**Status.** A bounded evidence check of two shipped language-decipherment games against the team's language-play proposals. Sources are developer statements in official material, interviews and one developer-authored article; secondary summaries did not fill gaps. Nothing here shows that our version will be fun, readable or long enough. Four kinds of claim are kept apart: an implemented mechanic the developer describes, a stated intention, evidence the developer reports from players, and our own proposals. Neither game's sources supply player data beyond the developers' reports.
+
This comparison uses primary accounts of two shipped language-decipherment games. It separates described mechanics, design intentions, developer-reported observations and proposals for Hold The Horde. It does not establish that our version will be enjoyable, readable, affordable or long enough.
-
## Heaven's Vault (inkle, 2019)
+
## Heaven's Vault
-
Sources: [inkle's official page](https://www.inklestudios.com/heavensvault/); [Game Developer interview, 17 April 2019](https://www.gamedeveloper.com/design/how-inkle-developed-its-own-ancient-language-for-i-heaven-s-vault-i-) (Jon Ingold, Joseph Humfrey); [MCV/DEVELOP interview, 16 April 2019](https://mcvuk.com/development-news/heavens-vault-how-inkle-wrote-its-own-herstory/) (Ingold, Humfrey); [TheSixthAxis interview, 27 April 2018](https://www.thesixthaxis.com/2018/04/27/interview-talking-with-inkles-about-deciphering-ancient-language-in-heavens-vault/) (Ingold); [PCWorld preview, 18 April 2018](https://www.pcworld.com/article/401842/heavens-vault-preview-inkle.html) (one Ingold quote, otherwise a journalist's account).
+
The player chooses English candidate meanings for unfamiliar words. Glyphs contain smaller elements that can suggest relationships; subsequent inscriptions are selected or generated with the player's vocabulary progress and location in mind. The developers describe translation choices that can advance the story without waiting for a correctness gate. Guesses affect interpretation and narrative consequences; further context can expose a mistaken reading. This is not evidence that every guess is consequence-free or impossible to test. The interviews do not document every later dictionary-confirmation or companion-hint rule. [Developer interview, Game Developer, 17 April 2019](https://www.gamedeveloper.com/design/how-inkle-developed-its-own-ancient-language-for-i-heaven-s-vault-i-), [Ingold and Humfrey interview, MCV, 16 April 2019](https://mcvuk.com/development-news/heavens-vault-how-inkle-wrote-its-own-herstory/), [inkle's official description](https://www.inklestudios.com/heavensvault/).
-
**Implemented, as the developers describe it.** Inscriptions are broken into words and the player selects, for each word, from a small set of English candidates the player character offers; words are built from smaller atoms, so related symbols hint at connections already made. The game never states whether a choice is right: "What we don't tell you is if you're right" (Ingold, MCV). Progress does not wait for a correct translation, but a choice has consequences: a wrong idea feeds into the story and into later translations (Ingold, MCV), "the wrong translation might send you down the wrong track" (inkle page), and as a player drifts the design narrows or eliminates candidate options (Game Developer paraphrase). Inscriptions are generated to suit each player's dictionary progress and location (Game Developer). Ingold expects players to be confident about roughly a third of the language by the end (TheSixthAxis). The player still decides which candidate to commit, whether to trust an earlier guess when a word recurs, and what a place therefore is. The exact rule by which a guess becomes a settled dictionary entry, and the companion's hinting role, are not described in these sources.
+
These accounts describe a goal of making the player feel they are translating and forming their own understanding. The MCV interview includes a developer impression that players' guesses are often good; that is not a published accuracy dataset. Its account of lengthy mechanic development indicates authoring burden, without measuring the cost of our proposal. [MCV interview](https://mcvuk.com/development-news/heavens-vault-how-inkle-wrote-its-own-herstory/).
-
**Stated intention.** To feel like translating rather than to model a language, the "Guitar Hero of linguistics" (Humfrey, Game Developer); ownership through uncertainty (Ingold, TheSixthAxis).
