example, the "real" situation or the "correct" rules of treatment).
Personalizing one's choices is a way to maintain more flexibility and freedom of movement
in strategizing. That is, it is always easier to change one's own construals and choices than to
change an "externally determined" situation.
Another major dimension of strategizing is the size of the unit of activity that the
therapist is strategizing about. Obviously, this is related to the level of strategic focus (choice
of specific nonverbal movement, type of question to ask, general therapeutic technique to
employ, conceptual posture to adopt, and so on), but it is not exclusively determined by [pg
5] level. For instance, if the therapist is strategizing at the level of specific therapeutic
techniques or strategies, he or she could be formulating a specific question to get past an
apparent impasse, or could be strategizing about a whole sequence of questions that might
occupy a major portion of the interview. It is beyond the scope of this paper to discuss the
way in which the posture of strategizing supports the implementation of particular treatment
techniques. My primary purpose here is to introduce the notion of strategizing as a foundation
for interventive interviewing.
One task in embracing this perspective on interviewing would be to strategize about
developing competence in maintaining a constellation of carefully refined conceptual
postures so that one's spontaneous responses at any particular moment would likely be
therapeutic. To do this with conscious deliberation, a therapist would have to critically
examine what his or her current inclinations are (preferably with the help of a supervisor or
colleague) and decide on modifying and/or strengthening specific areas. For instance, if one
decided to strengthen skills in circular hypothesizing, one might join a clinical team that is
committed to systemic brainstorming. However, if one wanted to develop substantive
expertise in this area (especially after a history of prolonged immersion in a culture disposed
toward lineal thinking), one might have to undergo considerable theoretical study, self
exploration, and perhaps some "corrective" personal experiences. As one's expertise and
security in holding a certain stance develop, there is a natural shift in focus from decisions
about the posture to its behavioral products, that is, to the specific questions, sequences, and
nonverbal activity that flow from it.
A second task would be to organize a heuristic direction for the flow of the therapist's
consciousness. For instance, a logical sequence for focused attention is to examine the
products of circularity, then those of hypothesizing, then of strategizing, then of neutrality,
and back to circularity. In other words, therapists can begin by drawing distinctions about the
family in the recursive interaction of circularity and take these observations into
hypothesizing. Having developed a hypothesis of some sort (possibly including the
hypothesis that what one still lacks is a clear hypothesis about the family), they make some
strategic choices about what to pursue (for example, first elicit more information) and how
to do so (perhaps explore how they decided to come for therapy). These decisions become
the basis for purposive actions (like asking about the initiative for a referral). Having
intervened, one jumps back (conceptually and behaviorally) to a position of neutrality to
accept whatever happens. One observes the family for differences in their responses (the
father may interrupt the mother to point out that the pediatrician sent them) and a new circuit
begins. The new observations are taken into the ongoing process of hypothesizing, and on
the basis of a modified hypothesis (for example, the husband is minimizing family initiative
for help), the therapist once again begins strategizing about what to do (Should I ask the wife
who first thought about and is most interested in therapy, or should I respect the husband's
sensitivity and ask about the pediatrician's views?). Thus, while the interview is taking place
the therapist may attend to the products of circularity, hypothesizing, strategizing, neutrality,
and circularity in a recursive circuit that is parallel to the scientific method. Disciplined
application of this recursive pattern of thought and action would probably significantly
enhance the therapeutic effectiveness of interventive interviewing.
Another task may be to develop a special sensitivity to cues in the therapeutic system
that suggest that a major shift in posture is indicated. For instance, when the atmosphere in
the interview has become constrained or is oppositional, it is likely that the therapist is
leaning too heavily in the direction of strategizing. The client(s) may be experiencing the
therapist as highly judgmental or as demanding too much change. This should be a cue for
the therapist to shift posture and to become more neutral. On the other hand, if the session
seems rather dull or boring, there is probably a need for more vigorous strategizing. When
an interview seems to lack direction, more hypothesizing (including hypotheses about the
therapeutic system) is clearly indicated. If the therapist does seem to have clear hypotheses,
yet the session does not seem to be very fruitful, one can give more refined attention to what
the clients are actually doing and experiencing by focusing on the feedback in circularity.
New "differences" or distinctions need to be drawn from the experiences of family members
that may intervene in the therapist's existing hypotheses. In addition to learning to pick up
and respond to such cues, a therapist should remain open to intermittent re-evaluation and
refinement of established postures. Some degree of inadvertent drifting as a result of the
continuous interventions from family members usually does occur. For instance, if a therapist
does not have a sensitivity to deceptiveness, caring circularity could drift into naivety when
clients have well established skills in exploiting the good will and trust of others. Here,
perceptiveness to changes in self (as well as in the family and therapeutic system) is required.
Ultimately, the strategies for mobilizing, maintaining, and altering these postures will "sink"
into nonconscious process, along with the conceptual postures themselves, so that the
therapist's consciousness can "float" freely to where it is most needed to maximize the clinical
effectiveness of the interview.

References
1. Bateson, G., Steps to an ecology of mind. San Francisco: Chandler, 1972.
2. Bateson, G., Mind and nature: A necessary unity. New York: E.P. Dutton, 1979.
3. Deissler, K., Recursive creation of information: Circular questioning as information production.
Unpublished manuscript, translated from the German by S. Awodey, 1986.
[pg 6]
4. Fleurides, C., Nelson, T. S. and Rosenthal, D. M., The evolution of circular questions: Training
family therapists, Journal of Marital and Family Therapy, 12, 113-127, 1986.
5. von Foerster, H., Observing systems. Seaside CA: Intersystems Publications, 1981.
6. Hoffmann, L., Foundations of family therapy: A conceptual framework for systems change. New
York: Basic Books, 1981.
7. Kenney, B. P., Aesthetics of change. New York: Guilford Press, 1983.
8. Lipchik, E. and de Shazer, S., The purposeful interview, Journal of Strategic and Systemic
Therapies, 5, 88-9, 1986.
9. Maturana, H. R. and Varela, F. J., Autopoiesis and cognition: The realization of living. Boston:
Reidel, 1980.
10. Mendez, C., Coddou, F. and Maturana, H., The bringing forth of pathology, submitted for
publication.
11. Penn, P., Circular questioning, Family Process, 21, 267-28, 1982.
12. Penn, P., Feed-forward: Future questions, future maps, Family Process, 24, 299-310, 1985.
13. Selvini-Palazzoli, M., Boscolo, L., Cecchin, G. and Prata, G., Hypothesizing-circularity-
neutrality: Three guidelines for the conductor of the session, Family Process, 19, 3-12, 1980.
14. Tomm, K., One perspective on the Milan systemic approach: Part II. Description of session
format, interviewing style and interventions, Journal of Marital and Family Therapy, 10, 253-271,
1984.
15. Tomm, K., Circular interviewing: A multifaceted clinical tool. In D. Campbell & R. Draper (eds.),
Applications of systemic family therapy: The Milan approach. London: Grune & Stratton, 1985.
16. Tomm, K., Lannamann, J. and McNamee, S., No interview today: A consultation team intervenes
by not intervening, Journal of Strategic and Systemic Therapies, 2, 48-6, 1983.
17. Viaro, M. and Leonardi, P., Getting and giving information: Analysis of a family interview
strategy, Family Process, 22, 27-42, 1983.

Manuscript received September 25, 1985; Revisions submitted July 3, 1986; Accepted July 3, 1986.

[pg 7]
Fam Proc 26:167-183, 1987

Interventive Interviewing: Part II. Reflexive Questioning as a Means

to Enable Self-Healing
KARL TOMM, M.D.a
aFamilyTherapy Program, Department of Psychiatry, University of Calgary, 3330 Hospital Drive, N.W., Calgary, Alberta,
Canada T2N 4N1.

Abstract
Reflexive questioning is an aspect of interventive interviewing oriented toward enabling clients or families
to generate new patterns of cognition and behavior on their own. The therapist adopts a facilitative posture
and deliberately asks those kinds of questions that are liable to open up new possibilities for self-healing.
The mechanism for the resultant therapeutic change in clients is postulated to be reflexivity between levels
of meaning within their own belief systems. By adopting this mode of enquiry and taking advantage of
opportunities to ask a variety of reflexive questions, a therapist may be able to augment the clinical
effectiveness of his or her interviews.

Introduction
A major stimulus for the work summarized here came from an interesting experience in
Rotterdam, Holland, in 1981. I happened to be behind a one-way mirror observing a family therapy
session being conducted by a trainee. The family consisted of middle-aged parents and eight
children (latency to teen age). They were referred because the father had been unduly violent in
disciplining the older boys. A series of circular questions quickly revealed that there was a split in
parenting functions, with the mother taking the warm nurturing role and the father the firm
disciplinary role. Indeed, the children described their father as quite a tyrant. He was regarded as
an uncaring person who was always angry and unreasonable in his parental demands. The
nonverbal behavior of the children indicated a strong coalition with their sympathetic and
supportive mother. As the session progressed, the father became increasingly tense and withdrawn.
Becoming somewhat concerned about the tension that had developed in the session, I
interrupted the interview and suggested that the trainee ask each child: "If something were to
happen to your mother so that she became seriously ill and had to be hospitalized for a long time,
or perhaps even died, what would become of the relationship between your father and the rest of
the children?" When the trainee resumed the interview and asked this question, the first child
exclaimed: "Oh, he would get even worse! He would become more violent!"; the next responded:
"But he might see another side of us because we would have to get him to help us with our
schoolwork"; another remarked: "Yes, he would probably help us with the cooking and the cleaning
Fam Proc 26:167-183, 1987

too”. By the time all the children had answered, the father was being talked about in warm,
nurturant terms and, of course, he relaxed and began participating in the discussion. The intent of
the question had been achieved and the trainee moved on to explore other areas of the family's
functioning.
Later, during the intersession discussion, the team elaborated a hypothesis about the
interpersonal dynamics of the family. There was consensus that the father was heavily blamed and
was relatively isolated in the family. This position left him vulnerable to excessive anger and
punitiveness. His hostility, in turn, had the effect of bringing the mother and children together,
triggering their collective blaming, and maintaining his isolation in a circular fashion. An end-of-
session intervention was developed that focused on disrupting this pattern. It took the form of a
paradoxical opinion positively connoting the father's uncaring, tyrannical behavior as helping the
mother and children to get closer and support one another (for the time being) because he knew
how much they would miss each other when the children left home. On hearing this opinion, the
children immediately protested, saying that their father was not uncaring or tyrannical. They
insisted that he was very affectionate and helpful! This response of the family was a surprise to the
team, especially after the father had been described so negatively during the initial part of the
session. On further reflection, however, it became apparent that while the team had remained
preoccupied with the information elicited at the beginning of the interview, the children had altered
their views of the father during the course of the session. In other words, the family's orientation
toward the father had changed more than that of the team! In retrospect, there was, in fact, no need
for the final intervention8.
How had this change in the family come about? It seemed that the question addressed to the
children about the effects of the hypothetical absence of the mother had been instrumental in
interrupting the malignant process of blame and had enabled the children to "bring forth" (5) a
construal of their father as a caring parent. This altered "reality" not only allowed the interview to
proceed more smoothly, but it also had healing potential for family members in that it was then
easier for them to explore new patterns of interaction. Thus, the question itself seemed to have

8
It was partly because of this incident that I first came to the conclusion, as indicated in Part I (9) of these papers on
Interventive Interviewing, that one could answer "yes" to the question posed by the Milan team: "Can family therapy
produce change solely through the negentropic effect of our present method of conducting the interview without the
necessity of making a final intervention?" (8, p. 12).
Fam Proc 26:167-183, 1987

functioned as a therapeutic intervention during the process of interviewing. But why was this
particular question so therapeutic? How was its impact mediated in the family?
As I pondered these issues, I began looking for other questions that seemed to have similar
therapeutic effects. To my [pg 1] delight, it was possible to identify a large number of them. Indeed,
it seems that most clinicians use these kinds of questions from time to time, albeit in differing ways
and with varying degrees of awareness. After discussing the nature of these questions with a
number of colleagues and exploring various possible explanations, I decided to call them
"reflexive”. Giving these questions a name turned out to be very useful. Reflexive questions
became more "tangible and real" to me. I subsequently began employing them more frequently in
my practice. In time, I noticed that therapeutic interventions were being introduced in the form of
reflexive questions in most of my sessions. The necessity of the formal end-of-session intervention
began to pale. Sometimes it seemed quite irrelevant, occasionally even contraindicated. What
transpired moment to moment during the interview became more important. Although I often still
use a carefully prepared final intervention, I now regard it as only one component of the treatment
process and not as the essential therapeutic agent, as I once did.

