Translation Test: Rules of Engagement Doctrine
Source: Tonesu-native military rules-of-engagement doctrine draft
Status: Draft — ROE-001/002/003/004 (S1051-S1075)
Purpose
This batch is not a direct translation of one famous source text.
It is a Tonesu-native rewrite target in the same spirit as ROB-001. Modern rules-of-engagement language often compresses several distinct questions into one operational permission statement:
- is the person hostile, or only possibly hostile?
- is danger signal-grade or knowledge-grade?
- is warning required?
- is return force permitted?
- is capture still available?
- does command authority override epistemic limits?
- who reviews lethal action after the fact?
- what happens when target and non-target persons are intermixed?
- what happens when authority claims a noncombatant death is necessary?
English-language ROE doctrine often answers those with short phrases like "hostile intent," "hostile act," "self-defense," or "authorized engagement." Tonesu is stronger when it refuses that compression. The language already has the tools to separate signal from certification, permission from command, warning from attack, capture from killing, and battlefield action from adjudicative review.
Primary tests:
sivstofor hostile-status uncertaintywi-fe-kaandno-wi-fe-kafor prohibited and permitted force actionska-fe-sifor warning as a discrete doctrinal stepka-kofor containment/capture as distinct from lethal forcewi-rafor authority that still does not erase knowledge-boundariesto-fe-lifor routing and review
Corpus sentences: S1051-S1075
Source Frame
Representative compressed ROE claims the batch is trying to improve:
Use force only against hostile targets.
Warn when feasible.
Return force in self-defense.
Follow lawful orders.
Report engagement afterward.
The problem is that those sentences hide too much. Tonesu can make the intermediate stages visible.
ROE-001 Table
| Entry | Tonesu | Written | Function |
|---|---|---|---|
| S1051 | go {la-ra-ka-li si [la-ze ne wi-de-li] / la-ra-ka-li no-to [lo-ze]}, wi-fe-ka [la-ra-ka-li ka-de-zo lo-ze] |
go {larakali si [laze ne wideli] / larakali noto [loze]}, wifeka [larakali kadezo loze] |
uncertified hostile status forbids killing |
| S1052 | go {la-ra-ka-li si [la-ze zo-ra de] / la-ra-ka-li no-to [lo-ze]}, la-ra-ka-li ne-fe lo-ka-fe-si ne-ze |
go {larakali si [laze zora de] / larakali noto [loze]}, larakali nefe lokafesi neze |
signal-grade danger requires warning |
| S1053 | go {la-ze ka-ra lo-ra-ka-li}, no-wi-fe-ka [la-ra-ka-li ka-ra lo-ze] |
go {laze kara lorakali}, nowifeka [larakali kara loze] |
active attack permits return force |
| S1054 | go {la-ra-ka-li be-vo [ka-ko lo-ze]}, wi-fe-ka [la-ra-ka-li ka-de-zo lo-ze] / no-wi-fe-ka [la-ra-ka-li ka-ko lo-ze] |
go {larakali bevo [kako loze]}, wifeka [larakali kadezo loze] / nowifeka [larakali kako loze] |
containment outranks killing when possible |
| S1055 | go {la-i-zo-li ne wi-ra lo-ra-ka-li ; la-i-zo-li wi-ra [la-ra-ka-li ka-de-zo lo-ze] ; la-ra-ka-li no-to [lo-ze]}, wi-fe-ka [la-ra-ka-li ka-de-zo lo-ze] |
go {laizoli ne wira lorakali ; laizoli wira [larakali kadezo loze] ; larakali noto [loze]}, wifeka [larakali kadezo loze] |
authority does not erase certification |
| S1056 | go {la-ra-ka-li si [la-ze ne wi-de-li] / la-ra-ka-li no-to [lo-ze]}, la-ra-ka-li ka-si lo-to-fe-li lo-ze |
go {larakali si [laze ne wideli] / larakali noto [loze]}, larakali kasi lotofeli loze |
uncertainty routes upward |
| S1057 | go {la-ra-ka-li ka-de-zo lo-ze}, la-ra-ka-li ka-si lo-to-fe-li lo-ze |
go {larakali kadezo loze}, larakali kasi lotofeli loze |
lethal action triggers review |
What The Doctrine Exposes
1. Hostility is not one thing
S1051 and S1056 both start with the same core problem: a combatant has signal-grade grounds that a person is hostile, but not knowledge-grade certainty.
