Week 5: Whose Apology Is It?

Four things an apology can accomplish for the speaker, and a test for telling them apart that doesn't require reading anyone's mind

22 March 2027

Slides: pending upload

Start with a limit: nothing here claims to know what anyone privately felt when they apologised. “They didn’t really mean it” isn’t a claim this course can check. What can be checked is what an apology’s wording, timing, and cost accomplish — regardless of what was going on internally when it was written.

Two cases.

A public figure issues a carefully worded, apparently contrite statement the day before a scheduled interview that would otherwise have opened with the same topic. Whether the regret is genuine is not observable. What is observable: the timing closes out a news cycle on the speaker’s terms rather than the interviewer’s. That’s a real effect of the statement, sincere or not.

A colleague, asked what happened to a shared project, gives their manager a specific, self-incriminating account when a vaguer one would have protected them just as well with the person they’d let down. The specific version costs them something with their manager that the vague version wouldn’t have. That cost is evidence of what the apology was built to prioritise, independent of what the colleague felt while giving it.

Four things an apology can do for the speaker

  1. Relational repair

    Restores the relationship on the harmed party’s terms.

  2. Self-protection

    Limits damage to the speaker’s own standing.

  3. Reputation management

    Controls how the speaker is perceived by people who weren’t harmed.

  4. Closure-seeking

    Ends the speaker’s discomfort, regardless of whether the harmed party is satisfied.

These aren’t mutually exclusive, and none of them is the same axis as “was this a real apology” from Weeks 1–4. An apology can be component-complete in Week 2’s sense and still be primarily doing reputation management. Having all four components doesn’t guarantee repair is what’s happening.

A test that doesn’t require telepathy

Instead of asking what someone meant, ask: what would have to be true of the wording if this function were the primary one? Reputation management predicts wording vague enough to survive many audiences, timed to news cycles. Self-protection predicts minimal concrete admissions. Closure- seeking predicts language asking for the matter to be considered settled. Repair predicts wording answerable to what the harmed party specifically needed, which usually means it can’t be drafted without knowing what they needed. Check the actual statement against the prediction; that’s evidence, not a guess about interiority.

Week 4’s “sorry you feel that way” fails as repair — nothing is acknowledged as the speaker’s doing. It can still succeed at closure-seeking: the sentence ends the exchange, for the speaker, whether or not the other person is satisfied. Failing on one function doesn’t mean failing on all of them.

When the wording predicts two functions at once

The test isn’t always decisive. A manager who mishandled a promotion decision writes a private note that is specific, names the exact process failure, and offers to redo the process properly — wording that predicts relational repair. It’s also wording that predicts self-protection: specific, procedural language is exactly what a careful person would write if they were mainly managing their own exposure to a formal complaint. Both predictions are satisfied by the same sentence, and nothing in the wording alone breaks the tie.

What can break it is what happens next, which isn’t part of the utterance but is still checkable: does the offered redo actually happen, on a timeline that doesn’t wait for a complaint to force it? A function can be underdetermined by the wording at the moment it’s spoken and still become determinable once there’s a next action to check it against — the test doesn’t fail here, it just needs more evidence than one sentence provides.

All of this has assumed one listener. Once there’s an audience — people who weren’t harmed but are watching — a single sentence can pursue different functions for different addressees at once. That’s Week 6.

Teaching team