Journal of Marital and Family Therapy

doi: 10.1111/j.1752-0606.2011.00238.x
October 2012, Vol. 38, No. 4, 667687

MAKING SENSE OF MULTI-ACTOR DIALOGUES IN

FAMILY THERAPY AND NETWORK MEETINGS
Jaakko Seikkula and Aarno Laitila
University of Jyvaskyla

Peter Rober
University of Leuven

In recent years, a number of family therapists have conceptualized psychotherapy as a

dialogical activity. This view presents family therapy researchers with specic challenges, the
most important of which is to nd ways of dealing with the dialogical qualities of the multiactor dialogues that occur, for example, in family therapeutic conversations. In this article,
we propose some preliminary ideas concerning qualitative investigations of multi-actor dialogues. Our aim is to work toward an integration of Bakhtins theoretical concepts with good
practices in qualitative research (e.g., dialogical tools and concepts of a narrative processes
coding system) in order to make sense of family therapy dialogues. A specic method that we
have called Dialogical Methods for Investigations of Happening of Change is described. This
method allows for a general categorization of the qualities of responsive dialogues in a single
session, and also for a detailed focus on particular sequences through a microanalysis of specic topical episodes. The particular focus is on the voices present in the utterances, the positioning of each speaker, and the addressees of the utterances. The method is illustrated via an
analysis of a couple therapy session with a depressed woman and her husband.
As family therapists, we have conceptualized psychotherapy as a dialogical activity (Rober,
2005b; Seikkula & Arnkil, 2006). Conceiving therapy in this way poses research challenges, the
most important of which is to nd ways to see the dialogical qualities of therapeutic conversations. Several authors have inspired us in this endeavor, as in recent years, a number of
research methods have been developed aimed at a deeper understanding of the dialogical qualities of therapeutic conversations (e.g., Leiman, 2004; Salgado & Goncalves, 2007; Stiles, Osatuke, Click, & MacKay, 2004; Wortham, 2001). These methods have addressed the complex
and multilayered nature of dialogue and introduced the possibility of analyzing the polyphony
presented in an individuals utterances. Thus, Stiles et al. (2004) revised the Assimilation model,
moving from a schema formulation to a voices formulation, seeking to discover how the voices
develop in the process of individual psychotherapy. For his part, Leiman (2004) developed a
Dialogical Sequence Analysis specically to examine the polyphonic process in individual psychotherapy. However, these methods all focused on dyadic dialogues, i.e., the dialogue between
one therapist and one client. As family therapists, we are faced with the need to develop methods for the study of multi-actor dialogues, such as occur in family therapy, network meetings,
and group discussions. However, there is no literature on dialogical research on multi-actor
meetings, except for the work of Markova`, Linell, Grossen, and Orvig (2007), which deals with
the dialogical analyses of focus group interviews. The main interest of that research group was
on identifying the dierent voices present in the utterances within focus group dialogues.
In our research on family sessions, we have been especially interested in the responses
generated in family therapy dialogues, in which different voices are present, both in the outer
Jaakko Seikkula, PhD, Professor of Psychotherapy, Department of Psychology, University of Jyvaskyla;
Aarno Laitila, PhD, Senior Assistant, Department of Psychology, University of Jyvaskyla; Peter Rober, PhD,
Professor, University of Leuven.
This article results from two meetings aimed at developing the new method called DialogicalMethods for
Investigations in Happenings of Change. We are especially grateful to our colleagues.
Address correspondence to Jaako Seikkula, Department of Psychology, P.O. Box 35, 40014 University of
Jyvaskyla, Finland; E-mail: jaakko.seikkula@jyu.

October 2012

JOURNAL OF MARITAL AND FAMILY THERAPY

667

conversation between the family members and in the inner speech of each participant in the session. This article is our rst report on our joint efforts to develop further research methods for
the investigation into multi-actor dialogues. While we acknowledge the importance of looking
at the inner voices of each participant (Laitila, 2009; Rober, 2005a), in this article, we shall
focus on the outer dialogues and on the responses of each interlocutor in the present moment
of the dialogue encounter.
Here, we shall rst present some conceptual tools that can offer family therapists an interesting perspective on storytelling in family sessions. Second, we shall provide an overview of
our main theoretical concepts, including categories that can be useful for a dialogical analysis
of family sessions, namely voice, words actions, positioning, and addressees. Finally, we shall
concentrate on the way we dialogically make sense of multi-actor dialogues in family sessions.

THE DIALOGICAL PERSPECTIVE IN THE FIELD OF FAMILY THERAPY

In recent years, a new perspective has emerged in the family therapy eld. This perspective
puts the spotlight not on what is told but rather on how things are told and responded to in the
dialogical process (Rober, Van Eesbeek, & Elliott, 2006; Seikkula, 1995). While stories may
dene who the narrator is, in the telling of the story the relationship between the speaker and
the interlocutors emerges (e.g., Andersen, 1995; Anderson & Goolishian, 1988; Homan, 1991;
White, 1991). Telling stories is seen as a performance in context, implying that stories only exist
through the presence of others who listen to the stories. In that sense, storytelling is a dialogical
phenomenon (Bakhtin, 1981, 1984, 1986).
While we describe therapeutic conversations as dialogues, we have to keep in mind that
they are not readymade stories or narratives, but rather broken stories. Clients are speaking
about particularly sensitive and emotional issues, and while some of what the clients say can be
understood as narratives about their lives, in many respects, their narratives may be incomplete,
multivoiced, and contradictory (Gergen, 2009). Clients do not have exact words or phrases to
utter their most sensitive experiences and their most pressing concerns. Indeed, it could be
argued that it is precisely in this impossibility to express the reason that one may discover the
reason for seeking help in therapy: psychological symptoms can be seen as our clients way of
dealing with the absence of expression concerning what moves them. Clients try to convey
strong embodied experiences, but the experiences may not be connected to exact words or constitute a coherent narrative.
Viewed from this perspective, psychotherapy can be seen as a process toward nding
words for those experiences in ones life that have not yet been given words. In family therapy, to search for these words, we include the family members as real living persons who are
present in the actual session. Dialogically this is very important becauseas was pointed out
by Mikhail Bakhtin (1984)interlocutors are active coauthors of an individuals utterances.
Every word said by participants in the actual dialogue becomes part of the tension-lled
network of previously said words, alien to the words already uttered. Every new word
contributes to an elastic environment of already spoken words on the same subject. In this
process, the word cannot fail to become an active participant in social dialogue (Bakhtin,
1981, p. 276). Everything said or done is a response to what has been said or done before
(Bakhtin, 1981).
In family therapy, new words new language develops especially in the answers that family
members give to each other in response to the questions, concerns, and worries that are
expressed in the session. Moreover, therapists are in specic answering positions, in a situation
of responsive responsibility (Steiner, 1989). In dialogue, we are part of a joint project for
increasing the understanding of things related to the specic situation in which help is needed.
Understanding is gained in an active process of answering, which is an act of taking responsibility for the other and for the situation. In an open dialogue, utterances are constructed to
answer previous utterances and also to wait for an answer from utterances that follow (Seikkula, 2002, p. 268). Hearing is witnessed in our answering words. A family session can be seen
as a forum where stories develop gradually and in unpredictable ways out of the contradictionridden and tension-lled interactions of all the interlocutors present (Bakhtin, 1981). Every
668

JOURNAL OF MARITAL AND FAMILY THERAPY

October 2012

utterance is implicitly or explicitly evaluated by the others, and their verbal and nonverbal reactions invite new utterances in a complex dialectical dance of dierences and similarities (Baxter,
2004).