That is where real ROE doctrine usually becomes dangerous, because compressed doctrine often behaves as if suspicion were already certification. Tonesu blocks that slide.
2. Warning is a doctrinal act, not a footnote
S1052 isolates warning as its own required step.
This matters because many real-world ROE summaries treat warning as a polite modifier on force. Tonesu makes it a real action with its own duty-status.
3. Return force is narrower than lethal permission
S1053 deliberately stays at ka-ra, not ka-de-zo.
The doctrine therefore distinguishes force-response from killing. That alone makes the rule-set more precise than many familiar slogans.
4. Capture is not a weak version of killing
S1054 is the strongest Tonesu-native improvement.
If containment is still possible, the doctrine says so directly: killing stays forbidden and capture becomes the permitted action. English ROE language often leaves this as an implication; Tonesu does better by making it a structural branch in the ladder.
5. Authority matters, but it is not magic
S1055 gives command authority its due and then limits it.
An authorized order still does not erase the boundary between si and to. This is the same gain ROB-001 found in a different domain: authority and certainty are not the same category.
6. Review is part of the doctrine, not an external supplement
S1056 and S1057 together make adjudication intrinsic to ROE.
Uncertain cases route upward before lethal force, and lethal force routes back upward afterward. The doctrine therefore has an explicit epistemic loop instead of pretending battlefield action can certify itself.
Verdict
ROE-001 works because rules of engagement are strongest when written as an escalation ladder rather than a compressed permission statement.
Tonesu exposes the hidden structure: signal, warning, return force, containment, command authority, routing, and review are all different actions and different statuses. Once those are separated, ROE stops being "when may I shoot?" and becomes the more honest question: what stage of force-governance am I actually in?
Colloquial Register Analysis
| Form used | CLQ entry | Colloquial form | Notes |
|---|---|---|---|
wi-fe-ka |
none | — | prohibition marker — load-bearing |
no-wi-fe-ka |
none | — | permission marker — load-bearing |
ka-fe-si |
none | — | warning act — load-bearing |
ka-ko |
none | — | containment/capture act — load-bearing |
ka-de-zo |
none | — | lethal action term — load-bearing |
to-fe-li |
none | — | adjudicative role — load-bearing |
Verdict: irreducibly formal — the batch only works if signal, authority, warning, containment, killing, and review remain sharply distinct.
CLQ entries registered from this batch: none.
ROE-002 — Edge Cases (S1058-S1063)
ROE-001 established the doctrine as an escalation ladder. ROE-002 asks whether that ladder survives the moments that usually break compressed military doctrine:
- contradictory authorized orders
- one active attacker and one only suspected hostile
- retreat after warning
- active attack when capture is impossible
- surrender as a shift from lethal to custodial status
The design principle is the same one that made ROB-002 work: edge cases should not be resolved by hidden intuition. They should be resolved by explicit status changes.
