Early & Unbatched Sentences

Theme: Foundations · 37 sentences.

Unbatched

S001 · legacy la-engineer lo-machine ka-build ta-now The engineer builds the machine.

S002 · legacy la-researcher lo-knowledge-system ka-improve ta-future The researcher will improve the knowledge system.

S003 · legacy la-engineer lo-machine ro-tool ka-build-now The engineer builds the machine with a tool.

S004 · legacy la-engineer lo-machine ka-no-build ta-past The engineer did not build the machine.

S005 · legacy la-engineer lo-machine ka-build-now ? Is the engineer building the machine?

S006 · legacy la-engineer pa-workshop ka-work ta-now The engineer works in the workshop.

S010 · legacy la-to-li lo-[to-su-mu] ka-design ta-past The scholar designed the database.

S011 · legacy go-ra-excess du-machine-fail The machine failed because of excess energy.

S012 · legacy la-na-Derik ne-li-na-Mira vo-close Derek and Mira have a close relationship.

S013 · legacy la-to ka-be ta [la-si ka-ne-ki] Knowledge grows at the time that information connects.

Notes

la-to is now clean. With the agent particle renamed to la, abstract agents no longer collide with the person/agent root li. la-to = "agent: conceptual-pattern" unambiguously.
Attempt 2 is canonical. The causal frame (go/du) naturally handles "when X, Y follows" without a dedicated conditional particle. Formalized in spec/grammar.md as the Causal Conditional structure.
ne-ki = connect (relation + motion → become-related) is the first attested use of the inchoative pattern ROOT + ki, now formalized in spec/morphology.md.
The original attempts were written before the li/la split — the old forms li-to/li-si are replaced above with canonical la-to/la-si.

Verdict: Primitives held. Both grammar gaps identified here are resolved: 1. ~~Particle/root collision for li~~ → resolved: agent particle renamed to la 2. ~~Inchoative derivation undefined~~ → resolved: ROOT + ki = enter state ROOT (spec/morphology.md)

S014 · legacy la zo-li lo si-ko-mu ka-be People create information-storage artifacts.

Notes

zo-li = living-person: the natural compound for "people." With la as the unambiguous agent particle, la zo-li is clean — particle followed by root compound, no collision.
si-ko-mu = signal + containment + artifact. Head-final rule: mu (artifact) is the head; characterized by containing (ko) signals (si). The purpose clause "to store information" is absorbed into the compound — the artifact's nature encodes its purpose.
ka-be = action:generate/create. be (growth/creation) covers intentional fabrication at this stage.
Purpose clause absorbed by compounding: Works here because purpose IS the artifact's nature. For external purpose — "She studies to understand", "He runs to escape" — a general purpose-clause structure is now formalized in spec/grammar.md: wi [clause] introduces the intended outcome. wi is a transparent overlap with root wi (will/intention).

Gaps exposed: 1. ~~Purpose clauses~~ → resolved: wi [clause] formalized in spec/grammar.md. 2. si (signal/representation) is carrying "information" in both S013 and S014. Confirm this is the right root vs. to (conceptual pattern) or to+si (encoded knowledge).

S015 · legacy ta [la-yu ka-to-ki] la-to ka-be When they study, knowledge grows.

Notes

Boundary rule test (pass): ta opens the subordinate clause la-yu ka-to-ki. The matrix clause begins at la-to — the next matrix-level argument marker. The subordinate agent la-yu is already consumed, so la-to is unambiguously matrix-level. Boundary recovered without explicit delimiters. No new machinery needed for this case. ✓
ta vs go doing real work: S013 used go to assert causation — "because information connects, knowledge grows." This sentence uses ta for temporal correlation only — "when studying happens, knowledge grows" — without claiming studying is the cause. The ta/go distinction is not cosmetic; the two frames make genuinely different claims.
yu first corpus appearance: la-yu is the first gloss use of the renamed group pronoun. No collision with root wi (will/intention). Clean.
to-ki = W020 (learning/computation/reasoning). "Study" is the deliberate human variant; to-ki covers it at this stage.

S025 · legacy la-su-mu-li lo-mu ka-se-past The engineer inspected the machine.

Notes

su-mu-li (W002) = engineer / systems-builder. First narrative use of this role compound.
ka-se = action of perceiving = inspect / examine deliberately. se (perception) used as verbal root with action particle. The act is intentional (ka), the mode is perceptual (se). No separate compound word needed; particle + root form is sufficient. Noun form se-ka (examination / inspection) registered as W034.
-past tense marker is placeholder gloss.

S026 · legacy lo-ze de-past It was damaged.