A theoretical rationale
The term "reflexive" was borrowed from the "Coordinated Management of Meaning"
(CMM), a theory of communication proposed by Pearce and Cronen (6). In CMM theory,
reflexivity is regarded as an inherent feature of the relationships among meanings within the belief
systems that guide communicative actions. A brief description of Cronen and Pearce's theory will
help explain what they mean by reflexivity and why I chose this term to characterize these
questions.
CMM theory regards human communication as a complex interactive process in which
meanings are generated, maintained, and/or changed through the recursive interaction among
human beings. That is, communication is not taken to be a simple lineal process of transmitting
messages from an active sender to a passive receiver; rather, it is a circular, interactive process of
co-creation by the participants involved. Pearce and Cronen originally set out to differentiate and
describe the rules that organize this generative process. Two major categories of rules were
delineated: regulative (or action) rules and constitutive (or meaning) rules. Regulative rules
determine the degree to which specific behaviors ought to be enacted or avoided in certain
situations. For instance, a regulative rule in a particular communication system might specify that
Fam Proc 26:167-183, 1987

"when one's integrity is challenged, it is obligatory to defend oneself”. Constitutive rules have to
do with the process of attributing meaning to a particular behavior, statement, event, interpersonal
relationship, and so on. For instance, a constitutive rule might specify that "in the context of an
argumentative episode, a compliment constitutes sarcasm or hostility rather than friendliness or
respect”. CMM theory proposes that a network of these regulative and constitutive rules guide the
moment-to-moment action of persons in communication.
Of particular relevance to the notion of reflexive questioning is the organization of
constitutive rules. Building on Bateson's (1) application of Russell's theory of logical types, Cronen
and Pearce suggest that the communication systems in which human beings are immersed entail a
hierarchy. They outline an idealized hierarchy of six levels of meaning rather than just two (report
and command levels), as popularized by Watzlawick, Beavin, and Jackson (10) and the Mental
Research Institute (MRI) group. These six levels include: content (of a statement), speech act (the
utterance as a whole), episode (that is, the whole social encounter), interpersonal relationship, life
script (of an individual), and cultural pattern. Following Bateson further, they postulate a circular
relationship between the levels in the hierarchy (not a lineal one as originally implied by Russell
and the early MRI group). For example, not only does the relationship (command level) exert an
influence in determining the meaning of the content (report level), but the content of what is said
also influences the meaning of the interpersonal relationship. The organizational relationships
between any two levels of meaning⎯content and speech act, content and episode, relationship and
life script, cultural pattern and episode, and so on⎯are circular or reflexive. The meaning at each
level turns back reflexively to influence the other. Thus, the Cronen and Pearce hierarchy is not
just a simple vertical organization, but a self-referential network.
Cronen and Pearce went on to describe the nature of this reflexive relationship among
constitutive rules. At any one moment, the influence of one level of meaning on another, for
instance, of item A at one level on item B at a lower level, may appear stronger than, vice versa,
the influence of B on A. In this case, Pearce and Cronen would say that A exerts a downward
"contextual force" within the hierarchy, with A determining the meaning of B. However, they point
out that while the relationship between these levels may appear lineal and stable, with B responding
passively to the dominance of A (as if in a vertical hierarchy), the relationship actually remains
circular and active. That is, B always continues to exert an upward "implicative force" on A. The
circular nature of the relationship becomes more apparent as the implications of B for A become
Fam Proc 26:167-183, 1987

more noticeable. For instance, the implicative force of B may be potentiated when connections are
made between aspects of B and certain meanings at levels higher than A. Furthermore, if the
implicative force of B increases in significance, its influence will eventually exceed the contextual
force of A. When this happens, the levels in the hierarchy suddenly become reversed. B then
becomes the context, and what previously was B's upward "implicative force" now becomes B's
downward "contextual force," which then redefines the meaning of A. Depending on the nature of
B, such a reversal may result in a dramatic change in the meaning of A. This could produce a
sudden change in communicative behaviors because a different constitutive rule now applies.
For example, suppose two individuals have an interpersonal relationship that they consider
friendly. Each person would expect to have a friendly episode of interaction if they happened to
meet. Thus, their initial actions would tend to be friendly [pg 2] and each would be oriented to
interpret the other's actions as friendly. In other words, the meaning attributed to the relationship
would provide the contextual force that determines the nature and meaning of the initial behaviors
in the interactive episode. But let us suppose that during the episode they entered into a discussion
and began to disagree about some issue. If the contextual force of friendliness continued to
predominate, they would regard the articulation of the incompatibilities of their respective positions
as helpful efforts to clarify and resolve their differences. Their discrepant points of view would,
however, still have implications for their relationship; the friendship might become strained.
However, if the incompatibilities widened and the disagreement evolved into an angry conflict
(perhaps because an ethnic or life-script issue at a higher level became implicated), the significance
of the episode could outweigh the original friendliness of the relationship. If this occurs, a reversal
takes place in the hierarchy and the episode of conflict then becomes the context for redefining the
relationship. With this recontextualization, the contextual force of the conflictual episode could
redefine the relationship as one of competitiveness or perhaps even as one of enmity. When this
happens, even a conciliatory statement or an apology is liable to be viewed with suspicion because
of the new context. Future episodes of interaction would then begin with different assumptions
about the relationship and with different behaviors.
A reversal of this type may have been triggered by the question addressed to the Dutch
family. By introducing the hypothetical scenario of the mother's absence (in the form of a reflexive
question), the relationship between the children and father was isolated from the mother and the
implications of the father's parenting in the family became more apparent. When the "implicative
Fam Proc 26:167-183, 1987

force" of the father's positive contributions as a parent became strong enough (perhaps partly
because all eight children were asked the same question and each built on the answers of the others),
a reversal took place between levels in the children's hierarchy of meanings so that their construal
of their relationship with their father changed from an uncaring one to a caring one. Such a change
is therapeutic and potentially healing because it places the father and children in a context that is
much more favorable for working toward a mutually acceptable solution.
More recent work in CMM theory has explored two variations in this reflexive relationship
between levels of meaning. Cronen, Johnson, and Lannamann (2) suggest that when the contextual
and implicative influences become relatively equal, through the activation of inherent reflexivity,
a "reflexive loop" is created. Two types of loops are described: strange loops and charmed loops.
A strange loop denotes a reflexive process in which a reversal of levels results in a major change
of meaning, that is, an opposite or a complementary constitutive rule is activated. A charmed loop,
on the other hand, denotes a reflexive process in which a reversal results in the meanings remaining
basically the same.
The change "from friends to enemies" described above, illustrates the effects of a reversal
mediated by a strange reflexive loop. It would appear that a similar kind of reversal occured in the
Dutch family, "from uncaring to caring”. In other words, the therapeutic effect of the question
addressed to the children could have been mediated by a strange loop. In both of the examples
cited, the change in meaning mediated by reflexive activity and recontextualization was followed
by a dramatic change in behavior: the "friends" became hostile, while the children and father
relinquished their pattern of escalating blame. In clinical terms, these changes could be referred to
as second-order change (11). The change associated with charmed reflexive loops is different.
Because meanings remain basically the same (despite reflexive recontextualization), only first-
order change occurs in the ongoing behavior. For example, there is little difference in behavior if
an amicable episode serves to redefine a friendly relationship as amicable. Similarly, not much
changes when a hostile relationship is recontextualized by an episode of confrontation. The changes
with charmed loops are not major or dramatic; they tend to be small and subtle. The activation of
reflexivity mediated by charmed loops only results in patterns becoming somewhat more
generalized or more deeply entrenched. However, the process of generalization and/or
entrenchment is extremely important. A therapist can ask questions to facilitate an extension of
healthy patterns that already exist in the family, or ask questions to stabilize new therapeutic
Fam Proc 26:167-183, 1987

developments that are still tenuous. In other words, some reflexive questions may realize their
healing potential through charmed loops. For example, during the interview with the Dutch family,
the trainee conceivably could have gone on to strengthen the change triggered by the initial
reflexive question, by asking a further series of reflexive questions as follows: (to mother) "When
they are at home, which of the children would be the most likely to see how much your husband
does to help them?... Who would be second most likely to notice?... Who third most?"; (to children)
"If your father was convinced that, deeper down, you recognized and appreciated the things he
does for you, would it be easier or more difficult for him to tolerate some of your mistakes?... When
you think of your father as a caring parent, are you more, or less, inclined to do what he asks of
you?"; (to father) "If you decided that as a father you wanted to convince Jan that you really cared
for him, how would you go about it?... If you were to apologize afterwards, when you recognized
that you had gone too far in your discipline, do you think he would respect you more as a caring
parent, or less? ... If your wife decided to try to help him see more of your positive contributions
to the family, what might she do?" These questions might have enabled further consolidation of
the "new reality" by orienting the family toward perceptions and actions that reflexively supported
the new construal of the relationship between the father and the children.
Thus, from a theoretical point of view, the therapeutic effects of reflexive questions may be
mediated by strange loops or by charmed loops. The questions themselves remain as probes,
stimuli, or perturbations. They only trigger reflexive activity in the connectedness among meanings
within the family's own belief systems. This explanation acknowledges the autonomy of the family
with regard to what change actually occurs; that is, the specific effects of the questions are
determined by the [pg 3] client or family, not by the therapist. Change occurs as a result of
alterations in the organization and structure of the family's pre-existing system of meanings. Given
this formulation, the basic mechanism of change is not insight, but reflexivity. The organizational
alterations do not enter consciousness (even though family members subsequently may become
aware of the effects or consequences of reflexive changes). It is on the basis of this possible change
mechanism that these questions are referred to as reflexive9.
By definition then, reflexive questions are questions asked with the intent to facilitate self-
healing in an individual or family by activating the reflexivity among meanings within pre-existing

9
Although the choice of the adjective "reflexive" was not based on grammatical usage, as with reflexive verbs (where
the subject does something to itself), the similarity is compatible and apt.
Fam Proc 26:167-183, 1987

belief systems that enable family members to generate or generalize constructive patterns of
cognition and behavior on their own. It is important to note that the designation of certain questions
as reflexive is based on the therapist's intent in asking them, that is, to facilitate the family's own
self-healing. The significance of intentionality in distinguishing reflexive questions from other
kinds of questions, such as circular, lineal, or strategic ones, will be discussed in Part III. Suffice it
to say here that these questions are not defined on the basis of their semantic content or syntactic
structure, but on the nature of the therapist's intentions in employing them. The process of asking
them is referred to as reflexive questioning. It implies a carefully considered and deliberate use of
language that entails a conceptual posture of strategizing that is facilitative rather than directive.

