Heidelberg systemic therapy

1. Introduction

Version followed: the systemic (post-1978, Milan-influenced) Heidelberg family therapy of Helm Stierlin's department, in the session format taught in Simon & Rech-Simon, Zirkuläres Fragen (14th ed., 2021: interview, pause, closing intervention). Stierlin's earlier relational concepts (binding and expelling, delegation, related individuation) are used as lenses for hypotheses. The method catalogue follows the later consolidation by Heidelberg member Jochen Schweitzer with Arist von Schlippe, Lehrbuch der systemischen Therapie und Beratung I (4th ed., 2025/26).

What it is. The "Heidelberg school" is a school and research tradition, not a single manualised therapy. In 1974 Stierlin founded the Abteilung für psychoanalytische Grundlagenforschung und Familientherapie at Heidelberg University. Its members over time included Ingeborg Rücker-Embden-Jonasch, Michael Wirsching, Gunthard Weber, Fritz B. Simon, Gunther Schmidt, Hans-Rudi Fischer, Arnold Retzer and Jochen Schweitzer. Stierlin saw himself as a mentor rather than the owner of a doctrine, and the group treated opposing positions as a source of new ideas. [source: Ritscher 2026, pp. 5, 15, 18; Reitz 2014, p. 9]

The tradition developed in two phases:

• Psychoanalytic-relational phase (1970s). Stierlin wrote on the separation of adolescents from their parents (1970/1974) and on delegation (1978), building on Wynne, Lidz and Boszormenyi-Nagy.
• Systemic-constructivist phase (from the late 1970s). The group worked closely with the Milan team from 1974 and took over its guidelines of hypothesising, circularity and neutrality, then turned to constructivism and circular interventions.

[source: Ritscher 2026, pp. 6–9, 15]

Heidelberg-specific contributions include:

• research and therapy with families of people with schizophrenic, schizoaffective and manic-depressive psychoses, and with psychosomatic and eating disorders;
• the distinction between "hard" and "soft" realities;
• a strong emphasis on clinical epistemology.

[source: Ritscher 2026, pp. 14–17; DGSF Wissensportal, sect. 1; Simon 1990/2025 sample]

How change happens. Problems and symptoms are treated as part of recurring patterns of interaction and meaning, captured in Stierlin's phrase that one person's doing is the other's doing. Change comes from bringing in news of difference:

• questions that make each member's constructions visible to the others and open alternative possibilities;
• closing interventions that reframe the pattern or prescribe it in a new context.

The family, not the therapist, then reorganises itself. The developmental goal underneath is "related individuation": becoming autonomous in relation to the family, not against it. [source: Ritscher 2026, pp. 10–11, 15–17; Simon & Rech-Simon 2021, Vorbemerkung, ch. 2; Methodenpool Köln]

2. Framework

The sources do not give a fixed number of phases or a session protocol. The structure below follows the ideal-type session sequence of Simon & Rech-Simon (interview, pause, closing intervention) and the contracting and assessment steps in von Schlippe & Schweitzer. [source: Simon & Rech-Simon 2021, contents and ch. 2; von Schlippe & Schweitzer 2026, chs. 13–19]

Setting (classic research-clinic version). Two therapists sit in the room with the family, and two colleagues watch from behind a one-way mirror. Sessions are video-recorded with consent; patients may refuse recording or have tapes erased. The authors call a team a luxury that is mostly affordable in research projects. A single therapist can run the same structure, using a self-imposed pause instead of a team break. [source: Simon & Rech-Simon 2021, Vorbemerkung, ch. 2]

[source: Simon & Rech-Simon 2021, ch. 2; von Schlippe & Schweitzer 2026, sects. 13.3, 14.1, 14.4, 16.7, 18.5; Ritscher 2026, p. 16]

Session count and spacing.

• Spacing [uncertain]. The Heidelberg and Milan teams are widely described as using long intervals of several weeks between sessions, the so-called "long brief therapy". This was not confirmed in any source, and Retzer's psychosis follow-up studies (1989, 1991), which report the actual figures, are not drawn on.
• A related German team for comparison (Stuttgart, not Heidelberg). Sessions were usually 14 days apart, longer later in treatment and shorter in psychotic crises. The interventional-team format averaged 24.4 sessions over 14 months; the reflecting-team format averaged 8.7 sessions over 7.7 months. [source: Helmes et al. 1999, pp. 65, 68]
• Germany today. At the Heidelberg university outpatient clinic (HIP), statutory insurance covers a first session plus up to three follow-ups. Further therapy is covered under the Psychotherapy Guideline if a family member has a treatable disorder. [source: HIP Ambulanz page]

Who attends. Ideally all family members, but the constellation can vary from session to session. [source: HIP Ambulanz page; Helmes et al. 1999, p. 67 for variable constellations]

Certification and licensing.

• There is no licence or fidelity programme for "Heidelberg" therapy as such.
• In Germany, practitioners train in systemic therapy through DGSF- or SG-accredited institutes. The Helm Stierlin Institut in Heidelberg is one of the originating training sites.
• Delivering it as statutory psychotherapy requires psychotherapist licensure (Approbation) with systemic therapy as the specialist method.