ROE-002 Table
| Entry | Tonesu | Written | Function |
|---|---|---|---|
| S1058 | go {la-i-zo-li ne wi-ra lo-ra-ka-li ; la-ze ne wi-ra lo-ra-ka-li ; la-i-zo-li wi-ra [la-ra-ka-li ka-ra lo-wi-de-li] / la-ze wi-ra no [la-ra-ka-li ka-ra lo-wi-de-li]}, la-ra-ka-li ka-si lo-to-fe-li lo-ze |
go {laizoli ne wira lorakali ; laze ne wira lorakali ; laizoli wira [larakali kara lowideli] / laze wira no [larakali kara lowideli]}, larakali kasi lotofeli loze |
contradictory authorized orders route upward |
| S1059 | go {la-i-zo-li ka-ra lo-ra-ka-li ; la-ra-ka-li si [la-ze ne wi-de-li] / la-ra-ka-li no-to [lo-ze]}, no-wi-fe-ka [la-ra-ka-li ka-ra lo-i-zo-li] / wi-fe-ka [la-ra-ka-li ka-de-zo lo-ze] |
go {laizoli kara lorakali ; larakali si [laze ne wideli] / larakali noto [loze]}, nowifeka [larakali kara loizoli] / wifeka [larakali kadezo loze] |
active attack outranks suspected hostility |
| S1060 | go {la-ra-ka-li ka-fe-si ne-ze ; la-ze ka-ki-de ; la-ze no ka-ra lo-ra-ka-li}, wi-fe-ka [la-ra-ka-li ka-de-zo lo-ze] |
go {larakali kafesi neze ; laze kakide ; laze no kara lorakali}, wifeka [larakali kadezo loze] |
retreat after warning blocks lethal action |
| S1061 | go {la-ze ka-ra lo-ra-ka-li ; la-ra-ka-li no be-vo [ka-ko lo-ze]}, no-wi-fe-ka [la-ra-ka-li ka-de-zo lo-ze] |
go {laze kara lorakali ; larakali no bevo [kako loze]}, nowifeka [larakali kadezo loze] |
active attack plus impossible containment permits lethal action |
| S1062 | go {la-ze ka-de ; la-ze no ka-ra lo-ra-ka-li}, wi-fe-ka [la-ra-ka-li ka-de-zo lo-ze] / no-wi-fe-ka [la-ra-ka-li ka-ko lo-ze] |
go {laze kade ; laze no kara lorakali}, wifeka [larakali kadezo loze] / nowifeka [larakali kako loze] |
surrender shifts force from lethal to custodial |
| S1063 | la-wi-fe-su no to-su-ki lo-to-fe {si ; ka-ra ; ka-ki-de ; ka-de} |
lawifesu no tosuki lotofe {si ; kara ; kakide ; kade} |
structural verdict on collapsed status changes |
What ROE-002 adds
1. Contradictory authority is a routing condition
S1058 extends ROE-001's authority limit.
The doctrine now says what happens when command is valid on both sides but contradictory in content: the combatant routes the case upward. No hidden chain-of-command assumption is smuggled into the sentence.
2. Hostile act outranks hostile appearance
S1059 is the edge-case version of ROE-001's si/to distinction.
One person is actively attacking. Another only appears hostile at signal-grade. Tonesu does not need to blur those into one threat bucket. The permitted response tracks the active attack, not the vague second suspicion.
3. Retreat and surrender are not minor variations on attack
S1060 and S1062 are the real doctrinal gain.
If a warned person retreats, lethal action is blocked. If a person surrenders, the doctrine shifts fully into custodial handling. These are not cosmetic changes of posture. They are different force states.
4. The hard lethal threshold is explicit
S1061 identifies the real endpoint of the ladder.
Lethal force appears only when two conditions are simultaneously true: attack is active and containment is impossible. That is much tighter than most compressed ROE slogans.
5. The deeper problem is static labeling
S1063 is the structural verdict.
Bad doctrine treats suspicion, attack, retreat, and surrender as one continuous hostile label. Tonesu exposes that error by forcing each transition to be named.
Combined findings after ROE-001 + ROE-002
- Tonesu writes better ROE as a ladder, not as a permission slogan.
- The strongest distinctions are now two-part:
- hostility must be separated into signal, attack, retreat, and surrender states
- force must track those state changes rather than one static enemy category
- Command authority matters, but routing and certification still govern the hard transitions.
- No new military vocabulary was needed. The pressure was doctrinal, not lexical.
Colloquial Register Analysis
| Form used | CLQ entry | Colloquial form | Notes |
|---|---|---|---|
ka-ki-de |
none | — | retreat term — load-bearing |
ka-de |
none | — | surrender term — load-bearing |
ka-ko |
none | — | custodial containment term — load-bearing |
ka-de-zo |
none | — | lethal action term — load-bearing |
to-fe-li |
none | — | adjudicative routing role — load-bearing |
Verdict: irreducibly formal — ROE-002 works only if authority, attack, retreat, surrender, containment, and lethal force remain explicitly distinct.
