Module 3 "Deepening" — Lesson 17 · Reading time approx. 12 min

A premature hypothesis feels just as coherent in the moment as a sound one. Coherence alone proves nothing — seven criteria that make this testable in the middle of a session.

Building On — What This Lesson Won't Repeat

Module 2 already established what separates a hypothesis from a diagnosis. That lesson focused on the speech act itself: holding several readings in parallel, not committing too early, letting clients offer their own interpretations. It also introduced Cecchin's core stance — "exploratory, not conclusive." This lesson assumes all of that. It won't repeat it.

Module 4, Lesson 17, "Reliability and Self-Critique of the Genogram Method," asks a different question: whether the genogram as a tool reliably depicts what it claims to depict at all. That methodological critique of the instrument itself is not the topic here.

This lesson's question is narrower and more practical: how do you tell, in the middle of a session, whether a hypothesis you've just formed actually holds up — or merely sounds good? That calls for criteria you can apply in seconds — not criteria for the quiet of your desk after the session, but ones you can use while the genogram is still on the table and the family is waiting for a response.

Why Coherence Is Not a Criterion

Putting a good hypothesis into words often brings a sense of relief. At last, the picture makes sense. The symbols on the genogram come together into a story, and the session gains a direction.

That very feeling can be deceptive. A premature hypothesis produces the same sense of relief as a sound one. The difference doesn't show up at the moment of forming it. It only shows up once you hold the hypothesis against criteria that don't depend on how it feels.

One reason lies in the genogram itself. It reveals patterns very quickly: identical symbols across several generations, similar intervals between events, recurring line patterns. The eye finds order before the mind has checked it. This speed is one of the genogram's strengths — and, at the same time, its trap. A pattern that looks immediately convincing is not yet a tested explanation.

Selvini Palazzoli, Boscolo, Cecchin, and Prata (1980) made hypothesizing itself a core methodological principle of the Milan School. For them, a hypothesis isn't a random idea but a tool with a demanding job: it must account for the entire system, not just one symptom or one person. That demand can be translated into concrete test questions. The seven that follow are a proposal, not a closed catalog. Each can be applied on its own. Still, it's usually worth checking all seven, because weaknesses rarely occur in isolation — a linear hypothesis, for instance, is often also hard to falsify, because it asserts a trait rather than a process.

Criterion 1: Circular, Not Linear

A linear hypothesis looks for a cause. It often names a person to blame, too. A circular hypothesis instead describes an interplay: what A does, how B responds, and how A responds in turn.

The difference often shows up in the grammar. "The mother is overprotective of the child" is linear. "The more anxious the mother appears, the more the child withdraws — and the more the child withdraws, the more anxious the mother becomes" is circular. Both can rest on the very same genogram data. Only one of them lets you keep thinking with the system.

On the genogram itself, this often shows up in where your gaze fixes. A linear hypothesis fastens onto a single person or generation and explains everything else from there. A circular hypothesis moves back and forth between at least two positions — and can therefore be grounded in at least two places in the picture, not just one.

Criterion 2: Falsifiable

A sound hypothesis can name its own downfall: which observation would disprove it? Anyone who can't answer that question has probably not formulated a hypothesis, but a conviction.

The same principle applies here as to any other test: a result that would look identical whether it failed or succeeded tests nothing at all. A hypothesis that fits any conceivable response from the family ultimately explains nothing. It can only be confirmed, never seriously challenged.

A simple test helps in the session itself: briefly ask yourself which answer from the family would overturn your own hypothesis. If nothing comes to mind, that's a warning sign — not evidence of a particularly robust hypothesis.

Criterion 3: Open to Continuation

Can the family keep thinking with the hypothesis? Or is all that's left to them agreement or objection? A hypothesis open to continuation creates space: "Could it be that...?" One that isn't sounds like a verdict.

Cecchin (1987) later sharpened this point. He reframed neutrality as curiosity — as "an invitation to curiosity." A hypothesis that only allows for yes or no has already lost that curiosity. It has arrived at its destination before the conversation got there.

Criterion 4: Non-Pathologizing, Open to Resources

A good hypothesis doesn't only describe what a pattern prevents. It also asks what it has accomplished, or still accomplishes. Withdrawal can protect. Conflict can force a closeness that would otherwise be silence.

Without that second look, a deficit narrative quickly takes hold. The family becomes the carrier of a problem, rather than people who once found a solution that made sense.

A simple follow-up question sharpens this criterion: when did this pattern last help rather than harm? If nothing on the genogram answers that, it's worth a second look at the wording. Often, swapping out a single verb is enough — "refuses closeness" becomes "protects against too much closeness."

Criterion 5: Grounded in the Material

A sound hypothesis can be anchored in concrete genogram data: dates, intervals between events, repetitions across generations. A weak hypothesis, by contrast, hangs mainly on a theory the practitioner brought into the room.

The test question is: exactly where in the picture do I see this? If it can only be asserted, not shown, caution is warranted.

This applies especially to theories a practitioner brings from their training. A theory is valuable for knowing what to look for in the first place. But it doesn't substitute for evidence in the concrete genogram. Two practitioners from different theoretical schools should be able to point to the same data points on the same picture, even if they interpret them differently.

Criterion 6: Situated in Time

A hypothesis should hold for a specific phase, not for some supposed essence of the family. "In the year after the father's death, the eldest daughter took on parental responsibilities" is situated in time. "This family just works this way" claims something timeless — and is therefore almost impossible to test.