[uncertain: training and Approbation details are from general knowledge and are unverified, except that the HIP page confirms delivery under the Psychotherapy Guideline]

3. Guiding principles

• Hypothesise, and keep several hypotheses. Meaning: before and during each session, write down several systemic explanations of how the symptom fits the pattern. Use them to choose questions, and drop any that stop generating information. You cannot not hypothesise. [source: von Schlippe & Schweitzer 2026, sects. 11.5, 13.3; Simon & Rech-Simon 2021, fn. 1 citing Selvini Palazzoli et al. 1981]
• Circularity. Meaning: ask about relationships and differences, not about isolated traits. Treat each answer as feedback for the next question. [source: von Schlippe & Schweitzer 2026, sect. 11.6; Methodenpool Köln]
• Information comes from differences. Meaning: when someone uses a label, ask what difference it makes. Simon & Rech-Simon: "Nur wenn man nach Unterschieden fragt, gewinnt man Informationen." (Only by asking about differences do you gain information.) [source: Simon & Rech-Simon 2021, ch. 2]
• Neutrality, all-partiality and curiosity. Meaning: do not take sides between members, between explanations, or between change and no change, and never act as judge. Watch out for terms that one member will hear as siding. [source: Simon & Rech-Simon 2021, ch. 2; von Schlippe & Schweitzer 2026, sect. 11.7]
• Respect for people, irreverence towards ideas. Meaning: question any belief (including diagnoses and your own theories) while staying warm towards the person who holds it. [source: von Schlippe & Schweitzer 2026, sect. 11.8]
• Make not-knowing your partner. Meaning: ask rather than explain, and let the family's answers lead. [source: von Schlippe & Schweitzer 2026, sect. 11.10]
• Context first. Meaning: early on, clarify the institution, the referral, any parallel helpers, previous treatments and diagnoses, and what others have promised the family. [source: Simon & Weber 2004, via Simsa 2004 review; Simon & Rech-Simon 2021, ch. 2]
• Do not devalue other helpers or methods. Meaning: replace "either the other therapist or us" by asking which complementary or competing functions each helper is expected to serve. [source: Simon & Rech-Simon 2021, ch. 2]
• Systemic humility about responsibility. Meaning: take responsibility only for what you can influence (the frame, your questions, your interventions). Still ask yourself whether what you do widens or narrows clients' room to act. [source: Simon & Weber 2004, via Simsa 2004 review]
• Take the clients' language seriously. Meaning: treat a medical word used by the family (e.g. "manic-depression") as their private usage, to be explored, not corrected. Notice personifications ("the depression came between us"). [source: Simon & Rech-Simon 2021, ch. 2]
• Soft versus hard reality. Meaning: relationship and illness definitions are "soft" constructions. Avoid cementing unchangeable labels such as "chronically schizophrenic", and keep "hard" constructions for physical facts. [source: Ritscher 2026, pp. 16–17]
• Circular causality. Meaning: one person's doing is another's doing, so describe behaviour as moves in a pattern rather than as individual guilt. [source: Ritscher 2026, p. 11; Ritscher 2022, sect. 1.2.1]
• Related individuation as the developmental yardstick. Meaning: aim for each member to become more themselves while staying connected. Be wary of goals that buy autonomy by cutting off, or connection by giving up the self. [source: Ritscher 2026, pp. 10–11]
• Conflict is productive; reconciliation is a process. Meaning: treat opposites as mutually dependent. Where reconciliation is not possible, help the family find ways of coping instead of forcing harmony. [source: Ritscher 2026, pp. 12–13; Ritscher 2022, sect. 1.2.1]
• Fairness and merit across generations. Meaning: listen for the "ledger" of guilt, debt and merit between generations, and for imbalances in fairness. Acknowledging merit can loosen deadlocks. [source: Ritscher 2026, pp. 13, 21]
• Emotions are interactional. Meaning: explore the interaction patterns around an emotion (who feels what when whom does what) rather than only the inner state. [source: Simon & Weber 2004, via Simsa 2004 review]
• Transparency and confidentiality. Meaning: show the setting openly. Tell participants that each person stays responsible for guarding their own secrets in a joint session: "Jeder Teilnehmer an solch einer Sitzung behält die Verantwortung für die Bewahrung seiner Geheimnisse." [source: Simon & Rech-Simon 2021, ch. 2]
• Psychotic behaviour is behaviour in context. Meaning: explore what "crazy" behaviour communicates in its context, including incongruent family communication, without claiming a family cause of psychosis. [source: Simon 1990/2025 sample, ch. 7; DGSF Wissensportal, sect. 5.1]

4. Intervention inventory

The sources name these interventions, but full descriptions are given only for the setting, context work and question types. For several closing interventions only chapter titles were visible. Those entries say so, and their steps are marked [synthesis] or left undescribed rather than invented. [source: Simon & Rech-Simon 2021, contents; von Schlippe & Schweitzer 2026, contents]

A. Preparation, frame and assessment

4.1 Pre-session letter and transparent setting

• Purpose / when: before the first session, so the family is not taken by surprise by the team, the mirror or the video. [source: Simon & Rech-Simon 2021, ch. 2]
• Steps:
  1. Send a short letter explaining that two therapists will talk with the family, that colleagues will watch from behind a mirror, that the session will be video-recorded with consent, and that there will be a break followed by a comment from the team.
  2. At the start, introduce everyone and offer to show the observation room. A Stuttgart team also had the whole team say goodbye in person at the end, partly to ease paranoid fantasies about the people behind the glass. [source: Helmes et al. 1999, p. 65]
  3. Ask for video consent. State that recording can be refused, or erased later.
  4. Give the confidentiality rule: each person decides for themselves what they share.
• Example language (illustrative): "Behind this mirror are two colleagues who help us think. Would you like to meet them first? You can say no to the video, now or later."
• Watch for: suspicion or shame about being watched, especially in psychosis. Offer to drop the video, or for the team to come into the room. Videos taken home can trigger further family discussion; this can be useful, but say that it may happen. [source: Simon & Rech-Simon 2021, ch. 2]

4.2 Team pre-session hypothesising

• Purpose / when: before every session. It turns referral data and the previous session into several working hypotheses that steer the questions. [source: von Schlippe & Schweitzer 2026, sects. 11.5, 13.3; Simon & Rech-Simon 2021, fn. 1]
• Steps:
  1. Collect the facts: referral source and wording, who is in the household, symptoms, diagnoses, previous treatments and other helpers.
  2. Each team member proposes a hypothesis about how the symptom fits the relationships. Use the relational lenses in 4.6 as prompts.
  3. Draft the first circular questions that would test each hypothesis.
  4. In the session, drop hypotheses that produce no new information, and add new ones.
• Example language (team, illustrative): "Hypothesis 2: the son's absence from school keeps the mother busy at home since the father's job move. Test it by asking the sister who worries most since the move, and who least."
• Watch for: one hypothesis turning into the truth, which is the loss of neutrality. Hypotheses that blame a member. Hypotheses are judged by usefulness, not truth. [synthesis]