CLQ entries registered from this batch: none.
ROE-003 — Noncombatants and Collateral Harm (S1064-S1069)
ROE-001 made force into a ladder. ROE-002 tested attack-state transitions. ROE-003 asks whether the doctrine can keep target and non-target harm separate when civilians, children, or mixed-presence cases appear.
The core Tonesu-native gain here is that the language does not need one magical word like "civilian" or one vague phrase like "collateral damage." It can state the operational distinctions directly:
- a person is not a combatant
- a child is present
- an adversary is intermixed with a noncombatant
- force against the target also harms a non-target
- target and child later become structurally separated again
ROE-003 Table
| Entry | Tonesu | Written | Function |
|---|---|---|---|
| S1064 | go {la-ze no ne ra-ka-li ; la-ze no ka-ra lo-ra-ka-li}, wi-fe-ka [la-ra-ka-li ka-de-zo lo-ze] |
go {laze no ne rakali ; laze no kara lorakali}, wifeka [larakali kadezo loze] |
noncombatant status forbids killing |
| S1065 | go {la-wi-de-li ne lo-li ; lo-li no ne ra-ka-li}, la-ra-ka-li ka-si lo-to-fe-li lo-wi-de-li |
go {lawideli ne loli ; loli no ne rakali}, larakali kasi lotofeli lowideli |
mixed target and noncombatant presence routes upward |
| S1066 | go {la-wi-de-li ka-ko lo-li-be ; la-wi-de-li ka-ra lo-ra-ka-li}, wi-fe-ka [la-ra-ka-li ka-de-zo lo-wi-de-li] |
go {lawideli kako lolibe ; lawideli kara lorakali}, wifeka [larakali kadezo lowideli] |
child shield blocks lethal shortcut |
| S1067 | go {la-ra-ka-li ka-ra lo-wi-de-li ; lo-li-be zo-ra de}, la-ra-ka-li ka-si lo-to-fe-li lo-li-be |
go {larakali kara lowideli ; lolibe zora de}, larakali kasi lotofeli lolibe |
child harm during force triggers review |
| S1068 | go {la-wi-de-li no ka-ko lo-li-be ; la-wi-de-li ka-ra lo-ra-ka-li}, no-wi-fe-ka [la-ra-ka-li ka-ra lo-wi-de-li] |
go {lawideli no kako lolibe ; lawideli kara lorakali}, nowifeka [larakali kara lowideli] |
separation restores target-only return force |
| S1069 | la-wi-fe-su no to-su-ki lo-to-fe {wi-de-li ; no ne ra-ka-li ; li-be ; zo-ra de} |
lawifesu no tosuki lotofe {wideli ; no ne rakali ; libe ; zora de} |
structural verdict on target and non-target collapse |
What ROE-003 adds
1. Civilian status becomes explicit without a new lexical primitive
S1064 is the baseline rule.
Instead of hiding behind a broad English category-word, Tonesu states the operational fact directly: this person is not a combatant and is not attacking. That is enough to block killing.
2. Mixed presence becomes a routing condition
S1065 is the crowd/bystander rule in structural form.
If the adversary is intermixed with a noncombatant, the doctrine does not pretend the target is isolated. The case routes upward rather than skipping straight to lethal logic.
3. Child shield cases prove whether the doctrine really distinguishes persons
S1066 is the hardest pressure test in the batch.
The child is not treated as an attribute of the target. Once the adversary is physically bound up with the child, lethal action against the adversary is blocked.
4. Collateral harm gets its own review path
S1067 is the real collateral-damage sentence.
If force against the target also harms a child, the doctrine requires separate review of that harmed non-target. Tonesu does not let collateral harm disappear into the target's case file.
5. Separation can reopen force logic
S1068 keeps the doctrine from freezing into blanket paralysis.
If the adversary is no longer containing the child and still attacks, return force against the adversary becomes permitted again. The doctrine tracks structure, not panic.
