Review / Sindala dialogue workshop — testing the evolution proposal

mcp proposal · package #203

Sindala dialogue workshop — testing the evolution proposal

User-authorized proposal-stage dialogue workshop. Local companion: languages/sindala/dialogue-workshop-proposal.md. Linguistic sources: languages/sindala/evolution-proposal.html (proposal), source/sindala.export.html and source/sindala.vulgar.txt, generated/lexicon.csv and generated/language.json; languages/sigtaara/generated/lexicon.csv. Canon grounding: species.firgol.sentient.lahkiin.kaemisaedhel, species.firgol.sentient.lahkiin.kaekreaedhel, species.firgol.sentient.lahkiin.zaetaedhel. All scenes, constructions, and historical extrapolations are proposed. No canon changes requested.

In Review 0/2 items resolved

Proposal workspace — not canon. Accepting an item records your review decision. Writing it into canon requires a separate exact preview and typed confirmation. Overview items are never imported.

Review the whole package · 2 items

Mechanical consistency

Canon impact

Targets and conflicts are derived without a model call. Overview items are omitted — they never touch canon.

Five dialogue scenes and provisional grammar findingsmapping required

Kind: Prose
Target: mapping required

✓ No mechanical conflict detected

Selected item

Five dialogue scenes and provisional grammar findings

Candidate Prosev1

Manual revision

Title Text

# Sindala dialogue workshop

**Proposal — not canon. · 14 September 2026**

Five small scenes to test the [evolution proposal](/home/ai/mine/nixxia/languages/sindala/evolution-proposal.html). All speakers, encounters, customs introduced in these scenes, and grammatical innovations are invented for this workshop. They establish no events or language rules. The original proposal and imported language remain unchanged.

The working direction is to preserve verb-final speech, test subject-first order alongside inherited object-first order, and let the historical form acquire a meaning of received account. Ordinary conversation should still accommodate impatience, courtesy, evasion, and affection.

## How to read the examples

The English exchanges are complete dialogue sketches. **Sindala lines translate only the short clauses explicitly paired with them**, not every sentence in the English exchange. They are constructed examples, not attested sentences or finished translations.

Lowercase words follow the export's romanization. Hyphens expose grammatical boundaries for review; they do not propose a new spelling convention. The source prints, for example, `dalpanvu`, not `dal-panvu`.

| Material | Established in the language files | Workshop treatment |
|---|---|---|
| Word order | Object–indirect object–subject–verb | Retain object-first as a baseline; try subject-first for introducing or contrasting an actor. This distribution is proposed. |
| Past | Exported `dal-`, from inherited `tɑhl-` | General past, with no guarantee about evidence. |
| Historical | Exported `cir-`, from inherited `zir-` | Test a received-account reading for past events, including recent ones. |
| Remote past | Exported `duul-`, from inherited `tuul-` | Leave its meaning unchanged in this workshop; “used to” needs a separate test. |
| Pronouns | Exported `cuu` “I”; `hayar` masculine singular “you”; `hayara` masculine plural “you” | Use these forms as exported. Respectful singular use of `hayara` is a proposed extension. |
| Negation | Dictionary `dan` “not,” an adverb | Its position before the verb in scene 3 is provisional. |

The normalized profile preserves ancestral forms in its grammar tables, while the HTML export shows their transformed Sindala realizations. The latter supplies the surface forms here. The complete clauses still require validation: attaching an exported prefix to another stem does not establish every boundary effect, valency pattern, or discourse rule.

Two corrections to keep in view: first-person pronouns already share forms across the gender-labeled rows; no new merger is needed there. Also, Sindala already has `dun` “obey,” alongside `hon` “hear, heed” and `vu` “hear.”

## 1. An eyewitness account: who opened the gate?

**Situation:** Two neighbors discuss an ordinary incident. One saw it; the other heard an account. No official testimony system is assumed.