+
For our design, candidate meanings and useful feedback need not eliminate interpretation or application. A static station layout also does not logically rule out adaptive content selection. The relevant constraint is that previously recorded evidence and established world state remain coherent. We should not silently rewrite an inscription already observed in order to correct the player's vocabulary.
-
**Reported evidence.** Only a developer impression that guesses are "usually pretty good" (MCV paraphrase). No player data.
+
## Chants of Sennaar
-
**Transfer limit.** Heaven's Vault's consequences are narrative and arrive later through story and further inscriptions. In our game a reading can be applied in the world at once, under oxygen and horde pressure, while journal reading is paused; a wrong reading can cost supply as well as understanding. Fixed geometry limits regenerating placed inscriptions to fit progress, though adaptive selection of which persistent clues a player meets remains possible if established clues stay coherent.
+
Moya describes a design rule that each word appears in at least two different situations, alongside notes, hypotheses and further exploration. He names validation inspired by Return of the Obra Dinn and automatic translation of complete sentences. That interview does not specify the exact validation triggers, batch policy or uncertainty that remains after confirmation; those details cannot be supplied by assumption. [Moya interview, Game Developer, 29 February 2024](https://www.gamedeveloper.com/design/immersing-players-in-the-culture-of-a-people-with-language-puzzler-chants-of-sennaar).
-
## Chants of Sennaar (Rundisc, 2023)
+
The same interview reports substantial playtesting, difficult words or sections, and additional occurrences, clues, locations or characters introduced in response. This is qualitative developer-reported evidence of an iterative process. The cited material provides no public dataset that validates a particular redundancy rule for Hold The Horde. Moya's separate account connects word introduction, culture, grammar, puzzles and level layout, describing a progression from more concrete language toward abstraction. [Moya interview](https://www.gamedeveloper.com/design/immersing-players-in-the-culture-of-a-people-with-language-puzzler-chants-of-sennaar), [Moya's authored Deep Dive, 23 November 2023](https://www.gamedeveloper.com/design/deep-dive-the-visual-tapestry-of-i-chants-of-sennaar-i-).
-
Sources: [Game Developer interview, 29 February 2024](https://www.gamedeveloper.com/design/immersing-players-in-the-culture-of-a-people-with-language-puzzler-chants-of-sennaar) (Julien Moya); [Moya's Deep Dive, 23 November 2023](https://www.gamedeveloper.com/design/deep-dive-the-visual-tapestry-of-i-chants-of-sennaar-i-); [developer store description](https://store.steampowered.com/app/1931770/Chants_of_Sennaar/); [Rundisc's page](https://rundisc.io/chants-of-sennaar/). Two further interviews ([80.lv](https://80.lv/articles/postmortem-chants-of-sennaar-s-story-road-to-success), [Pocket Tactics](https://www.pockettactics.com/chants-of-sennaar/interview)) contained no mechanic detail.
+
The per-word rule is a useful precedent, not a proven minimum per sign or a guarantee that every missed clue can be recovered. Our survivors can supply testimony and context; whether any speak an alien language is still a design question. Gathering and applying information can happen under oxygen and horde pressure, while reading the journal pauses the world. The sources do not establish that observation in the comparison games is always safe.
-
**Implemented, as the developer describes it.** A redundancy rule per word: "every word you encounter in the game must be seen at least twice in two different situations" (Moya, Game Developer), usually more. Players are not expected to understand a word at once; they "jot down a few hypotheses and keep exploring" in a notebook (Moya; store text: "note each symbol, keep track of your surroundings"). Words are introduced from the concrete and simple toward the abstract, in an order set by gameplay and staging (Deep Dive). Two systems are named but not explained in the reachable sources: an "Obra Dinn-inspired validation system" and "automatic translation of complete sentences" (Moya, Game Developer). The widely reported picture-page batch confirmation could not be verified from a developer source and is treated as unverified.
+
## Implications for this project
-
**Stated intention.** The feeling of a traveller who does not speak the language and comes to understand it through immersion and practice; confronting players with seemingly impossible challenges while discreetly giving them the tools and clues to succeed (Moya, Game Developer).