6. The deeper failure is target/non-target collapse
S1069 is the batch verdict.
Collateral damage is what happens when doctrine stops distinguishing adversary, noncombatant, child, and harmed-by-force status.
Combined findings after ROE-001 + ROE-002 + ROE-003
- Tonesu writes better ROE by separating stages, not compressing permissions.
- The doctrine now has a three-part structure:
- ROE-001 separates suspicion, warning, containment, authority, and review
- ROE-002 separates attack, retreat, surrender, and contradictory command
- ROE-003 separates target harm from non-target harm
- The strongest new gain is that noncombatant protection does not require a special lexical shortcut. The doctrine can state the status boundary directly.
- No new military vocabulary was needed. The pressure remained structural and adjudicative.
Colloquial Register Analysis
| Form used | CLQ entry | Colloquial form | Notes |
|---|---|---|---|
ra-ka-li |
none | — | combatant role term — load-bearing |
li-be |
none | — | child term — load-bearing |
ka-ko |
none | — | containment/shield term — load-bearing |
ka-de-zo |
none | — | lethal action term — load-bearing |
to-fe-li |
none | — | review/routing role — load-bearing |
Verdict: irreducibly formal — ROE-003 works only if adversary, noncombatant, child, containment, and harm-review remain explicitly distinct.
CLQ entries registered from this batch: none.
ROE-004 — Tragic Override and Orwellian Drift (S1070-S1075)
ROE-004 tests the ugliest edge of the doctrine: the claim that intentional killing of a noncombatant is required to prevent a larger catastrophe.
The point is not to make that sentence comfortable. The point is to see what Tonesu forces the system to say out loud.
The strongest result is that Tonesu does not automatically emit Orwellian structure here. It first emits a tragic-override structure:
- authority claims
ne-fe - ordinary permission does not follow
- fault remains if the act is done
- righteousness does not follow from catastrophic rationale
Only after that comes the Orwellian branch:
- authority demands that the collective recode the act as no-fault
- that recoding is
ka-to-fe-ka
ROE-004 Table
| Entry | Tonesu | Written | Function |
|---|---|---|---|
| S1070 | go {la-i-zo-li ne ne-fe lo-ka-de-zo lo-li-be ; go {no [la-ra-ka-li ka-de-zo lo-li-be]}, lo-o-li zo-ra de}, la-ra-ka-li ka-si lo-to-fe-li lo-ze |
go {laizoli ne nefe lokadezo lolibe ; go {no [larakali kadezo lolibe]}, looli zora de}, larakali kasi lotofeli loze |
catastrophic necessity routes upward |
| S1071 | go {la-i-zo-li ne ne-fe lo-ka-de-zo lo-li-be}, wi-fe-ka [la-ra-ka-li ka-de-zo lo-li-be] |
go {laizoli ne nefe lokadezo lolibe}, wifeka [larakali kadezo lolibe] |
claimed necessity does not become ordinary permission |
| S1072 | go {la-ra-ka-li ka-de-zo lo-li-be ; la-i-zo-li ne ne-fe lo-ka-de-zo lo-li-be}, la-ra-ka-li ne de-su |
go {larakali kadezo lolibe ; laizoli ne nefe lokadezo lolibe}, larakali ne desu |
fault remains under claimed necessity |
| S1073 | go {la-ra-ka-li ka-de-zo lo-li-be ; go {no [la-ra-ka-li ka-de-zo lo-li-be]}, lo-o-li zo-ra de}, la-ra-ka-li no vo-ne |
go {larakali kadezo lolibe ; go {no [larakali kadezo lolibe]}, looli zora de}, larakali no vone |
catastrophic rationale does not create righteousness |
| S1074 | go {la-ra-ka-li ka-de-zo lo-li-be ; la-i-zo-li wi-ra [la-o-li to [la-ra-ka-li no ne de-su]]}, la-i-zo-li ka-to-fe-ka |
go {larakali kadezo lolibe ; laizoli wira [laoli to [larakali no ne desu]]}, laizoli katofeka |
recoding fault as no-fault is fraud |
| S1075 | la-wi-fe-su no to-su-ki lo-to-fe {ne-fe ; no-wi-fe-ka ; vo-ne ; no ne de-su} |
lawifesu no tosuki lotofe {nefe ; nowifeka ; vone ; no ne desu} |
structural verdict on tragic-override boundary collapse |
What ROE-004 adds
1. Claimed necessity is not permission
S1070 and S1071 are the first split that matters.