4.3 Clarifying the context and referral

• Purpose / when: first session, and whenever another helper or institution comes into play. [source: Simon & Rech-Simon 2021, ch. 2; Simon & Weber 2004, via Simsa 2004 review]
• Steps:
  1. Ask whose idea the appointment was, what the referrer said would happen, and what each person was told or promised.
  2. Map parallel helpers: psychiatrist, inpatient unit, school, youth office.
  3. Ask which functions each helper is meant to serve, and whether these complement or compete with family therapy.
  4. Never devalue the other helpers.
• Example language: questions in the sources' own style (paraphrased from Methodenpool Köln): "Whose idea was it to come here?" "Why do you think your teacher gave you our number?" Illustrative: "Your doctor is treating the illness. What do you think she hopes we will add?"
• Watch for: promises the referrer made that you cannot keep; a family who comes only because a third party insists (see 4.4); rivalry with the inpatient team.

4.4 Contracting (Auftragsklärung) and the expectation carousel

• Purpose / when: first session, and again when therapy stalls. It makes "who wants what from whom" explicit, including absent people. [source: von Schlippe & Schweitzer 2026, sects. 14.1, 14.4, 16.7; Methodenpool Köln]
• Steps:
  1. Ask each member in turn what they want from the sessions, what they fear will happen, and what the others want.
  2. Ask what the referrer and absent members expect.
  3. For the expectation carousel, list every stakeholder and that stakeholder's expectations side by side. [uncertain: von Schlippe & Schweitzer describe the procedure in sect. 16.7]
  4. Agree a provisional mandate: what would be different, for whom, by when.
  5. If no one wants anything, make that the topic: who would be most relieved if nothing changed?
• Visitor / complainant / customer (de Shazer's distinction, as reproduced in Methodenpool Köln, not a Heidelberg original):
  • visitor: compliments only;
  • complainant: observation tasks;
  • customer: change tasks.
• Example language (illustrative): "If these sessions were a success, what would your mother notice first? And your brother? Who would doubt it longest?"
• Watch for: the therapist adopting a mandate nobody gave (for example, "cure the schizophrenia"); conflicting mandates that should be made visible rather than resolved.

4.5 Genogram

• Purpose / when: early sessions, when multigenerational themes, losses, migrations or repeated patterns are suspected. [source: von Schlippe & Schweitzer 2026, sect. 13.4]
• Steps:
  1. Draw three generations with the family on a large sheet.
  2. Mark births, deaths, separations, illnesses and moves.
  3. Ask circular questions about each figure: who resembles whom, who carries whose mission, whose merits went unacknowledged.
  4. Use it to generate hypotheses about legacies and delegations (see 4.6).
• Example language (illustrative): "Who in this family would grandfather have said is most like him? Who would disagree?"
• Watch for: turning it into a fact-finding exercise. Its value lies in the questions it prompts. [synthesis]

4.6 Stierlin's relational lenses (for hypotheses and questions)

• Purpose / when: as a template for hypotheses about separation and loyalty problems, especially in adolescence and young adulthood (eating disorders, psychosis, school refusal, leaving home). [source: Ritscher 2026, pp. 9–13]
• The lenses:
  • Binding versus expelling, as centripetal versus centrifugal forces. Binding can work on three levels:
    • id-binding: care, closeness and gratification that tie the child;
    • ego-binding: the child adopts the parents' perception of reality;
    • superego-binding: guilt and shame block a life of one's own.
  • Delegation. Missions handed to a child are normally meaning-giving. They "derail" when they are contradictory or impossible, or when their fulfilment goes unacknowledged. A healthy delegation lets the child refuse or reinterpret the mission.
  • Related individuation. The goal: autonomy in relation to the family.
  • Merit and the intergenerational ledger. After Boszormenyi-Nagy: invisible loyalties and accounts of guilt and merit between generations.
  • Mutuality. Positive mutuality versus the "malignant clinch".
  • [source for all lenses: Ritscher 2026, pp. 7, 9–13; Reitz 2014, pp. 10–11]
  • [uncertain: Stierlin et al. 1977, Das erste Familiengespräch, the primary text presenting these as perspectives for the first interview, is not drawn on]
• Steps (applying them):
  1. In the pre-session, ask of each lens: is there evidence of binding, expelling or delegation here, and on which level?
  2. Translate each into circular questions. For example: who would feel most guilty if the daughter moved out; what mission does the son fulfil by staying ill; whose merits have not been acknowledged.
  3. Use the answers to frame the closing intervention, for example by giving positive connotation to a loyalty.
  • [synthesis: the translation into questions is my own; Ritscher describes the concepts, not the question wording]
• Example language (illustrative): "Suppose Anna did leave home next year. Who would miss her most? Who would feel they had failed? Who would secretly be proud?"
• Watch for: using the lenses as diagnoses of the family, which conflicts with neutrality and the "soft reality" principle. Use them only as hypotheses.

B. Systemic questions

4.7 Circular (triadic) questions

• Purpose / when: throughout the interview. Ask one member about the relationship between two others, or about how another member thinks or feels. People then hear each other's views of themselves, and speculation becomes information. [source: Simon & Rech-Simon 2021, Vorbemerkung fn. 1; Methodenpool Köln]
• Steps:
  1. Pick a hypothesis and a relationship to explore.
  2. Ask a third person to describe it: "What does X think when Y does Z?"
  3. Ask the others whether they agree.
  4. Follow up on differences in the answers.
  5. Move through members in turn, so no one is the permanent informant or target.
• Example language: Simon & Rech-Simon's own examples are of the kind: what does your husband think when you go to judo with your son; when you cry like this, how does your mother-in-law feel. Illustrative: "When Dad goes quiet at dinner, what does your sister think is going on with him?"
• Watch for: answers that are new to other members (signs of surprise, laughter, protest). These show the question produced difference. Also watch for one member always being talked about; rotate.

