Belonging

Telling apart 'I lack knowledge I can acquire' from 'this environment is hard for people like me' — because they feel identical, have opposite remedies, and one of them is not your problem to fix.

Before you start


Read this	Before you need it. It is much harder to think clearly about this from inside it.
It will not tell you	That you belong. A page cannot establish that, and being told it by a stranger is worth nothing.
It will tell you	How to work out which of two very different problems you actually have, and what each one responds to.

The distinction that does all the work

Two experiences feel identical from the inside:

“I do not know things that everyone here seems to know.”
“I do not belong in this room.”

They produce the same sensation — exposure, the sense that you are about to be found out, the urge to say less. But they are completely different problems with opposite remedies, and treating one as the other is the most expensive mistake in this whole area.

The first is a knowledge gap. It is finite, nameable, and closes with time. Everyone who has ever entered this field has had one. The remedy is to name the specific missing thing and go get it.

The second is a fit or environment problem. It does not close with effort, because effort is not what it is responding to. The remedy is different and sometimes involves leaving.

Almost always, at the start, it is the first one. New entrants systematically misread a knowledge gap as a belonging problem, because the knowledge gap is invisible to them — they cannot see that the person across the table also had it four years ago, and had it quietly.

The test

When the feeling arrives, write down the specific thing you did not know.

You can name it — “I did not know what a materialization version was”, “I did not know you were supposed to read the methods first”, “I did not know postdocs get paid from the grant, not the department” — then it is a knowledge gap. It is on this site or in a colleague’s head, and it will be gone in a week. Go and get it, and notice afterwards that the feeling went with it.
You cannot name anything specific — you just felt wrong in the room — then look harder before concluding anything. Nine times out of ten there is a specific unnamed thing and you have not yet caught it. Ask someone: “What was the thing everyone else in that meeting seemed to already know?” That is a completely reasonable question and most people enjoy answering it.
You can name it, you got it, and the feeling did not move — that is worth taking seriously, and the rest of this page is about that case.

What it actually looks like day to day

Not dramatic. That is why it goes unaddressed.

You have a question and you do not ask it. Then someone else asks it and it turns out to have been a good question.
You do extra preparation for a meeting that others walk into cold — and read your own thoroughness as evidence you are behind, rather than as unpaid labour you are doing to feel safe.
You wait to be certain before speaking, so you speak late or not at all. Meanwhile the norm in the room is to think out loud at 60% confidence.
You attribute your successes to luck or to the problem being easy, and your failures to yourself. Other people’s successes get the reverse treatment.
You do not put your name on things.
You read a rejection or a critical comment as a verdict on whether you should be here, rather than as information about a specific piece of work.

The pattern underneath: you are applying a stricter standard to yourself for entry into a conversation than the people already in it apply to themselves. That is a real, measurable cost — it is time and airtime you are paying and they are not.

The first-generation and career-changer cases specifically

These are different from each other and both differ from the general case.

If nobody in your family did this

What you are missing is not intelligence or preparation. It is exposure to the ordinary unglamorous mechanics — that a PhD is a job with a salary, that professors are mostly writing grant applications, that everybody’s first paper gets rejected, that the confident person in the seminar is often wrong and does not mind. People who grew up around academics absorbed all of this at dinner tables, without effort and without noticing.

The gap is real and it is not about you. It is also fully closable, and closing it is what this whole section is for — start with career mechanics and lab norms, which are the two with the highest concentration of things everybody assumes you know.

One concrete thing that helps more than it should: ask people how they got here. Most senior people’s paths were far messier than their biography implies, and hearing three of those in a row does more for calibration than any amount of reassurance.

If you are arriving from industry or another field

Your problem is usually the opposite one and gets misdiagnosed as belonging.

You are competent, you are used to being competent, and you have landed somewhere your competence does not immediately transfer. That is disorienting in a way that reads as not belonging but is not. What is actually happening is that you have a vocabulary and pace mismatch — a large share of the apparent difficulty is terminology, and it is fixable in a week with the dictionary.

The specific trap: deciding the biology is a detail you can pick up later. It is not, and the field will notice. The other trap: reading academic slowness as incompetence rather than as a different set of incentives.

Building legitimacy through contribution, not through readiness

The instinct is to wait until you feel ready before participating. This is backwards, and the ordering matters more than any amount of encouragement.

Legitimacy is a consequence of contribution, not a prerequisite for it. Nobody feels ready and then contributes. People contribute something small, get treated as a contributor, and the feeling follows some months later.

