There's a game running every time someone tells a story to an audience. The storyteller pretends to be making it up on the spot. The audience pretends to believe that's what's happening. Both sides know the truth — the storyteller has thought about this before, probably has told it before, definitely has structured it — and both sides also need the pretense to hold for the piece to work. Matt Dicks calls this the game we play with our audience. They pretend that we're making it up and we pretend that we're making it up.1 The audience wants to believe they're getting a one-of-a-kind moment that exists only because they're in the room tonight. The storyteller wants to deliver that experience. The game is the shared fiction that makes it possible.
The discipline that makes the game playable is remember the story, don't memorize it. Memorized stories break the game. Improvised stories break it too — in the other direction. The skilled position sits between the two: knowing the beats, the dialogue, the location, the destination — but not the exact sentences that will carry you across them.
Dicks distinguishes three states a storyteller can occupy:
Memorized. You have learned the piece word-for-word. You can recite it. The sentences come out the same each time. The audience can tell — if you'd memorized your story, really you're just a word caller. You've just memorized a series of words and hopefully they'll come out properly.2 The piece is dead because the storyteller is delivering a recording rather than speaking. The game breaks because the audience can feel the recording quality.
Improvised. You have no preparation. You walk on stage with a rough idea and spew the first thing out of your mouth. The piece lacks structure, the engagement engines don't deploy on rhythm, the destination is unclear because you don't know it yourself. The piece is rough because nothing has been thought through.
Remembered. You know what's going to happen — the location, the action, the beats, the dialogue that matters, the destination. You don't know exactly which sentences will carry you between them. We remember them, meaning we understand the beats, we understand what's going to happen. But if I was to tell you a story now and then tell you a story five minutes later, the sentences are absolutely going to be different even though the events, the dialogue, the descriptions will all be there in some way.3 The piece lives because the storyteller is speaking, not reciting, and the structure does its work because everything that matters has been pre-located.
The remembered position is the only one that lets the game work. The audience needs to believe the storyteller is present — actually choosing words in this room, in this moment, in response to who is sitting in front of them — and remembering-not-memorizing is what makes that presence real. The storyteller is genuinely choosing words on the fly. The structural skeleton is fixed; the surface language is live.
The reason the remembered position is harder than either memorization or improvisation is that it requires trusting the structure enough to improvise the surface. Memorizers don't trust the structure; they substitute word-memorization for it. Improvisers don't have the structure; they're winging the whole thing. The remembered position demands you do the structural work — Big Three sequence locked, engagement engines mapped, opening and closing rehearsed enough to be solid — and then let go of sentence-level control during delivery.
That trust-the-structure-and-let-the-surface-vary move is what enables real-time room-responsiveness. The best game players stand in the middle and say I know what I'm going to say but I don't know exactly how I'm going to say it. And that allows you to read the audience and figure out is this landing? Do I need to pivot right now? Do I need to pull an anecdote out of my pocket?4 You can only read the room if your attention isn't locked into reciting the next memorized sentence. You can only pivot smoothly if you know what beat comes next without needing to remember the exact words.
The mechanism behind this is cognitive load. Memorization consumes working memory for sentence-retrieval. Improvisation consumes it for structure-construction. The remembered position consumes very little working memory for either — because the structure is so internalized it runs in the background, and the surface-language production runs in the foreground in the way ordinary speech does. That freed working-memory becomes available for reading the audience and responding to them. The freed bandwidth is the point of the discipline.
The remembered position is the delivery-side complement to the no-one-wants-to-hear mindset. The mindset says the audience owes you nothing; the remembered position is what lets you respond in real-time to whether you're losing them. Memorized storytellers can't read the room because they're locked into sequence; remembered storytellers can pivot the surface to keep the room engaged.