4.8 Difference questions: ranking, percentage, agreement, subsystem comparison, before/after

• Purpose / when: to turn fixed labels and global complaints into gradations and comparisons. [source: Methodenpool Köln; von Schlippe & Schweitzer 2026, sect. 15.2; Simon & Rech-Simon 2021, ch. 2]
• Steps:
  • Ranking: "Who would be first to... and who last?"
  • Percentages: "What percentage is this illness, and what percentage is it something else?"
  • Agreement: "Do you agree, or do you see it differently?"
  • Subsystem comparison: parents versus children; who is closer to whom.
  • Before/after: "What changed in relationships when the problem began?"
  • Label difference: ask what the difference would be if a label did not apply. Simon & Rech-Simon's example asks what the difference is to not being manic-depressive.
• Example language (illustrative): "On a scale where 100% is 'only illness' and 0% is 'only stubbornness', where would your father put Tom's not going out? And you?"
• Watch for: members adjusting their numbers to each other, which shows coalitions; and relief when a total label ("he is ill") breaks into graded parts.

4.9 Questions about reality constructions and explanations

• Purpose / when: to make visible how each person explains the problem and what follows from that explanation, including illness concepts. [source: Methodenpool Köln; von Schlippe & Schweitzer 2026, sect. 15.3; Simon & Rech-Simon 2021, ch. 2]
• Steps:
  1. Ask each person for their explanation of how the problem arose.
  2. Ask what that explanation implies for what one should do.
  3. Ask what others' explanations are, and who agrees with whom.
  4. When a medical term is used, explore its private meaning instead of correcting it.
  5. Check whose opinion is being voiced. Simon & Rech-Simon ask a husband whether a view is his own or his wife's, using his capacity to take her perspective.
• Example language (illustrative): "When you say 'manic phase', what exactly does Peter do differently? Who notices first?"
• Watch for: rigid "hard" constructions of soft matters ("he will always be like this"). Your questions should introduce time and conditions.

4.10 Invitations to speculate ("Vermuten Sie mal")

• Purpose / when: when a member says "I don't know" about another's thoughts. Behaviour depends on what we believe others think, so guesses are clinically useful information. [source: Simon & Rech-Simon 2021, ch. 2]
• Steps:
  1. Accept the "don't know".
  2. Invite a guess anyway.
  3. Ask the person guessed about whether the guess fits.
• Example language: "Vermuten Sie mal!" ("Take a guess!") [source: Simon & Rech-Simon 2021, ch. 2]. Illustrative: "If you had to guess, what does your son think you worry about most?"
• Watch for: guesses that are hostile attributions. Ask the guessed person to respond, and frame it as "how you see each other", not as facts.

4.11 Possibility constructions: miracle/fairy, improvement and exception, worsening, as-if, time, function of the problem

• Purpose / when: to open alternatives without asking anyone to change yet. People can rehearse change, discover their part in maintaining the problem, and see what the problem does for the system. [source: Methodenpool Köln; von Schlippe & Schweitzer 2026, sects. 15.3–15.4; Simon & Rech-Simon 2021, Vorbemerkung]
• Steps by type (all illustrative wording unless cited):
  • Miracle/fairy: suppose a good fairy took the problem away overnight; what would be different tomorrow, and who would notice first? (A good-fairy question appears in Simon & Rech-Simon's foreword.)
  • Exceptions and resources: "When was the problem absent, or weaker? What were you doing then?" "What do you like about each other?"
  • Worsening: "What could each of you do to make things worse quickly?" This shows that the problem can be influenced, so stopping is possible.
  • As-if / relapse simulation: "How would you have to behave so that the others thought the problem was back?"
  • Hypothetical intention: "Suppose, purely hypothetically, he wanted her to stay at home whatever happened. What would he have to do?"
  • Time: "How long will the problem stay with you? When will you say goodbye to it?"
  • Function: "What would be the advantage for the family if the problem stayed a little longer?"
• Watch for:
  • Worsening and function questions can sound cynical. Frame them explicitly as thought experiments, as the Methodenpool recommends.
  • Answers that name concrete behaviours mean the question worked.
  • In psychosis, keep relapse-simulation questions hypothetical and pair them with early-warning planning (see 4.25). [synthesis]

4.12 Externalising and personifying the symptom

• Purpose / when: when the family already speaks of the symptom as an agent ("the depression came between us"), or to loosen identification with a label. [source: Simon & Rech-Simon 2021, ch. 2; von Schlippe & Schweitzer 2026, sect. 15.5]
• Steps:
  1. Pick up the family's personification.
  2. Ask about the symptom's "behaviour": when does it come, whom does it visit most, who is best at keeping it out.
  3. Ask how each member relates to it.
• Example language (illustrative): "When the depression comes between you, where does it sit, nearer to you or to your husband?"
• Watch for: the family using externalising to deny responsibility for harmful acts. Name the acts separately. [synthesis] von Schlippe & Schweitzer place this with narrative-deconstructive questioning, which is a later addition to the Heidelberg repertoire. [uncertain]

4.13 Reflexive questions about the therapeutic relationship

• Purpose / when: when the therapy itself becomes part of the pattern (stalemate, members who keep missing sessions, idealising or devaluing the therapists). [source: von Schlippe & Schweitzer 2026, sect. 15.6 (heading only)]
• Steps: ask circular questions about the therapy: who expects most and least from it, what each thinks the therapists believe about the family, what would show that coming here makes things worse. [synthesis]
• Example language (illustrative): "Who here thinks we are on Mum's side? What have we done that gave that impression?"
• Watch for: answers that reveal lost neutrality. Correct your stance openly.