Situating a hypothesis in time has a practical advantage: it automatically makes the hypothesis falsifiable, since it can be checked again at a later point. If the situation changes but the pattern doesn't, the time frame was drawn incorrectly — and the hypothesis needs adjusting, not the observation.

Criterion 7: Neutral Toward All Members

Does the hypothesis quietly favor one person's viewpoint? Often that's the viewpoint of whoever sat down first or spoke loudest. A neutral hypothesis can be followed from the position of every family member — not just one.

A simple perspective shift reveals this: say the hypothesis out loud from the position of the quietest person in the room. If it still sounds plausible from there, it's probably neutral. If it suddenly sounds accusatory or out of place, it was only ever thought through from one direction.

A Case to Test Along the Way: The Attachment Anxiety That Isn't

Hypothetical case example: A counselor is working with Ms. Bergmann, 34, who has spent months thinking about a recurring pattern of breakups. The genogram shows something striking: grandmother, mother, and client each ended a relationship after roughly seven years. The counselor formulates a hypothesis: "In this family, the women carry an attachment anxiety passed down across three generations." The sentence sounds coherent, and the data visibly line up. Yet the hypothesis names a trait the women supposedly "have" — not a pattern that arises between people. Any future breakup would confirm it, and any stable relationship could be explained away as "not there yet."

This hypothesis fails at least three criteria at once. It's linear, not circular: it describes a trait within the women, not an interplay between people. It's barely falsifiable: almost any breakup could be read after the fact as confirmation. And it pathologizes: "attachment anxiety" names a deficit without asking what a breakup after seven years might also have made possible in this family — independence, say, or a deliberate no to a relationship that no longer worked.

A sounder version might read: "Perhaps, in this family, year seven marks the point where a partnership needs renegotiating — and across three generations, that renegotiation has proven difficult." This is more tightly framed in time, leaves room for exceptions, and invites joint examination rather than pronouncing a verdict.

Common Failure Patterns

Some hypotheses don't fail on a single criterion but on a recurring pattern in the practitioner's own thinking.

The theory goes looking for its own confirmation. Anyone who enters a session with a preferred explanation already in mind — a favorite theory from their training, say — tends to read the genogram selectively. Fitting details stand out; unfitting ones get overlooked.

A cluster of symbols gets turned into an invented pattern. Three divorces in one family stand out. But they're not yet proof of a "divorce pattern" if the gaps in the data allow other explanations just as easily — missing information about further generations, for instance, that could make the picture look entirely different.

The first plausible explanation becomes the only one. Once a hypothesis feels relieving, the motivation to look for a second one often drops. This is exactly where falling back on Module 2 helps: deliberately holding several readings in parallel.

The hypothesis unknowingly adopts one family member's language. When one person dominates the genogram conversation, the practitioner easily picks up their choice of words — "the troublemaker," say, or "the sensible one." These labels then slip unnoticed into the practitioner's own hypothesis, violating the neutrality criterion without anyone noticing.

Checklist for the Session

These seven questions can be run through in a few seconds, while the hypothesis is still fresh in mind:

• Does it describe an interplay — or does it look for a cause?
• Which observation would disprove it?
• Can the family keep thinking with it — or only agree or object?
• Does it also name what the pattern accomplishes or has accomplished?
• Exactly where in the genogram does this show up?
• Does it hold for a phase — or does it claim a permanent trait of the family?
• From whose point of view is it formulated — and does it still hold from another?

This list also works well for peer consultation. Two practitioners who independently hold the same hypothesis up against the seven questions rarely arrive at exactly the same result. That very discrepancy is informative: it shows which criterion the hypothesis is weakest on.

When a Hypothesis Fails the Test: Discarding It Isn't Failure

A hypothesis that fails on one criterion wasn't worthless. It sharpened the view of the genogram and proposed a first sense of order. Cecchin's curiosity principle calls for exactly this: hypotheses come and go, and discarding one isn't a defeat.

In practice, that means naming a failed hypothesis openly, if necessary to the family as well. "At first I thought of X — now I see that doesn't capture the whole picture" often builds more trust than clinging to a hypothesis just because it's already been said out loud. A practitioner who lets go of a hypothesis quickly shows the family that their own view counts for more than a thought the practitioner once had.

The order matters here. Test first, then say it out loud — not the other way around. The seven criteria are meant as a tool for your own thinking, not as a lecture to deliver to the family. What ultimately gets said out loud is usually only the part of a hypothesis that passed the test — carefully worded, as a question, not as a finding.

A discarded hypothesis doesn't diminish its longer-term value. Every hypothesis that gets tested and then discarded meaningfully narrows the space of remaining possibilities. That's exactly the kind of progress a single session can achieve — not the one correct explanation, but a smaller, better-grounded set of possible explanations.

Cross-References

• Module 2, Lesson "Developing and Testing Hypotheses" — fundamentals: hypothesis vs. diagnosis, holding several readings in parallel, Cecchin's stance of "exploratory, not conclusive." Assumed as known here.
• Module 4, Lesson 17, "Reliability and Self-Critique of the Genogram Method" — a methodological critique of the genogram as a tool itself, not of any individual hypothesis.

Sources

• Selvini Palazzoli, M., Boscolo, L., Cecchin, G., Prata, G. (1980): „Hypothesizing — Circularity — Neutrality: Three Guidelines for the Conductor of the Session." Family Process 19(1), S. 3–12.
• Cecchin, G. (1987): „Hypothesizing, Circularity, and Neutrality Revisited: An Invitation to Curiosity." Family Process 26(4), S. 405–413.
• McGoldrick, M., Gerson, R. (1985): Genograms in Family Assessment, Norton.