LOOKING AT VOICES, ADDRESSEES, AND POSITIONING

Voices
It is difcult to give a precise denition of voices, although there have been various
attempts. For instance, Stiles et al. (2004) had a cognitive, ecological view of voices when they
attempted to operationalize the idea of voices, noting that Voices are traces and they are activated by new events that are similar or related to the original event (p. 92). This is a formulation that carries traces of Piagetian constructivism in it, even if it is dialogically formulated.
Piaget himself emphasized the active role of the developing child as a learner. This activity is
especially recognizable in the processes of assimilation and accommodation. Stiles et al. went
on to say that all our experiences leave a sign in our body, but only a minimal proportion of
these signs ever become formulated into spoken narratives. Although this view is valuable, in
adopting a dialogical framework, we believe that voices only make sense when viewed within a
dialogical process. They are always relationally addressed. Voices are the speaking consciousness (Bakhtin, 1984; Wertsch, 1991) that becomes visible in an interchange. Experiences formulated into words in a dialogical context become the voices of our lives.
Every utterance has an author whose position it expresses (Bakhtin, 1981, 1984). The
words I utter can be my very own words, or they may include also the words of other people.
In this way, I speak in dierent voices, allowing me to take dierent positions toward these
voices. As Bakhtin notes, small children in school already learn the polyphonic quality of language while reading aloud some given text. Take, for instance, a conversation within supervision, when a therapist is describing what his client has said and then commenting on that. In
this utterance, at least two voices are present, i.e., the clients and the therapists. Likewise, in
the family therapy session, the voices the family members call upon while speaking may refer
to real persons who are present, but they may also refer to absent persons or even to ctional
or abstract characters. Thus, although one speaker pronounces the words of the story, storytelling involves dierent voices in a dynamic interaction with each other. That is why, in our dialogical analyses, we focus on voices, not on persons.
The richness of the family therapy conversation becomes evident if we focus on those
voices which are not seen but which are present in each persons inner dialogues. While the
therapist does not have access to the inner voices of the clients, he or she has access to part of
the polyphony of his or her own inner voices. As Rober (1999) wrote, these voices can refer to
the therapists professional self or to his or her personal self. Therapists participate in the
dialogue in the voices of their professional expertise, which involves being a doctor, being a
psychologist, having training as a family therapist, and so on. In addition to their professional
voices, the therapists participate in the dialogue in their more personal, intimate voices. If a
therapist, for instance, has experienced the loss of someone loved or near, the voices of loss
and sadness become part of the polyphony, not in the sense that the therapist would speak
about the personal or intimate experience of death, but in the way the therapist adapts to the
present moment. The therapists inner voices containing their own personal and intimate experiences become a powerful part of the joint dance of the dialogue (Seikkula, 2008), involving
aspects such as how the therapists sit, how they look at the other speakers, how they change
their intonation, at which point they break in, and so on. These inner voices, evoked in the session, will not merely be present in the story; they will also become part of the present moment.
The inner dialogue between the voices of our lives is not so much a matter of focusing on the
meaning of the words uttered, but of sensing the nuances of the present moment. As Vygotsky
(1962), referring to the peculiarities of inner speech, noted, the rst [of these peculiarities] is
the preponderance of the sense of a word over its meaning (p. 146). What is more important
than the stable meaning of the words said is the sense of the words in the actual present
moment. The same words can generate very dierent types of meanings when used in dierent
conversations, even if the same people are present.
October 2012

JOURNAL OF MARITAL AND FAMILY THERAPY

669

Positioning
While the concept of voice refers to the question who is speaking? the concept of
positioning refers to the question from where is the person speaking? (Hermans, 2004,
2006; Markova` et al., 2007; Rober, 2002, 2008). Positioning implies a spatial metaphor, linking
a voice with a point of view from which one takes part in the dialogue. Each point of view
gives the person a perspective that allows that person to see, hear, and experience. At the same
time, such a point of view has inherent limitations: from each point of view, some things can
be seen, while other things remain out of focus, in the shadows, or out of sight. Dialogue (inner
or outer) consists of the meeting of dierent points of view, within which each voice expresses
something from its perspective, activating another voice speaking from another point of view in
a continuous play of agreement disagreement (content), or identication dierentiation (position). Such a play of agreement disagreement can often be observed in the rst words in a
speakers turn. These are often words like yes, no, of course, exactly, indeed, or combinations of
these words with other words, such as but or however: yes, but . . . , no, but . . . , indeed,
however . . . (Rober, 2008).
Wortham (2001) distinguishes between representational positioning (the positions of the subject in the story) and interactional positioning (the speaker, the addressee, and the audience in the
storytelling situation). In family therapy, interactional positioninghow the family members
position themselves in the present moment of the sessionneeds careful consideration. It should
be kept in mind, however, that positioning is usually not a voluntary act by the speaker aimed at
manipulating others in the dialogue. Rather, positioning happens unreectively, in the process of
continuous responses to what is uttered. As therapists, we prefer to think that rst come utterances as a response to something previously uttered, with the utterance thus positioning us in the
dialogues. In multi-actor dialogue, interactional positioning can be analyzed by looking at who is
taking the initiative, both regarding the content and the process of speaking (Linell, 1998). As
family members contrast their perspectives with the positions that they attribute to other interlocutorsas well as with the positions they are invited into by other interlocutorsconict and disagreement are phenomena that are interesting to observe in family sessions. They refer to the
continuous dance of the changing positions in the session, giving the therapists some sense of
what is at stake for the family members. However, the therapists are also part of the dialogue.
This means that they are also invited to take positions in the familys performance. The task for
the family therapist is to remain sensitive to the familys invitations and to guard his or her mental
space to reect on this positioning: do these invitations open up a space for stories untold? Do
they add to the security of the session? Do they leave enough space for other family members to
move exibly in the family performance? (Rober, 2005a).
Addressees
Every utterance has both an author and the person to whom it is addressed, as every utterance is a response to what has previously been said (Bakhtin, 1986). The utterance may be
addressed to someone who is present in the same room. We can state our opinion directly about
the issues under scrutiny; we can agree with what was previously said, we can object to it, we
can partly agree, adding our own point of view to what has been said, and so on. In multi-actor
dialogues, we typically speak to one person, but at the same time, we are very aware of those
others who are present, and our speaking is modied because of their presence. In this sense,
those others who are present are part of the addressed audience and become part of the utterances. Bakhtin (1986) calls these people the addressees. But there is more, as in dialogue, a third
party is always present, even if only two persons are speaking to each other. They are speaking
in the present moment, and at the same time, they are addressing their words to those who participated in discussion of the same issue in the past. Furthermore, when they are speaking about
emotional issues, they may be addressing their words to those nearest to them: their mother,
father, or loved one. Bakhtin also speaks of the super-addressee present when we address our
words to some ideology related to our life.
In analyzing dialogues, it is not always easy to recognize the addressees. If the addressee is
dened unanimously as a person sitting in the same room, no difculties emerge. The addressees may be referred to openly and can thus be dened, or they may be present in the speakers
670