Practically, in this field, contributions are unusually available at the low end:

Proofread something and be right about it. Proofreading is real, checkable work, and being demonstrably good at it is a form of standing that does not require you to have opinions about anything yet.
Ask the question in the seminar that three other people also had.
Write the documentation nobody wrote. This is genuinely valuable and chronically under-supplied.
Reproduce someone’s analysis and report a discrepancy carefully.
Answer a question you can answer, on a mailing list or a forum, where you are the fifth-most-qualified person present. Do it anyway.

Note that all five are checkable. That is the point: they produce evidence that does not depend on how you feel.

What a mentor can and cannot do

Can: state the norms out loud, name the specific gap, say “everyone finds this hard and here is why”, introduce you to people, put your name on things, tell you their own messy path, and — most usefully — say “the thing you did not know in that meeting was X, and I did not know it either until 2019.”

Cannot: make you feel that you belong. That is not a thing one person can install in another.

If you are mentoring someone in this position, the highest-value move is not reassurance. It is specificity: convert the diffuse feeling into a named, closable gap. “You are doing fine” changes nothing. “The thing you were missing is that nobody actually reads the whole paper, here is how people really read them” changes something that afternoon.

When the answer is that the environment is bad

Not every case is a knowledge gap, and this section exists because most writing on this topic pretends otherwise.

Some environments are genuinely hostile, and some are hostile specifically to people with your background. Persistence is not always correct. Signals that you are in one:

The same treatment is not applied to peers doing the same things.
You are asked to prove basic competence repeatedly, well past the point where it has been demonstrated.
Your contributions get reattributed, more than once, and correcting it is treated as the problem.
People leave, and the ones who leave say similar things on the way out.
The behaviour is known and the response is that you should manage around it.

If several of those hold, the thing to internalise is that staying is not a measure of your worth, and leaving is not a failure of resilience. Grit is a real construct and it is also routinely used to move responsibility for a bad environment onto the person suffering it. This site’s own evidence base hedges that literature deliberately; hold the same line about yourself.

Practical steps, in order: document contemporaneously, find one person outside the situation who can sanity-check your read, identify what you need from the environment to finish what you are doing, and treat exit as one legitimate option among several rather than as the last resort. Conflict covers the mechanics of recourse and what protection it does and does not offer.

Check yourself

You leave a lab meeting feeling out of your depth. What is the first thing to do?

Write down the specific thing you did not know — before the feeling generalises into a conclusion about yourself.

Nine times out of ten there is a nameable item: a term, a convention, a piece of context about the project’s history. Once it is named it is a task, and a task is a different object from a verdict.

If you genuinely cannot name anything, ask someone who was there: “What was everyone else already assuming in that discussion?” This is a normal question, it makes you look engaged rather than lost, and the answer is usually one sentence.

The failure mode to avoid is going home, feeling bad, and drawing a conclusion about fit from a feeling whose actual cause was one undefined acronym.

Someone you mentor says "I don't think I'm cut out for this." What is the least useful response, and what is a better one?

Least useful: “Of course you are, you’re doing great.” It is unfalsifiable, they will not believe it, and it closes the conversation without producing information.

Better: get specific. “What happened this week that made you think that?” Then work out with them whether it is a knowledge gap (closable, common, name it and close it) or something about the environment (in which case reassurance is actively wrong and you should be helping them assess their options honestly).

If it is a gap, the useful move is to name your own version of it: the thing you did not understand for two years, the meeting you left confused. Not as comfort — as calibration data. They are working from a sample in which everyone else appears to have always known everything.

Common errors and how to recover

Reading a specific gap as a general verdict. Recover: name the gap in writing. Notice that it is one item, and that it has a fix.

Waiting to feel ready. Recover: pick one of the five contributions above and do it this week. The feeling is downstream of the act, not upstream.

Doing invisible extra preparation and treating it as evidence you are behind. Recover: notice it is unpaid labour. Sometimes the right response is to do less of it and speak at the confidence level the room actually uses.

Assuming everyone else’s path was linear. Recover: ask three people how they got here. This is the cheapest intervention on this page and among the most effective.

Treating persistence as always correct. Recover: apply the environment checklist above honestly, with one outside opinion.

What this page does not cover

Clinical anxiety and depression, which are real, common in research populations, and not what this page is about — if the description here matches something that persists regardless of context, that is worth taking to someone qualified rather than to a curriculum page. It also does not cover formal recourse for discrimination or harassment, which is in conflict.

Lab norms — the specific unwritten rules whose absence produces most of this feeling
Career mechanics — the machinery that people with academic families absorbed for free
Conflict — when the problem is someone else’s behaviour
Meta-learning — telling improvement from familiarity, which is the other thing that gets misread as inadequacy
Learner personas — Julian and Amir are the two cases this page splits