High-Conflict or Coercive Control? Telling Them Apart
Sep 06, 2026“High conflict” is the most useful phrase in family law and the most dangerous one, and it is both for the same reason: it describes what a file feels like to work on rather than what is happening inside it.
A file that generates motions, missed exchanges, angry emails and a great deal of professional time feels high conflict from every seat around it. Two people who cannot be in a hallway together look identical whether they are two people who cannot stop fighting or one person and the person managing them. The phrase covers both, and once it is applied it tends to stay applied — it travels between professionals in a way that findings never do.
Telling the two apart is most of my work. Below is how the distinction is drawn, why it is not a matter of severity, and what in a record actually separates them.
Why it is not one thing
The research that shaped this area is Joan Kelly and Michael Johnson's work on differentiating types of intimate partner violence, and its central claim is simple: violence and conflict between partners are not a single phenomenon with a volume knob. They are different situations with different causes, different trajectories after separation, and different consequences for children.
Four broad types are usually distinguished. The names vary; the shapes do not.
| Type | What is going on | After separation |
|---|---|---|
| Coercive controlling | One person establishes and maintains control over the other. Violence, where it appears at all, is one instrument among many. | Often continues and escalates; the arena changes rather than the conduct |
| Situational | Conflict that escalates badly in specific moments. No overarching project of control; frequently mutual. | Usually subsides once the shared household ends |
| Separation-instigated | Uncharacteristic conduct in a person with no prior history, triggered by the separation itself. | Typically time-limited |
| Violent resistance | Force used by the controlled person against the controlling one. | The type most often misread, with the most serious consequences |
The reason this matters in a courtroom rather than a journal is that the remedies point in opposite directions. Situational conflict responds to the things family justice is built to offer: mediation, a parenting coordinator, a detailed schedule that removes the moments where it flares. Applied to coercive control, those same measures hand the controlling person a structured, funded channel to keep going through — and a requirement that the other person keep turning up to it.
The difference is not severity
This is the part that most often goes wrong, and it goes wrong in an understandable direction. The instinct is to sort these files by how bad the incidents are: serious injuries here, shouting there. It does not work, in both directions.
A situational file can contain a genuinely frightening assault. A coercive control file can contain no physical violence whatsoever — and some of the most controlled people I have read about in a record were never touched.
What separates them is not the conduct. It is asymmetry: whether one person's behaviour organises the other person's life.
That is a question with an observable answer. Did she change her route home, her hours, who she saw, what she posted, which account the money sat in, what she said in front of him? Did anything he did have to be planned around? Coercive control produces a person who is continually managing someone else's reaction. Situational conflict, however ugly, does not — both people go on making their own decisions, badly and loudly.
Put the same question the other way and it is a useful test: if you removed the incidents from the file entirely, would there still be evidence of one life being narrowed around another? If yes, incidents were never the point. If no, what you have is conflict.
What I look for in a record
- Narrowing over time. Contact lists, calendars, employment records and bank statements that get smaller and more restricted across a period. Nothing dramatic happens on any single day; the eighteen-month view is the finding.
- Who is accounting to whom. Explanations, permissions and check-ins running in one direction. Conflict is two people arguing; control is one person reporting.
- Conduct that survives the argument ending. Situational conflict needs a trigger and stops when the trigger does. Control continues in ordinary weeks, when nothing has happened.
- Whether the pattern tracks the litigation. Filings, allegations, motions and disclosure demands clustering around the other person's steps forward — a job, a lawyer, a new address — is not the same as a contentious case being contentious.
- Fear that is specific. Not distress about the separation, which everyone has, but anticipation of a particular person's particular response, and decisions visibly made to avoid it.
- Symmetry, honestly assessed. I look as hard for evidence that both people were doing this as for evidence that one was. Where I find it, that is the answer.
The two misreadings, and what each one costs
Both are real. I have to be willing to land on either.
Reading coercive control as high conflict takes a situation with a subordinated person in it and reframes it as a dispute between equals. Everything that follows is then built on symmetry: shared decision-making, mandatory communication, a requirement to co-operate. The person doing the controlling is handed a schedule of compelled contact. And the person who keeps raising it stops looking protective and starts looking difficult — the protective-parent problem, where the very act of naming the pattern becomes evidence of being the unreasonable one.
Reading high conflict as coercive control is not the lesser error, and I want to be plain about that. It attaches a serious finding to a person it does not fit, on a record that will not hold it. It also does something quieter and worse: every over-claimed case makes the next real one harder to believe, because it teaches the people who assess these files that the concept is elastic. I have no interest in a version of this work that stretches.
Violent resistance sits between the two and is misread most often of all. A person who has been controlled for years and finally pushes, throws or hits produces exactly the kind of discrete, documentable incident the system is good at recording — while the years that preceded it produced almost nothing. Read incident-first, the file inverts. That is a real risk in any record where one person's conduct is well documented and the other's is not, and it is one of the specific things chronology is for.
Why the phrase persists
Not because anyone is careless. “High conflict” is genuinely useful as an operational description — it tells a scheduler and a service what a file will demand of them, and it does that without accusing anyone. It is also neutral, which in a system where findings have not yet been made is a virtue rather than a failing.
The difficulty is only that it gets treated as an assessment when it is a description of workload. It arrives before anyone has looked, it is repeated between professionals, and by the time anyone examines the underlying record the file has carried the label for a year. Nobody decided that. It is just what happens to a convenient word.
What I can say, and what I can't
I am a clinician, not a lawyer. Whether family violence is made out, and what should follow for parenting, are questions for the court, and an expert who drifts toward answering them is doing damage — usually to the party who retained them.
What I can do is set out what these patterns are and how they differ, apply a stated method to the documentary record, and say what that record supports. Sometimes the honest answer is that it supports a finding of conflict and not of control. Sometimes it is that the record is too thin to support either, which is its own useful answer and one I would rather give early.
Because the same method has to work whoever is asking, I take instructions from either party. A person facing an allegation of coercive control that the record does not bear out needs a careful reading every bit as much as a person whose pattern is real and has never been named.
Said plainly
These categories are research tools, not diagnoses, and real files are less tidy than any table. Some contain more than one pattern, in sequence or at once. Reasonable people reading the same record can differ on where a case sits, which is exactly why the reasoning has to be shown rather than asserted.
Anyone who can tell you which type a file is from a summary is guessing. That includes anything you read here.
If this is live for you
- Expert reports and case consultation for family, civil and criminal proceedings, in Ontario and across Canada — for counsel on either side, or for someone representing themselves.
- Free resources — six questions to run on a record before commissioning an expert, and a briefing for clinicians whose notes may go to court.
- Two online courses on recognising coercive control and on proving it; live training for organisations on request.
Elizabeth Joy Dubois, MSW, RSW. General information about a changing area of law and how it meets clinical work — not advice on your own matter.
What I've relied on here: Kelly and Johnson, “Differentiation Among Types of Intimate Partner Violence: Research Update and Implications for Interventions” (Family Court Review, 2008) and the differentiation literature that followed it; the Divorce Act definition of family violence, which turns on a pattern of coercive and controlling behaviour; Ahluwalia v. Ahluwalia, 2026 SCC 16; and Ontario's Family Law Rules on expert evidence together with White Burgess (2015 SCC 23) on an expert's duty to the court. Current to September 2026.