---
title: "A shared shape is not a shared stake — the field keeps invoking the 1871–1932 'War of the Dots' as the modern Nemeth/UEB split's precedent, but the two disputes share only a decentralized, institution-by-institution shape, not the stake, and no one has ever overlaid the maps"
type: "moc"
writer_model: "warden/claude-opus-4.8"
tags: ["disability-history","braille","tactile-technology","standards-wars","accessibility-tech","nemeth-code","unified-english-braille","cross-time-bridge","historiography","false-analogy"]
date_created: "2026-09-19T00:00:00.000Z"
updated: "2026-09-19T00:00:00.000Z"
provenance: "Warden pass 2026-09-19 (warden/claude-opus-4.8), run per 00-meta/specs/seek-warden-spec.md on a maintenance engine kept separate from the notes' writers (claude-sonnet-5, claude-opus-4-8). Discharges the 2026-09-19 [entity] flag 'the braille / War of the Dots / Nemeth-UEB cluster is now 8+ claim-notes ... past the spec §6 threshold ... worth a deliberate moc-braille-standards-war pass' (00-meta/seek-flags.md). Named for the argument, not 'braille-standards-war' (the topic the flag proposed), per the 2026-07-25 lesson. Grounded in a direct read of all nine member notes and the open question this pass, not in cosine — the tag-based ranker undercounts this silo (Cali's 2026-09-01 directive), and it did not appear in the 2026-09-19 constellation's unnamed-clusters list at all."
audit_status: "2026-09-19 Warden self-check only (warden/claude-opus-4-8): all nine member links, the open question, and the four entity hubs resolve; tiers and statuses read against the member notes (spine note claim-nemeth-ueb-split-... Tier 2, cross-model audited 2026-08-23; Irwin institution-by-institution Tier 2 with the 40% figure carrying [unverified-quant]; BANA-2012 Tier 1 verified_verbatim; Cook-Walker and Jacobson Tier 2 NFB Braille Monitor; BEAR-2022 Tier 3 [unverified-quant]; gradually-2015-2022 [unverified-quant] on both endpoints; war-of-the-dots Tier 3; New York Point designer Tier 4). NO independent cross-model reading has run on this map yet — it is a first-pass build by a single engine and a future cross-model audit target, the same posture the moc-a-shared-slogan-is-not-a-shared-model build carried before its next-day audit. | 2026-09-20 cross-model audit (claude-fable-5; builder warden/claude-opus-4.8): the independent reading the line above anticipated. All nine member links, the open question, both sibling/neighbor MOC pointers, and all four entity hubs re-checked and resolve; every tier, status, and caveat description in this file re-read against the member notes' own frontmatter and bodies and confirmed exact — including the spine note's Tier 2 / 2026-08-23 BANA→NCEO correction history, the Irwin 40%-by-1910 [unverified-quant] flag, BANA-2012's Tier 1 verified_verbatim, the Cook-Walker 14-of-21 roster and 'War of the Dots' quote, the Jacobson 'turning the clock back one hundred years' quote, the BEAR 31/7/11/1 count with its 'according to our research' caveat, and every member sitting at seedling. No corrections needed; a clean first cross-model pass."
seek_code_commit: "21947c9"
audits: ["2026-09-20 claude-fable-5"]
---


The recurring argument in this cluster is not "the history of braille codes" and
not "history repeated itself." It is a discipline about *self-invoked historical
parallels*: **the blindness-education field reaches for the 1871–1932 "War of the
Dots" — by name, self-consciously, at the very moment the modern Nemeth-vs-UEB
split is forming — as the precedent to fear. But the two disputes share only a
*shape* (decentralized, unresolved, settled institution-by-institution and
state-by-state, accreting over years rather than decided at once), not a *stake*.
The first war was over which physical tactile alphabet a blind reader's fingers
would learn to read at all; the modern dispute is over where within one
already-standardized six-dot cell a number is placed. And the literal comparison
the parallel implies — do the modern state lines fall where the old institutional
ones did? — has never actually been performed by anyone.**

That distinction — *the process rhymes; the stakes don't* (the member notes' own
words) — is why this is a map and not a pile of notes that both say "braille."
It is a sibling to [[moc-a-shared-slogan-is-not-a-shared-model]]: there, a
borrowed *slogan* ("run to stay in place") travelled across domains while the
*formalism* behind it mostly did not; here, a borrowed *historical precedent*
travels across time within one field while the *stake* behind it does not. Both
maps refuse to let a shared surface pass as a shared substance. Titled for the
argument, not for the recurring entity or episode, per the 2026-07-25 lesson.

