You keep score. They don't know it's a game.

The receipt you never showed them

You know the exact number of times you drove to the airport. You could name the last three birthdays you planned alone. Somewhere behind your ribs there's a spreadsheet, and it is meticulous. The other person has never seen a single row of it.

That's the strange part. You've been keeping the books on a relationship the other party doesn't know is being audited.

The dilemma

What follows is a composite — stitched from the kind of case we see often in this category, not a real person.

Call her Devika. Six years with Sameer. On paper, fine — no cheating, no shouting, no single crime you could point to in court. But Devika keeps a count. When Sameer forgot to call her back the week her project shipped, that went in. When she nursed his flu for four days and he took a half-day off when hers came, that went in. The dinners she books, the family visits she remembers, the apology she gave last March that he never returned — all filed, all dated, all quotable.

She's never said any of this out loud. When Sameer asks "are we good?", she says "yeah, we're good," and adds the question itself to the ledger, because he asks instead of noticing.

Now she's deciding whether to leave. And here's the detail she doesn't say to friends: the ledger isn't helping her decide. It's already decided. She's not weighing him — she's collecting enough entries to make the verdict feel earned. The tally isn't measuring the relationship. It's building the case for the sentence she's quietly wanted to hand down for a year. She just wants the debt to get large enough that walking looks like justice instead of a choice.

Is this a boundary? Or a punishment with a longer fuse?

The read

We score every dilemma on eight fixed lenses. Four of them bite hard here.

Honesty — what did she actually tell him?

Honesty asks whether your words match your inner state. Devika's do not. Every "we're good" is technically a sentence and functionally a lie — she is not good, and she knows the exact reasons. The ledger only exists in the dark. Sameer is being tried in a court he was never summoned to, on evidence he was never shown. Score: -3. Not because she's cruel, but because a verdict reached in secret can't call itself fair. He never got to enter a defense, or, more importantly, to change.

Non-harm — who does the counting hurt?

Non-harm looks at damage done, including the slow kind. Scorekeeping harms in two directions at once. It denies Sameer the one thing that could fix any of it — knowing the relationship is broken. And it quietly reshapes Devika, who now experiences her own relationship as an accounts-receivable department, logging every warm moment against the balance instead of feeling it. Score: -2. The tally protects her from disappointment by converting affection into evidence before she can enjoy it. That isn't safety. It's harm she's absorbing to avoid a harder conversation.

Motive — what is the ledger actually for?

Motive asks what engine is really running underneath the stated reason. Devika would say the ledger is about fairness — keeping honest track of an uneven load. But she admitted the tell herself: it's already decided. A real fairness tool gets shown to the other party and used to negotiate. This one is used to justify. Its true job isn't to measure the relationship — it's to manufacture permission to leave without feeling like the one who left. Score: -3. The point 16.13 keeps making is that a mind running mostly on inventory — what I'm owed, what's still outstanding, what's coming to me — has stopped seeing the person in front of it. Devika isn't reading Sameer anymore. She's reading a balance sheet with his name on it.

Discernment — is this a boundary or a sentence?

Discernment is the capacity to tell one thing from a thing that only looks like it. And here's the tension that makes this hard: a boundary and a stored-up verdict wear the same clothes. Both involve noticing a pattern. Both can end a relationship. The difference is direction. A boundary faces the other person — it's spoken, it names the imbalance, it asks for change and watches what happens. A verdict faces away — it collects quietly and delivers the ruling only once it's too late to appeal. Score: -2. Devika has a real grievance. An uneven six years is a legitimate thing to leave over. But she's chosen the form that never risks the conversation — which means she's protecting herself from an outcome she might not like: him actually trying.

The verdict: The relationship may be worth leaving — but the ledger is not a reckoning, it's a sentence handed down without a trial, and it has to be spoken aloud before it can be trusted.

Why this isn't a chatbot nodding along

Lean into a general AI chatbot with "he never notices, he never reciprocates, should I leave?" and it will hand you a warm, well-organized yes — sycophancy is a documented behavior of these models, and your framing is the prompt it's built to flatter. Tilt the same paragraph toward "but I love him" and watch it pivot to a gentle stay. It argues whichever way you already lean. KarmaLens landed honesty at -3 and motive at -3 not because that's the verdict Devika wanted, but because the score aggregates the same eight lenses in the same order every time, and the number doesn't get renegotiated when the reader flinches at it. The call above — grievance possibly real, secret ledger not — is one committed position, and it cites 16.13 verbatim instead of paraphrasing scripture into whatever the mood wanted. You can disagree with it. It won't quietly agree back.

The takeaway: the read-back test

You don't need us to run this one tonight. Take your ledger — the actual list, the one you can recite — and try to read it back to the person it's about. Not as an ultimatum. As information: "here are the five things I've been counting, and I've never told you."

Now notice what happens in your body before you even open your mouth.

If the thought makes you want the conversation — if part of you is relieved to finally put the list on the table and see what he does with it — you're holding a boundary. It faces him. It leaves room for repair, or for a clean, spoken ending.

If the thought makes you want to skip straight to the exit — if reading it aloud feels like it would ruin the case, because he might actually respond — you're holding a verdict. The secrecy isn't an accident. It's load-bearing. The ledger stays hidden precisely because daylight would give him a chance to change the number, and some part of you needs the number to stay high.

Either way, the list stops being evidence the moment it's spoken. It becomes an ask, or an ending — but an honest one. Browse how other committed calls land, or bring your own to a full relationships read.

Here's the question the ledger is built to avoid: has this person ever been told the score — or only sentenced by it?