C. Action and spatial methods

4.14 Family sculpture

• Purpose / when: when words circle without movement, with children, or to make closeness, distance and hierarchy visible. Heidelberg members Schweitzer and Weber published on the sculpture as a diagnostic, therapeutic and training technique. [source: von Schlippe & Schweitzer 2026, sect. 16.1; uncertain: Schweitzer & Weber 1982, Familiendynamik]
• Steps (standard sequence; detail [synthesis]):
  1. Ask for consent.
  2. Ask one member (the sculptor) to place every member in the room, without talking, so that distance, direction of gaze, height and posture show how relationships are now.
  3. Ask each placed person what they notice in this position.
  4. Ask the sculptor to build the feared future and the wished-for picture.
  5. Let others build their own version.
  6. De-role everyone and discuss.
• Example language (illustrative): "Put everyone where they belong at the moment, closer or further, facing towards or away. No talking yet." "Standing there, what do you see, and what can't you see?"
• Watch for:
  • strong affect or physical discomfort: let people step out at any time;
  • avoid sculpting coercive closeness, and use no touch beyond lightly guiding a position, with consent;
  • a different sculpture by another member is information, not a contradiction.

4.15 Who-is-who or position sculpture

• Purpose / when: a quicker, more symbolic variant: positions instead of whole bodies, or objects on a table. [source: von Schlippe & Schweitzer 2026, sect. 16.2 (heading only)]
• Steps: use coins, figures or chairs to stand for members, and run the same questions as in 4.14. [synthesis]
• Example language (illustrative): "Choose a figure for each person and put them on the board as things are now."
• Watch for: suitable when bodily sculpture would be too exposing (psychosis, trauma, conflictual couples). [synthesis]

4.16 Timeline

• Purpose / when: to place symptoms, life events and transitions in time; it links naturally to the family-life-cycle and separation themes (4.6). [source: von Schlippe & Schweitzer 2026, sect. 16.5 (heading only)]
• Steps (illustrative procedure, [synthesis]): lay a rope or tape on the floor, mark years, and have members step to key events. At each point ask: what changed in relationships here, and who coped best?
• Example language (illustrative): "Stand at the year the problem started. What else happened in the family that year?"
• Watch for: grief surfacing at losses. Slow down and acknowledge it.

4.17 Family constellations (Aufstellungen): caution

• Context: von Schlippe & Schweitzer list family and organisational constellations among the spatial methods. Gunthard Weber was a key early publisher of Bert Hellinger's work. [source: von Schlippe & Schweitzer 2026, sect. 16.3; uncertain: Weber's editing of Hellinger's Zweierlei Glück (1993) is not confirmed]
• Clinical position [synthesis]:
  • Use constellations, if at all, only in the systemic-constructivist form: as hypotheses offered for the client to use or reject, by trained practitioners, with screening and aftercare.
  • Do not use the Hellinger-style form, in which the facilitator declares fixed "orders" of family life and directs clients to bow to perpetrators or accept a fate.
  • Be especially careful in large-group workshops without follow-up.
  • Screen out acute psychosis, active suicidality and recent trauma.
  • Debrief afterwards and arrange a follow-up contact.
• Watch for: clients who leave with authoritative "truths" about hidden entanglements, guilt over a dead relative, or instructions to break off contact. Revisit these as hypotheses.
• Note: no Heidelberg procedure for constellations is drawn on. This entry gives only the scope and the risks.

D. Pause and closing interventions

4.18 The pause (team consultation)

• Purpose / when: every session, after about 1 to 1.5 hours of interview. The team, or the therapist alone, steps out to decide what to advise. [source: Simon & Rech-Simon 2021, ch. 2]
• Steps:
  1. Announce the break at the start.
  2. In the break, review which hypotheses held up and what news of difference emerged.
  3. Choose one closing intervention type (4.19–4.24), and draft its exact wording together.
  4. Decide whether to deliver a single team view or a split view.
  5. Return and deliver it briefly, then end without discussion, or with minimal discussion. [uncertain: the source does not say whether the Heidelberg team allowed discussion after the closing intervention]
• Example language (illustrative): "We'll take fifteen minutes to talk about what we've heard and come back with our thoughts."
• Watch for: pauses that run too long and raise anxiety. Families often treat the closing comment as the core of the session, so prepare it carefully.

4.19 Compliment and appreciative (positive) connotation

• Purpose / when: opening almost every closing intervention. It acknowledges each member's good intentions and contributions, including the symptom's protective side, so that no one is blamed and the next intervention is acceptable. [source: von Schlippe & Schweitzer 2026, sect. 18.1 (heading); Ritscher 2026, p. 15 on the Milan influence]
• Steps:
  1. Name a specific, credible strength or effort of each member.
  2. Connote the problem behaviour as serving a shared value (loyalty, protection, keeping the family together) rather than as pathology.
  3. Include everyone, so no one is the exception.
• Example language (illustrative): "We are impressed by how much each of you invests in keeping this family together: Mum by watching over Lisa, Dad by keeping calm, and Lisa by not leaving her parents alone while they are still finding their way after Grandma's death."
• Watch for: connotations that sound sarcastic, or that endorse violence or abuse. Never connote harm to others positively; connote the intention or the dilemma instead. [synthesis]

4.20 Reframing (Umdeutung)

• Purpose / when: to offer a different, plausible meaning for the problem pattern that opens new options. [source: von Schlippe & Schweitzer 2026, sect. 18.2; Simon & Rech-Simon 2021, ch. 13 heading "Umdeutungen"]
• Steps:
  1. Take the family's own facts.
  2. Choose a frame that fits them at least as well as the old one but implies different behaviour: for example, "stubbornness" as "standing up for independence", or "illness" as "a way of slowing down a too-fast separation".
  3. Offer it tentatively, as the team's idea.
• Example language (illustrative): "We wondered whether Max's refusal to get up is less laziness than a way of telling you he isn't ready to leave yet. But we may be wrong."
• Watch for: relief or surprise and later reports of trying something new, which suggest it worked. Flat rejection: drop it, since it was a hypothesis. A frame should not compete with the medical treatment. [synthesis]

