Co-supervision without stepping on each other

Co-supervision is now the default in most doctoral programmes, and for good reasons: broader expertise, a second perspective, and some protection for the scholar if one relationship goes badly. It also fails in a specific, predictable way, and the failure lands on the person with the least power to fix it.

The failure is not conflict. Open disagreement between supervisors is usually fine and sometimes valuable — watching two experienced researchers disagree about method is an education. The failure is unresolved disagreement, transmitted to the scholar as two sets of instructions, with no mechanism for deciding between them. The scholar then either picks one and disappoints the other, or tries to satisfy both and does twice the work for a worse result.

Three questions to settle before the scholar starts

Almost every co-supervision problem traces back to one of these going unasked.

Who decides when you disagree? Not who is senior, not who is on paper as primary — who actually resolves it. The answer can be anything you both accept: the primary decides on methodology and the secondary on domain questions; whoever is closer to the specific problem; you always resolve it between yourselves before speaking to the scholar. What matters is that an answer exists before it is needed, because the moment it is needed is the worst possible time to negotiate it.

What is each of you actually responsible for? Vague answers here are the root of the most common co-supervision pathology, which is not conflict but neglect. Both supervisors assume the other is handling pastoral care, progress monitoring, or the difficult conversation about the chapter that is not working. Nobody handles it. This is far more common than two supervisors fighting over the same territory, and much harder to notice, because neither party experiences themselves as having dropped anything.

How often do the two of you talk without the scholar? If the answer is never, you are not co-supervising. You are two independent supervisors sharing a student, which is a materially worse arrangement than either of you supervising alone.

Meet together more often than feels necessary at the start

The instinctive arrangement — the scholar meets each supervisor separately, and all three meet occasionally — puts the scholar in the position of being the only channel between two people who need to be aligned. That is a heavy job to hand someone with no institutional power, and they will do it badly, not through any fault of theirs but because they will unconsciously soften each supervisor's position when relaying it to the other.

Joint meetings should be frequent early, when direction is being set and the risk of divergence is highest, then taper. A reasonable pattern is all three together monthly for the first six months, then each term, then whenever a decision affects both domains.

The cost is real. Three diaries is more than twice as hard as two. But the alternative cost is paid in the scholar's fourth year, when a chapter has to be substantially reworked because one supervisor was never told what the other agreed to.

Give the scholar an explicit escalation route

Tell the scholar, in plain words and early: if we give you contradictory advice, do not try to reconcile it yourself. Tell us, and we will resolve it.

Scholars do not do this by default. Reporting a contradiction feels like telling on someone, or like admitting you cannot handle a normal part of research. So they absorb it — and absorbing it means either doing both, or silently picking the supervisor they find more intimidating, which is not a research judgement at all.

Saying it out loud converts a social problem into a procedural one. It works far better when both supervisors say it, in the same conversation, so the scholar can see it is genuinely shared.

Keep one shared record, not two sets of notes

The most practical fix in co-supervision is also the least interesting: both supervisors should be looking at the same record of what was decided.

Two separate sets of notes guarantee divergence, because each supervisor records what they thought was important. Six months later, one remembers the meeting as having settled the sampling strategy and the other remembers it as having deferred the question. Both are recording honestly. Neither can be checked.

A shared record also fixes the neglect problem. If both supervisors can see that the scholar has an open action from three meetings ago, and that it belongs to neither of them, someone picks it up. If each can only see their own half, the gap is invisible from both sides.

This matters most where the co-supervisor is external, part-time, or based elsewhere — exactly the cases where the shared record is least likely to exist, and where the scholar is most exposed.

When the relationship is the problem

Sometimes co-supervision is not a coordination problem. Sometimes one relationship has broken down, and the arrangement is being used — by the scholar, or by a department — as a way of avoiding saying so.

This is worth naming, because the coordination advice above will not fix it and may prolong it. A scholar who has stopped bringing real problems to one supervisor and takes everything to the other is telling you something. The honest response is to find out what, not to schedule more joint meetings.

The structure of co-supervision genuinely does protect scholars from a single bad relationship. That protection only works if someone is willing to notice when it is being used.