JOURNAL OF MARITAL AND FAMILY THERAPY

October 2012

inner dialogue, affecting his or her intonation, choice of words, body gestures, and many other
thingsbut without being openly recognizable.
Applying all this in the family therapy setting involves our basic assumption that the presence of the therapists in a responsive relationship with their clients can be the basis for positive
change in the clients life. Perhaps, more importantly, we nd that family members also enter a
responsive relationship with each other and in this way become major contributors to this positive change. The specic quality of any kind of family therapy consists of family members
becoming embodied real persons who are present in all the dialogues. Hence, everything that is
said is colored by the others who are present, and affects everyone, including the therapists.

EXPLORING RESPONSE PROCESSES

Responsive dialogical processes have been studied in individual psychotherapy (Leiman,
2004; Lysaker & Lysaker, 2004; Stiles et al., 2004), and these studies have given a new picture
of the psychotherapy processes. In their interpretations of their ndings, all these authors focus
primarily on the inner experiences of the patients and on how the change in the dialogue with
the psychotherapist illustrates the improvement in the process. None of these approaches have
actually focused on the dialogue and response between the patient and the therapists, even
if the dialogue is supposed to be that which produces new insight and understanding. Thus,
Leiman (2004) described the Dialogical Sequence Analysis as a micro-analytic method to
examine the dialogical organization of client and therapist utterances (p. 267). The unique
aspect of our work, and our addition to these methods, is that, in our approach, we focus especially on what therapists actually do in the session and try to explore the dialogical qualities of
conversations in multi-actor therapy sessions.
Designing the Studies
Before the analysis can be started, the multi-actor session has to be video-recorded and
transcribed. Depending on the focus of the study, we transcribe a specic part of the session or
else the complete session. To make a multi-actor perspective possible, the transcript of the therapy conversation is printed in columns, one column for each speaker. Utterances are written in
the columns in temporal order (Table 1). For a successful exploration, one has to be able to
read the text simultaneously with a video or audio recording of the session.
The research process proceeds in steps, as follows:
STEP I: Exploring topical episodes in the dialogue
Dening topical episodes means taking them as the main object for analysis (Linell, 1998).
Topical episodes are dened in retrospect, after the entire dialogue of one session has been
divided into sequences. Episodes are dened by the topic under discussion and are regarded as
a new episode if the topic is changed. The researcher can choose, out of all themes, some specic important topics for further analysis. After dividing the session into topical episodes,
within each episode, certain variables are identied, as specied below.
STEP II: Exploring the series of responses to the utterances
In each sequence, the way of responding is explored. Responses are often constructed
within a series of utterances made by each participant in the actual dialogue. Within each topical episode, the responses to each utterance are registered to gain a picture of how each interlocutor participates in the creation of the joint experience in the conversation. A three-step
process is followed. The meaning of the response becomes visible in the next utterance to the
answering words. It can start with whatever utterance is regarded as the initiating utterance
(IU). The answer given to this IU is categorized according to the following aspects:
1) The participant who takes the initiative (i.e., who is dominant) in each of the following
respects:
- Quantitative dominance: this simply refers to who does most of the speaking within a
sequence.
- Semantic or topical dominance: this refers to who is introducing new themes or new words
at a certain moment in the conversation. This individual shapes most of the content of the discourse.
October 2012

JOURNAL OF MARITAL AND FAMILY THERAPY

671

672

JOURNAL OF MARITAL AND FAMILY THERAPY

October 2012

Yes. (turns to Matti) Well,

what about you, after youve
been involved with this . . . I
mean involved in Arjas life for
quite a while, how would you
comment on that, or is there
something else you would like
to add?

. . . to get rid of
the reality?

Yes, exactly. It
doesnt always
work out the
way you want?

Yes.

T1

T2

Yes.
Interactional dominance
by inviting new speaker
Responding to M;
encouragement to talk
about anything at all

New initiative, involving

something not previously
spoken
Dialogical; symbolic
meaning

Not always . . . but

you can . . . in a
way . . . to get rid . . .

Semantic; interactional
and quantitative
dominance

Response
categories

Symbolic meaning
[reexive mode]
Dialogical dialogue;
responds to the position
of not succeeding

Matti

That doesnt always

work out.

Yes . . . well . . . when

Im feeling bad, it
will . . . I try to
have . . . I really try
to get a better
feeling . . . (sobbing)

Arja

NEW TOPICAL EPISODE

Address to M, but also to A

Positioning as the one who

tries and succeeds

Address to M+T2

feeling bad voice,

positioning as weak in
attempts to be better;
address to other voice
inside

Voices, addressees,
positioning

Table 1
Sequence I: Arja (A) and Matti (M) With Therapists 1(T1) and 2 (T2), First Session, Topical Episode 9 (187214)

October 2012

JOURNAL OF MARITAL AND FAMILY THERAPY

673

T1

Hmh . . .
hmh.

Yes, yes . . . hmh.

T2

Table 1
Continued

Arja

. . . so that we should have . . . were

living in a at . . . the thing is that
Arja wont . . . she easily
withdraws . . .

That she should let things go in one ear

and out the other. And Ive been
worried, also, about her alcohol use . . .
and . . . well . . . Ive tried to discuss it
with her. And now weve rented a
house and . . . this . . . in a way . . .
has cheered her up . . .

Well, Ive tried to

support Arja and . . .
and to discuss these things. Ive
tried to . . . in a way . . . give
my support to her work, to say, Just
try to have control of your own work
and do your own business as well as
possible . . . and the thing to do is look
after your own aairs and not think
about the boss so much.