4.21 Prescribing the problem pattern

• Purpose / when: when the family is caught in a pattern that resists direct change, and especially when change advice has repeatedly failed. Prescribing the pattern, often with a reason or a schedule, changes its context and meaning. [source: Simon & Rech-Simon 2021, ch. 14 heading "Verschreibung des problematischen Musters"; Ritscher 2026, p. 15 citing Paradoxon und Gegenparadoxon]
• Steps (general Milan-derived procedure; [uncertain]: the Heidelberg case chapter is not drawn on):
  1. Give positive connotation to the pattern first (4.19).
  2. Advise the family not to change too quickly, or to carry out the pattern deliberately at set times, giving a reason tied to the connotation.
  3. Ask them to observe what happens.
• Example language (illustrative): "Given how much is still uncertain for the family, we advise you not to change anything for now. Please keep arguing about homework, but only on Tuesday and Thursday evenings for half an hour, and notice what it feels like."
• Watch for:
  • Do not prescribe violence, self-harm, suicidal behaviour, substance misuse, starvation or other dangerous behaviour, or anything that endangers a child.
  • Use the prescription respectfully, never as a trick or punishment, and be ready to explain your reasoning if asked.
  • It is working when behaviour changes, or loses its compulsive quality.
  • [synthesis: the safety exclusions are mine]

4.22 Rituals

• Purpose / when: to bring a new pattern into everyday life as a fixed, framed sequence, or to mark a transition or loss. [source: von Schlippe & Schweitzer 2026, sect. 18.6; Simon & Rech-Simon 2021, chs. 15–16 (one case chapter titled "ein Ritual")]
• Steps (general; [uncertain] for Heidelberg specifics):
  1. Define who does what, when, where and for how long.
  2. Give it a frame ("as an experiment").
  3. Ask for observation rather than evaluation.
  4. A Milan-origin example is the odd-days/even-days ritual, in which parents take turns being in charge on alternating days while the other parent only observes. [uncertain: Selvini Palazzoli et al. 1978]
• Example language (illustrative): "On Mondays, Wednesdays and Fridays, Mum decides everything about bedtime and Dad only watches and notes what he sees; on the other days it is the reverse. On Sundays everyone does as they like."
• Watch for: rituals that are too complicated to carry out. Partial or changed performance is also information: ask about it at the next session.

4.23 Split-team message (therapist splitting)

• Purpose / when: when the family is ambivalent about change, or the team genuinely disagrees. Both sides are voiced, so the family does not need to defend either one. [source: Ritscher 2026, p. 16 (named as a Heidelberg technique, not described)]
• Steps (general; [synthesis]):
  1. Report that part of the team thinks one thing (for example, that the family is ready for the daughter to leave) and part thinks another (that it is too early and the symptom is still needed).
  2. Do not resolve the split.
  3. Leave the family to observe which view proves right.
• Example language (illustrative): "Half of our team believes you are ready for Sara to move out. The other half thinks the family still needs her at home, and that her symptoms remind everyone of that. We couldn't agree."
• Watch for: families who try to find out which therapist is "right". Stay neutral about the split.

4.24 Metaphors, jokes and cartoons

• Purpose / when: to convey a reframe or a pattern lightly, in a memorable image. [source: von Schlippe & Schweitzer 2026, sect. 18.4 (heading)]
• Steps [synthesis]:
  1. Draw the image from the family's own language or interests.
  2. Offer it briefly within the closing comment.
• Example language (illustrative): "It sounds as though you are two drivers each with a brake pedal, and whenever one accelerates the other brakes."
• Watch for: humour that feels mocking to the identified patient, and literal thinking in acute psychosis.

E. Heidelberg application areas and relational work

4.25 Systemic work with psychotic and manic-depressive behaviour

• Purpose / when: families in which a member has a schizophrenic, schizoaffective or manic-depressive diagnosis. This was the Heidelberg group's core research field in the 1980s. [source: DGSF Wissensportal, sect. 1; Ritscher 2026, p. 15; Simon & Rech-Simon 2021, ch. 2]
• Core ideas from the sources:
  • Treat "crazy" behaviour as behaviour and communication in a context. Context marking matters, because the same act means different things in different contexts. [source: Simon 1990/2025 sample, ch. 7, pp. 138–139]
  • Some, not all, of these families show incongruent, ambivalent communication. This is a description, not a cause of illness. [source: DGSF Wissensportal, sect. 5.1, citing Simon et al. 1989]
  • Treat illness labels as soft constructions, and avoid chronicity labels. [source: Ritscher 2026, p. 17]
  • Explore the family's private use of diagnostic words through difference questions. [source: Simon & Rech-Simon 2021, ch. 2]
  • Retzer's follow-up studies are reported to show lower medication use after this therapy. [source: Helmes et al. 1999, p. 65, citing Retzer; primary not drawn on]
• Steps (clinical application, [synthesis] built from the above):
  1. Coordinate with the treating psychiatrist. Do not compete over the explanation or the medication (4.3).
  2. Ask what each member means by the diagnosis, and what changes when someone is "ill" versus "well" (who decides, who is excused, who worries).
  3. Map early signs circularly: who notices first that something is changing, what do they do, and what do others do then?
  4. Agree a family plan for early warning signs and crisis contacts with the psychiatric team. Under the Stuttgart protocol, sessions were brought closer together in crises to avoid hospital admission. [source for spacing: Helmes et al. 1999, p. 65]
  5. In closing interventions, connote everyone's caretaking positively. Address ambivalence about separation (4.6) with a split message (4.23) rather than taking a side.
• Example language (illustrative): "When Jan is in what you call a 'manic phase', who is the first to notice? What does that person do, and what does Mum do then?" "What is the difference between Jan being ill and Jan being difficult? Who sees that difference most clearly?"
• Watch for:
  • acute risk (suicidality, aggression, severe disorganisation): switch to standard risk assessment, crisis planning and psychiatric care;
  • families who hear systemic questions as blame for causing the psychosis: say explicitly that the team does not think families cause psychosis;
  • pressure to stop medication: do not endorse it, and refer medication decisions to the prescriber.
  • [synthesis]