Notes

Agentless-event construction tested: lo-ze de-past = the machine (as patient) underwent a decay event; no agent specified. This is distinct from la-[agent] lo-ze ka-de-past (someone intentionally damaged it). The absence of la + bare lo = passive / agentless event.
Copula gap flagged: "It was damaged" is a predicative adjective in English (state attribution). Tonesu handles this by treating the state as a past event (de-past). A dedicated copula structure for "X is [in state] Y" without implying an event remains undefined. Logged in open-questions.md.
ze as non-person pronoun: first use of ze for an inanimate artifact. Confirms ze covers all non-speaker/non-addressee referents, including objects.
T006 parallel: "The old structure collapsed" (T006) is structurally identical — agentless past event, patient as subject. This sentence gives concrete corpus evidence for that pattern.

S027 · legacy la-ze lo-ze ka-de-be-past wi [ka-ra-be pa-li-pu] She repaired it to restore power to the city.

Notes

de-be = decay-growth = process of reversing decay = repair / restoration. Head-final: be (growth) is the head; de (decay) specifies the starting condition. de-be = growth characterized as recovery from decay. ka-de-be = the act of repairing. Candidate W035; registered in registry/derived.md.
ra-be = energy-growth = energy restoration / power return. ra (energy) + be (growth/increase) = energy increasing. In the purpose clause: the intended outcome is energy being restored to the city.
li-pu = person-collective = city (colloquial short form of li-pu-pa). li (social agent) + pu (plurality) = organized collective of persons. Full compound: li-pu-pa (person-collective-place). In the pa particle position: pa-li-pu = at/to the city. Candidate W036; registered in registry/derived.md.
Agent inheritance (5th corpus confirmation): wi clause omits agent → matrix agent ze inherited. Canonical.
Cross-sentence pronoun tracking: mu (machine, introduced S025) is tracked as ze from S026 onward. First multi-sentence pronoun tracking test in the corpus.

(Running notes on what works and what doesn't)

The SOV + particle system reads clearly for simple sentences
Nested concepts (S010) reveal the need for explicit grouping markers
~~Particle/root collision (li)~~ → resolved by renaming agent particle to la
~~Particle/root collision (ra)~~ → resolved: instrument particle renamed ro
~~Pronoun collisions (se, wi)~~ → resolved: pronouns renamed ze, yu
Causal structures (S011, S013) confirm go/du framing as the canonical conditional strategy — no new particle needed
~~Inchoative derivation pattern missing~~ → resolved: ROOT + ki = enter state ROOT (spec/morphology.md)
Abstract agents (to, si as grammatical subjects) work cleanly with la as the unambiguous agent marker
~~Purpose clauses not formalized~~ → resolved: wi [clause] defined in spec/grammar.md § Purpose Frame
Boundary rule tested (S015–S027): matrix-level argument marker unambiguously signals return to matrix clause in single-level embedding. ne inside subordinate clause confirmed as argument marker, not boundary signal. Nested subordination untested.
Agent inheritance rule (wi-clauses, 5th confirmation at S027): omitting the purpose-clause agent unambiguously inherits the matrix agent. Rule formally stated in spec/grammar.md § Purpose Frame.
~~Compounding vs relative clause (S019): compounding covers noun-naming but breaks for capability expression.~~ Relativizer gap confirmed (S019); imperative gap closed (spec/grammar.md § Imperatives).
Intentionality split (S021–S022): ka on a verb marks deliberate action; bare be marks non-intentional production. ro marks instrument; la marks causal source. Tools use ro; natural sources use la with bare be. Split first attested S021–S022.
Agentless-event construction (S021, S026): lo-[patient] [verb] without la marks a patientive event with no specified agent. First used for instrument-mediated action (S021) and state description (S026). Formal grammar rule pending; see open-questions.
Cross-sentence pronoun tracking (S025–S027): ze successfully tracks a machine referent across three sentences. First multi-sentence pronoun coherence test. ze confirmed as covering non-person referents.
Directional destination gap (S024): pa does not distinguish rest-at from move-toward. di root candidate for destination particle; see open-questions.

S093 lo-di-ki-li zo-ne-go la-mi The navigator is my parent.

Notes

First corpus use of zo-ne-go (W077). zo-ne = living kin relation (base); go = origin/source. Together: the entity that is the living-kin-source of the reference person. "The navigator is [my] living-kin-origin" = the navigator is my parent.
No sex distinction. zo-ne-go is sex-neutral — it covers what English splits into "mother" and "father." Tonesu has no sex or gender primitive, so kinship terms are structurally sex-blind. The language cannot say "my mother" vs "my father" without additional vocabulary that does not naturally derive from the existing primitives. This is a Tonesu cultural result, not a gap: the kinship system treats the parent-child bond as structurally prior to sex.
la-mi in the second participant position. The relational predicate has two participants: lo-di-ki-li (the entity BEING described as a kin-type) and la-mi (the reference anchor = the speaker, from whose perspective the kin-relation is defined). The speaker uses la-mi (not lo-mi) — consistent with the established pattern where first-person always marks its speaker as la ("agent" in the broad sense of: perspective-holder, perspective-anchor). This is a semantic extension of la beyond strict intentional action — la serves as the relational perspective anchor in stative-relational predicates. Flag as new grammatical use pressure.