Matti

As problems
[Reexive mode]

Interactional dominance
Opens up the emotionally
loaded issue of
drinkingstill indicative
[Internal mode]
[External mode]
Back to practical life

Semantic dominance;
monological: giving a
report of his concerns,
lists six issues; indicative
language [External mode]

Response
categories

Address to therapists
and A

Positioning as active
male voice
Positioning as responsible
Addressee: therapists and A
I as agent
Positioning himself as not
succeeding in talking about
the drinking; positioning
changes: we

Positioning himself as
supporter; voice of
husband and supervisor
A as addressee; voice of a
severely depressed spouses
husband

Voices, addressees,
positioning

674

JOURNAL OF MARITAL AND FAMILY THERAPY

October 2012

T1

Table 1
Continued

Aha.

Was it this
summer you
got the idea
for this house?

Was it this
summer you
got the idea
of . . .

Yeah . . .

Hmh.

T2

Arja

Yes, we were
proposing to act as
caretaker of this
house.

What?

But now Ive noticed

after being three
weeks in the house,
shes been a lot more
active.

. . . into the home and

doesnt easily go out.

Matti

Indicative language;
interactional and
semantic dominance;
dialogical
Dialogical

Response to the cottage

theme, out of six possible
topics; not to Ms
concerns about As
drinking or her other
problems

Response
categories

Surprised
Positioning for choice of
empowered parts of what
was said
Address to both A and M
Adapts to the empowering
topic of the question

Voice of husband of
depressed wife; having
more agency after nding
some means to deal with
matters; address to A

Voices, addressees,
positioning

October 2012

JOURNAL OF MARITAL AND FAMILY THERAPY

675

Hmh.

T1

Table 1
Continued

T2

Well . . . one thing

aecting my
physical condition
is that Ive not been
able to walk
properly.

Arja

But that . . .

The life [there]

consists of doing
dierent things. . . .
We stay outside a lot
. . . and it seems to
be . . . To live in a
at is a bad thing for
Arja . . . especially
when she is on sick
leave and doesnt go
out and . . . Of
course she hasnt
been able to walk so
much either.

Matti

A responds, even if not

directly addresssed
Semantic dominance
Indicative

Indicative
Answering topic
previously dealt with
[Reexive mode]
[External mode]

Response
categories

NEW TOPICAL
EPISODE

Positioning himself as a
supervisor; voice of a
worried husband
Address to A
Voice of a worried husband
with a critical tone

Voices, addressees,
positioning

- Interactional dominance: this refers to the inuence of one participant over the communicative actions, initiatives, and responses within the sequence. It is possible that this individual
will have more inuence on other parties than that exerted by the actual interlocutors (Linell,
1998; Linell, Gustavsson, & Juvonen, 1988). For instance, when a family therapist is inviting a
new speaker to comment on what was previously said, he or she can be said to have interactional dominance; however, someone who is very silent also can have interactional dominance by evoking solicitous responses from others.
Rather than identifying the person who is dominant in the family session, the main focus
of our research is on the shifting patterns of these three kinds of dominance.
2) What is responded to? The speakers may respond to
- their experience or emotion while speaking of the thing at this very moment (implicit
knowing)
- what is said at this very moment
- some previously mentioned topics in the session
- what or how it was spoken
- external things, outside this session
- other issues (If so, what?)
These are not mutually exclusive categories, as in a single utterance, many aspects can be
presented. The special form of answers in a situation in which the speaker introduces several
topics is considered to form one utterance. We look at how the answer helps to open up a
space for dialogues in the response to that answer.
3) What is not responded to?
What voices in the utterance (bearing in mind that a single utterance by a single participant can include many voices) are not included in the response of the next speaker?
4) How is the utterance responded to?
Monological dialogue refers to utterances that convey the speakers own thoughts and
ideas without being adapted to the interlocutors. One utterance rejects another one. Questions
are presented in a form that presupposes a choice of one alternative. The next speaker answers
the question, and in this sense, his or her utterance can be regarded as forming a dialogue;
however, it is a closed dialogue. An example would be when the therapist asks for information
about how the couple made the contact, and the couple answers with information about their
actions leading up to participation in the therapy session. In dialogical dialogue, utterances are
constructed to answer previous utterances and also to wait for an answer from utterances that
follow. A new understanding is constructed between the interlocutors (Bakhtin, 1984; Luckman,
1990; Seikkula, 1995). This means that in his or her utterance, the speaker includes what was
previously said and ends up with an open form of utterance, making it possible for the next
speaker to join in what was said.
5) How is the present moment, the implicit knowing of the dialogue, taken into account?
When one looks at videos of dialogues in which there are sequences of responses, one observes
body gestures, gazes, and intonation. Often this includes (for example) observing tears or anxietyaspects not seen when one merely reads the transcript. The present moment becomes visible also in the comments on the present situation (e.g., comments on the emotions felt
concerning the issue under scrutiny).
STEP III: Exploring the processes of narration and the language area
This step can be conducted in two alternative ways:
1) Indicative versus symbolic meaning
This distinction refers to whether the words used in the dialogue are always being used to
refer to some factually existing thing or matter (indicative language) or whether the words are
being used in a symbolic sense; in other words, whether they are referring to other words, rather
than to an existing thing or matter (Haarakangas, 1997; Seikkula, 1991, 2002; Vygotsky, 1981;
Wertsch, 1985). Each utterance is categorized as belonging to one of these two alternatives.
2) Narrative process coding system
The preliminary development of this coding system was undertaken by Agnus, Levitt, and
Hardtke (1999) within individual psychotherapy. Laitila, Aaltonen, Wahlstrom, and Agnus
(2001) further developed the system for the family therapy setting. Three types of narrative
676

JOURNAL OF MARITAL AND FAMILY THERAPY

October 2012

processes are distinguished. The speaker uses (a) external language, giving a description of
things that happened or (b) internal language, describing his or her own experiences in the
things he or she describes; or (c) reective language, exploring the multiple meanings of things,
the emotions involved, and his or her own position in the matter.
After analyzing the responses in the chosen topical episodes, a conclusion is reached concerning how the chosen topic is handled in this specic therapy process.
One nal comment should be made before we illustrate our approach with a case example,
namely that we prefer to work with a team of researchers. This is not only because it enhances
the credibility of the research but also because team investigations into multi-actor dialogues
seem appropriate when one is considering analysis as a multi-actor process. One possible way is
to start the analysis of the transcript with a single researcher. After the preliminary categorization (using the three steps described previously), the research team comes together to review the
video of the session and also the transcript. In the course of that meeting, as a check on the
trustworthiness of the rst authors analysis, the coresearchers review the categorization, focusing more on their points of disagreement than on their points of agreement. In dialogue with
each other, the different voices enrich the picture of the dialogue in focus.