4.26 Delegation, fairness and new relationship rules

• Purpose / when: separation crises, symptomatic young adults who "cannot leave", and couples or families locked in a malignant clinch over guilt and merit. [source: Ritscher 2026, pp. 11–13, 21]
• Steps:
  1. Using circular questions, make visible the missions (delegations) each child carries, and whether they are contradictory, impossible or unacknowledged.
  2. Acknowledge the merit of fulfilling them.
  3. Explore whether the young person may refuse or reinterpret a mission. Legitimating refusal is what keeps delegation from derailing.
  4. Where guilt and merit are out of balance across generations, ask about the ledger: who has given, who owes, whose merits were never recognised.
  5. Introduce new relationship rules that secure positive mutuality. Ritscher gives listening and limited speaking time as examples.
  6. Where reconciliation is not possible, work on coping instead.
  • [source for steps 1–5: Ritscher 2026, pp. 11–13; the sequencing is synthesis]
• Example language (illustrative): "If Lena became the doctor her father never could be, who would feel most satisfied? And if she chose something else, who would she be letting down?" "Let's try a rule for the next ten minutes: each of you speaks for two minutes, and the other summarises before answering."
• Watch for: delegations that involve exploitation or abuse. These need protection and reporting, not acknowledgment of merit. [synthesis]

4.27 Reflecting team and reflecting positions (later development)

• Purpose / when: an alternative to the team pause and closing intervention. The team discusses its impressions in front of the family, and the family then comments on what it heard. von Schlippe & Schweitzer present it as part of the current systemic repertoire. [source: von Schlippe & Schweitzer 2026, sects. 19.1, 19.3 (headings); Helmes et al. 1999, pp. 65–67]
• Steps (Stuttgart protocol after Andersen):
  1. Interview.
  2. The team reflects aloud while the family listens. Families could choose whether the team observed from behind the mirror or sat in the room.
  3. The family discusses the reflections with the therapist.
  4. Keep reflections short, since phases over about 10 minutes overloaded families.
  5. Offer several ideas, without insisting on any.
  • [source: Helmes et al. 1999, pp. 65–67]
• Example language (illustrative): "I was struck by how carefully everyone avoided upsetting Dad. I wonder what would happen if... But I may be completely off track."
• Watch for: in the Stuttgart follow-up survey, families rated the interventional (closing-intervention) format as more effective for symptom improvement, especially with a schizophrenic member, and remembered the reflecting format more positively for its atmosphere, especially with affective psychosis. Choose by case. [source: Helmes et al. 1999, pp. 68–71]

4.28 Closing interventions named in the Heidelberg casebook (not drawn on)

• Zirkuläres Fragen has case chapters on two further closing interventions: "Fürsorgliche Belagerung" (caring siege) and "Mein Joghurt, dein Joghurt" (my yoghurt, your yoghurt). [source: Simon & Rech-Simon 2021, contents]
• Their content is not drawn on. They are not described here, to avoid inventing technique. Read chs. 13–16 of the book before using them. [uncertain]

5. Handling problems

The sources address only a few of these problems directly: context and referral conflicts, the risk of seeming partisan, video and confidentiality, and responsibility. Everything else in this section is marked [synthesis].