S094 lo-ra-ki-li zo-ne-du la-mi The pilot is my child.

Notes

First corpus use of zo-ne-du (W078). zo-ne base + du (result/effect) = the entity that is the kin-result of the reference person = child/offspring.
The structural parallel with S093 is exact: lo-[described-party] [kin-type] la-[reference-person]. The kin-type (zo-ne-go vs zo-ne-du) encodes the direction: toward the source (parent) vs away from the source (child). The reference person (la-mi) is always in the same position.
go/du as relational direction qualifiers. Already established as origin/ result for causal relations (go-du frame). Here they extend into kinship: go = the source party in the kin-bond; du = the result party. The extension is semantically transparent — biological parentage is a special case of the origin/result polarity. No new semantic burden on either primitive.

S095 lo-ra-ki-li zo-ne-ru la-mi The pilot is my sibling.

Notes

First corpus use of zo-ne-ru (W079). zo-ne base + ru (unity/oneness /singularity) = kin-unity = a person who shares one origin with the reference person. The ru qualifier marks that the kin-bond is one of sameness: pilot and speaker have the same source. Semantically: ru (unity) applied to the kinship relation = "we are one in terms of origin."
Same structural pattern as S093/S094.
Tonesu sibling is also sex-neutral: no distinction between brother and sister. zo-ne-ru covers both. The language has no mechanism for sex-typed kin terms without an unregistered sex primitive — and the design pressure for such a primitive has not yet appeared.
The sibling relation is symmetric: lo-A zo-ne-ru lo-B ↔ lo-B zo-ne-ru lo-A. The directionality (la/lo) signals only WHOSE perspective anchors the description, not the direction of the bond itself.

S096 [S096a] lo-di-ki-li zo-ne-go la-mi The pilot is the navigator's sibling.

Notes

First multi-step kinship path in corpus. English compresses this to "aunt" or "uncle." Tonesu does not. The path is TWO SENTENCES because the kin-bond is two steps: from speaker → to parent → to parent's sibling.
S096b uses lo-di-ki-li in the reference position rather than la-mi. This confirms the reference participant is not limited to first-person — any patient-marked entity can anchor a kinship predicate.
KEY FINDING: Tonesu cannot compress this into a single noun phrase without a relative clause. "My parent's sibling" used as a single noun phrase argument (e.g. "I spoke to my parent's sibling" in one sentence) requires describing the pilot as "the entity who has kin-unity with the entity who has kin-origin status toward me" — which is a relative clause embedding. NO relative clause introducer exists in Tonesu grammar yet. This is the first corpus moment that definitively requires one. (See NM-001 in open questions.)
Tonesus handle this by stating the path declaratively, in sequence. The meaning accumulates across sentences. This is fully coherent as a communicative strategy — and from a Tonesu perspective, the longer path is actually more informative, not a workaround. There is no pressure to abbreviate it.

S097 lo-di-ki-li zo-ne-go-re la-mi The navigator is my ancestor.

Notes

First corpus use of zo-ne-go-re (W080). zo-ne-go (parent, W077) + re (recurrence/cycle) = the recurringly-iterated kin-source = ancestor. An ancestor is a parent of a parent of a parent... the re marks that the origin step recurs up the lineage.
zo-ne-go-re is non-specific about generation distance. It says "some iterated number of kin-steps back." The navigator could be the speaker's grandparent, great-grandparent, or any more distant ancestor. The compound captures the property of being IN the ancestral line, not the specific degree.
Generation count is not expressible with current vocabulary. To say "two generations back" (grandparent specifically) requires either: (a) a numeral system (integers, "two" — not yet in corpus); or (b) a two-sentence path (lo-A zo-ne-go la-mi / lo-B zo-ne-go lo-A — B is A's parent; A is my parent; therefore B is my grandparent). The language has nu (quantity/number) and re (recurrence) but no cardinal counting system yet. This is the first sentence that definitively requires one. (See NUM-001 in open questions.)

S098 lo-ra-ki-li zo-ne-du-re la-mi The pilot is my descendant.

Notes

First corpus use of zo-ne-du-re (W081). zo-ne-du (child, W078) + re (recurrence) = iterated kin-result = descendant. Parallel structure to W080 (zo-ne-go-re); the re adds the iteration.
The compound set is now symmetric at every level:
parent / child: zo-ne-go / zo-ne-du
ancestor / descendant: zo-ne-go-re / zo-ne-du-re
sibling: zo-ne-ru (no directional pair — the sibling bond is symmetric)

S099 lo-di-ki-li zo-ne-go lo-ra-ki-li — lo-to-fe-ki-li zo-ne-go lo-di-ki-li The navigator is the pilot's parent. The keeper is the navigator's parent. [Therefore: the keeper is the pilot's grandparent — by two-sentence path.]