Types of Reflexive Questions

The variety of questions that could be employed reflexively is enormous. They can be as
varied as the hypotheses a therapist can formulate about the problems of an individual client or
family and the strategies he or she might consider useful in enabling family members to find
alternatives in their problem-solving activities. In presenting the notion of reflexive questioning to
colleagues, I have found it useful to provide examples of reflexive questions that seem to fall into
natural groups: future-oriented questions, observer-perspective questions, unexpected context-
change questions, embedded-suggestion questions, normative-comparison questions, distinction-
clarifying questions, questions introducing hypotheses, and process-interruption questions.
Although the questions within these groups are linked by one or two basic concepts, there is
considerable overlap among them. Their sequence and classification does not provide a recipe for
the conduct of an interview. The specific examples are offered only to illustrate the kinds of
questions that could be employed to utilize momentary opportunities for therapeutic intervention
while respecting the autonomy of the family to generate solutions on their own. To be appreciated
fully as reflexive, each question would have to be placed in the context of a therapeutic scenario
like that of the Dutch family and analyzed in terms of the reflexivity of CMM theory.

Future-Oriented Questions
This constitutes an extremely important group. Families with problems are sometimes so
preoccupied with present difficulties or past injustices that, in effect, they live as if they "have no
future”. That is, they focus so little on the time ahead of them that they remain impoverished with
respect to future alternatives and choices. By deliberately asking a long series of questions about
Fam Proc 26:167-183, 1987

the future, the therapist can trigger family members to create more of a future for themselves.
Members of "present-bound" or "past-bound" families may not be able to answer these questions
during the session. But this alone should not deter the therapist from asking them. Family members
often "carry the questions home" and continue to work on them on their own. Future eventualities
do, of course, have significant implications for present commitments and behavior. It is through
these implications that future questions realize their reflexive effects10.
One can delineate several subtypes of future-oriented questioning. The most straightforward
and simple is to cultivate family goals: collective goals, personal goals, or goals for others. For
example, one might ask an adolescent daughter who is underachieving at school: "What plans do
you have for a career?... What else have you considered?... How much formal education do you
think you would need?... What kinds of work experience would be useful in getting that sort of
job?... How will you go about getting it?"; (to parents) "What accomplishments do you have in
mind for your daughter?... What would be reasonable for the next year?... Are there any goals that
you all agree on and see yourselves working toward together right now?... How do you plan to help
her reach these goals?" If the therapist thought it useful for family members to operationalize vague
goals, he or she might ask: "How will you know when that goal has been achieved?... What would
she have to do to show that she had completed it?... What specific behavior would be the most
convincing to you?" In asking these questions reflexively, the therapist is less interested in the
particular content of the answers than in the fact that family members entertain the questions and
begin to experience the implications that the answers might have. Nevertheless, the answers do
become a useful source of data for the therapist's ongoing hypothesizing and strategizing about
what further questions to ask.
Further future-oriented questioning that follows naturally might be to explore anticipated
outcome: "How much progress do you think she actually will make in the next month?... In six
months?... Who would be the most surprised if she exceeded that objective?... Who is most liable
to be disappointed if she fell short?... How would this disappointment show?" If the therapist
wanted to highlight potential consequences that might arise when certain patterns continued, she
or he might ask: "If your husband continued to show his disappointment the way he does now,
what do you expect would happen to their [pg 4] relationship?... What about in 5 years from now?...
What kind of father-daughter relationship would have resulted by then?" To explore catastrophic

10
Using a different theoretical frame, Penn (7) has described the use of future questions as a "feed-forward" technique.
Fam Proc 26:167-183, 1987

expectations is a way to facilitate the exposure of covert issues so that they can be dealt with more
overtly. For instance, one might ask overprotective parents: "What are you worried might happen
when your daughter stays out so late?... What is the worst thing that comes to mind?"; (to daughter)
"What do you imagine your parents are most afraid of?... What terrible things do they expect might
happen that keeps them awake at night?" When family members are reticent to be open, these
questions can be followed by others to explore hypothetical possibilities: (to daughter) "Do you
imagine that your parents might be worried about your getting into drugs or alcohol?... Are they
afraid that you might get pregnant?... Are they even too afraid to mention this fear because they
think it might offend you?"; (to parents) "If you did raise these worries with her, do you think she
would take it as a lack of trust?... As an intrusion into her privacy?... Or as an indication of your
caring as a parent?" Additional questions could be used to suggest future construal and/or action:
(to parents) "If you decided that you really cannot control her sexual behavior, felt that she needed
to know more about the risks of pregnancy, and suggested she see the family doctor about birth
control pills, would she take this as permission for sexual promiscuity or as an indication of your
support for her to take more responsibility for her own life and behavior?... If she became indignant,
or even furious, when some guy got fresh and tried to take advantage of her, would you be
surprised?; (to daughter) Would your parents support you if you raised assault charges against
him?"
Future-oriented questions that introduce hypothetical possibilities allow the therapist to
share his or her own ideas in a process of co-creating a future together with the family. They can
be used to stimulate families to entertain possibilities that they may never have considered on their
own, yet are compatible with their pre-existing values and beliefs: (to parents) "Can you imagine
that her heavy commitment to being with her friends and, in so doing, developing excellent social
skills, could result in a successful career in the field of promotions?... With her talent for talking,
how do you think she would do in sales?... How do you think she would score on 'human relations'
in an aptitude test?... Is that kind of testing available at school?... Where could you get it?" What
is so enticing about hypothetical future questions is that they offer unlimited opportunities for a
therapist's creative imagination. The question format can even be used to introduce stories and to
pose dilemmas: (to daughter) "Let us imagine your sister meets a young man that she likes a lot,
and he cares enough about her to try to get her to stop drinking, do you think she would be more
willing to listen to his advice than that of your parents?... What do you think your parents might do
Fam Proc 26:167-183, 1987

if they did discover that he had more influence on her than they?... Would they still refuse to let
her go out, or would they encourage her to spend time with such a friend?" Future questions can
also be employed to instill hope and to trigger optimism: (to parents) "When [not if] she does find
a way to take better care of herself, who will be the first to notice?... In what way will your relief
or gratitude show?... How will it improve your relationship?... Who would be the first to suggest
the change be celebrated?"

Observer-Perspective Questions
This group of questions is based on the assumption that becoming an observer of a
phenomenon or a pattern is a necessary first step toward being able to act in relation to it. For
instance, it is impossible to empathize with another person when one is unable to make some
observations of the experiential conditions of the other. In addition, when family members do not
recognize how they are inadvertently hurting each other and themselves in the process, they cannot
apply their good will to correct their own behavior. Observer-perspective questions are oriented
toward enhancing the ability of family members to distinguish behaviors, events, or patterns that
they have not yet distinguished, or to see the significance of certain behaviors and events by
recognizing their role as links or connections in ongoing interaction patterns. Asking a series of
these questions often helps family members to "open their eyes" and develop a new awareness of
their situation. It is, of course, possible to make direct statements and point out certain
circumstances to family members, instead of trying to achieve this indirectly by asking questions.
This may be much more efficient and more desirable on some occasions. However, there are
advantages to creating a context in which they can generate the new distinctions for themselves.
First, family members are stimulated to develop better observational skills when asked to reflect
on their own behavior and interaction patterns. Second, when they actually draw new distinctions
on their own, they experience the heuristic observational resources in themselves and other family
members and develop greater confidence in their own healing potential. Consequently, they
develop less dependency on the therapist and on therapy.
Observer-perspective questions may be categorized according to the person being asked to
comment and the person(s) or relationship(s) being asked about. For instance, questions addressed
to an individual may be used to enhance self-awareness, that is, to become a better observer of
oneself: "Just how did you react?... How did you interpret the situation that triggered those
feelings?... When you responded the way you did, how did you feel about your reaction?... What
Fam Proc 26:167-183, 1987

else could you have done?... If you had the chance, what would you do differently?" Questions
about another's experiences may encourage "other" awareness: "What did he think about it?...
What do you imagine he experiences when he gets into a situation like that?... When he thinks that
way, how does he feel?" These are sometimes referred to as mind-reading questions. They may be
elaborated further to explore interpersonal perception: "What does he think that you think is going
on when he threatens suicide?... If he got the impression that you thought he was not really that
upset and was just trying to get attention, do you think he would feel less suicidal or even more?"
Questions asked to explore [pg 5] interpersonal interaction focus on behavior patterns and may
include or exclude the person being addressed. They are extremely useful in drawing attention to
the recursiveness of behavior patterns in dyadic, triadic, or more complex relationships. For
instance, to help a married couple to see the circular nature of their interaction one might ask the
wife: "What do you do when he gets depressed and withdraws?... And when you get frustrated and
angry, what does he do?" and then the husband: "What do you do when she gets frustrated and
angry?... And, when you get depressed and withdraw, what does she do?" It is easier for a couple
to interrupt such a pattern when they can see the circular patterning of it than when they are limited
to seeing only their own lineal reactiveness. In systemic therapy, "triadic questioning" refers the
use of a series of questions that are addressed to third parties about interactions between two (or
more) other persons. In other words, triadic questions explore interpersonal behavior patterns that
do not include the person being addressed, thus enabling that person to become a more neutral
observer: "When your father gets into an argument with your sister, what does your mother usually
do?... Does she get involved or stay out of it?... When she gets involved, does she usually take his
side or hers?... When she takes your sister's side, what does your father do?... Does he feel betrayed
by her or does he appreciate her involvement to help him realize he has gone too far?" These kinds
of questions are often used for assessment purposes, but they may also be used reflexively.
One advantage of seeing family members together in family therapy, rather than alone in
individual therapy, is that the process of asking questions of one family member in the presence of
others always places the others in the position of being observers. These "passive" observers obtain
a great deal of information. Not only do they see and hear the overt responses of the person being
addressed and see the nonverbal responses of others, they also obtain information from their own
private responses to the questions, from "the differences" between their private responses and the
addressee's actual responses, and from "the differences" between how the addressee did respond
Fam Proc 26:167-183, 1987

compared to what the observers may have anticipated. These phenomena are always taking place
in marital and family therapy, but they may be used deliberately (through the use of observer-
perspective questions) to help family members to see or hear certain things. To do so effectively,
the therapist must become sufficiently coupled with family members to see what they are seeing
and are not seeing, and hear what they are hearing and are not hearing. In other words, therapists
should strive to observe their clients' observing, and listen to their clients' listening, as they
strategize about precisely what questions to ask.
It is interesting to note that individuals do not necessarily have to become conscious of an
observation for it to have an effect on their behavior. The phenomena and the connections implied
in the therapist's questions or in the family's answers may be recognized nonconsciously and still
trigger a change in patterns of thought and action. On the other hand, explicit awareness of an
object or process is necessary for family members to act on it with conscious intent. Thus, observer-
perspective questions may operate at two levels of complexity with respect to the observer/listener.