• Family arrives with no shared mandate, or under outside pressure (court, school, psychiatry).
  • Make the referral context the topic. Ask what the referrer expects, what would satisfy them, and who most and least wants to be here (4.3–4.4).
  • Do not pretend there is a mandate. With "visitors", give compliments only and no tasks.
  • [source: Simon & Rech-Simon 2021, ch. 2; Methodenpool Köln; synthesis for the court and school cases]
• A member experiences the therapist or team as partisan.
  • Notice which words or questions carried a side, for example a term that belongs to one member's context.
  • Ask circularly who sees you as on whose side (4.13), and rotate your questions.
  • Next time, voice the opposite position in the closing comment, or use a split-team message.
  • [source: Simon & Rech-Simon 2021, ch. 2 on the risk of seeming partisan; the remedy is synthesis]
• Rupture with another helper (for example, a psychiatrist who distrusts family therapy).
  • Avoid "either them or us". Ask the family, and ideally the colleague, which complementary functions each helper serves.
  • Inform the colleague, with consent.
  • Never devalue their method in front of the family.
  • [source: Simon & Rech-Simon 2021, ch. 2]
• Stuck therapy: the same pattern every session, no news of difference.
  • Revisit the hypotheses in the pre-session, and generate at least one hypothesis that includes the therapists.
  • Ask for the function of the problem ("what would be lost if it went?") and worsening questions (4.11).
  • Consider prescribing the pattern (4.21) or a split message (4.23).
  • Renegotiate the mandate.
  • Consider lengthening the interval, so the family has time to react. [uncertain: the long-interval rationale]
  • [synthesis]
• The family does not carry out the task or ritual.
  • Do not reproach them. Ask circularly: who thought it would not help, who would have done it, what happened instead.
  • Treat non-performance as information about the pattern.
  • Simplify or drop the task, or reframe the non-performance as a sensible caution about changing too fast.
  • [synthesis]
• Overwhelm or high emotional arousal in session (crying, shouting, dissociation).
  • Slow down and switch from triadic questions to direct contact with the overwhelmed person.
  • Allow time out, and protect anyone who is being attacked.
  • Resume with less exposing methods (4.15 instead of 4.14).
  • Shorten reflections; Helmes et al. found more than about 10 minutes overloaded families.
  • [source: Helmes et al. 1999, p. 66 for reflection length; otherwise synthesis]
• Paranoid fear of the team or the video.
  • Show the observation room. Introduce, and later say goodbye with, the team in person.
  • Offer to stop recording or erase tapes, or to bring the team into the room.
  • [source: Simon & Rech-Simon 2021, ch. 2; Helmes et al. 1999, p. 65]
• Secrets and disclosures in joint sessions.
  • State the rule that each person guards their own secrets.
  • Offer individual contact when disclosures may endanger someone, for example abuse or intimate-partner violence.
  • Child-protection and safeguarding duties override neutrality: report as the law requires, and tell the family how and why unless that would increase risk.
  • [source: Simon & Rech-Simon 2021, ch. 2 for the rule; safeguarding is synthesis]
• Violence, abuse or coercive control discovered.
  • Neutrality does not extend to violence. Do not connote it positively, do not prescribe the pattern, and stop conjoint work if it puts anyone at risk.
  • Move to individual safety planning and the relevant services.
  • [synthesis]
• Suicidality, acute psychosis or an eating disorder with medical risk.
  • Interrupt the systemic format for a standard risk assessment, crisis plan and medical or psychiatric involvement.
  • Bring sessions closer together in a crisis rather than keeping long intervals. [source: Helmes et al. 1999, p. 65 for crisis spacing]
  • Lethal-means counselling with the family: safe storage and removal, as in standard practice.
  • No paradoxical or worsening interventions about the risk behaviour.
  • [synthesis]
• The family hears systemic questions as blame for the illness.
  • Say plainly that the team does not hold the family responsible for causing the disorder, and is interested in how everyone copes and what helps.
  • Use appreciative connotation (4.19).
  • The DGSF author stresses that double-bind ideas are not a causal theory of schizophrenia.
  • [source: DGSF Wissensportal, sect. 5.1; synthesis]
• Paradoxical interventions backfire (the family feels tricked or mocked).
  • Acknowledge it openly and explain the team's thinking in plain terms.
  • Switch to direct, transparent interventions: compliment, reframe, observation tasks.
  • Note for next time that this family needs more transparency.
  • [synthesis]
• Team splitting or conflict behind the mirror.
  • Use the disagreement as a resource (4.23) rather than hiding it.
  • Agree before the session who decides in the break; in the Stuttgart team, the therapist in the room decided.
  • [source: Helmes et al. 1999, p. 65; synthesis]
• Working without a team (most private practice).
  • Keep the structure: written hypotheses beforehand, a self-imposed pause of 5–10 minutes to draft the closing comment, and supervision or intervision in place of the team.
  • The authors note that single therapists can do family therapy.
  • [source: Simon & Rech-Simon 2021, ch. 2; synthesis for details]
• Client is drawn to constellation or "entanglement" explanations from workshops.
  • Explore what the client took away. Re-frame the "truths" as hypotheses.
  • Check for distress, guilt, or ordered cut-offs that came from the event, and offer follow-up (see 4.17).
  • [synthesis]

Sources

Sources

• Ritscher, W. (2026). Über Helm Stierlin (systemagazin PDF)
• Simon, F. B. & Rech-Simon, C. Zirkuläres Fragen. Systemische Therapie in Fallbeispielen: Ein Lernbuch, 14th ed., Carl-Auer 2021 – publisher reading sample
• von Schlippe, A. & Schweitzer, J. Lehrbuch der systemischen Therapie und Beratung I, 4th ed., Vandenhoeck & Ruprecht 2025/26 – publisher sample (contents, prefaces)
• Simon, F. B. Meine Psychose, mein Fahrrad und ich, 15th ed., Carl-Auer 2025 – publisher sample
• Reitz, M. (2014). Helm Stierlin – Zeitzeuge und Pionier der systemischen Therapie, Carl-Auer – publisher sample
• Ritscher, W. Einführung in die systemische Soziale Arbeit mit Familien, 2nd ed., Carl-Auer 2022 – reading sample
• Simsa, R. (2004). Review of Simon & Weber, Vom Navigieren beim Driften, socialnet
• DGSF Wissensportal: Systemische Familientherapie im psychiatrischen Kontext
• Methodenpool Universität zu Köln (ed. K. Reich): Zirkuläres Fragen
• Helmes, D., v. Bebenburg, M. & v. Keyserlingk, L. (1999). Instruieren oder dialogisieren – was wirkt wie? System Familie 12, 64–73 (Stuttgart team, comparison only)
• Heidelberger Institut für Psychotherapie, Ambulanz: Systemische Psychotherapie für Familien und Paare

Further reading (not drawn on) [uncertain where relied on]

• Stierlin, H. (1974). Eltern und Kinder im Prozess der Ablösung. Suhrkamp.
• Stierlin, H. (1978). Delegation und Familie. Suhrkamp.
• Stierlin, H., Rücker-Embden, I., Wetzel, N. & Wirsching, M. (1977). Das erste Familiengespräch. Klett.
• Simon, F. B., Stierlin, H. & Wynne, L. C. (1984/1985). Die Sprache der Familientherapie / The Language of Family Therapy.
• Simon, F. B., Weber, G., Stierlin, H., Retzer, A. & Schmidt, G. (1989). Schizoaffektive Muster: Eine systemische Beschreibung. Familiendynamik 14, 190–213.
• Retzer, A. et al. (1989/1991). Follow-up studies of manic-depressive and schizoaffective psychoses after systemic family therapy (Familiendynamik; Family Process).
• Retzer, A. (1991). Die Behandlung psychotischen Verhaltens. Carl-Auer.
• Weber, G. & Stierlin, H. (1989). In Liebe entzweit: Die Heidelberger Familientherapie der Magersucht. Rowohlt.
• Schweitzer, J. & Weber, G. (1982). Beziehung als Metapher: Die Familienskulptur als diagnostische, therapeutische und Ausbildungstechnik. Familiendynamik 7.
• Selvini Palazzoli, M., Boscolo, L., Cecchin, G. & Prata, G. (1981). Hypothetisieren – Zirkularität – Neutralität. Familiendynamik 6; and (1977) Paradoxon und Gegenparadoxon. Klett-Cotta.
• Simon, F. B. & Weber, G. (2004). Vom Navigieren beim Driften. Carl-Auer (known via review).