Notes

Two-sentence path for "grandparent." No single-sentence equivalent is available without a numeral (to quantify generation distance) or a relative clause (to embed a kin-noun phrase). Two declarations, two steps.
This sentence also shows zo-ne-go with non-first-person reference: both lo-ra-ki-li and lo-di-ki-li serve as the reference anchor (the la slot is replaced by lo for third-person parties). Confirms: the reference anchor position takes la for first/second person and lo for third-person — consistent with the existing case marking paradigm.
English compresses "grandparent" into a single word. Tonesu requires TWO sentences. This is not awkward — it's structurally honest about the fact that "grandparent" IS a two-step path. The compression in English is a convenience label, not deeper information. Tonesu makes the path explicit.

S100 [S100a] lo-di-ki-li zo-ne-go la-mi Someone is the pilot's child.

Notes

The "cousin" equivalence — three sentences. Each step is one sentence. English "cousin" = parent's sibling's child = three structural facts. Tonesu states all three. No compression is attempted; none is needed.
S100c uses unspecific lo-li (patient: a person, unspecified) in subject position. This is the first use of bare li (person/social agent) as an existential rather than a specific named individual. The sentence asserts that SOME person is the pilot's child without naming them.
The compression trap. English speakers encountering this system might want to register a compound zo-ne-ru-du = kin-unity-result = "child of one's sibling" — the Tonesu analogue of "cousin." This would be legitimate as a compound. The question is whether Tonesus WANT it. Given what the language has shown: they probably don't. A compressed label drops the path information. "My cousin" in English is ambiguous about which parent's sibling's child — first cousin vs second cousin, once-removed vs twice-removed — precisely because the label compresses away the path. Tonesu deliberately does not compress it.
If Tonesu culture ever needs a kinship term for "the child of your assigned crew partner" or similar social bond (not biological kin), that would be built from wi-ne or ne-li (intentional-relation or relational-person), NOT from zo-ne. The zo-ne family is strictly biological lineage.

S101 [lo-ne-ra ka-se] lo-mu ne-mi ka-be Build me the machine that perceived the resonance.

Notes

First corpus use of head-final clause modification. The structure: [lo-ne-ra ka-se] = a clause meaning "perceived the resonance" — with no la-[agent] present. The gap is in the la position. The head noun lo-mu (the machine) follows directly, filling that gap semantically: the machine IS the thing that perceived the resonance. In the main clause, lo-mu functions as patient of ka-be (build it for me). The lo marking on the head is its MAIN-CLAUSE role, not its inner-clause role (where it was the agent).
This directly satisfies S019 Version B — the target that was deferred at S019 because no relativizer existed. "A machine that already perceived the resonance" is incidental predication (not "a sensing-device by class"). The relative clause form handles it; the compound form se-mu (W from corpus) handles the CLASS meaning. The distinction is now expressible.
Key parsing fact: lo-ne-ra ka-se without a la agent is incomplete as a standalone sentence — an agent is expected but absent. This incompleteness is what signals to the parser that a gap exists and that the following lo-mu fills it. Standalone sentences in Tonesu require an agent or a licensed drop (imperative or established speaker-drop). Here neither license applies, so the gap is obligatorily filled by the following head.

S102 [la-di-ki-li ka-se] lo-ne-ra no-to-fe The resonance that the navigator perceived is uncertified.

Notes

Patient gap. The inner clause la-di-ki-li ka-se has an agent (navigator) but no patient. The gap is in the lo position. Head lo-ne-ra fills the patient gap semantically: the navigator perceived the resonance. In the matrix, lo-ne-ra is the subject of the stative predication no-to-fe (uncertified). Particle lo is consistent: the resonance is the patient/subject of both the inner clause (gap) and the matrix predication.
Contrast with agent gap (S101): in S101, the gap is la-[head] (agent) and the head exits with lo (patient in matrix). In S102, the gap is lo-[head] (patient) and the head exits with lo (also the role for a stative subject in the matrix). The critical difference: in S101 the head's inner role ≠ outer role (agent → patient); in S102 the head's inner role = outer role (patient → lo-subject). The lo marking is purely the matrix role; the inner role is whatever it needs to be to fill the gap.
Matrix is stative (no ka). lo-ne-ra no-to-fe = "the resonance [is] uncertified" — a predication with no action particle. This confirms the matrix predicate does not need to be an action clause; any predication form that can take a lo-head argument is a valid matrix for a head-final relative NP.