CASE EXAMPLE OF RESPONSES

In what follows, a case is presented to illustrate how the method can be used in making
sense of different aspects of responding dialogues. We shall rst show how we use the concepts
introduced above to conduct a global evaluation of the dialogical qualities of the session. We
shall then introduce a microanalytic exploration of the session around this question: How can
therapists in dialogue aect the future direction of the conversation? In the microanalytic investigation, the focus is on voices, addressees, and positioning. For this purpose, two short episodes
will be reported here.
The case of Arja and Matti is part of the research project Dialogical and Narrative
Processes in Couple Therapy for Depression. Arja and her husband, Matti, were referred to a
mental health outpatient clinic by the doctor in the occupational health care because of Arjas
depression. In exploring the responses in the dialogue, the rst sessions are important in the
sense that no dialogue habits have yet been cocreated by the therapists and the clients. Two
therapists, one man and one woman, (each with more than 20 years experience as family therapists) conducted the therapy. Altogether, Arjas treatment took 11 months and extended over
eight family therapy sessions. Arjas depression eased o to such an extent that while at the
outset her Beck Depression Inventory score was 32, at the 18-month follow-up, it had fallen to
2. The therapy can thus be seen as successful.
The research was conducted by a team of researchers. First of all, students of psychology
conducted the narrative process coding system as part of their masters thesis. Thereafter, the
rst two authors of the present article conducted a dialogical investigation. The specic analysis
of the case was discussed in two seminars with senior researchers from outside our research
team. This provided a rich picture in which different aspects of the dialogical process of the
session with Arja and Matti were articulated. We shall now focus on the rst session of the
therapy with Arja and Matti, looking at how the concepts we introduced above can (a) help us
to look at the session as a whole and evaluate its dialogical qualities and (b) help us to focus
on some topical episodes, to conduct a microanalytic study of certain aspects of the session that
are of interest to researchers.
Evaluation of the Dialogical Qualities of the Session as a Whole
In the rst therapy session with Arja and Matti, altogether 33 topical episodes emerged, of
which 26 had therapeutic quality. In seven episodes, the topic involved therapy agreements and
other concrete practical issues. Arja had semantic dominance in 13 sequences, Matti in four
sequences, and the therapists in 13. Arja had interactive dominance in 12 sequences and her husband in four sequences. The variation in dominance seems to suggest that, in the session, there
was enough exibility in alternating initiatives to allow all parties the possibility of addressing
issues they considered important.
October 2012

JOURNAL OF MARITAL AND FAMILY THERAPY

677

The rst four episodes consisted of information on the study and treatment context.
Episodes 5 and 6 already had some therapeutic quality, but were still informative. In episodes
79, Arja reported on several health problems. The language area remained indicative, and the
conversation proceeded in a monological way. In the ninth episode, however, when the spouse,
Matti, talked about his worries concerning his wife, the language changed to symbolic and dialogic. During episodes 1021, Arja often positioned herself as the victim of a heavy workload
in her job and also of misinterpretations by her boss and by the general practitioner to whom
she had tried to talk about her health concerns. Matti positioned himself as the one supporting
Arja, both in practice and in the storytelling. In the episodes from 22 to 25, on the therapists
initiative, the focus was on Arjas heavy drinking. The emotional tension in the session
increased, and the clients started to use symbolic language in a dialogical way. In many ways,
this episode seemed to be of specic importance for the therapy process. For this reason, this
particular episode was selected (together with episode 9 in which Arjas health problems were
discussed) for the microanalysis, which we shall report on later in this article.
When we looked at the way they told their story and at the creation of meaning, it was
noted that Arja and Matti described their lives mostly in indicative ways, with symbolic meaning being present in only nine of the sequences. They mostly talked about what happened in
their lives in a traditional question answer format, and only nine of the topical sequences in
the session took a dialogical form. The therapists respected the clients way of being in language and did not challenge it. In that way, they succeeded in promoting an open atmosphere
in which it was possible to speak about sensitive issues. The responses were almost entirely
given to previously spoken issues, existing within the present moment. There were only a few
instances when the therapists wanted to return to themes that had been spoken about previously (each time connected with alcohol use). In the two phases of the session in which they
conversed in the symbolic language area, the therapists responded to relational aspects of Arjas
and Mattis stories. On three occasions, the therapists did not respond to the emotional reactions of Arja, two of these being connected with the theme of Arjas somatic illness.
As regards the rest of the therapy process, no signicant changes took place in the dialogical processes within the session. The topical episodes mainly focused on incidents in the couples life, with indicative language use. Most of the parts consisted of answering questions. The
most remarkable change was in the polyphonic quality of the couples utterances. For example,
in the last session of the therapy (the eighth session), they both talked a great deal about what
the other one had said, and thus, the inner voice of the spouse became evident in their talking.
This probably meant that the words were directed to the other spouse, who was present at the
same time.
Microanalysis of Two Dialogue Episodes
We shall now focus on two topical episodes (9 and 22) of the rst therapy session to conduct a microanalysis aimed at gaining an understanding of the way the therapist can make
room for certain themes to be discussed in the session. These episodes were chosen because
both of them illustrate a change in the dialogue and thus are informative for our purposes. We
shall focus in particular on positioning, addressees, and voices (Table 1).
In the rst sequence, Arja described how she felt bad, and T1 joined this in a dialogical
way in the symbolic language area. We can assume that she addressed her words to Matti also.
Even if she positioned herself as being powerless, after T1 responded to her utterance, she
found some empowering ideas (you can get rid). After that, T1 turned to Matti, asking about
his concerns in the present situation: how would you comment on that or is there something
else you would like to add? In inviting a new speaker, the therapist was taking interactional
dominance. In response, Matti talked about his concerns over Arjas heavy workload. He positioned himself as the one supporting Arja and searching for new options in life. He seemed to
address his words both to the therapists and to Arja. This became evident when Arja continued
uently from this point, with Matti stopping at this point. He explained that he had tried to
support Arja and that he was worried about Arjas alcohol use. He spoke about the new house
that they had hired for the summer and about Arjas withdrawal from social contacts. During Mattis speech, the second therapist (T2) became more active, as if the male voice was
678