**Neighbor A:** “The mother opened the gate. I saw her do it.”

**Neighbor B:** “The mother opened the gate—that is the account I heard. I wasn't there.”

**Neighbor A:** “Then we agree about the gate. We got there differently.”

The shared claim, with three possible presentations:

| Constructed Sindala | Literal order | Intended reading |
|---|---|---|
| `Virhnasol nehnison dal-bex.` | mother · gate · PAST-open | “The mother opened the gate.” Trial subject-first introduction of the actor. |
| `Nehnison virhnasol dal-bex.` | gate · mother · PAST-open | The same claim in inherited object-first order. Trial reading: the gate is already under discussion. |
| `Nehnison virhnasol cir-bex.` | gate · mother · ACCOUNT-open | “According to an account, the mother opened the gate.” Proposed extension of the historical form. |

`virhnasol` “mother,” `nehnison` “door, gate,” and `bex` “open” are exported vocabulary. Articles in English are supplied for readability; these examples establish no article system.

**What earns its place:** The `dal-/cir-` choice lets two people agree on an event while presenting different relationships to it. The object-first/subject-first pair may organize the conversation without requiring two separate dialect grammars.

**What does not follow:** Neighbor B could use `dal-` without necessarily lying. General past does not assert direct observation. The explicit English statement “I saw her do it” supplies that claim for A; the prefix alone does not.

**Provisional decision:** Keep the received-account experiment. Leave basic word order undecided. We have a useful contrast to test, not yet evidence that subject-first should become obligatory or unmarked.

## 2. A disputed inherited memory: who is “I”?

**Situation:** A speaker describes a supposed ancestral memory. For this scene only, assume the experience includes a first-person perspective. The fidelity, transfer mechanism, and reliability of that experience remain open lore questions.

**Rememberer:** “I opened the gate. I remember the weight of it.”

**Examiner:** “You weren't alive. Whose hands do you mean?”

**Rememberer:** “Hers. The memory comes to me as ‘I.’”

**Examiner:** “Then tell me separately what you experience and what you claim she did.”

| Constructed Sindala | Literal order | Clause represented |
|---|---|---|
| `Nehnison cuu dal-bex.` | gate · I · PAST-open | “I opened the gate.” |
| `Pirbusod cuu daman.` | memory · I · remember | “I remember the memory.” Deliberately plain scaffolding for the present experience. |
| `Nehnison virhnasol cir-bex.` | gate · mother · ACCOUNT-open | “By the received account, the mother opened the gate.” A possible external retelling. |

`pirbusod` “memory” and `daman` “remember” are exported. The memory's sensory detail and the examiner's questions remain English; no possessive, quotation, or subordinate-clause grammar is smuggled into the Sindala fragments.

**What earns its place:** The dispute concerns both information source and the referent of `cuu`. Changing `dal-` to `cir-` would not by itself explain who “I” is. Quoted ancestral speech and the current speaker's own assertion need to be distinguishable.

**Useful disagreement within a community:** A senior memory practitioner could insist on `cir-` precisely because they respect the difference between receiving and originating an experience. A younger practitioner might defend first-person reenactment. Neither position needs to belong to an entire people.

**Provisional decision:** Give the developing marker the narrow claim “I present this past event through a received account.” Do not make it mean “this account is authentic,” “this is uncertain,” or “the institution approves it.” Leave quoted first-person usage unresolved until quotation grammar is designed.

## 3. A household argument: hearing is not obedience

**Situation:** Two adult brothers disagree about whether one should remain at home. The masculine singular pronoun therefore requires no new gender-neutral paradigm.

**Older brother:** “You hear every word and do nothing I ask.”

**Younger brother:** “I hear you.”

**Older brother:** “That is not what I meant.”

**Younger brother:** “I heed you. I do not obey you. I am still leaving.”

| Constructed Sindala | Literal order | Clause represented |
|---|---|---|
| `Cuu hayar vu.` | I · you.SG.M · hear | “I hear you.” |
| `Cuu hayar hon.` | I · you.SG.M · heed | “I heed you.” |
| `Cuu hayar dan dun.` | I · you.SG.M · NOT · obey | “I do not obey you.” Trial negation placement. |

Subject-first is being tested here to foreground the younger brother's own stance. Object-first alternatives are `Hayar cuu vu`, `Hayar cuu hon`, and, under the same provisional negation rule, `Hayar cuu dan dun`.

