A writer drafts a scene with dialogue. The dialogue reads off. The reader can sense something is wrong before she can name what. The lines are technically grammatical. The lines do not violate any obvious rule. The lines simply do not land. The reader moves on, slightly less trusting of the writer than she was before.
Tucker has a one-line diagnosis. Bad dialogue is usually someone trying to make a character sound a certain way. I almost never see bad dialogue in memoir unless they're lying to themselves.1
The diagnosis is specific. Bad dialogue is not a craft failure. Bad dialogue is an honesty failure. The writer who is honestly transcribing what she remembers people saying produces dialogue that lands. The writer who is constructing dialogue to fit the character she wants the person to be produces dialogue that does not land. The reader detects the construction. The construction is the bad-dialogue signature.
The frame has three parts.
First: dialogue is not a craft module. Most writing-training treats dialogue as a craft skill — the writer learns to make characters sound different, to use dialogue tags efficiently, to embed exposition without exposition-dumping. Tucker is making a different claim. In memoir, dialogue is not a constructed object. Dialogue is what the writer remembers people saying. The remembering is the operation. The craft is just the recording.
Second: the falsifying motive is making the character sound a certain way. The writer who has decided what a character means in the story tries to make the dialogue match the meaning. The decision-to-fit is the falsification. The character is no longer producing what the character actually said; the character is producing what the meaning requires. The reader detects the production.
Third: the discipline is render what you remember, even if it doesn't fit. The remembered line may be banal. The remembered line may contradict the character-portrait you were building. The remembered line is what was said. The discipline is to put what was said on the page. In memoir, if someone in my group, the people I work with, if they really are dedicated to truth, if they're writing dialogue, it's just because they remember what the person said and it's what they said.1
Perell raises a wrinkle. If you don't remember the exact words but you definitely remember the conversation, the tenor of the conversation, how do you write a conversation out of that?1
Tucker accepts the wrinkle. I do that for sure. Like, I for sure have stories that I wrote where I wrote out a lot of the dialogue, and I'd be a liar to say I have a photographic memory of every single thing everyone said. But I wrote the dialogue the way I remember it.1
The discipline holds with a soft edge. The writer rendering half-remembered dialogue is still rendering her memory — what the tenor was, what the gist was, what each party seemed to be doing. The rendering is to the writer's actual recollection, not to a constructed character-portrait. The other party may remember the dialogue differently — that's not at all how I remember it, as Doug said. The two memories coexist. Neither is lying.
The boundary is internal. The writer knows the difference between I am rendering the dialogue as I remember it and I am constructing dialogue to fit my agenda for the character. The first is the discipline. The second is the failure.
The page sits next to Three Truths: Yours, Mine, What Actually Happened (Tucker). The three-truths frame is the broader epistemic position. This page is the dialogue-specific application. The writer renders her remembered dialogue (first truth). The other party may dispute (their first truth). Neither party is contradicting the third truth, which is unrecoverable.
It connects to Rule One: Write True Sentences (Tucker). The rule-one discipline is the universal principle. This page is the dialogue-scale instance. Dialogue is one of the places the rule-one discipline either holds or fails. The reader detects the failure faster in dialogue than in many other places because dialogue is the most-surveilled prose surface in fiction-reader training. Readers know what real speech sounds like. Constructed speech registers immediately.
It sits in productive relation to Fingerspitzengefühl: Two Paragraphs to Truth (Tucker). The two-paragraph diagnostic detects honesty in small samples. Dialogue is one of the prose surfaces the diagnostic most reliably reads. The developed reader hears the constructed dialogue in seconds. The two pages combined explain why the dialogue tell is so reliable as a diagnostic — dialogue is where construction-pressure on the writer is highest, so the falsification is most likely to surface there.
You're writing a dialogue scene. Read the dialogue back. Ask: do I actually remember this person saying this, or am I making them say this because the chapter needs them to? The question is the diagnostic.
If you remember the line, keep it. The line may sound flat. The line may contradict your character-portrait. The line is what was said. Trust the rendering.
If you don't remember the line and you're constructing it, mark the moment in your draft. The mark is to be revisited. The construction is the failure-zone. The revisit is to either find the actual memory or render the half-memory as half-memory rather than as quoted dialogue. Something like, "What if you tried something different?" — at least, that's what I remember her saying is honest. "What if you tried something different?" she said with a smile is constructed and the reader will detect the construction.
A second discipline. Read your dialogue aloud. Speech that does not work as speech is usually constructed speech. The writer who can hear it aloud and hear it land has rendered honestly. The writer who hears it land flat is hearing the construction.
You've collapsed into constructed dialogue when characters' lines do not break or hesitate. Real people break, hesitate, repeat. Constructed dialogue is unnaturally clean. Real people deliver odd lines that don't quite fit. Constructed dialogue is too on-the-nose. The cleanliness and on-the-noseness are the signature. The reader detects them as fakeness even if she cannot name what she is detecting.
You're rendering honestly when characters say odd things, break their own grammar, repeat themselves, refuse to make the point cleanly. The mess is the signature of actual speech. The actual speech is what your memory contains. Render the mess.
The frame opens into two adjacent territories where the dialogue diagnostic does generative work.
The handshake into psychology covers what the fakeness-detection mechanism actually operates on.
The handshake into behavioral mechanics covers the use of dialogue in influence-detection.
The Sharpest Implication Most contemporary writing-instruction teaches dialogue as a craft module separate from honesty. Students learn dialogue-tag conventions, voice-differentiation exercises, exposition-embedding techniques. The training is not wrong but the training is missing the load-bearing element — in memoir, dialogue is recording, not construction. Students who treat dialogue as a craft module produce competent constructed dialogue. Readers detect it. The students cannot understand why their dialogue does not land. The fix is curricular. Some teachers do teach this. Most do not. The form's overall dialogue-quality suffers as a result.
Generative Questions