Unexpected Context-Change Questions

Every quality, meaning, or context may be regarded as a distinction that is made in contrast
to some other distinction, that is, to an opposite or a complementary quality, meaning, or context.
Yet, the act of drawing a particular distinction often masks its complement or opposite. It is easy
to forget that "the bad" exists only in relation to "the good" and that sadness and despair exist only
in contrast to happiness and hopefulness. Questions to trigger an unexpected change in context
focus on bringing forth that which has been masked or lost. Family members often get themselves
locked into seeing certain events from one perspective, and their behavioral options are constrained
accordingly. They may need help to see the reciprocal view in order to open up new possibilities
for themselves. A few well-placed questions can sometimes do this, that is, pry them free of a
limiting cognitive set and enable them to entertain other perspectives.
One subtype of unexpected context-change questioning is to explore opposite content. For
instance, a couple came in complaining about the wife's depression. They explained how they had
endured a long series of serious physical illnesses in various members of the nuclear and extended
families over the past few years. The wife had been deeply involved with problems posed by these
illnesses, and she continued to be heavily preoccupied with them. Her despondency was easy to
understand. A reflexive enquiry along the following lines triggered a transformation: "When was
the last occasion that the two of you had a good time together? ... What do you do these days that
Fam Proc 26:167-183, 1987

you find enjoyable?... What kinds of events do you usually celebrate?... What about together as a
whole family?... What kinds of things are you most grateful for?" The wife suddenly realized that
they were all still alive, they had a good income, a comfortable home, and so on. At the next session
the couple cheerfully announced that they had decided to terminate therapy and were taking a
holiday "for the first time in years”.
An intermittent question or two that introduces the opposite or complementary side of an
issue can enhance the interest of family members in the proceedings as well as loosen fixed patterns
of perception and thought. For instance, in the context of complaints about incessant arguing and
fighting (which are taken for granted as unwanted), one could explore an opposite context: "Who
in the family enjoys the fighting the most?... Who would experience the greatest emptiness and
loss if it all suddenly stopped?"; or explore opposite meaning: "Who would be the first to recognize
that father gets angry because he cares too much rather than too little?" Similar kinds of questions
may be formulated to explore a need to [pg 6] conserve the status quo: "Let us assume that there
was an important reason for you to continue in this uncomfortable pattern, what could it be?... What
is happening in your family that needs this kind of behavior?... What other more serious problem
may this difficulty be solving or preventing?" The latter line of enquiry is, in fact, a method of
triggering the family to generate their own positive connotation of problematic patterns.
These questions also can be used to introduce paradoxical confusion: "How good are you
at stealing?... How come you get caught so easily?... Can you not steal any better than that?" The
implications of such questions stir up a paradox: stealing is good, yet it is bad; getting caught is
bad, yet it is good. With care, these questions can even be used to join feared impulses transiently:
"Why is it that you have not killed yourself already?... Which ideas and thoughts need to die?...
Are there some patterns of behavior that do, in fact, need to be destroyed and buried?" When
addressed to a client who has become entrapped in a struggle against suicidal thoughts, these
questions may be experienced as liberating, and facilitate a fresh re-evaluation of the situation.

Embedded-Suggestion Questions
These questions are helpful when family members need to be nudged along with a little
more specificity. In each question, the therapist includes some specific content that points in a
direction he or she considers potentially fruitful. However, when the therapist begins to push a
client too hard, for instance, to see problems or solutions the same way he or she does, these
questions become strategic (see Part III). This may not necessarily be problematic for the therapy
Fam Proc 26:167-183, 1987

but it sometimes leads to quasi-lecturing. The temptation to "drive home" the therapist's "truth" can
be minimized if, immediately after having asked the question, the therapist moves quickly back to
a posture of neutrality and accepts the family's responses, whatever they might be.
A wide variety of suggestions may be embedded in a question. For example, one may embed
a reframe: "If, instead of your thinking that he was being willfully stubborn, you thought that he
was just confused, so confused he did not even know he was confused, and that he simply did not
understand what you wanted of him much of the time, how do you imagine you might treat him?";
embed an alternative action: "If, instead of withdrawing or leaving when she got upset, you simply
sat with her or perhaps even put your arm around her shoulder, what would she do?... If you
persisted for a few minutes in a quiet and gentle manner despite a half-hearted rejection, would she
be more likely to accept your caring initiative as genuine?"; embed volition: (regarding an
anorectic) "When did she decide to lose her appetite?... When she decides to stop eating, what is it
that she is on strike about?"; embed an apology: "If, instead of not saying anything and avoiding
her, you admitted you made a mistake and apologized, what do you think might happen?"; embed
forgiveness: "When the time came that she was ready to forgive you, would she do so silently or
would she be explicit about it?... To what extent would you be able to forgive yourself?”.
Any question may be analyzed on a post-hoc basis as containing one or more embedded
suggestions. However, to be considered a reflexive question, the embedding would not have
occurred inadvertently but, rather, deliberately as a part of the therapist's therapeutic intent.

Normative-Comparison Questions
Individuals and families with problems tend to experience themselves as deviant or
abnormal. They inevitably develop a longing to become more normal. A therapist may take
advantage of this desire and help family members orient themselves toward healthier patterns by
asking them to make relevant comparisons. For instance, if conflict is typically suppressed in a
family, one might ask questions to draw a contrast with a social norm: "Do you think that you are
more open about your disagreements than most families, or less?... Do you know some healthy
families that are able to express their frustration and anger openly?... Can you imagine that they
actually find it useful to express their frustration in order to clarify important underlying issues?"
Questions also could be used to raise a contrast with a developmental norm: "In most families at
this stage in life, boys are closer to their fathers. What is keeping Juan so close to his mother?"; or
to contrast a cultural norm: "If you were an English-American family, do you think there would
Fam Proc 26:167-183, 1987

be less involvement between your wife and son?" The latter would, of course, only be appropriate
if the family were of a different ethnic origin and were interested in becoming more acculturated.
By drawing attention to specific ways in which the family deviates from a norm, the therapist helps
connect relevant lower-level meanings to higher-level cultural patterns, thus triggering changes in
the reflexive organization of the family's own belief systems.
The implications of normality may be employed in another manner. Rather than focus on
differences, the therapist could highlight similarities. This would be indicated if the therapist felt
that the perceived deviation from normal was generating progressive isolation and alienation. To
emphasize differences in such a situation could risk further alienation and actually interfere with
the family's ability to draw on "normal" social solutions. Thus, instead of drawing a contrast with
normality, one could work on helping family members redefine themselves as normal. For instance,
some questions could be oriented toward social normalization: "All families have problems dealing
with anger. When did you first realize that you had the [pg 7] same difficulty?" toward
developmental normalization: "Since most families eventually have to struggle with the problem
of children leaving home, who do you know of that would understand your situation most readily
because they had just been through it? ... Which parent do you imagine typically has the most
difficulty?"; or toward cultural normalization: "If your mother found out that most American
mothers have a terrible time when the last child leaves home, would she be surprised?"
It is useful for therapists to think in terms of generating a process of inclusive normalization
when formulating questions to facilitate a sense of belonging for an alienated individual. For
example, when someone is suicidal, one might ask another family member: "Do you imagine she
feels isolated and disconnected from everyone when she feels suicidal?... Would she be surprised
to find out that most people have suicidal thoughts at some time during their lives?... Say one of
her friends confided in her and admitted that she also had suicidal feelings, would she believe
her?... Say she found out that an acquaintance actually did attempt suicide once, do you think she
would be shocked?... If she realized how common these issues were, would she more likely be able
to talk about them?... Would you be surprised if some day she mustered up the courage to ask
someone else how he or she got through a similar difficult time?... What do you imagine helps most
people find solutions other than suicide?" By addressing these questions to another person in the
presence of the suicidal individual, the latter is given more space to entertain the questions and
Fam Proc 26:167-183, 1987

their implications. This is desirable when the social expectations for an explicit response from an
isolated individual might inadvertently generate further alienation.
If the alienated individual is a child, it is useful to orient the inclusiveness to the family:
"Say everybody in the family stole something at some time in their lives, who do you imagine may
have stolen the most?... The second most?... And then who?... Some people are so good at lying
and stealing that no one ever knows. Who in the family do you imagine might have been the best
at it?... Second best?... Who had the most difficulty stopping?... Second most?" A series of
questions like these could enable a child, who has become isolated, defensive, or defiant because
of judgmental family reactions toward lying and stealing, to become reconnected as a "normal"
member so that corrective efforts are more likely to be heard, accepted, and heeded.

Distinction-Clarifying Questions
Introducing or clarifying a key distinction can have major implications in any system of
beliefs. These implications may be quite therapeutic, especially when there is considerable
confusion surrounding the issues related to the problem. For instance, when family members' causal
attributions are not clear, the chances of being consistent or coordinated in problem-solving efforts
is unlikely. A therapist could ask a series of questions with the intention of helping to clarify causal
attributions that are already held by family members but that are inconsistent or unclear. When
such confusion is covert or pervasive, it is often useful to ask the same question of several family
members and to approach the same issue from different points of view in order to give family
members ample opportunity to entertain the ramifications of the distinctions. In a recent case that
was referred because an adolescent girl was apprehended during a major theft after recurrent
stealing episodes, the same basic question was addressed to each family member about every other
member's views and finally about their own: "Do you think that your father (your mother, your
brother, your sister, or you yourself) sees stealing as more 'socially bad,' more 'psychologically
sick,' or more 'sinful'?" This series of questions helped clarify underlying assumptions about the
nature of the problem and inconsistencies in their corrective efforts. One unexpected consequence
was the father's initiative (after the session) to mobilize some helpful religious resources. Another
was the daughter's clear recognition of the legal risks involved, which she then used successfully
to curtail the temptations she often faced. Similar questions may be used to clarify family members'
assumptions about the degree to which various biological, psychological, or social factors are
Fam Proc 26:167-183, 1987

operating in the maintenance of a variety of problematic behaviors. Different assumptions do, of

course, have different implications for problem solving.
A variety of questions may be used to clarify categories: "When she is crying, is it because
she is whining to get her way or is she weeping out of emotional pain?... Do you think your father
has even more difficulty telling the difference between whining and weeping," to clarify sequences:
"Did you take the pills [regarding an overdose] before or after the discussion about leaving home?"
and to clarify dilemmas: "What is really most important for you, being highly successful in your
career or having a rich family life?... If it were impossible to have both, in which would you prefer
to invest your limited time and energy?... Who would be the first to recognize that in an effort to
avoid facing this dilemma you might, indeed, be sacrificing both?" Clarifying questions may
operate either by separating components of a pattern and thereby decomposing vagueness or by
connecting elements into a pattern thereby creating new units of distinction. The latter can
sometimes be achieved with questions that deliberately introduce a metaphor: "Is he getting to be
more and more like a porcupine, the closer you get the sharper and more prickly he becomes?... Or
is he getting more like a watermelon seed, the harder you press him, the further he flies away?", or
introduce hypotheses, a major group that will be discussed below.
A therapist's attention to the distinctions made by family members may be useful in another
way. When families have been stuck in problematic patterns for a long time, it is reasonable to
assume that some family members are probably holding some crucial distinctions with too much
clarity or too much certainty. This would, of course, constrain their ability [pg 8] to entertain
alternative distinctions. The therapist may be able to assist the family in opening up new domains
by identifying the crucial underlying presuppositions and ask questions to invite uncertainty: "How
long have you had these ideas?... When did you first begin to think that way?... If you did happen
to be mistaken, how could you find out?... How long would it take for you to see that the situation
may not, in fact, be as it appears to be?... If you were blind to what kept these things happening,
who would be the first to see your blindness?... Is there anyone who would bother to try to convince
you that your views were mistaken?... Would you ever actually invite someone else to help you see
what you cannot see? ... Whom do you respect enough that you could believe, if they had ideas
different than yours?" To be reflexive, the tone with which these questions are asked would have
to be neutral and the posture of the therapist one of acceptance. Otherwise, they could constitute a
strategic confrontation.
Fam Proc 26:167-183, 1987