+
- **Testability is broader than machine feedback.** Later inscriptions, testimony and environmental consequences can challenge a reading. A small set of examples can support an inference without proving a unique grammar.
+
- **Plan further learning opportunities.** The developers describe repeated contexts and adaptive content intended to support learning. Neither account guarantees recovery for every player or establishes that any two examples plus an exception are sufficient.
+
- **Define exactly what assistance confirms.** Candidate meanings, visual comparisons, attributed survivor help and factual outcome records can leave substantial interpretation and action to the player. Withholding a correctness label does not uniquely create discovery or ownership.
+
- **Keep evidence limits visible.** The existing games' mechanics and development experiences inform alternatives; they do not validate this game's fun, balance, content burden or campaign duration. The rejected death-count-to-creature example cannot be inherited as a settled mechanism.
-
**Reported evidence.** Moya reports "a lot of playtesting", that some words or sections "were definitely too difficult", and that the fix was to "add a few occurrences, a few clues, an additional place or character" using the word in a new context; the result, in his account, was levels that "could be completed by just about anyone" (Game Developer). This is qualitative and developer-reported, with no public dataset and no causal proof for our game.
-
-
**Transfer limit.** Sennaar's context is largely conversation with living speakers; our proposals have no living speakers of the alien languages, but they do have survivors whose testimony and labels supply context, and whether Sennaar's observation is ever under pressure is not established by these sources. A batch-validation system also grants a certainty our proposals withhold; adopting one is an explicit change of stance, not a detail.
-
-
## What this says about the released positions
-
-
**"Every inference testable", "costs little", "cannot be solved by dialogue".** Heaven's Vault never states correctness, yet wrong guesses surface through later inscriptions and story consequences, so testing need not mean a machine responding: text, context and testimony test a reading too. "Costs little" is unsupported; inkle's "about a year and a half" for the mechanic (Ingold, MCV) and Rundisc's iteration signal burden without measuring ours.
-
-
**"Any two comparable instances plus one that breaks the reading will do."** Sennaar's stated safeguard is a per-word rule of at least two situations, tuned by adding occurrences where players struggled, and Heaven's Vault compensates for misses by generation. Neither shows that two-plus-one guarantees recovery, which still depends on the player noticing the exception. Context-dependent signs also still yield reusable rules: inkle's atoms are precisely reusable context, so an indexical language does not exclude a lookup.
-
-
**"A hint, match or confirmed gloss removes interpretation."** Heaven's Vault hands the player candidate words for every inscription and still leaves choice and consequence; Sennaar's automatic sentence translation is assistance the developer describes as part of the puzzle. Help does not by itself dissolve interpretation or ownership, and required help is a gate only where a policy makes it one.
-
-
One inherited example should be dropped: a ledger tally matching the player's death count assumes a count-based mechanism no direction has adopted.
-
-
## Design implication and the open choice
-
-
Both shipped designs give the player something firmer than open inference, candidates to choose among or validation once enough is consistent, and both survive missed clues by redundancy or generation. The transferable safeguard for a station priced in oxygen is a stated redundancy rule per learnable unit, taking Sennaar's per-word rule as precedent rather than a validated minimum, with adaptive selection of persistent clues where coherence allows; the paused journal can present a sign's own recorded instances side by side without confirming a meaning. The choice the sources cannot make for us: withhold correctness entirely and let readings feed consequences, or confirm bounded outcomes once observed. The first fits discovery first and personal ownership; the second gives certainty and pacing. Neither source shows which holds up when reading is applied under threat.
+
The practical choice is what feedback appears, when it appears, and which decisions remain. Bounded factual confirmation can coexist with uncertain interpretation. [Language and Journal](/Hold%20The%20Flood/Language%20and%20Journal) proposes recording an observed destination or incomplete attempt without grading arbitrary notes; [Language Expedition](/Hold%20The%20Flood/Language%20Expedition) supplies a written use case. Their precise rules remain proposals, subject to the owner brief and the focused review process.