**What earns its place:** The existing vocabulary supports three different commitments: perceiving words, taking them seriously, and complying. The proposed specialization of `hon` can stop at attentive consideration. The older brother may pragmatically demand compliance, but that expectation need not become the verb's definition.

**Provisional decision:** Keep `vu / hon / dun`. Test `hon` with advice, warnings, and affectionate requests before narrowing it further. This gives Sindala ordinary emotional precision without adding a new grammatical category.

## 4. A formal petition: respect can travel both ways

**Situation:** A male resident petitions a male registrar for water. This is an invented institutional encounter, not a newly established Thalonian office or legal procedure.

**Resident:** “I hear you, honored registrar. We request water.”

**Registrar:** “I hear you.”

**Resident:** “Then let the request be heard before you ask for another seal.”

| Constructed Sindala | Literal order | Clause represented |
|---|---|---|
| `Cuu hayara vu.` | I · you.PL.M · hear | “I hear you,” addressing one man respectfully. Proposed singular use. |
| `Ninir mu lan.` | water · we · request | “We request water.” |
| `Cuu hayar vu.` | I · you.SG.M · hear | “I hear you,” addressing one man with the ordinary singular. |

`ninir` “water,” `mu` “we,” and `lan` “want, request, ask, desire” come from the export. “Honored registrar,” the seals, and the last request have not been translated.

The registrar has a small but consequential choice: return the resident's `hayara`, or answer with `hayar`. In this proposed usage, reciprocal plural address can recognize the petitioner's standing. A singular reply could sound familiar, reassuring, or condescending depending on their relationship; grammar alone cannot select the interpretation.

**Counter-scene:** An enclave elder uses the respectful plural for a younger healer, while the healer asks to be addressed in the singular. This is an optional scene premise, not a claim about Kaekreaedhel custom. It tests whether chosen respect can reuse a form associated elsewhere with hierarchy.

**Provisional decision:** Keep respectful plural as a social resource. Do not yet choose one generalized plural for all genders, prescribe institutional etiquette, or declare that any community universally accepts or rejects it.

## 5. A joke: the cat acquires a reputation

**Situation:** Two adult brothers find the fish missing. One has food on his sleeve.

**First brother:** “The cat ate the fish, according to accounts.”

**Second brother:** “Which accounts?”

**First brother:** “Accounts are still being collected.”

**Second brother:** “You ate the fish. Your sleeve has provided its account.”

| Constructed Sindala | Literal order | Clause represented |
|---|---|---|
| `Pinuss can cir-nan.` | fish · cat · ACCOUNT-eat | “According to an account, the cat ate the fish.” |
| `Pinuss hayar dal-nan.` | fish · you.SG.M · PAST-eat | “You ate the fish.” |

`pinuss` “fish,” `can` “cat,” and `nan` “eat” are exported. The questions and punchline remain English. This is a proposed joke, not an established Sindala idiom.

**What earns its place:** A received-account form can be borrowed for comic evasiveness. Its ordinary use need not be solemn or legal. The first brother is pretending to have an account; the marker does not magically supply one.

The second brother uses general past while making an inference from a sleeve. That is intentional: `dal-` remains available for claims based on inference. There is no need to invent an inferential suffix just to complete this exchange.

**Provisional decision:** Keep the possibility of playful misuse. Do not broaden `cir-` into a universal marker of uncertainty or make every ordinary sentence specify its evidence.

## A small morphology check: witness and witnesses

The source gives Sigtaara `da'suk` /dæˈsuk/ and Sindala `da'sun` /dæˈsun/, “witness.” The inherited noun plural is `-ɪ`; the Sindala export illustrates its surface descendant `-ya` with `luya`, “dogs.”