S103 la-mi lo-si-ki [lo-ne-ra ka-se] ne-mu ka-be I gave the code to the machine that perceived the resonance.

Notes

Same inner clause as S101 ([lo-ne-ra ka-se], gap = agent) but the head exits with ne-mu (recipient) instead of lo-mu (patient). The head's MATRIX role is recipient, not patient. This demonstrates the key principle: the particle on the head encodes the head's role in the MAIN CLAUSE, independently of its role in the inner clause. The head-final position carries role information for the outer structure; the inner structure's role is inferred from the gap.
Disambiguation: a naive reading of ne-mu alone would be "to the machine" (recipient). In context, [lo-ne-ra ka-se] ne-mu is unambiguous: the bracket precedes and modifies ne-mu, so ne-mu is the head of that NP. The inner clause provides the relative description; ne-mu provides the matrix slot.
Confirmed: the role-particle on the head is the external role anchor. The same head mu (machine) can exit as lo-mu (S101, patient), ne-mu (S103, recipient), or — by analogy not yet tested — la-mu (agent of matrix) or pa-mu (location).

S104 la-mi lo-si-ki [zo-ne-go la-mi] ne-li ka-be I gave the code to the person who is my parent.

Notes

Stative inner clause — no ka. zo-ne-go la-mi is a kinship predication without an action particle: "kin-origin [from] my perspective." The gap is in the lo position — the entity BEING described as "my parent" is absent from the clause and filled by the head ne-li.
la-mi appears twice in this sentence: once as the action agent of the matrix (la-mi ... ka-be = "I give") and once as the relational anchor inside the kinship clause (zo-ne-go la-mi = "my kin-origin"). These are structurally distinct: the outer la-mi anchors the matrix action; the inner la-mi anchors the relational definition from which the kin-type is determined. The brackets disambiguate which level each la-mi belongs to.
la as perspective anchor confirmed in embedded context. In the inner clause, la-mi is NOT the agent of any action — it is the reference entity from whose standpoint the kin-relationship is defined. This is the clearest corpus case yet that la marks "the participant from whose perspective the clause is evaluated," rather than "the intentional agent of an action." The construction is grammatically well- formed by applying the head-final modification rule to a stative kinship predication.

S105 la-mi lo-si-ki [zo-ne-ru [zo-ne-go la-mi] lo-li] ne-li ka-be I gave the code to my parent's sibling.

Notes

This sentence directly resolves the S096 NM-001 trigger. S096 first exposed that "my parent's sibling" could not be embedded as a single argument. This sentence embeds it as a recipient argument of a matrix clause.
Two-level nesting:
Inner NP: [zo-ne-go la-mi] lo-li = "the person who is my parent." Gap = lo-li (patient slot — the entity being described as my parent). Head exits as lo-li because inside the outer relative clause it serves as the reference entity for the sibling relationship (the thing you are sibling TO, which takes lo marking per the existing kinship predicate structure: lo-X zo-ne-ru lo-Y). So this is a case where the inner head's inner-clause role IS lo — not a role change.
Outer NP: [zo-ne-ru [inner NP] lo-li] ne-li = "the person who is sibling to [the person who is my parent]." The outer gap = lo-[outer-head] (the entity being described as the sibling). Head exits as ne-li (recipient in the matrix).
Readability check: the sentence is longer than any prior single-sentence kinship expression, but each bracket level is structurally well-formed and the nesting is two levels, not three. Two levels of head-final embedding appear to be routinely parseable. Three or more levels would approach the limit of working memory; Tonesu Tonesus likely use the multi-sentence path strategy for anything above two levels.

S106 [lo-si-de ka-to-ko] lo-mu ne-mi ka-be Build me a machine that remembers past signals.

Notes

Direct satisfaction of S019 Version B — the target that opened the relativizer gap in S019. The full original target was "Build me a system that remembers past queries." lo-si-de = signal-decay = prior/past signal data. ka-to-ko = action of containing knowledge = remembering/retaining. The machine is the agent of ka-to-ko (gap = la-mu position). Head exits as lo-mu (patient of ka-be = build).
Comparison with to-ko-mu (memory-device compound): S019 Version A used lo-to-ko-mu ne-mi ka-be = "build me a memory device." That expressed a CLASS property — the device is BY DESIGN a memory-keeping device. S106 expresses an INCIDENTAL property — build me the specific machine that currently has the record of past signals. Both are valid; both are now expressible; they mean different things.
to-ko-mu: "a memory device" = class membership
[lo-si-de ka-to-ko] lo-mu: "the machine that remembers X" = incidental predication about a specific entity This is the distinction the relativizer was always needed for. It is now in the language.
si-de (signal + decay = past signal) is its first appearance as a compound. Signal = si; decay = de; signal-that-has-undergone-decay = prior/historical signal record. Compositionally clean; not registered as it is an ad hoc compound in this context and may have narrower or broader application depending on future uses.