JOURNAL OF MARITAL AND FAMILY THERAPY

October 2012

responding to the voice of another man. T2 was here positioning himself as active. He followed
Mattis speech, giving short encouraging comments. In the middle of Mattis story, he asked,
Was it this summer you got the idea of the summer cottage? It is interesting that out of Mattis six dierent topics, the therapist selected the summer cottage. Later on, this turned out to
be a selection that supported the couples stories in the frame of the positive resources available. By selecting the topic of the house, the therapist succeeded in connecting with some of the
positive resources of Arja and Matti, instead of focusing on the problem elements in Mattis
answer.
In addition, this response can be seen as a choice of positioning. Knowing the narrative
background of the therapist, this might be an intentional positioning, inviting the clients to
speak with empowered voices about their life in connection with some of their resources. The
therapists answer is interesting because at this specic point, he used the word you in the plural
form. In that way, he was addressing both Matti and Arja. The therapists response was initiated by Mattis relating of what has happened, but it was directed to both of the spouses simultaneously, and thus, both of them remained agents in the story.
The voices emerging most often in the session were those of Arja and Matti talking about
each other. Arja recounted many details about her fellow employees, about her boss with
whom she repeatedly had diculties, and about the doctors and nurses in the occupational
health care system. A specic voice entered when Arja and Matti talked about their wish for a
child. Arja said, with a touch of sarcasm, that it was good that they did not have a baby, as
the baby did not need to suer from Arjas misery. Although these words were addressed to
the therapist, the baby can also be seen as an addressee, someone to whom the words were
directed. Other outside addressees were not recognized in the topical episodes.
The second episode we wish to focus on here is episode 22, in which Arjas drinking is
discussed and in which more emotions emerge (Table 2).
In this episode, T1 asked the clients about the use of alcohol. The therapists seemed to
have the idea that Arja wanted to stop drinking and that she had already decided to quit. However, it appeared that this was not the case. The therapist responded to this surprising information by accepting Arjas comment and starting to explore the signicance of alcohol use. In
answering the therapists questions, Arja reacted very emotionally, which, however, was not
responded to by the therapists. As in the ninth sequence, they seemed to prefer to focus more
on everyday aspects of becoming bored.
In Arjas emotional utterance, many voices seem to be present, even if they were not
directly claried. Her husband, Matti, was both one voice and her addressee, which became visible when she slightly smoothed his hand. Some voices of her life, related to her need to dull
her emotions through alcohol use, became visible, even if they were not dened precisely. Arja
positioned herself as powerless concerning the issue of drinking (representational positioning),
but at the same time, she demonstrated that she was an agent in relation to the therapist by
disagreeing with T1s suggestions (interactional positioning). One can speculate about what
would have happened if the therapist had focused on Arjas emotional reaction (her crying)
and encouraged her to speak more about her experience at the present moment. Would that
have made it possible for these experiences to be voiced and for these voices to be heard by the
therapists, Matti, and herself? One possible voice seemed to speak of the loss of children, which
Arja mentioned in her next utterance. The never-born baby became one voice in her utterance.
In addition, the therapists professional voices can be observed in this transcript. The choice
of responding to the renting of the house perhaps illustrates the therapists specic family therapy orientation. They seemed to be working from systemic premises, enriched by a narrative
orientation. Here, also, may be the explanation of the therapists choice not to respond to the
emotional reaction of Arja, but rather to emphasize the positive experience in Mattis utterance.
As the narrative therapist Michael White (1991) has noted, in dealing with dicult and traumatic issues, it is important to rst construct a frame of the client as a competent individual, to
guarantee resources for dealing with experiences in which the client has not been so competent.
As we mentioned earlier, in our analyses, we only focus on how therapists, through their
answers, guide the dialogue process. In the sequence focused on here, the therapists choice of
talking about the house meant that other choices were left out. This kind of point is always
October 2012

JOURNAL OF MARITAL AND FAMILY THERAPY

679

680

JOURNAL OF MARITAL AND FAMILY THERAPY

October 2012

Yes.

Under control . . .

If I could return to
the alcohol issue,
when you said . . .
you said in a way . . .
that I understood
by implication that
. . . that in a way
you . . . you feel
powerless . . . or do
you have a feeling
that it is under . . .

T1

How do you know

that it isnt under
control . . . do you
have . . . ?

T2

Well . . . it isnt under

control.

Under control?

Thats true.

Responds to control;
dialogical

Responds to powerlessness
issue
Dialogical
Empathic responsiveness;
dialogical
Responds by denying

Interactional and semantic

dominance; dialogical, even
if it increases the persons
own interpretation of what
was said; symbolic meaning

Response
categories

Table 2
Sequence II: Arja (A) and Matti (M) With Therapists 1(T1) and 2 (T2), First Session, Topical Episode 22 (904944)

Voice of not having

control, but
positioning herself as
the one opposing T1s
assumption

TOPICAL EPISODE
ALCOHOL ISSUE
Two voices: her own,
As
Positioning herself as
somewhat uncertain;
positioning A as
powerless

Voices, addressees,
positioning

October 2012

JOURNAL OF MARITAL AND FAMILY THERAPY

681

Aha, yes, (nodding

her head), exactly

. . . have a beer after

sauna, so that . . .

. . . suggestions
to . . .

(nodding her head)

Hmh . . . That in a
way even if youve
decided that you
wont drink, you
will drink. . . . That
these . . .

T1

Table 2
Continued

Hmh.

T2

My decision hasnt
been made. . . .

But I havent decided

that . . . I havent yet
managed to decide
that . . .

Hmhm.

Hmh.

Well . . . I . . . I drink
too much.

Responds to objection, accepts

Semantic and interactional

dominance
Monological

Responds by taking up word

for word what the client has
said, but still holding on to
her prejudgment
(youve decided)

Indicative

Response
categories

Double positions: not

managed but not
decided either;
positioning as the one
who can disagree with
T1 + 2

Positioning herself as
the one who can make
judgments about the
others response

Voices, addressees,
positioning

682

JOURNAL OF MARITAL AND FAMILY THERAPY

October 2012

To dull your
emotions?
About what?

Yes.

Yes.

Well . . . whats the

problem? That if
you see that youre
drinking too much
...

Yes?

Your decision
isnt made.

T1

Table 2
Continued

Hmh.

T2

(6) (crying) I dont know (6)

(sobbing). About my whole
life, I suppose (9) (weeping
more freely).

Perhaps I only want to

dampen myself down.

. . . by my side.

. . . and I have an ally

here (touching the
arm of her husband)
...

. . . isnt yet made . . .

My decision . . .

Symbolic; semantic and

interactional dominance

Dialogical

Interactional dominance; symbolic

Responds to As undecided
comment; dialogical

Dialogical

Response
categories

Intimate voice of her

lifeimplicit, not
explicitly denable

Address to M

Voice of a person
(whats the
problem)

Voices, addressees,
positioning

October 2012

JOURNAL OF MARITAL AND FAMILY THERAPY

683

Yes, yes . . .

Yes, yes . . .