## The old war — the historical shape

- [[claim-war-of-the-dots-competing-braille-systems-1829-1932]] — the frame. For
  over a century the US had no single tactile-literacy standard: Braille, New York
  Point, and Modified American Braille (plus Moon type) competed in incompatible
  embossings, resolved only at the 1932 London Type Conference. Uncontested
  institutional history; Tier 3 (APH history blog), corroboration not independently
  re-checked.
- [[claim-1871-1932-braille-standards-war-fought-institution-by-institution-not-fixed-lines]]
  — **the shape that will recur.** Robert B. Irwin's firsthand 1955 memoir: the
  unit of allegiance was the individual school, not a regional bloc, and a public
  switch could cost a superintendent his seat at the convention luncheon table.
  Tier 2 for Irwin's testimony; the embedded "40 percent by 1910" figure is
  `[unverified-quant — needs primary]` (a round number recalled 45 years later).
- [[claim-william-bell-wait-sighted-designer-of-new-york-point]] — the losing
  side, and a foil that belongs half to another thread (see *What this map is
  not*): New York Point was designed top-down by a sighted superintendent, its
  variable-width cell a *physically different* encoding from Braille's fixed cell —
  the concrete fact that makes the old war a fight over scripts. Tier 4 (Wikipedia),
  uncontested biographical claim.

## The disanalogy — shape, not stake (the spine)

- [[claim-nemeth-ueb-split-is-a-notation-convention-not-a-competing-script-dispute]]
  — **the load-bearing finding, and the one the writer trusts most because it does
  not need the state-count number to be true.** Per a 2019 NCEO brief, UEB and
  Nemeth use the *identical* six-dot cell and differ only in whether a number sits
  in the upper or lower four dots — a convention choice within one standardized
  alphabet, not a competing script. Where New York Point demanded a reader relearn a
  tactile geometry, the modern dispute demands nothing of the sort. Tier 2, quotes
  cross-model audited verbatim 2026-08-23 (which also corrected the source
  attribution from "BANA guidance" to the NCEO/D'Andrea brief and the tier from 1 to
  2). This is the seam: same decentralized shape, categorically smaller stake.

## The field invoking its own past — by name, in real time

The evidence that the parallel is *reached for*, not merely available in hindsight
— and reached for at the moment of formation, which is what makes it interesting
rather than obvious.

- [[claim-cook-walker-2015-named-war-of-the-dots-and-logged-earliest-nemeth-ueb-state-roster]]
  — one month *before* UEB's January 2016 rollout, an NFB braille-education manager
  named the historical episode directly ("We do not need a repeat of the 'War of the
  Dots'") and logged the earliest located state roster (14 of 21 declared states
  retaining Nemeth). Tier 2 (NFB *Braille Monitor*, named author); the roster is an
  advocacy canvass of *declared intentions*, not enacted policy — caveat carried.
- [[claim-nfb-2016-warned-ueb-rollout-could-repeat-1871-1932-fragmentation]] — the
  same month rollout went live, Jacobson posed the parallel as a rhetorical warning
  ("are we turning the clock back one hundred years?"). Tier 2. Evidence the parallel
  was *feared* in real time by a witness with no reason to perform worry for an
  audience a decade early — a smaller, more honest finding than "the pattern
  repeated."

## The modern mechanism, and the soft counts it produced

Honestly tiered: the mechanism is Tier 1; every attempt to *count* the resulting
split is soft, and the map keeps it that way rather than quoting a tidy figure as
settled.

- [[claim-bana-2012-ueb-adoption-kept-nemeth-code-parallel]] — the mechanism.
  BANA's 2012 UEB vote did not replace Nemeth; the US still has no single mandated
  math/science braille code, only a standing choice pushed down to states and
  readers. Tier 1, `verified_verbatim`.
- [[claim-bear-2022-count-most-us-states-default-to-nemeth-for-technical-braille]] —
  the single most citable modern figure (31 states + DC default Nemeth, 7 UEB, 11
  IEP-team, 1 undecided) and the least load-bearing on its own terms: a Tier 3
  advocacy blog whose stated method is "according to our research."
  `[unverified-quant — needs primary]`.
- [[claim-nemeth-ueb-state-split-settled-gradually-2015-2022-not-at-rollout]] —
  what the two soft snapshots can *jointly* support without either being
  load-bearing: the split did not snap into place at rollout, it accreted state by
  state (Nemeth-default column growing from a floor of 14 to 31+DC as previously
  silent states declared) — a resemblance of *process* to Irwin's
  institution-by-institution history that survives even with both counts wobbling.
  The two canvasses measure different objects (declared intention 2015 vs. enacted
  default 2022) and cannot be subtracted; `[unverified-quant]`, only the direction
  survives.

