A field guide that only catalogues danger is a bad field guide. It trains the eye to find one thing, which means the eye finds it everywhere, and after a month of that a reader is no better at telling a warning sign from an ordinary bad evening than when they started.

A healthy relationship is one where the ordinary friction of two lives — a refusal, a mistake, a bad week, a disagreement about money — is handled without anyone paying for it later. That is a description of conduct under pressure, not a mood and not a feeling of certainty. It is also deliberately unglamorous, because the glamorous version is the one that is easy to perform.

This section covers the same scenes as the red half of the site, from the other side.

What is a green flag, exactly?

Green flags are behaviours that hold up under friction and over time — not gestures, and not declarations. The distinction is the whole of the definition.

The internet renders green flags as inventory: flowers, replies within the hour, remembering the coffee order. All of that is cheap and all of it is performable, which is why an inventory tells you almost nothing about a person in week three. What is expensive to fake is conduct when there is nothing to be gained. Being wrong about something in front of other people. Hearing a no that costs them a plan they wanted. Sitting with your bad news when there is nothing to fix and no credit available for fixing it.

For instance, the difference between “he brings me coffee” and “he was wrong about the restaurant booking and said so before I could” is not a difference of degree. The first is a gesture. The second is a sample of how being wrong is handled, taken when nothing was at stake — which is the only time you get an uncontaminated sample.

What does a good argument actually look like?

It ends with both accounts still standing, and something follows it.

Repair is the move back toward each other after friction: the returning to a subject an hour later, the apology without a grievance attached, the joke that is offered and accepted. It is the single most useful thing to watch for, because it is the mechanism that keeps ordinary disagreement from accumulating. A couple who argue and repair are in better shape than a couple who have simply stopped raising things.

Two things are worth naming about a workable argument. The subject stays the subject — it does not widen into a review of your character or a list of previous offences. And losing is survivable for both people: someone can be persuaded without it costing them standing. John Gottman and Nan Silver’s The Seven Principles for Making Marriage Work is the best-known practical book built around that observation. It is a couples’ book, not a diagnostic instrument, and reading it will not tell you what your own arguments mean.

The mirror image of all this is on toxic vs healthy, where the same five scenes are set side by side.

What does a no sound like when it is heard?

Like nothing much, which is the point. It is accepted the first time, it is not revisited later that week, and it does not appear again as evidence in an unrelated argument.

Consider the smallest possible version: you would rather stay in tonight. Heard and kept looks like mild disappointment and a different plan. It may include a genuine wobble — that is human, and a partner who never minds anything is not a higher grade of person. What it does not include is renegotiation, a chill, agreement followed by the same suggestion on Thursday, or the request resurfacing a month later as proof of something about you.

This is also where the two halves of the site meet most directly. The green flags page describes the accepted version scene by scene, and it is worth reading beside whichever red entry brought you here, because the comparison is what makes either one legible.

Is any of this backed by research?

Some of it is, in the same narrow way as everything else on this site: there is a body of work describing how couples behave, and there are no numbers here taken from it.

The clearest example is the therapeutic literature on secure connection — on what it looks like when one partner reaches for the other and is answered. Sue Johnson’s Hold Me Tight is the popular-audience account of that work; Johnson developed emotionally focused therapy and writes for couples rather than for clinicians. Imagine the pattern it describes at its simplest: a signal of distress, met rather than managed, often enough that sending one stops feeling like a risk.

Stan Tatkin’s Wired for Love covers similar ground from the day-to-day angle — how two people handle each other’s alarm systems in ordinary weeks. Both are books, not evidence about anyone’s relationship, and neither can tell you what yours is.

What this section will not do

It will not score your relationship. There is no checklist here that a relationship must pass, no threshold, and no count that means something. An instrument like that, applied by a website to a relationship it has never seen, would be wrong often enough to do real damage.

It also will not promise that a baseline protects you. Green behaviour can be performed for a while, and time is the only test nobody can rehearse for. What a baseline does is smaller and genuinely useful. It gives you something to compare against, so that unease has a shape and a sentence you can say out loud — rather than staying a feeling you cannot argue with at two in the morning.

If reading this half mostly produces a sinking feeling — that very little of it describes your relationship — that is worth taking seriously. It is also worth taking to someone who can talk it through with you, rather than to another page.