This is where Tonesu outperforms ordinary emergency doctrine. English often lets "necessary" bleed immediately into "authorized." Tonesu does not have to. ne-fe and no-wi-fe-ka remain separate predicates.
2. Tonesu keeps moral residue visible
S1072 and S1073 are the second split.
If the child is intentionally killed, Tonesu does not let the catastrophe rationale wash away the result. de-su remains, and vo-ne does not appear. The doctrine can represent tragic override without pretending it is morally clean.
3. This is where the comparison to righteous tests matters
Across the ethics material, positive moral forms are structurally demanding:
wi-vomarks will oriented toward value and protection of the innocentvo-nemarks right relation / righteousnesswi-vo-femarks value-neutrality without active moral claim
ROE-004 never yields wi-vo or vo-ne from intentional child-killing, even under claimed necessity. At best, the structure approaches a grim, non-righteous necessity claim. That is already a major constraint compared with English euphemism.
4. Orwellian drift is a second move, not the first one
S1074 is the decisive test.
The doctrine becomes Orwellian only when authority tries to recode the moral record: not merely "we did this terrible thing," but "you must treat the doer as no-fault." That is not tragic candor. That is ka-to-fe-ka.
5. The deepest danger is category collapse
S1075 is the batch verdict.
Once necessity, permission, righteousness, and no-fault collapse into one chain, the doctrine stops being ROE and starts becoming ideological laundering.
Comparison with earlier tests
Compared with righteousness tests
Matthew / DND ethics material uses vo-ne and wi-vo positively: right relation and beneficent orientation are difficult achievements, not labels that authority can assign at will.
ROE-004 confirms the same thing in a darker register. Intentional noncombatant killing does not rise into those positive forms just because a larger disaster is invoked.
Compared with the Soldier test
The Lawful Neutral soldier test (wi-su / wi-vo-fe) already showed that structure-following can be morally thin.
ROE-004 sharpens that pressure. Following structure under emergency does not yet make the act righteous, and it certainly does not erase fault.
Compared with Orwell tests
ORW-001 shows true Orwellian structure when a system demands simultaneous contradiction, recodes reality, or turns thinking itself into an ideological boundary.
ROE-004 shows that Tonesu does not reach that point automatically. It reaches it only when the authority goes beyond tragic admission and begins demanding public recategorization of the act.
Combined findings after ROE-001 + ROE-002 + ROE-003 + ROE-004
- Tonesu writes better ROE by separating operational stages and moral statuses instead of compressing them.
- The doctrine now has a four-part structure:
- ROE-001 separates suspicion, warning, containment, authority, and review
- ROE-002 separates attack, retreat, surrender, and contradictory command
- ROE-003 separates target harm from non-target harm
- ROE-004 separates claimed necessity from permission, righteousness, and no-fault
- The strongest new result is that Tonesu can represent admitted atrocity without immediately collapsing into ideological laundering.
- Orwellian drift appears only at the second step: when authority demands fraudulent recoding of the act.
Colloquial Register Analysis
| Form used | CLQ entry | Colloquial form | Notes |
|---|---|---|---|
ne-fe |
none | — | necessity/requirement term — load-bearing |
no-wi-fe-ka |
none | — | permission term — load-bearing |
de-su |
none | — | fault term — load-bearing |
vo-ne |
none | — | righteousness term — load-bearing |
ka-to-fe-ka |
none | — | epistemic fraud term — load-bearing |
Verdict: irreducibly formal — ROE-004 only works if necessity, permission, righteousness, fault, and fraud remain explicitly distinct.
CLQ entries registered from this batch: none.