The AI Journal

Open a historian’s edition of almost any diary and you meet the same sentence of intent: this is where I can finally be honest. Samuel Pepys kept his in shorthand partly to keep his household out of it, and still wrote as though the truth were the point. The form carries a promise of candour the way a courtroom carries a promise of truth, and for much the same reason: the reader assumes that nobody was performing.

That assumption is what needs examining. Honesty is not a property of a text; it is a relation between a speaker and a listener who could have been told something else, and who could refuse the version on offer. A sentence is honest when it takes the risk of a listener’s judgement. The diary removes the listener who could judge. Its only reader is the diarist’s later self, and the later self is not a judge but a continuation: it inherits the reasons for what was written, shares the writer’s stake in the account, and cannot walk away from it or refuse it.

Linguists have known for decades that speakers design what they say for whoever is listening, down to the smallest choices of form. Allan Bell’s account of audience design showed that a speaker’s style shifts measurably when the audience shifts, including when the audience is only imagined (Bell, 1984). The diary is that finding carried to its limit. The diarist holds a real audience in mind, the self of tomorrow or the “dear journal” that diarists so often address as a you (Lejeune, 2009), and writes for it. But it is an audience the writer selects, edits, and can dispose of at will. It is the one listener that cannot answer back that this is not what happened.

This is why sincerity in a diary is so easily mistaken for honesty. Lionel Trilling drew the distinction half a century ago: sincerity is a virtue of the relation between a person and others, the effort not to appear to be what one is not, while authenticity turns inward and asks only whether one is true to oneself (Trilling, 1972). The diary lives on the second term and borrows the credit of the first. It asks the writer to be true to a self, and asks nobody to accept the performance. What looks like the absence of an audience is really the presence of an audience that has been disarmed.

The objection writes itself. Diarists are not fools, and many know exactly what they are doing: some keep diaries precisely to break a habit of self-flattery, and the form has produced documents of extraordinary self-accusation, from Pepys’s recorded cruelties to a diarist’s near-clinical inventories of his own cowardice. If a diary can hold a brutal sentence, what more could honesty require? The answer narrows the claim rather than abandoning it. A diary can record a true sentence; it cannot test one. A test requires the possibility of failing, and the page grants no such possibility. The confession that costs its author nothing and the confession that costs everything are, from inside the diary, indistinguishable, and nothing in the form can tell them apart.

Nor does this dismiss what the diary does well. Expressive writing about emotional experience reliably improves measurable outcomes, from mood to the frequency of medical visits (Pennebaker, 1997). But those gains are counted in health and feeling, not in accuracy; the writing helps whether or not it is fair to what happened. A form can be therapeutic and still be incapable of the thing its readers praise it for.

So the diary’s candour is not a settled fact about the genre. It is something readers project onto a text that never had to survive a listener. The most honest page most people keep is honest in the way any answer is honest when the person answering has also written the question.