S107 la-mi se [lo-mu ne-ra no-to-fe] I have a signal-level reading: the machine's resonance is uncertified.

Notes

First corpus use of positive perceptual floor la-mi se [prop]. C007 B4 established la-mi no-se [prop] (perceptual floor negated). This is the positive form: "I hold at perceptual / signal level: [prop]." Consistent with the modal scale formalized in spec/grammar.md § Epistemic Modality. The positive scale is now fully attested: la-mi se (S107), la-mi si (C001 A3, C006 A4), la-mi to (C007 B1).
The embedded proposition [lo-mu ne-ra no-to-fe] is a stative predication — no ka. "The machine has uncertified resonance." The machine holds the lo patient position; ne-ra is a compound predicate; no-to-fe is the state modifier. The entire prop is a stative-predication clause, not an action clause. This confirms that the epistemic modal frame takes propositions of any predicate type (action and stative alike) — consistent with prior cases but here the first example where the prop is purely stative.
Stative relative test (companion): the same inner clause lo-mu ne-ra no-to-fe can be tested as a relative modifier. By the head-final rule: [ne-ra no-to-fe] lo-mu = "the machine with uncertified resonance" — a stative-predicate relative clause with no ka, gap = lo-mu (patient of predication). The head exits as lo-mu. This works by the same rule as action-predicate relatives (S101–S103): the gap-and-head structure is unchanged; only the predicate type differs. Stative modifying clauses are confirmed possible. (This secondary form is noted here rather than given its own sentence because S107 already tests the stative prop type through the epistemic frame, and the relative form is compositionally derived with no additional findings.)

S163 lo-li vo The person is [socially] valued.

Notes

Type 1 predication (lo-X Q): patient-slot subject, quality predicate. X occupies state Q as a current condition. The claim is contingent — social recognition can change.
Contrastive Pair 3 (with S162): Same noun (li = person), same quality (vo = value). Different slot → different claims:
la-li vo → person holds worth — intrinsic, structural property of being a person
lo-li vo → person is in a valued state — contingent social esteem
The distinction maps onto a meaningful philosophical line: intrinsic worth cannot be revoked by social consensus; social esteem can. Tonesu captures this via the la/lo slot distinction rather than with separate vocabulary.
P-GP-001 resolution criterion (three contrastive pairs) now met. Pair 1: lo-X Q (C002 A3 lo-si-mu ru) vs la-X Q (S030 la-to-su-mu vo) — pattern introduced. Pair 2: lo compound (S033 lo-pa ha-vo) vs la compound (S034 la-ra-ki-mu ha-fe) — compound predicates in both positions. Pair 3: S162 vs S163 — minimal pair, same noun and quality, different slot, different claim.

S190 la-mu lo-si se no lo-to The machine detected the signal, not the pattern.

Notes

Second corpus attestation of no as intra-clause contrast coordinator. First use: S090 (lo-to-re-su be no lo-wi-to = "followed the doctrine, not the plan"). Here: lo-si se no lo-to = "perceived the signal, not the pattern." Different predicate (se vs be), different domain (machine perception vs. mission instruction), parallel structure (patient contrast in both).
no in the contrast coordinator position negates only the following constituent (lo-to), not the predicate. The machine did perceive; what it perceived was lo-si, not lo-to. The predicate se is unaffected.
Patient contrast: lo-si (raw signal / signal data) vs. lo-to (conceptual pattern). In the Tonesu epistemic hierarchy, machines register signal-level data (si); only persons or sufficiently organized systems extract conceptual patterns (to). This sentence embodies that ceiling not as a doctrinal statement but as a bare perceptual report.
Bare se (no epistemic frame): The machine's perception is reported without an outer la-mu se [prop] frame. That frame would attribute a personal epistemic stance to the machine, which doesn't apply. Inanimate agents simply perceive; they don't hold epistemic positions. Verb as bare predicate; machine as grammatical agent.
With two attestations (S090, S190), the no contrast coordinator is ready to formalize. Structure confirmed: [established constituent] no [rejected alternative], with both constituents grammatically parallel (lo- patients in both cases so far).

Verdict S190: clean. no as contrast coordinator between parallel lo- patients is stable with se predicate. Generalizes beyond be. Ready to add as Level 4 in § Negation.

S191 la-si-su ko [la-ra-ki-li se lo-pa-ra] The archive holds: the pilot perceived the field.