Hmh . . . well, when I

was listening to
you talking about
Mattis keep-t
hobbies, and . . .
about his training
programs, I was
thinking that if you
feel bored . . . that
if you have . . .

T1

Table 2
Continued

T2

are no children who

should suer from . . .

. . . that we dont have

children . . . that there

Hmh . . . well, I should

be more active . . . that
there are other
hobbies available, as
well. Not only
walking. Perhaps it
depends on the fact . . .

Semantic dominance

Responds to indicative part

of T1s utterancesteps
away from emotions

Does not respond to the emotion

at the present moment
Responds to something
previously discussed
Interactional dominance
Indicative language

Response
categories

Address to M
Voice of the
unborn baby

. . . children; private
voice of sorrow

Shifting toward a NEW

TOPICAL EPISODE
of not having . . .

Ms voice; address
to M
Positioning herself as
the one who can
disacknowledge
emotions

Voices, addressees,
positioning

684

JOURNAL OF MARITAL AND FAMILY THERAPY

October 2012

T1

Table 2
Continued

T2

Yes . . .

Well, it depends on
these . . .

Responds to T1 and the wife

at the same time
Indicative language

You are very

slow . . . slow to make visits
to anyone.

. . . nothing works out when

Im trying to get her out to go
somewhere.

It is . . .

Response
categories

Loss of agency

Voice of A

Voices, addressees,
positioning

crucial in a therapy session. If the therapists had had a dialogical orientation instead of a
narrative one, their answer might have been different and more focused on Arjas emotional
reaction (e.g., Seikkula, 2008). This would have led to a dierent type of dialogue after the
comment, compared with the dialogue that actually occurred.
What Can We Learn About the Case Using This Method?
In the course of the general categorizing of the responses, some illustrative elements of this
therapeutic process became evident. The couple wanted to talk at length about emotionally
loaded experiences concerning their work, their friends, and stressful aspects of their intimate
relationssuch as not yet having a child. They did this mostly using indicative language,
reporting how things had happened and how they had reacted.
From one specic session, we selected two decisive episodes for microanalysis. In these
episodes, symbolic meanings emerged. From a more detailed description using the Narrative
Process Coding System, it was possible to observe shifts between the internal and reective
mode in the couples stories. The therapists acted in a dialogical way, with their main aim being
to follow the stories of the clients and thus focus on responding to their utterances. Mostly,
they managed to do this in such a way that the clients could go on in their experiences and nd
new words for those parts of their stories that did not have any exact narrative form. In their
ne tuning, the therapists acted according to their systemic and narrative ideas on family
therapy. They chose to answer those parts of the utterances that consisted of descriptions of
actions and happenings. On the other hand, they chose not to respond to embodied implicit
utterances, such as emotions that became evident in crying. After the rst session, the therapeutic process continued following the same main lines. In the clients stories, more voices emerged,
which can be seen as a sign of having more inner dialogues concerning all the issues in their
life. This seemed to be associated with an easing of the depressive mood, with a diminishing of
the need for further therapy.
In this investigation, we can easily see the elements of the family therapy that may explain
why the process was successful. As is known, the relationship between therapists and clients is
one of the most important elements for positive change. In this case, the foundations for good
listening and empathetic relations were present from the very start of the rst sessions. The
clients could use their familiar language and take initiatives regarding both the content of the
stories and the way it was spoken about. They could freely correct the therapists suggestions
or preunderstanding and thus affect the therapeutic process. The most profound issues in their
critical situation were put into words from the very beginning, and they thus received more
agency within the experiences they underwent.
For family therapists, the method introduces interesting possibilities for looking at our
personal ways of participating in therapeutic dialogues. In becoming aware of how we take the
initiative regarding the form of the dialogue (interactional dominance), we may nd new
resources for inviting more speakers into the specic subject. In looking at the response categories, we may develop our dialogical sensitivity to a multiplicity of forms of answers; thus, we
may avoid repeating the same patternsrepetitions that could hinder new perspectives from
being opened in emotionally charged topics. We may nd, for instance, that it is our (repeated)
answer form that is causing the dialogue to focus on negative and pathological aspects of the
couples interaction, rather than on nding new resources in the midst of all the difculties that
the couple may have. And in becoming aware of the indicative quality of the language of the
family, we may gain a better understanding of the uneasy feeling that we may experience in a
session when we strive for a new understanding and nothing new seems to come out of the conversation. When we are present in the actual dialogue, it is difcult for us to separate ourselves
from the context and to look at ourselves as outsiders; nevertheless, it may be that some of the
aspects that are analyzed are capable of being observed at the actual moment of dialogue.

DISCUSSION
In this article, we have proposed some preliminary ideas on qualitative investigations into
multi-actor dialogues. Our aim has been to work toward an integration of Bakhtins theoretical
October 2012

JOURNAL OF MARITAL AND FAMILY THERAPY

685

concepts with good practices of qualitative research (e.g., dialogical tools and concepts of a
narrative processes coding system) to make sense of family therapy dialogues. While communication research has been an important pioneering line in the development of family therapy
and family therapy research (e.g., Bateson, 1972), the perspective of the dialogical self, multiple
voices, and positioning adds interesting new dimensions to the behavioral level of systems
thinking. The social reality of therapy conversation is not a mere strategic game, but rather a
multidimensional negotiation of living persons participating in both outer and inner dialogues;
it is a negotiation where new meanings are generated (Rober, 2005b).
We have taken on the challenge of operationalizing Bakhtinian concepts, moving from
heuristic ideas to tools for empirical qualitative research. So far, we have found that this way
of analyzing the data provides us with new and exciting alternatives for reading marital therapy
transcripts. Our methodology involves a way of categorizing the data from the entire session,
allowing us to see the quality of the session as a whole. Furthermore, it includes also microanalyses of specic topical episodes that seem to be important for the therapeutic process. We
have tried to use the possibilities afforded by recorded data, therapy transcripts, and research
team negotiations to reach multiple descriptions of therapy sessions and identify some crucial
moments in these sessions. The testing of this procedure has turned out to be particularly fruitful with data from marital therapy. The three-step analytical procedure shows the importance
of looking at the therapists response and seeing how the horizontal client-informed or clientcentered process intertwines with the therapists various agendas. Using all the sources of data
within research, team negotiations have also helped us to take notice of all the participants and
of their individual ways of participating in therapy discourse. While we have used this research
tool in research on couple therapy sessions, the next step in the development of this research
tool could be to try out these analytical tools in family therapy sessions with children or in the
treatment meetings of a social network. In the future, we would also envision doing this kind
of analysis and then comparing sessions with good and poor outcomes. And overall, it will
always be our aim to look for ways in which dialogues can promote change for the better in
communication between couples and within families.