Applying the listed sound changes in order to the ancestral singular and a plural assembled **before** those changes gives this limited hypothesis:

| Stage | Singular | Plural |
|---|---|---|
| Ancestral input, stress omitted | `dæsuk` | `dæsuk-ɪ` |
| After `k > c` | `dæsuc` | `dæsucɪ` |
| After word-final `c > n` | `dæsun` | `dæsucɪ` — consonant protected by the suffix |
| After `ɪ > ya` | `dæsun` | `dæsucya` |
| Working romanization | `da'sun` | **`da'sucya` — reconstructed, unattested** |

A later leveled plural could instead be **`da'sunya`**, extending the singular stem. Neither plural is being added to the dictionary. This calculation supports the proposed n/c alternation for one number pair; it does not establish an entire historical paradigm, stress assignment, export engine behavior, or frequency-based exceptions.

My preference is to retain both as competing hypotheses until several other consonant-final nouns have been checked. There is no reason yet to declare “witness” irregular simply because it matters to the setting.

## What this workshop changes in the recommendation

1. **Develop `cir-` as an extension of historical past into received-account past.** Keep ordinary `dal-` neutral about evidence. Recent reports provide the clearest test of the innovation. Present and future reports, tense stacking, negation scope, and questions remain open.
2. **Preserve object-first as a live option.** Subject-first produces useful contrasts in these sketches, but the examples do not justify replacing the inherited default. Next tests should use two human participants, explicit recipients, questions, and subordinate clauses.
3. **Retain the three hearing/attention/compliance verbs.** The dictionary offers more nuance than the original two-way split suggested.
4. **Test respectful address through relationships and reciprocal choices.** Cultural affiliation can affect expectations without determining every speaker's behavior.
5. **Leave pronoun leveling, habitual `duul-`, and legal meanings of memory words pending.** These scenes do not supply adequate tests for those changes. The workshop also does not establish new grammatical agreement or generalize neuter pronouns to people.

The best next grammar decision is the scope of the received-account construction, especially in quoted memories. The best next narrative test is an exchange between two people who share a community but disagree about appropriate usage.

## Source grounding and limits

- [Original evolution proposal](/home/ai/mine/nixxia/languages/sindala/evolution-proposal.html): design starting point; proposal, not canon.
- [Sindala HTML export](/home/ai/mine/nixxia/languages/sindala/source/sindala.export.html), Grammar and Dictionary: transformed pronouns and affix examples; source-attested vocabulary. Its source labels are inconsistent with the profile's display name, as documented in [Sindala notes](/home/ai/mine/nixxia/languages/sindala/notes.md).
- [Sindala save](/home/ai/mine/nixxia/languages/sindala/source/sindala.vulgar.txt): grammar settings and ordered sound rules; noun gender disabled.
- [Sindala lexicon](/home/ai/mine/nixxia/languages/sindala/generated/lexicon.csv) and [Sigtaara lexicon](/home/ai/mine/nixxia/languages/sigtaara/generated/lexicon.csv): lexical meanings and the witness pair.
- [Sindala profile](/home/ai/mine/nixxia/languages/sindala/generated/language.json): imported configuration; ancestral grammar-table forms must not be confused with surface export forms.
- [Kaemisaedhel canon](/home/ai/mine/nixxia/canon/json/species/firgol/sentient/lahkiin/kaemisaedhel.json), `culture.traditions.memory-chants` and `traits.lifespan`: communal preservation of history and law; lifespan 300–340 years. These do not specify the mechanics of transmitted first-person memories.
- [Kaekreaedhel canon](/home/ai/mine/nixxia/canon/json/species/firgol/sentient/lahkiin/kaekreaedhel.json), `culture` and `taboos`: merit-based elders, rejection of imposed gender roles, ancestral-memory practices, and restrictions on memory manipulation. These do not prescribe pronouns or authorize a new memory-transfer custom.
- [Zaetaedhel canon](/home/ai/mine/nixxia/canon/json/species/firgol/sentient/lahkiin/zaetaedhel.json), `culture.society` and `culture.traditions`: Dream-Mothers, oral shellsong, and coral records. These ground the broader memory theme, not a single shared evidential grammar.

The checked lifespans differ among these peoples; no common generation interval or automatic rate of language change is inferred. All dialogue, social interpretations, and historical derivations above remain reviewable proposals.

What changed?

AI action · one call

Action Personality Provider Model Instruction