## The open question this map is built around — still open, and load-bearing on the honest side

- [[question-verify-nemeth-ueb-state-mandates-split]] — do current state-level
  Nemeth/UEB mandates split along the same lines as the 1871–1932 adoption map?
  **Still `open`.** No primary Tier-1 roster exists on either end, and — the crux —
  *no source anywhere has overlaid the two maps.* This is not a gate on the map's
  argument; it *is* part of the argument. The field invokes a precedent whose
  literal comparison nobody has run: the parallel is asserted more confidently than
  it has ever been checked. The historical side would need a real
  institution-by-institution roster (candidate: AAIB 1871–1880 convention
  proceedings) before an overlay is even resourceable.

## What this map is not

- **Not the origin-myth cluster.** The vault's Barbier/Braille notes
  ([[claim-barbier-night-writing-designed-for-blind-not-soldiers]] and kin) correct
  the "secret military night-writing" origin story and belong to
  [[moc-attribution-and-origin-myths]] as its disability-history leg — a *myth
  about origins*, not this map's *analogy about a standards dispute*. Kept there,
  cross-linked here, not folded in.
- **Not the blind-mathematicians thread.** That New York Point was designed *for*
  blind readers by a sighted administrator rather than *by* a blind one, and that
  Nemeth code was built by a blind mathematician, belong to
  [[moc-blind-mathematicians-working-methods]]; [[claim-william-bell-wait-sighted-designer-of-new-york-point]]
  sits in both threads and is included here only for the physical-script fact that
  grounds the disanalogy, not for the designer question.
- **Not `moc-a-shared-slogan-is-not-a-shared-model`, but its sibling.** That map
  tests whether a cross-*domain* metaphor carried its *formalism*; this one tests
  whether a cross-*time* precedent carries its *stake*. Same family (a compelling
  resemblance held more firmly than the evidence warrants), different specific
  argument — a pointer, not a member.

## Entity hubs

Built by prior promotions; listed, not built or edited by this pass.

- [[entity-nemeth-code]] — the parallel math/science code BANA's UEB vote left
  alive.
- [[entity-new-york-point]] — the losing rival of the first war; the physically
  different script.
- [[entity-william-bell-wait]] — its sighted designer.
- [[entity-robert-b-irwin]] — the AFB executive whose firsthand memoir supplies the
  institution-by-institution history.

## Open threads (honest caveats, not hidden)

- **Every member is `status: seedling`.** This map organizes the cluster's
  argument; it does not upgrade the notes.
- **The overlay has never been done, on either side.** Both the modern and the
  historical rosters are missing a primary; the map's own spine-adjacent claim
  ("the process rhymes") rests on two soft advocacy counts, and its strongest leg is
  the *disanalogy*, which needs no count at all.
- **The "invoked at the moment of formation" evidence is two NFB pieces** by
  advocates lobbying against All-UEB math — witnesses to the field's fear, not
  neutral observers of a correlation. That is exactly the size of finding the map
  claims, and no larger.
- **First-pass build, single engine.** No independent cross-model reading yet (see
  `audit_status`).

> [!note] Warden's commentary:
> What convinced me this was a map and not two notes that both say "braille" is the
> asymmetry the notes kept insisting on against their own most quotable material.
> The single most citable thing in the whole cluster is the 31/7/11/1 state count —
> and the writer who captured it flagged it, in her own commentary, as the *least*
> load-bearing thing in the file, precise enough to be dangerous. The finding that
> survives is the opposite of the tidy one: the field reaches back a hundred years
> for the name of the thing it fears, uses it without a gloss the way you name a
> family illness, and yet the two illnesses aren't the same disease. One was a fight
> over which alphabet would exist; the other is a fight over four dots' vertical
> position inside a cell nobody disputes. The process rhymes — decentralized, slow,
> most of the map blank at the start — and the stakes don't. I built the map so its
> spine is the disanalogy (which needs no roster to hold) and left the overlay
> question open on the page, because the honest shape of this cluster is that a
> parallel everyone invokes has never once been checked, and that absence is the
> finding, not a hole in it. If this file is shorter in two weeks it should be
> because a real silo finally has a home the tag-ranker never gave it — not because
> tonight drew a box around a number nobody has verified.
> — warden/claude-opus-4.8, 2026-09-19