Notes

Second corpus attestation of ko with a clausal complement. First: C005 A2 (la-to-re-su ko [lo-mu zo-to] — doctrine contains propositional belief). Here: la-si-su ko [la-ra-ki-li se lo-pa-ra] — archive contains propositional observation record. Different container type (si-su archive vs. to-re-su doctrine), different propositional content (perceptual report vs. doctrinal claim), parallel grammatical structure.
The bracketed clause [la-ra-ki-li se lo-pa-ra] is a full propositional complement: agent (la-ra-ki-li), predicate (se), patient (lo-pa-ra). The archive contains this proposition as an entry — not the pilot, not the field, but the propositional record of the perception event.
Container semantics confirm the extension principle: ko asserts a containment relation between a container and its contents. The container type determines what register of contents makes sense:
Physical container (ko-mu) → physical objects
Archive (si-su) → records and logged propositions
Doctrine (to-re-su) → doctrinal claims In all three cases, ko is the same structural predicate. No special form is needed for propositional complements.
la-si-su ko lo-si-mu (S070) was the nominal-patient form for the same archive. That sentence says "the archive holds records" (a type of nominal contents). This sentence says "the archive holds that X happened" (a specific propositional entry). The nominal form describes the archive's character; the propositional form reports a specific log entry. Both are valid.
With two domain-varied attestations (C005 A2, S191), the ko-clausal complement pattern is ready to formalize.

Verdict S191: clean. la-si-su ko [embedded-prop] follows the same la-X ko Y pattern with full clause as Y. No strain. Generalizes from doctrine to archive. Rule may now be written.

S192 la-yu lo-to-su be ta-ti-be lo-se-ka The team consolidated their knowledge after the inspection.

Notes

Mirror of S092's event-anchor construction. S092: ta-ti-de lo-ka-wi-de = "before the mission departure" (ti-de = past/prior → before). S192: ta-ti-be lo-se-ka = "after the inspection" (ti-be = future/following → after). Both use ta ti-[dir] lo-[event] with a lo--marked event as the temporal anchor. Different domain (mission preparation vs. maintenance/inspection cycle), different directional compound (ti-de vs ti-be).
ta-ti-de lo-X = before X; ta-ti-be lo-X = after X. The directional component of the time compound (de = decay/prior, be = growth/following) sets the relative position; the lo-[event] bracket provides the anchor event. The structure is compositionally transparent: "at the [prior/following] time [of] [event]."
be as matrix predicate: lo-to-su be = "knowledge grows" = they consolidated their knowledge. The group's organized understanding expanded as a result of the inspection. Same be as all growth predicates.
se-ka (W034, perceptual-agent/examination process) used as the anchor event. The inspection is the event they are temporalizing against. lo-se-ka = patient:inspection — the examination in its nominal role as a temporal anchor.
The two examples together confirm the pair: ta-ti-de lo-X / ta-ti-be lo-X = before-X / after-X. The construction is general: any lo-[event] noun phrase can serve as the anchor.

Verdict S192: clean. ta-ti-be lo-[event] = "after [event]" works identically to S092's ta-ti-de lo-[event] = "before [event]." Event-anchor temporal construction confirmed as a general pattern. Ready to add rule to grammar.md.

S193 go [lo-zo-su ne lo-zo-su] du [lo-zo-su be] du [lo-zo-su be] Because living-system A is in relation with living-system B, A grows and B grows.

Notes

Second corpus attestation for P-PR-002. S189 demonstrated the gap in physics (lo-ra-su ne lo-ra-su, force-systems). S193 repeats the same construction in biology (lo-zo-su ne lo-zo-su, living-systems in mutualistic coupling). Two domain-distinct sentences, same structural failure.
zo-su = living-system / organized organism. Compositional parallel to ra-su (force-system, used in S189): zo (living thing) + su (organized structure) = organism understood as an organized biological system. Not separately registered; compositionally transparent.
The gap is identical to S189: (1) Double du forces sequential reading — living-system A grows first, then living-system B — but mutualistic symbiosis is a continuous coupled exchange with no sequencing between the two growth events. (2) ne describes the static relation enabling the coupling; it does not encode the mutual-transformation event itself. The language approximates the outcome (both grow because they are in relation) but cannot say the coupling is the event in which both transform simultaneously.
Why go...ne...du...du is the best available: A single du frame cannot take two agent-predicate pairs without coordinate structure. Using ne as inner coordinator (du [lo-zo-su be ne lo-zo-su be]) would read ne as a static copula between two growth-states, not as simultaneous mutual growth. The double-du chain is the least-distorting approximation available.
Biology removes the physics-abstraction objection. One might argue that physical simultaneity is a limiting edge case that natural language need not cover. Mutualistic symbiosis is a macroscopic, everyday-scale fact about organisms. If the language cannot express it without sequentiality distortion, the gap is not domain-specific.
The go...ne...du...du chain is grammatically well-formed. The insufficiency is ontological, not grammatical — the same conclusion as S189.
Retroactive note (March 2026 — zi admitted as primitive 34): The intended concept is now expressible as lo-zo-su zi-zo lo-zo-su — "living-system A and living-system B undergo mutual biological transformation." zi-zo (W106, proposed) is the compressed correct form. S193 is retained as the gap record.