Questions Introducing Hypotheses

Clinical hypotheses are tentative explanations that serve to orient and organize the
therapeutic behavior of therapists. It is reasonable to assume that they also could serve to orient
and organize the self-healing behavior of family members. If there is no good reason to withhold
the therapist's working hypothesis, he or she may enrich the family's ability to find new solutions
on their own by introducing heuristic hypotheses in the form of questions. The question format
tends to convey the tentativeness that is important in systemic hypothesizing, compared to a direct
statement or explanation that implies more certainty. If the hypothesis is coherent and fits the
experiences of family members, immediate and dramatic changes may take place. If not, the family
often provides highly relevant information for the therapist to revise or elaborate the hypothesis.
To have an impact, the hypothesis need not be comprehensive or complete. Partial hypotheses can
be very useful. Indeed, the therapist and family can begin to function almost like a clinical team to
co-create a more systemic understanding of the situation.
The subtypes of this group may be extensive. Only a few examples will be included here to
illustrate how some aspects of clinical hypotheses can be introduced. Questions may be asked to
reveal recursiveness: "When you get angry and she withdraws, and when she withdraws and you
get angry, what do the children do?"; to reveal defense mechanisms: "When he can't tolerate his
own shame and guilt, but gets angry at you instead, what do you imagine might make it easier for
him to acknowledge and accept the pain?"; to reveal problematic responses: "If he does get angry
to cover up his vulnerability and you just can't reach out to connect with his underlying sadness,
does he see you as punishing and vindictive, or does he see you as simply protecting yourself, or,
perhaps, even as paralyzed by your fear?"; to reveal basic needs: "In order to grow and mature
naturally, what kind of protection and nurturing does she need the most?... Mainly some physical
and emotional space to exist and express herself?.... Being provided with comfort and support?...
Being given guidance and direction?" and to reveal alternative motives: "In looking for a mate
during courtship, what do you think your wife was looking for most? Was she looking more for a
companion for herself, for a father for her children, for someone to support her and the children
economically, for a sexual partner, or what?" Questions may also be formulated paradoxically to
reveal dangers of change: "If he were forced to acknowledge his own contributions to your
depression, even to himself, do you think he could handle it?... Or do you imagine him finding
himself overwhelmed with guilt and becoming suicidal?" A fully elaborated systemic hypothesis
Fam Proc 26:167-183, 1987

may be too complex to be incorporated into a question and may be more appropriate in the form of
a statement. Needless to say, no therapist should feel constrained to ask only questions.
Therapists and teams often formulate hypotheses about the treatment process as well as
about the family. Hence, questions may be asked in order to reveal hypotheses about the
therapeutic system: "If I began to relate to you more like a family member rather than like a
professional, how would this become apparent?... Who among us would be the first to notice?... If
I began siding with him again but did not realize it, would you point it out to me?"; or to expose a
therapeutic impasse: "Say it was impossible for me to be of any real help to you because my input
automatically disqualified your sense of self-sufficiency, what would you do?... If I decided that
only you could decide whether continuing therapy would be useful for you or not, could you accept
that?"

Process-Interruption Questions
There is an interesting group of questions that may be used to remark upon the immediate
process of an interview. For instance, if a conflictual couple began to argue during the course of
the session and the interaction appeared to be unfruitful and destructive, the therapist might address
the children with questions to expose the current process: "When your parents are at home, do they
argue as much as they do here?... Or is it even more intense?... Who among you is the most likely
to try to intervene?... To try to keep clear?" As the couple begin to follow the conversation about
them, which the therapist has initiated with the children, their arguing is interrupted and they are
triggered into assuming an observer perspective that helps curtail the process. This is certainly a
more elegant way to deal with this common problem in therapy than by asking or demanding that
the couple stop their fighting. The couple stop themselves reflexively.
The focus for these questions may also be to reflect on the therapeutic relationship: "Do
you think I may have offended your father by the way I have been asking these questions?... Could
it be that I've been getting caught up in seeing mainly [pg 9] your mother's side of things?"
Sometimes the therapist may wish to use a question to make an indirect therapeutic-process
comment. For example, if parents are giving the children cues (nonconsciously) to avoid the
disclosure of sensitive information, a therapist might choose to ask: "I know you would never do
this, but say you went to the neighbors and told them everything that was going on at home, who
would be the most upset?" Such a question helps reveal the source of the constraint and may trigger
the parent to give the child explicit permission to speak up because therapy is a different context.
Fam Proc 26:167-183, 1987

Nevertheless, unexpected disclosures during an interview may place family members at risk for
retaliation after the session. In this case the therapist could ask questions to minimize remote
reactions: "Do you think she might be frightened that you will be furious with her after you leave
the session, because of what she said?... If she was, would she admit it?... Even to herself?... Or
does she think that you recognize the need for her to get her complaints out, so they can be talked
about even though they are upsetting?" Finally, a series of questions may be asked to facilitate
readiness for termination: "Do you ever wonder if continuing therapy might actually interfere with
your ability to learn how to find solutions on your own?... If therapy did stop, who would be the
most upset?... Who would be the most relieved?... Do you ever hear yourselves asking the kinds of
questions we discuss here?"

Concluding comments
This sampling of reflexive questions is not intended to be comprehensive or complete.
Rather, it is intended to illustrate the variety of questions that could be used in this manner and to
provide sufficient examples to enable their distinctive character to emerge. Seasoned clinicians will
recognize many of these questions as familiar. Indeed, they probably have used some of them for
years, possibly in a similar manner, perhaps in a different one. However, it is not the specific
questions themselves to which I am trying to draw attention. It is the realization that they can be
carefully differentiated and intentionally employed to facilitate a family's self-healing capacity. If
this realization becomes part of a therapist's ongoing process of strategizing about what questions
to ask during an interview, his or her therapeutic impact may be enhanced substantially.
As noted in Part I (9), several other authors have examined the process of conducting a
systemic interview. Some of them also have explored the use of questions as therapeutic
interventions. For instance, Lipchik and de Shazer (4) describe "the purposeful interview" and
delineate a group of "constructive questions”. Fleuridas, Nelson, and Rosenthal (3) include
"interventive questions" in their listing of circular questions. White (12) describes "cybernetic
questions" and "complementary questions”. In some respects, all of these are similar to the
reflexive questions described here, especially those of White. There are, however, some
differences. Reflexive questioning focuses more heavily on an explicit recognition of the autonomy
of the family in determining the outcome. This has an important effect on both the therapist's choice
of question and his or her manner of asking. These issues will be explored further in Part III.
Fam Proc 26:167-183, 1987

References
1. Bateson, G., Steps to an ecology of mind. New York: Ballantine Books, 1972.
2. Cronen, V. E., Johnson, K. M. and Lannamann, J. W., Paradoxes, double binds, and reflexive loops: An
alternative theoretical perspective, Family Process, 21, 91-112, 1982.
3. Fleuridas, C., Nelson, T. S. and Rosenthal, D. M., The evolution of circular questions: Training family
therapists.
Journal of Marital and Family Therapy, 12, 113-127, 1986.
4. Lipchik, E. and de Shazer, S., The purposeful interview, Journal of Strategic and Systemic Therapies, 5,
88-99,
1986.
5. Maturana, H., personal communication, 1986.
6. Pearce, W. B. and Cronen, V. E., Communication, action and meaning: The creation of social realities.
New York:
Praeger, 1980.
7. Penn, P., Feed-forward: Future questions, future maps, Family Process, 24, 299-310, 1985.
8. Selvini-Palazzoli, M., Boscolo, L., Cecchin, G. and Prata, G., Hypothesizing-circularity-neutrality: Three
guidelines for the conductor of the session, Family Process, 19, 3-12, 1980.
9. Tomm, K., Interventive interviewing: Part I. Strategizing as a fourth guideline for the therapist. Family
Process,
26, 3-13, 1987.
10. Watzlawick, P., Beavin, J. H. and Jackson, D. D., Pragmatics of human communication: A study of
interactional
patterns, pathologies, and paradoxes New York: W.W. Norton, 1967.
11. Watzlawick, P., Weakland, J. H. and Fisch, R., Change: Principles of problem formation and problem
resolution.
New York: W. W. Norton, 1974.
12. White, M., Anorexia nervosa: A cybernetic perspective. In J. Harkaway (ed.), Family therapy and eating
disorders.
Rockville: Aspen Systems Corp., 1986.
Manuscript received September 30, 1986; Accepted September 30, 1986.
[pg 10]
[pg 11]
Fam Proc 27:1-15, 1988

Interventive Interviewing: Part III. Intending to Ask Lineal, Circular,

Strategic, or Reflexive Questions?
KARL TOMM, M.D.a
aFamily Therapy Program, Department of Psychiatry, University of Calgary, 3330 Hospital Drive, N.W., Calgary, Alberta, Canada
T2N 4N1.

Abstract
Every question asked by a therapist may be seen to embody some intent and to arise from certain
assumptions. Many questions are intended to orient the therapist to the client's situation and experiences;
others are asked primarily to provoke therapeutic change. Some questions are based on lineal assumptions
about the phenomena being addressed; others are based on circular assumptions. The differences among
these questions are not trivial. They tend to have dissimilar effects. This article explores these issues and
offers a framework for distinguishing four major groups of questions. The framework may be used by
therapists to guide their decision making about what kinds of questions to ask, and by researchers to study
different interviewing styles.