REFERENCES
Agnus, L., Levitt, H., & Hardtke, K. (1999). The narrative process coding system: Research applications and
applications for therapy. Journal of Clinical Psychology, 50, 12441270.
Andersen, T. (1995). Reecting processes: Acts of forming and informing. In S. Friedman (Ed.), The reecting
team in action (pp. 1137). New York: Guilford Press.
Anderson, H., & Goolishian, H. (1988). Human systems as linguistic systems. Family Process, 27, 371393.
Bakhtin, M. (1981). The dialogic imagination. Austin: University of Texas Press.
Bakhtin, M. (1984). Problems of Dostoevskys poetics. Minneapolis: University of Minnesota Press.
Bakhtin, M. (1986). Speech genres and other late essays. Austin: University of Texas Press.
Bateson, G. (1972). Steps to an ecology of mind: A revolutionary approach to mans understanding of himself.
Toronto: Random House.
Baxter, L. A. (2004). Dialogues of relating. In R. Anderson, L. A. Baxter, & K. N. Cissna (Eds.), Dialogues:
Theorizing difference in communication studies (pp. 107124). London: Sage.
Gergen, K. (2009). Relational being: Beyond self and community. New York: Oxford University Press.
Haarakangas, K. (1997). The voices in treatment meeting: A dialogical analysis of the treatment meeting conversations in family-centred psychiatric treatment process in regard to the team activity. Diss. English Summary. Jyvaskyla Studies in Education, Psychology and Social Research, 130, 119126.
Hermans, H. J. M. (2004). The dialogical self: Between exchange and power. In H. J. M. Hermans & G. Dimaggio (Eds.), The dialogical self in psychotherapy (pp. 1328). New York: Brunner Routledge.
Hermans, H. J. M. (2006). The self as a theater of voices: Disorganization and reorganization of a position repertoire. Journal of Constructivist Psychology, 19, 147169.
Homan, L. (1991). A reective stance for family therapists. Journal of Strategic and Systemic Therapies, 10, 417.
Laitila, A. (2009). The expertise question revisited: Horizontal and vertical expertise. Contemporary Family Therapy, 31, 239250.
Laitila, A., Aaltonen, J., Wahlstrom, J., & Agnus, L. (2001). Narrative process coding system in marital and
family therapy: An intensive case analysis of the formation of a therapeutic system. Contemporary Family
Therapy, 23, 309322.

686

JOURNAL OF MARITAL AND FAMILY THERAPY

October 2012

Leiman, M. (2004). Dialogical sequence analysis. In H. Hermans & C. Dimaggio (Eds.), The dialogical self in psychotherapy (pp. 255270). Sussex: Brunner-Routledge.
Linell, P. (1998). Approaching dialogue: Talk, interaction and contexts in dialogical perspectives. Amsterdam Philadelphia: John Benjamins.
Linell, P., Gustavsson, L., & Juvonen, P. (1988). Interactional dominance in dyadic communication: A presentation of initiative-response analysis. Linguistics, 26, 415442.
Luckman, T. (1990). Social communication, dialogue and conversation. In I. Markova & K. Foppa (Eds.), The
dynamics of dialogue (pp. 4561). London: Harvester.
Lysaker, P., & Lysaker, J. (2004). Dialogical transformation in the psychotherapy of schizophrenia. In H. Hermans & C. Dimaggio (Eds.), The dialogical self in psychotherapy (pp. 205219). Sussex: Brunner Routledge.
Markova`, I., Linell, P., Grossen, M., & Orvig, A. S. (2007). Dialogue in focus groups: Exploring socially shared
knowledge. London: Equinox.
Rober, P. (1999). The therapists inner conversation: Some ideas about the self of the therapist, therapeutic
impasse and the process of reection. Family Process, 38, 209228.
Rober, P. (2002). Some hypotheses about hesitations and their nonverbal expression in the family therapy session. Journal of Family Therapy, 24, 187204.
Rober, P. (2005a). The therapists self in dialogical family therapy: Some ideas about not-knowing and the therapists inner conversation. Family Process, 44, 477495.
Rober, P. (2005b). Family therapy as a dialogue of living persons. Journal of Marital and Family Therapy, 31,
385397.
Rober, P. (2008). The therapists inner conversation in family therapy practice: Struggling with the complexities
of therapeutic encounters with families. Person-Centered and Experiential Psychotherapies, 7(4), 245278.
Rober, P., Van Eesbeek, D., & Elliott, R. (2006). Talking about violence: A micro-analysis of narrative processes
in a family therapy session. Journal of Marital and Family Therapy, 32, 313328.
Salgado, J., & Goncalves, M. (2007). The dialogical self: Social, personal, and (un)conscious. In A. Rosa &
J. Valsiner (Eds.), The Cambridge handbook of social cultural psychology (pp. 608621). Cambridge: Cambridge University Press.
Seikkula, J. (1991). Perheen ja sairaalan rajasysteemi. [Family-hospital boundary system in the social network.
English summary]. Jyvaskyla Studies in Education, Psychology and Social Research, 80.
Seikkula, J. (1995). From monologue to dialogue in consultation with larger systems. Human Systems, 6, 2142.
Seikkula, J. (2002). Open dialogues with good and poor outcomes for psychotic crises: Examples from families
with violence. Journal of Marital and Family Therapy, 28, 263274.
Seikkula, J. (2008). Inner and outer voices in the present moment of family and network therapy. Journal of Family Therapy, 30(4), 478491.
Seikkula, J., & Arnkil, T. (2006). Dialogical meetings in social networks. London: Karnac Books.
Steiner, G. (1989). Real presences: Is there anything in what we say. Chicago: University of Chicago Press.
Stiles, B., Osatuke, K., Click, M., & MacKay, H. (2004). Encounters between internal voices generate emotion:
An elaboration of the assimilation model. In H. Hermans & C. Dimaggio (Eds.), The dialogical self in psychotherapy (pp. 91107). New York: Brunner Routledge.
Vygotsky, L. (1962). Thought and language. Cambridge, MA: MIT Press.
Vygotsky, L. (1981). The development of higher forms of attention in childhood. In J. Wertsch (Ed.), The concept
of activity in Soviet psychology (pp. 189240). New York: M. E. Sharpe.
Wertsch, J. (1985). Vygotsky and social formation of mind. Cambridge, MA: Harvard University Press.
Wertsch, J. V. (1991). Voices of the mind: A sociocultural approach to mediated action. Cambridge, MA: Harvard
University Press.
White, M. (1991). Deconstruction and therapy. Dulwich Centre Newsletter, 3, 121.
Wortham, A. (2001). Narratives in action: A strategy for research and analysis. New York: Teachers College
Press.

October 2012

JOURNAL OF MARITAL AND FAMILY THERAPY

687