The position is also the enabling condition for imperfection-is-the-beauty (Catherine Burns via Dicks). The small stumbles, the doubling-back, the moments of audible thought are what signal to the audience that the piece is being spoken rather than recited. Those imperfections are not failures of the remembered position; they're its evidence. It's in the imperfection that the beauty lies because the imperfection tells us you had not memorized that thing — you remembered it.5
The remembered position also gives five anecdotes always in pocket somewhere to land. The pocket-reserve is useless from a memorized state — there's no place to insert a 60-second deployable without breaking the memorized chain. From the remembered position, the pocket-anecdote drops in seamlessly because the storyteller knows what beat is coming next and can return to it after the digression. Pocket-reserve and remembered position are paired tools.
Two of Dicks's moments demonstrate the remembered position working at different scales.
The phone in the audience.6 Dicks is delivering a piece. Someone in the audience pulls out their phone. A memorized storyteller would either not notice or could not respond — both because they're locked into the next memorized sentence and because the response would derail the memorized chain. Dicks stops mid-piece and says: I really hope your kid is, like, on the way to the hospital right now. Big laugh. The audience member responds — kid getting to school, just checking. The piece continues. The moment landed because Dicks was in the remembered position. The structure could absorb the digression because Dicks knew where the next beat was; the surface language could absorb the digression because Dicks wasn't word-locked. The moment also tightened the room — the audience now knows the storyteller is actually present with them, not running a tape.
The Brooklyn Historical Society rewrite.7 Dicks arrives at the Brooklyn Historical Society with a story to tell — a single-shot Moth StorySLAM appearance. He has never been there before. When he gets in and looks at the audience, they are basically all blue-haired older women. The story he prepared was a runner-themed piece with humor calibrated for a different audience. The humor will not land in this room. He has roughly 15 minutes before he goes on. He spends those 15 minutes pulling all the humor he knew wouldn't land and leaning into the heart instead. He wins the slam that night.
What this case proves: the remembered position is not just useful in the small (mid-piece interruption) — it scales to the large (pre-delivery wholesale repositioning). A memorized storyteller could not do this rewrite at all, because the piece exists as a word-sequence rather than as a structure. A remembered storyteller can do it in 15 minutes because the piece exists as a structure that can be re-skinned with different surface language and different engagement-engine deployments while keeping its bones.
The case also proves that the remembered position is medium-specific. The next night, Dicks notes, the humor goes back in because the audience is different. The "right" version of the piece is not a fixed text; it's a structural skeleton that gets dressed differently for each room. That medium-specificity is impossible from the memorized position by definition.
Wednesday night. You're rehearsing the piece you're going to tell at Sunday's storytelling open mic. You have it drafted. You've done the four-engine sweep. The structure is solid. Now you need to get it into the remembered position.
You don't read the draft aloud. Reading aloud builds memorization. Instead you put the draft in a drawer. On a separate sheet of paper you write the skeleton: the location and opening action; the five or six scene-beats in order; the key dialogue lines (Jamie's "did you really think I was just going to leave it there for you?" — the dialogue that has to be exact); the engagement-engine deployment notes (where suspense builds, where the surprise lands); the destination (the 5-second moment as a single sentence).
You rehearse from the skeleton. First time through, you stumble. The sentences are clumsy because you're inventing them in real time around the skeleton-beats. You don't stop to fix the sentences. You let them be clumsy and finish. Second time through, fewer stumbles. Third time, the surface language is finding its shape — different sentences than last time, but smoother. By the tenth time through, you're delivering the piece without notes, with sentence-level variation each time, and the structure is so internalized that you could absorb a phone-in-the-audience moment without losing the next beat.
If you find yourself reciting the same sentences word-for-word by the tenth time through, you've drifted into memorization. The fix: re-shuffle the order of the scenes during the next rehearsal and let yourself find new connecting sentences. The point is not to vary the structure (the structure should be stable) but to keep the surface language flexible.
You can recite your piece from memory. Recitation-capacity is the marker of memorization. If you can write out your piece word-for-word from memory after a week, you're in the memorized position. The fix: stop reading the written version and rehearse from skeleton-notes only.