Introduction
From the perspective of an observer, the psychotherapies are essentially conversations.
However, they are not ordinary conversations. Therapeutic conversations are organized by the
desire to relieve mental pain and suffering and to produce healing. They occur between therapists
and clients within the context of consensual agreement that the therapist will contribute
intentionally toward a constructive change in the problematic experiences and behaviors of clients.
Whereas other conversations may have therapeutic effects (for instance, personal discussions
among family members, friends, work associates, acquaintances, and even strangers), these would
not be considered "therapy" unless there were some agreement that one participant accepted
responsibility for guiding the conversation to be therapeutic for the other. Thus, a therapist always
assumes a special role in a conversation for healing. This role entails a commitment to be helpful
with respect to the personal problems and interpersonal difficulties of the other.
The therapist's position in a therapeutic conversation not only implies special
responsibilities, it also confers special privileges. One example of the latter is that a therapist can
legitimately inquire about the clients' personal and private experiences. To do so can often expose
the clients' vulnerabilities. Consequently, the potential for further trauma exists alongside that for
healing. It is the manner in which such an inquiry is carried out that makes the difference. Some
Fam Proc 27:1-15, 1988

patterns of conversing are much more conducive to being therapeutic than others. One of the factors
contributing to such variations is the nature of the questions asked.
During a conversation that is intended to be healing, the therapist usually contributes both
statements and questions. These are quite different kinds of utterances. In general, statements set
forth issues, positions, or views, whereas questions call forth issues, positions, or views. In other
words, questions tend to call for answers and statements tend to provide them. At the same time,
however, these characteristics are not exclusive; there is considerable overlap between questions
and statements. For instance, questions can be posed in the form of statements. "You must have
had some reason to come to see me"; "Most people come because something is troubling them very
deeply”. Alternatively, statements can be made in the form of questions: "Isn't it interesting that
you came so late again?"; "Why didn't you leave earlier, when you knew the traffic would be so
heavy?"11. Despite this overlap, it seems reasonable to expect that the predominant linguistic form
of a therapist's contributions will have an important effect on the nature and direction of the
evolving conversation.
There seem to be some advantages for a therapist to ask mainly questions, especially in the
early and middle parts of an interview. For instance, doing so tends to assure a client-centered
conversation. The perceptions, experiences, reactions, concerns, goals, plans, and so on, of the
client are repeatedly called forth and take center stage. If the therapist responds to the client's
answers with further questions, the experiences and beliefs of the therapist remain in a supportive
role as the conversation unfolds. Thus, when the balance is in favor of questions over statements,
the "work" of the session naturally centers on the client, not on the therapist. Another advantage is
that questions constitute a much stronger invitation for clients to become engaged in the
conversation than do statements. The grammatical form of a sentence that poses a question arouses
the social expectation for an answer. The cadence, tone, and ensuing pause in the therapist's speech
add to the expectation for a response. When the therapist also conveys a clear commitment to listen,
and to hear the clients' answers, the expectancy is strengthened even further. Thus, through
questioning, clients are actively drawn into dialogue with the therapist. Indeed, even withdrawn

11
One could claim that every statement raises certain questions and every question implies certain statements. This
"reality" may be valid from the perspective of an observer performing an in-depth analysis of verbal transactions, but
it is usually not experienced by those who actively participate in the conversation. Nevertheless, the complexities of
what is being suggested or implied (in what is said or asked by the therapist) may be brought forth by the client upon
deliberate reflection.
Fam Proc 27:1-15, 1988

and/or mute clients find it difficult to escape entering into a process of silent conversing when
questions are addressed to them. A further advantage in therapists asking mainly questions, and
refraining from making statements, is that clients are thereby stimulated to think through their
problems on their own. This fosters client autonomy and allows a greater sense of personal
achievement for family members when therapeutic change takes place, rather than inducing
dependency on the "special knowledge" of the therapist.
[pg 1]
There are, however, limiting conditions to a predominance of questions over statements. A
therapist may, in effect, hide behind the perpetual questions and fail to enter into the relationship
as a real person. This could constitute a major disadvantage by limiting the development of a
therapeutic alliance. Clients usually need to experience the therapist as someone with coherence
and integrity in order to extend their confidence and trust. For this, the therapist does have to make
statements from time to time and take a position on certain issues (even if the position taken is
deliberately not to take one, such as whether a couple should separate or remain together).
Furthermore, the social expectancy for answers can be experienced as a demand and become an
imposition. Certain questions can be extremely intrusive or threatening. A long series of questions
may be experienced as an inquisition or as punishment. These possibilities highlight the need for
therapists to monitor the conversation continually and switch to making statements when their
questions become countertherapeutic. On the other hand, some of these difficulties may be dealt
with by changing the kind of questions being asked.
The balance between questions and statements, as utterances made by the therapist, tends
to vary with different schools of therapy. For instance, the Milan systemic approach depends
heavily on asking questions whereas the structural and strategic approaches depend on making
statements as well. Among the variables that influence the balance between questions and
statements in a particular session are the theoretical orientation and personal style of the therapist,
the types of problems, beliefs, expectations, and interaction styles presented by clients, and the
idiosyncratic pattern of interaction that evolves among them. As far as I am aware, the effects of
this balance have not yet been systematically explored in marital and family therapy research, nor
has the effect of deliberately altering the ratio of questions to statements during the course of the
interview been examined.
Fam Proc 27:1-15, 1988

Although this article focuses predominantly on questions and on the differences among
them, it is not intended to imply that a therapist should only ask questions. When clients are simply
unaware of basic information or do not have the knowledge resources to answer coherently, it is
appropriate that therapists provide answers for them. In addition, provisional "if-then" statements
that clarify mental process can contribute enormously to a family's awareness and understanding
of relevant events. For example, if parents repeatedly demand disclosure from a child, they
sometimes inadvertently teach the child to lie. The child may learn to invent any kind of answer
that might satisfy the parents' demands for an immediate response. Furthermore, ironic and
improbable statements by a therapist are sometimes the most effective means to awaken questions
in the minds of clients and to enhance their capacity to make pertinent discoveries on their own.

Therapist intentions and assumptions

Every question may be assumed to embody some intent. Whether consciously or not, the
therapist has some purpose in asking. This intent or purpose arises from the conceptual posture of
strategizing (see 4) that guides the therapist's moment-to-moment decision making during the
conversation. The most common intention behind the questions asked by a therapist is to find out
something about the clients or their situation. With the use of questions, the therapist invites clients
to share their problems, experiences, histories, hopes, expectations, and so on. The immediate
intent in the asking is to develop the therapist's understanding. The questions are designed to
trigger responses from clients that will enable the therapist to become coupled linguistically with
the clients, to draw relevant distinctions about their experiences, and to generate clinically useful
explanations regarding their problems. The questions are chosen to support the therapist's activity
in the conceptual postures of circularity and hypothesizing (see 4). Family members are expected
to answer according to the understanding they already have. They are not usually expected to
change as a result of these questions. In other words, during such questioning the primary locus for
intended change is the therapist, not the client or family. The goal at those moments in the interview
is for the therapist to become oriented to the problematic situation and the idiosyncratic experiences
of the client and family members. As the therapist constructs impressions and images from the
family's verbal and nonverbal responses, further questions are asked in order to fill in blanks, clarify
Fam Proc 27:1-15, 1988

ambiguities, and resolve inconsistencies that arise in the mind of the therapist. Thus, in the early
parts of an interview the therapist asks predominantly orienting questions12.
However, during the course of assessing the clients' situation, occasions frequently arise in
which therapeutic interventions seem particularly opportune. The therapist recognizes "a good
moment" or "an opening" in the conversation to influence the family's perceptions or beliefs. In
other words, the situation is conducive to an action on the part of the therapist that might enable
family members to change their views, and consequently their behavior. The therapist could alter
the pattern of asking questions and make some statements. If, however, the therapist decides to
continue the inquiry, he or she can still take advantage of these opportunities by introducing
therapeutic interventions in the form of questions. Indeed, for various reasons the therapist may
prefer to use questions to influence the client, rather than resort to making statements. The therapist
then formulates influencing questions, the kinds of questions that are liable to trigger therapeutic
change. In this case, the primary locus for intended change is the client or family, not the therapist.
This does not mean that the therapist is not open to further change in his or her understanding as a
result of the client's answers to these questions. On the contrary, the therapist always remains open
to change following an influencing question; otherwise the question becomes purely rhetorical.
However, this change in the therapist is secondary with respect to the therapist's predominant [pg
2] intent in formulating that particular question.
Thus, one basic dimension for differentiating questions is a continuum regarding the
intended locus of change that lies behind the question. At one extreme of the continuum is a
predominantly orienting intent, for change in oneself, and at the other end is a predominantly
influencing intent, for change in others. Orienting questions are designed to invite a response to
alter the therapist's own perceptions and understanding whereas influencing questions are designed
to trigger a response that might alter the family's perceptions and understanding. Any particular
question may, of course, entail mixed intentions and fall anywhere along the continuum. This
distinction between orienting questions and influencing questions constitutes an invitation for

12
In an earlier publication (3), I referred to these questions as "descriptive" because they invited clients to describe
their situation and experiences. However, the adjective "descriptive" could imply that family members provide
objective accounts of events and experiences and, thus, it can be misleading. I now prefer "orienting" because it is
more precise and coherent with a second-order cybernetic explanation of what takes place during an interview. The
family's answers simply orient the therapist in his or her subsequent actions; the answers are not necessarily taken as
statements about an objective "reality”.
Fam Proc 27:1-15, 1988

therapists to become more mindful of their intentions during the process of strategizing about what
to ask.
A second major dimension for differentiating questions has to do with varying assumptions
about the nature of mental phenomena and the therapeutic process. It seems reasonable to assume
that a network of assumptions and presuppositions concerning the issues being asked about exists
in the mind of the therapist as a foundation or rationale for the question. For the most part, these
underlying assumptions or presuppositions tend to remain nonconscious during the conduct of an
interview. They may, however, be brought into consciousness and deliberately be modified in one
direction or another. In other words, these assumptions may be plotted along a continuum as well.
At one extreme of this continuum might be predominantly lineal or cause-and-effect assumptions,
and at the other, predominantly circular or cybernetic assumptions.
The distinction between "lineal" and "circular" was imported into family therapy from
Bateson's pioneering work in exploring the nature of mind (1, 2). Since then, a rich network of
ideas, concepts, and associations has evolved around this distinction. These ideas now permeate
the family therapy literature. Lineal assumptions tend to be associated with reductionism,
dormative principles, causal determinism, judgmental attitudes, and strategic approaches. Circular
assumptions tend to be associated with holism, interactional principles, structure determinism,
neutral attitudes, and systemic approaches. These associations do not necessarily imply identity or
isomorphism within each cluster of concepts. Nor do they imply that lineal and circular
assumptions are mutually exclusive. Because the distinction between lineal and circular may be
regarded as complementary, and not just as either/or, these assumptions and their associations may
overlap and enrich one another. Most therapists have internalized these concepts to varying degrees
and probably operate with both sets of ideas, but in differing ways, with differing consistency, and
at different times. Although these assumptions and presuppositions tend to exert their effects
covertly and nonconsciously, they still have a significant effect on the nature of the questions asked.
Hence, this second dimension adds considerable depth to an understanding of differences among
the questions asked.
An intersection of these two basic dimensions (therapist intentionality and therapist
assumptions) yields four quadrants, which may be used to distinguish four basic types of questions.
This is indicated in the framework of Figure 1. The horizontal axis represents the degree to which
the therapist's intentionality is oriented toward changing the self, or toward changing the other. The
Fam Proc 27:1-15, 1988

vertical axis represents the degree of lineality or circularity in the therapist's assumptions about the
relevant mental process. If the therapist assumes that the events being explored occur
predominantly in a lineal or cause-and-effect manner, the orienting questions will reflect this and
may be considered "lineal questions”. If the therapist assumes that the events being explored are
circular, recurrent, or cybernetic, the orienting questions are labelled "circular questions”. If the
therapist assumes that it is possible to influence others directly through information input or
instructive interaction, then the influencing questions may be regarded as "strategic questions”. If
the therapist assumes that influence only occurs indirectly, through a perturbation of preexisting
circular processes in or among family members, the influencing questions are considered "reflexive
questions”.
[pg 3]

Figure 1.
A framework for distinguishing 4 major groups of questions.
Fam Proc 27:1-15, 1988

Because specific questions may reflect differing degrees of lineality and circularity as well
as varying intentionality, they could be plotted anywhere on the diagram. However, certain kinds
of questions will tend to fall into a particular quadrant. For instance, the common kinds of problem
definition questions and problem explanation questions tend to reflect a lineal inquiry. Difference
questions and a series of behavioral effect questions suggest an exploration of a circular process.
Leading questions and confrontation questions tend to be regulatory and strategic. Future oriented
questions and observer perspective questions tend to be reflexive. Different kinds and sequences
of questions may be expected to have very different effects in the evolving therapeutic
conversation. For instance, the manner in which a specific historical event is reported by the client
is influenced by the wording and tone of the therapist's question. A lineal question invites a lineal
description whereas a circular question invites a circular description. A further sketch of these
major groups of questions with a few examples of each will be provided before examining their
differential effects more closely.