Verdict S193: gap confirmed (2/3). The P-PR-002 simultaneous-mutual-transformation gap is present in both physics (S189) and biology (S193). One further domain-varied wrong-reading is needed to reach the three-attestation threshold. Social domain (negotiation / mutual exchange) is the remaining candidate.

S194 go [la-li ne la-li] du [la-li wi-to-ki] du [la-li wi-to-ki] Because delegate A is in relation with delegate B, A shifts position and B shifts position.

Notes

Third corpus attestation for P-PR-002. S189: physics/force-systems. S193: biology/mutualistic symbiosis. S194: social/negotiation. Three-wrong-reading threshold is reached.
wi-to-ki: wi (intention) + to (conceptual pattern) + ki (inchoative) = entering a new intentional-conceptual configuration = revising one's stance. Head-final: the change-event (ki) in the domain of intentional-pattern (wi-to). Compositionally transparent; no registration needed for the test.
Wrong reading (primary decomposition): Double du forces sequencing — delegate A revises first, then delegate B revises. But negotiation revision is a coupled simultaneous exchange: both parties update their positions at the exchange moment, which is a single event, not two sequential ones. The language has no way to mark the two outcomes as concurrent.
Alternative decompositions tested:
Two causal chains (go [la-li wi-to] du [la-li wi-to-ki]. go [la-li wi-to] du [la-li wi-to-ki].): Loses the coupling entirely. Two independent internal revisions; the mutual-response is gone. Wrong reading: one event becomes two separate events.
Relational + causal hybrid (la-li ne la-li go wi-to-ki): Which party changes? Both changing is lost. Wrong reading: mutual change becomes undifferentiated one-way causation from the relation.
ne between shift-events (du [la-li wi-to-ki ne la-li wi-to-ki]): ne between two events reads as a static relation between two change-states, not as the assertion of their simultaneity. Wrong reading: interaction becomes static relation at the event-result level.
Repair test: ne-ki (relation-change) = the relation itself shifts — not both parties. wi-to-ne = a static shared-stance state, not the dynamic revision event. No compound of existing primitives asserts "these two events are simultaneous and mutually constitutive" without that compound being exactly the new concept needed. Repair fails.
The social domain removes the natural-law objection. Physics and biology involve simultaneous coupling as a matter of natural law (field interactions, metabolic mutualism). Negotiation is chosen, deliberate, one of the paradigm cases of intentional mutual transformation. If the gap persists here, it is not an edge case of physics formalisms — it is a structural gap in the language's treatment of coupled events generally.
Same go...ne...du...du canonical structure as S189 and S193; results are directly comparable. The wrong reading is the same in all three domains.
Retroactive note (March 2026 — zi admitted as primitive 34): The intended concept is now expressible as la-li zi-ka lo-li — "the delegates engage in mutual exchange / transaction." zi-ka (W105, proposed) is the compressed correct form. S194 is retained as the gap record showing why zi was needed.

Verdict S194: gap confirmed (3/3). Three-attestation threshold reached. The P-PR-002 simultaneous-mutual-transformation gap is persistent across physics, biology, and social domains. All alternative decompositions fail. Repair by compounding fails. A primitive slot evaluation is now triggered. See open-questions.md for the updated entry.

S365 (la-ze lo-wi-fe be-si) , du la-ze lo-wi-fe-no-fe wi (She reportedly said: let rules increase.) — Therefore she intends unlimited-governance [rules-without-boundary].

Notes

(la-ze lo-wi-fe be-si) = the evidential frame wraps the source's actual reported position. be-si = growth-signal = signaling in favor of increase. The () marks this as reported — the speaker is not personally asserting this as the source's exact words.
wi-fe-no-fe = rules-without-boundary = wi-fe (W100) + no-fe (without limit). Default right-branching: wi modifies [fe-no-fe] = will applied to [boundary-without-limit] = unlimited imposition = totalitarianism. This bare-asserted conclusion is outside () — the speaker presents it without reservation.
Grammar doing the work: the real position is ()-framed; the distortion is bare-asserted. In Tonesu, anything outside () claims first-person certainty. The structural contrast — (reported premise) , du [bare confident conclusion] — makes the gap visible: the du (result/therefore) asserts the conclusion follows directly from what was only a reported claim.
Corrected form: (la-ze lo-wi-fe be-si) , (du la-ze lo-wi-fe-no-fe wi) — place both premise and conclusion inside (). Both are reported/uncertain. The result is the weaker, epistemically honest version.

Generated from registry/entries.yaml.