You panic when interrupted. If an audience member coughs or laughs at an unexpected moment or asks a question mid-piece and you lose your place, you're memorized. The remembered position absorbs interruption because the structure is what's holding the piece together, not the sequence of sentences.
Your piece sounds the same every time you tell it. Variations in surface language across deliveries are the marker of the remembered position. If the language is identical across multiple tellings, you've gone past remembered into memorized.
You're afraid to look at the audience. Memorized storytellers often deliver to the middle distance because eye contact threatens to derail the memorized chain. Remembered storytellers can hold eye contact because the structure does the holding-together work, freeing the eyes to read the room.
You can't pre-pivot the piece for a different audience. If you couldn't pull a 15-minute Brooklyn-style rewrite the way Dicks did — strip out the humor that won't land, lean into the heart — your piece exists as text rather than as structure. That's a tell that you've over-memorized at the sentence level.
The remembered-not-memorized discipline is named explicitly in the Knowledge Project interview and is consistent across both Dicks transcripts.8 It is also a discipline reported by working storytellers across the Moth StorySLAM community more broadly — Catherine Burns (former Moth director) names imperfection-is-beauty as the evidence-signature of the remembered position, suggesting it's a tradition-wide rather than Dicks-only principle.
A real tension exists with theatrical and political-speech traditions, which often require strict-memorization for reasons that are not about audience engagement but about message-control or text-fidelity. A Shakespearean actor cannot freely reword a line; a political speech writer needs the exact phrasing they pre-cleared. Those traditions treat the remembered position as a risk rather than a discipline. The tension is real and not resolvable in the abstract — both protocols are right for their respective contexts. Working storytellers in the Dicks tradition occasionally violate this distinction by treating Moth-style discipline as universally correct, which it isn't.
A subtler tension exists with the audience-shielding hacks in Karlsson's 5-step writing method: Karlsson's essay-writing practice asks writers to forget the audience during composition in order to write authentically. Dicks's remembered position asks storytellers to never forget the audience during delivery in order to deliver authentically. The contradiction is medium-specific: composition vs delivery. But the deeper question this raises is whether the remembered position's room-responsiveness is always a virtue, or whether some material is better served by the storyteller refusing to read the room (the way Karlsson refuses to imagine the audience during essay drafting). Working oral storytellers report mixed instincts on this; some pieces benefit from room-responsiveness, others land better when the storyteller commits to a fixed delivery regardless of audience signal.
Open question: how long does it take to build the remembered position for a new storyteller? Working storytellers report 20-50 deliveries of a piece before it locks into the remembered shape. The ramp varies by practitioner; some develop it faster, some slower. Whether the ramp can be compressed through specific practice protocols (rehearsal-with-shuffled-order, deliberate-stumble-allowance, mock-interruption drills) is an open empirical question.
Open question: does the remembered position generalize to argument-shaped pieces (lectures, panels, debates) or only to narrative-shaped pieces? Dicks teaches it specifically for narrative. The structural-skeleton-with-flexible-surface principle should transfer to argument-shaped pieces but the operational form would need translation — the "beats" of an argument are claims and evidence rather than scenes, and what counts as "the right surface language" is differently calibrated.
Dicks and Karlsson agree at a level both might not name. Karlsson's discovery writing as intimacy argues that essays come alive when the writer is genuinely finding the thought during composition, not reciting a pre-formed conclusion. Dicks's remembered position is the live-delivery equivalent: the piece comes alive when the storyteller is genuinely choosing words during delivery, not reciting a memorized sequence. Both writers are saying the surface of the work must be live, not pre-formed, even though the underlying structure can — and should — be designed. They're describing the same craft principle from opposite sides of the production cycle: Karlsson on the composition side, Dicks on the delivery side. What this agreement shows: the audience can detect pre-formed surface language and registers it as less alive, regardless of how good the content is.