Four major types of questions

Lineal Questions
These are asked to orient the therapist to the client's situation and are based on lineal
assumptions about the nature of mental phenomena. The intent behind these questions is
predominantly investigative. The therapist behaves much like an investigator or detective trying to
unravel a complex mystery. The basic questions are "Who did what?, Where?, When?, and Why?"
Most interviews begin with at least some lineal questions. This is often necessary in order to "join"
the family members through their typically lineal views of their problematic situation. With this
mode of inquiry, the therapist tends to adopt a reductionistic stance in trying to determine the
specific cause of the problem. Efforts are made to tease things apart so that the origin of the problem
eventually becomes clearly delineated.
For instance, a therapist may begin a session with a sequence of lineal orienting questions
as follows: "What problems brought you in to see me today?" (It's mainly depression); "Who gets
depressed?" (My husband); "What gets you so depressed?" (I don't know); "Are you having
difficulty sleeping?" (No); "Have you lost or gained any weight?" (No); "Do you have any other
symptoms?" (No); "Any illnesses lately?" (No); "Do you have a lot of morbid thoughts?" (No);
"Are you down on yourself about something?" (No); "There must be something troubling you.
What could it be?" (I really don't [pg 4] know); "Why do you think your husband gets depressed?"
Fam Proc 27:1-15, 1988

(I don't know either, he's just not motivated, he lies in bed all the time); "How long has he been so
depressed?" (Three months, he has hardly been out of bed in three months); "Did something happen
that started it all?" (I can't remember anything in particular); "Does anyone try to get him up?" (Not
really); "Why not?" (Well, I get fed up after a while); "Do you find yourself getting frustrated at
lot?" (Quite a bit); "How long have you been so frustrated with him?"; and so on.
The conceptual posture of lineal hypothesizing (see 4) contributes to the content issues and
subject focus for generating these lineal questions. Included in this posture is the habit of thinking
in dormative terms, that is, in maintaining the presupposition that certain characteristics, such as
depression, are intrinsic to the person, rather than to the distinctions made about the person.
Consequently, lineal questions about problems tend to convey a judgmental attitude, namely, that
something in the individual is wrong and ought not be the way it is. This often evokes shame, guilt,
and defensiveness in the client or family. Because people generally do not like to take blame onto
themselves, these questions may stimulate family members to become more critical of one another
as they provide answers.

Circular Questions
These are also asked to orient the therapist to the client's situation, but they are based on
circular assumptions about the nature of mental phenomena. The intent behind these questions is
predominantly exploratory. The therapist behaves more like an explorer, researcher, or scientist
who is out to make a new discovery. The guiding presuppositions are interactional and systemic.
It is assumed that everything is somehow connected to everything else. Questions are formulated
to bring forth the "patterns that connect" persons, objects, actions, perceptions, ideas, feelings,
events, beliefs, contexts, and so on, in recurrent or cybernetic circuits.
Thus, a more systemic therapist may begin the interview differently: "How is it that we find
ourselves together today?" (I called because I am worried about my husband's depression); "Who
else worries?" (The kids); "Who do you think worries the most?" (She does); "Who do you imagine
worries the least?" (I guess I do); "What does she do when she worries?" (She complains a lot,
mainly about money and bills); "What do you do when she shows you that she is worrying?" (I
don't bother her, just keep to myself); "Who sees your wife's worrying the most?" (The kids, they
talk about it a lot); "Do you kids agree?" (Yes); "What does your father usually do when you and
your mother talk?" (He usually goes to bed); "And when your father goes to bed, what does your
mother do?" (She just gets more worried); and so on. These questions seek to reveal recurrent
Fam Proc 27:1-15, 1988

circular patterns that connect perceptions and events. They tend to be more neutral and accepting.
The responses they elicit from family members are also less liable to be judgmental.
Circular questions tend to be characterized by a general curiosity about the possible
connectedness of events that include the problem, rather than a specific need to know the precise
origins of the problem. If the therapist has established a Batesonian cybernetic orientation toward
mental process, and has developed skills in maintaining a conceptual posture of circular
hypothesizing, these questions will come easily and freely. Two general types of circular questions,
"difference questions" and "contextual questions," have been associated with Bateson's
fundamental patterns of symmetry and complementarity. Several subtypes, including category-
difference questions, temporal-difference questions, category-context questions, and behavioral-
effect questions, have been described in an earlier paper (3).

Strategic Questions
These are asked in order to influence the client or family in a specific manner, and are based
on lineal assumptions about the nature of the therapeutic process. The intent behind these questions
is predominantly corrective. It is assumed that instructive interaction is possible. The therapist
behaves like a teacher, instructor, or judge, telling family members how they erred and how they
ought to behave (albeit indirectly in the form of questions). On the basis of hypotheses formulated
about the family's dynamics, the therapist comes to the conclusion that something is "wrong," and
through strategic questions tries to get the family to change, that is, to think or behave in ways that
the therapist thinks is more "correct”. The directiveness of the therapist may be covert, because the
corrective statement is packaged in the form of a question, but it is still conveyed through the
content, context, timing, and tone. Some families are offended by this mode of inquiry, but others
find it quite compatible with their usual patterns of interaction.
Giving examples of influencing questions is more difficult because hypotheses about some
of the mechanisms involved in the problematic situation are necessary for the formulation of the
question. But, continuing on with the hypothetical family being interviewed above, the therapist
might try to influence the couple by asking: "Why don't you talk to him about your worries instead
of the kids?" (He just won't listen, and stays in bed); "Wouldn't you like to stop worrying rather
than being so preoccupied by them?" (Sure, but what am I going to do about him?); "What would
happen if for the next week at 8 a.m. every morning you suggested he take some responsibility?"
(It's not worth the effort); "How come you're not willing to try harder to get him up?" (I'm tired and
Fam Proc 27:1-15, 1988

disappointed. He won't move and it just gets me more frustrated); "Can you see how your
withdrawal gets your wife disappointed and frustrated?" (What do you mean?); "Can't you see how
just going to bed instead of talking about what is bothering you is getting your family upset?"
(Well, I...); "Is this habit of making excuses [pg 5] something new?" (I didn't know I had one);
"When are you going to take charge of your life and start looking for a job?"; and so on.
It is quite apparent from these examples that by asking strategic questions the therapist is
imposing his or her views of what "ought to be" upon the client or family. Sometimes a directive
or confrontation by the therapist is needed to mobilize a stuck system, but too much directiveness
in this mode of inquiry may risk a disruption in the therapeutic alliance.

Reflexive Questions
These are intended to influence the client or family in an indirect or general manner, and
are based on circular assumptions about the nature of the process taking place in the therapeutic
system. The intent behind these questions is predominantly facilitative. It is assumed that family
members are autonomous individuals and cannot be instructed directly. Thus, the therapist behaves
more like a guide or coach encouraging family members to mobilize their own problem-solving
resources. One major presupposition behind these questions is that the therapeutic system is co-
evolutionary and what the therapist does is to trigger reflexive activity in the family's preexisting
belief systems. The therapist endeavors to interact in a manner that opens space for the family to
see new possibilities and to evolve more freely of their own accord.
Numerous examples of reflexive questions have already been provided in Part II (5) of this
series of articles. However, to provide an indication of what they might be like in this scenario, the
therapist could ask: "If you were to share with him how worried you were and how it was getting
you down, what do you imagine he might think or do?" (I'm not sure); "Let's imagine there was
something that he was resentful about, but didn't want to tell you for fear of hurting your feelings,
how could you convince him that you were strong enough to take it?" (Well, I'd just have to tell
him I guess); "If there was some unfinished business between the two of you, who would be most
ready to apologize?" (She would never apologize!); "Would you be surprised if she did?" (Sure
would!); "Suppose that it was impossible at this moment for her to recognize or to admit to any
mistakes on her part, how long do you think it would take before you could forgive her for being
unable to do so?" (Humm ...); "If this depression suddenly disappeared, how would your lives be
different?"; and so on.
Fam Proc 27:1-15, 1988

These questions are reflexive in that they are formulated to trigger family members to reflect
upon the implications of their current perceptions and actions and to consider new options. Even
though reflexive questioning is also intended to influence a family in a therapeutic direction, it
remains a more neutral mode of inquiry than strategic questioning because it is more respectful of
the family's autonomy. Well-developed skills in maintaining a conceptual posture of neutrality
contribute to the probability that an influencing question will be reflexive rather than strategic.
What is missing in all these examples is the emotional tone used in asking the questions.
The differences between these groups would become even more apparent if the therapist's vocal
cadence, tone, and accompanying nonverbal behaviors were present. What bears emphasis here is
that the differentiation of these questions does not depend on their syntactic structure or their
semantic content. It depends on the therapist's intentions and assumptions in the asking. Indeed,
the exact same sequence of words could constitute a lineal, a circular, a reflexive, or a strategic
question. For example, if a therapist asked a child "What does your mother do when your father
comes home late, and dinner has already gone cold?" only to find out how the mother responds
when provoked by the father, it would be a lineal orienting question. If it were asked as part of a
planned sequence of behavioral effect questions (to be followed by something like "And what does
your father do when your mother yells at him?") in order to explore the circular interaction between
the parents, it would be a circular orienting question. If the original question were asked to trigger
the parents to become observers of their own behavior and to mobilize their awareness to modify
their own behavior, it would be a reflexive question. If it were asked because the therapist
anticipated what the child probably would say, and wanted this information released at that moment
to confront either the mother or the father on their intolerant or inconsiderate behaviors, it would
be a strategic question. Thus, precisely the same words can mean and do very different things in
the course of a single interview. It is usually the therapist's emotional posture in the asking that
makes the difference in what the client hears in the question. These emotions are, in turn, associated
with the therapist's intentions and assumptions.

The effects of different questions

Before discussing the differential effects of these kinds of questions, it is important to
acknowledge the discontinuity between a therapist's intentions in asking certain questions and their
actual effects on clients. Recognizing and accepting this cleavage between intent and effect reduces
therapist frustration when therapy is not progressing well, and it opens space for the therapist to
Fam Proc 27:1-15, 1988

consider alternative courses of action. From the perspective of an observer of the therapeutic
process (who is usually the therapist observing himself or herself at work), there are two points,
one minor and one major, at which discontinuities occur. The first is between what the therapist
intends to do and what the therapist actually does do. This gap can be steadily narrowed as
therapists seek greater personal integration and develop better skills in implementing their
intentions. The second is the discontinuity between what the therapist actually asks and how this is
heard by family members. There is an absolute limitation here. The listening and responses of
clients are always determined by their own biological autonomy. At the same time, however, the
responses of family members are not arbitrary; they are triggered by [pg 6] and contingent to what
the therapist says and does. There is much that a therapist can do to improve the contingencies
between intent and effect by enhancing his or her linguistic coupling with clients through the
conceptual posture of circularity (4). But, ultimately, the intentions of a therapist in asking specific
questions never guarantee any specific effect on clients; nor could more highly refined precision
in the wording and tone of the questions do so. What actually happens to the client or family always
depends on the uniqueness of their own organization and structure at each moment. The importance
of recognizing and accepting this cleavage between intent and effect, between therapist action and
client responses, cannot be overemphasized. The actual effects are always unpredictable.
Nevertheless, a therapist can and does compute probabilities. For instance, it is more likely
that clients will become interested in their own interaction patterns through a series of circular
questions rather than lineal ones, or feel blamed more by strategic questions than by reflexive ones.
Because the therapist cannot know in advance what the actual effects of any particular question
will be, yet must make choices about what to ask before asking, these choices are made on the basis
of anticipated effects. The therapist can envisage the probable, possible, improbable, and
impossible effects of various questions. This process of anticipating is an important aspect of the
conceptual posture of strategizing. The following generalizations about the more probable effects
of different questions may be incorporated into a therapist's nonconscious habits of strategizing
and may guide the process of deciding what questions to ask.

Lineal Questions
These tend to have a conservative effect on the client or family. Because family members
usually think of their difficulties in lineal terms before coming to therapy, there is little "news of
difference" for the family when the therapist invites them to articulate their prior views (of what
Fam Proc 27:1-15, 1988

happened, who was involved, and how) with lineal questions. Family members answer the
questions but remain virtually unchanged13. However, one hazard of lineal questioning is that it
may inadvertently embed the family even more deeply in lineal perceptions by implicitly validating
preexisting beliefs. Unfortunately, this happens far more often than clinicians realize while they
are conducting ordinary "assessment" interviews. The interviewer is seldom aware of the fact that
further entrenchment of pathogenic perceptions and beliefs is taking place. This process is
particularly liable to occur if, during the course of the inquiry, the therapist does not ask the kinds
of questions (or make statements) that implicitly (or explicitly) challenge the family's prior beliefs.
Another risk with lineal questioning is that the reductionistic thinking involved tends to activate
judgmental attitudes. As the therapist brings forth "the cause" of a presenting problem or of an
undesired situation, negative judgments are automatically directed toward it because the problem
is unwanted. Thus, while lineal questions are necessary to develop a clear focus of the problem,
and are helpful in establishing an initial engagement, it is useful for therapists to remain mindful
of potential hazards as well.

Circular Questions
Circular questions, however, do have the potential of having liberating effects on the family.
As the therapist asks questions to identify patterns for a circular or systemic understanding of the
problematic situation, family members who are listening to the answers make their own
connections as well. Thus, they may be able to become aware of the circularity in their own
interaction patterns. With this increased awareness, they may be "liberated" from the limitations of
their prior lineal views and subsequently be able to approach their difficulties from a fresh
perspective. For instance, if through a series of behavioral effect questions a husband begins to see
that it is not simply his wife's worrisome complaints that activates his depression but also that his
depressiveness activates her complaining, he may be liberated to act differently rather than by just
becoming despondent when she worries and complains. He has more space to recognize that some
constructive initiative on his part may activate a different response from her. He is also likely to
become more accepting and less judgmental of her "worrying response" to his depressive behavior.
The main risk with circular questions is that as the therapist explores larger and larger areas of

13
Obviously, if the answer of the respondent includes information that other family members (who are listening) were
not aware of previously, this could be important news and have significant effects. However, this may occur with all
kinds of questions. It is a general effect of the method of conjoint interviewing in marital and family therapy, and not
specifically an effect of the kind of question asked.
Fam Proc 27:1-15, 1988

interaction, the inquiry may drift into domains that seem irrelevant to the immediate concerns and
needs of the family. Another risk is that clinicians who are learning to use circular questions may
use them in a rather stylized fashion. The questions then seem repetitive or trivial and, thus, can
become irritating to the family. On the whole, however, circular questions are more liable than
lineal ones to have inadvertent beneficial effects.

Strategic Questions
These tend to have a constraining effect on the family. The therapist tries to influence the
client (in a lineal fashion) to think or do what the therapist considers more healthy or "correct”.
The questions are intended to constrain the probability of family members continuing along the
same problematic path. A common side effect is for family members to feel guilty or ashamed for
having taken the path they are on in the first place. The constraint may be of two forms: not to do
something that the therapist thinks is "wrong" and is contributing to the problem or to do only what
the therapist thinks is "right" and would be helpful. Both tend to confine the family's options to
what the therapist thinks is best, whether it actually fits for them at that moment or not. Thus, these
questions tend to be more manipulative and controlling. In the extreme, they can be like the
questions a good lawyer might employ in cross-examining witnesses in a courtroom. The lawyer
uses strategic [pg 7] questions to lead, seduce, intimidate, or coerce a witness into saying precisely
what the lawyer wants the judge and jury to hear. Similarly, a therapist can "force" an individual
into saying things that the therapist wants to hear, or wants other family members hear, even when
the person really doesn't think or feel that way. Because of the potentially coercive nature of
strategic questions, too many of them could have inadvertent, counter-therapeutic effects.
On the other hand, occasional strategic questions can sometimes be extremely constructive
in the therapeutic process. These questions can be vigorously used to challenge problematic
patterns of thought and behavior without having to resort to direct statements or commands. If the
questions are carefully worded, clients often can be confronted with the limitations, constraints, or
contradictions in their own systems of belief. Alternatively, strategic questions sometimes can be
employed to lead the family quite directly to recognize and embrace an obvious solution.

Reflexive Questions
These questions are more liable to have a generative effect on the family. The therapist's
influencing intent is moderated by respect for the autonomy of clients and, hence, the tone of these
questions tends to be much softer. Family members experience themselves as being invited into
Fam Proc 27:1-15, 1988

entertaining new views instead of being pushed or pulled into them. The questions tend to open
space for family members to entertain new perceptions, new perspectives, new directions, and new
options. They also enable a reevaluation, without duress, of the problematic implications of the
family's current perceptions and behaviors. As a consequence, family members tend to generate
new connections and new solutions in their own manner and time. The most likely complication of
reflexive questioning is that it could foster disorganizing uncertainty and confusion. Opening a
multiplicity of new possibilities without providing adequate direction can easily become confusing.
However, such confusion may not necessarily be problematic for the overall therapeutic process.
Depending on the domain of the confusion, it may, in fact, be very therapeutic. For instance, when
certain family members "know the Truth" or "have all the answers" in a manner that keeps them
stuck in problematic patterns and blind to novel alternatives, the confusion can be quite liberating.
Finally, I would like to draw attention to the possible effects on the therapist of asking
different kinds of questions. The therapist is influenced by the questions as well. His or her thinking
is influenced not only by the assumptions and presuppositions aroused during the formulation of
the questions, but also by responding to the clients' responses to the questions. Lineal questions
tend to foster further lineal thinking in the therapist just as they do in the clients. Consequently, the
therapist is also more liable to become judgmental. The effect of circular questions on the therapist
is to enhance his or her neutrality and capacity to accept the client and family as they are. This
acceptance itself has healing potential in the therapeutic system by countering the immobilizing
effects of blame, which is so ubiquitous in symptomatic families. The effect of strategic questions
on the therapist is that they tend to lead him or her toward an oppositional stance with the family.
On the other hand, reflexive questions tend to guide the therapist toward becoming more creative
in the questions asked. If one question "doesn't work" in opening space for the family to evolve
more freely, the therapist searches for another one that is more likely to release the natural healing
capacity of the clients.
Figure 2 summarizes the predominant intent and the more probable effects associated with
each set of questions. Included in the diagram are the effects of the questions on the therapist as
well as on the family. The parentheses are intended to indicate that the actual effects always remain
unpredictable. Depending on the momentary structure of a family, a strategic question could have
a generative effect instead of a constraining one. A lineal question could have a liberating effect,
and a reflexive question could have a constraining one, and so on. All that one can say is that it is
Fam Proc 27:1-15, 1988

more likely that family members will experience respect, novelty, and spontaneous transformation
as a result of circular questioning and reflexive questioning, and judgment, cross-examination, and
coercion as a result of lineal and strategic questioning. If family members begin to feel judged or
manipulated, the session often becomes tense or "frozen”. This could become a cue to the therapist
to change the kind of questions to those that are more neutral and accepting (or temporarily to
abandon the process of questioning altogether). Alternatively, if family members have become too
comfortable and complacent in the therapy process, perhaps a few well-placed strategic questions
could stimulate them to consider new directions. What is being proposed here is that the use of
these distinctions could enable therapists to choose those kinds of questions that are more liable to
guide the interview to actually become a conversation for healing.
[pg 8]
Fam Proc 27:1-15, 1988

Figure 2.
Predominant intent and probable effects of differing questions.

Concluding comments
The impossibility of predicting actual effects points to the importance of the therapist's
ongoing activity of monitoring the immediate reactions of family members and revising hypotheses
as the session unfolds. However, the actual effects of a question often cannot be observed; the
reactions of family members are altogether too difficult to "read”. Sometimes the effects may not
even materialize at the time of the interview. The pertinent realization may dawn on family
members only after the session, perhaps the next day, or even later. There are some questions that
linger in the minds of clients for weeks, months, and occasionally years, and continue to have an
effect. To a large extent, a therapist always has to "work in the dark" and never knows the final
Fam Proc 27:1-15, 1988

outcome of specific questions. This leaves even more responsibility on the therapist's intentionality
in making decisions about what to ask. In other words, therapists need to take responsibility for the
questions being asked without ever knowing what their full effects might be. At the same time,
however, much can be done in personal professional development to increase the probability that
a therapist's spontaneous behavior in an interview is more liable to be therapeutic than
nontherapeutic or countertherapeutic. One has to bear in mind that, to a significant degree, the
question "prefigures" the response in that it structures the domain of an "appropriate" answer. That
is, a question presupposes a particular answer, or at least an answer in a particular domain. To ask
a particular question, then, is to invite a particular answer. The kinds of questions a therapist
chooses to ask depends on what kinds of answers the therapist would like to have heard. Whether
or not the client accepts the therapist's invitation to provide an answer in the "appropriate" domain
is quite another matter, but to select the question is to restrain the range of "legitimate" responses.
This selectiveness gives the therapist an enormous amount of influence in setting and maintaining
a direction for the conversation.
The distinctions in this article reflect the results of some qualitative research I have been
engaged in for the past several [pg 9] years. If an empirical researcher wanted to explore these
issues further and, for instance, establish whether a particular question was lineal, circular, strategic
or reflexive, he or she would have the problem of identifying the intentions and assumptions of the
therapist in asking it. The most direct route for this would be to ask the therapist to try to articulate
his or her thoughts while formulating questions. This could perhaps be achieved during a review
of a videotape immediately after the session. An outside observer could also evaluate each question
in its context. Subsequently, these ratings could be compared for degrees of fit and set alongside
descriptions of the moment-to-moment experiences of clients who also reviewed the tape. Further
studies along these lines may contribute a great deal to a deeper understanding of the process of
interventive interviewing.

References
1. Bateson, G., Steps to an ecology of mind. New York: Ballantine Books, 1972.
2. Bateson, G., Mind and nature: A necessary unity. New York: E.P. Dutton, 1979.
3. Tomm, K., Circular interviewing: A multifaceted clinical tool. In D. Campbell & R. Draper (eds.),
Applications of systemic therapy: The Milan approach. London: Grune & Stratton, 1985.
Fam Proc 27:1-15, 1988

4. Tomm, K., Interventive interviewing: I. Strategizing as a fourth guideline for the therapist. Family
Process, 26, 3-13, 1987.
5. Tomm, K., Interventive interviewing: II. Reflexive questioning as a means to enable self-healing. Family
Process, 26, 167-183, 1987.
Manuscript received May 11, 1987.; Accepted May 11, 1987.
[pg 10]