Dicks and Tucker Max split on the question of how much pre-fabrication is acceptable. Tucker's memoir methodology involves substantial rewriting and refining — the published sentences are very pre-formed compared to Dicks's remembered position. But the comparison reveals medium-specificity: pre-formed sentences in print are received differently than pre-formed sentences in live oral delivery. The reader of print expects craft-shaped sentences and rewards them; the listener of live oral expects spoken-shaped sentences and is alienated by craft-shaped ones. Same writer-craft principles (work hard on structure, trust the audience) producing opposite surface-treatments across media.
The remembered position is fundamentally about holding structural form loosely enough to remain responsive. That capacity has analogues across performance, contemplative practice, and behavioral mechanics — and tracing them out shows how widely the same craft-meta-skill operates under different surface vocabularies.
Eastern spirituality: Sahridaya — The Cultivated Reader — the Sanskrit aesthetic tradition treats the sahridaya (cultivated reader/listener) as actively co-creating the aesthetic experience with the artist. The remembered position is the artist-side complement: by remaining genuinely responsive during delivery, the storyteller invites the audience into co-creation rather than receiving them as passive listeners. Rasa-theory explains why the co-creation matters — it's what makes aesthetic experience real rather than transmitted; the experience happens between artist and audience rather than in either alone. Dicks supplies the operational discipline (remember, don't memorize) that keeps the artist-side of co-creation possible. Together you get both the depth-theory (why responsive co-creation works) and the working compression (what specifically the artist has to do to enable it). The cross-tradition convergence is striking: contemporary oral storytelling craft and classical Sanskrit aesthetics agree that responsiveness to the present audience is what distinguishes live art from recorded transmission. Both treat memorization as a failure of co-creation, even though the traditions don't talk to each other.
Behavioral mechanics: Six-Minute X-Ray (Hughes) — Hughes's elicitation work requires a structurally identical capacity: holding the protocol loosely enough to remain responsive to the target's actual behavior, while the structure of the protocol is solid enough to guide the conversation. Memorized elicitation fails for the same reason memorized storytelling fails — the operator cannot adjust to live conditions. The two handshakes together show that responsive structure is the mark of professional craft across orientations. Amateurs either memorize (and become brittle) or improvise (and lose structure). Professionals build durable structures that they can deploy loosely. Whether the goal is engaging an audience (Dicks) or extracting information (Hughes), the meta-skill is the same: hold the form, free the surface. Cross-craft transferability is high — once you have this skill in one domain, it tends to transfer to others. This is one reason working storytellers often make good interviewers and vice versa: both crafts demand the same operating disposition.
Eastern spirituality: The Rasa Weaver Archetype — the rasa-weaver tradition explicitly distinguishes the practitioner who can deploy state-modulation moves spontaneously from the one who has merely memorized a sequence of moves. The rasa-weaver classification has four levels (Novice through Master), and the mastery levels are defined by responsiveness to the actual audience state rather than by repertoire size. This is the rasa-tradition's articulation of exactly the Dicks remembered/memorized split, with a more developed apprenticeship-pathway between them. The cross-tradition insight: working practitioners across centuries and cultures converge on the same internal architecture for delivery — structure-stable, surface-responsive — but the contemporary working storyteller can borrow the rasa-tradition's developmental ladder to understand where they are on the path from novice to master, which Dicks's framework doesn't articulate explicitly.
The Sharpest Implication Most public speakers operate either memorized or improvised. Both positions are amateur. The remembered position is what working storytellers, working comics, working orators have always operated from — but most amateurs don't know it exists as a third option. The standard advice ("memorize your speech for confidence" or "just be natural up there") is both wrong, in opposite directions, and the right position requires more preparation than the improvisers do and more letting-go than the memorizers do. It's not a middle ground in difficulty — it's harder than either alone. The implication is uncomfortable: most professional speaker-training, most public-speaking courses, most TED-talk coaching teaches one of the two amateur positions because the remembered position requires more practice-time than the average client will commit to. The skill is gatekept by the practice-cost, not by any innate gift.
Generative Questions