Grammar Reference

Preface

Sorantoki is a simple, clear, and friendly auxiliary language designed to make communication easy for everyone. It occupies a middle ground in the auxlang spectrum: more regular and logical than natural languages, but more natural and intuitive than highly engineered or philosophical languages. Its goal is not perfection on paper, but ease of learning and ease of use in real conversations.

The language has a single fixed word order (SVO), a very small sound system, and a grammar built on general, learnable rules with no arbitrary exceptions. Instead of verb conjugations or complicated endings, Sorantoki uses a straight-line chain of small helper words to express tense, aspect, mood, and other meanings. This makes sentences easy to build and even easier to understand.

Although Sorantoki has not yet been tested with learners of different ages or abilities, it was designed with accessibility in mind. The simple sound system, predictable spelling, lack of clusters, optional pronunciation variants, and regular grammar may theoretically support learners with different cognitive or linguistic profiles, including those who benefit from reduced memorization or clearer patterns.

Most importantly, Sorantoki is meant to feel welcoming. Its structure encourages calm, direct, and cooperative communication, and its vocabulary favors everyday meanings over complexity. This grammar presents the language as a practical tool: something you can pick up quickly, use immediately, and grow comfortable with over time.

A note on this edition: This revision follows a full independent stress-test of the previous version. Every reported inconsistency was individually re-examined — some were fixed by correcting the affected words, some were fixed by correcting the rule, and a few were kept as deliberate, clearly labeled design choices rather than "fixed," because forcing them into false regularity would have made the grammar harder to learn, not easier. A full list of changes appears in the Revision Notes near the end of this document.

Your First Word

wasa — pronounced wasa, vasa, watha, or vatha — is the first word most learners discover in Sorantoki. It means "I didn't catch that" or "The message wasn't received clearly." It can refer to unclear hearing, unclear speech, or simply unclear understanding.

wasa is a gentle, non-blaming way to ask someone to repeat, slow down, or clarify. It signals cooperation, not error. In Sorantoki culture, communication is shared responsibility: if something isn't understood, the listener uses wasa without hesitation, and the speaker repeats without feeling corrected.

Use wasa anytime you need clarity. It keeps conversations friendly, patient, and clear — exactly the spirit Sorantoki is designed for.

Welcome to Sorantoki—an auxlang built to be kind to learners.

1. Phonology and Orthography

1.1 Consonants

Sorantoki has ten consonants:

• p – like English "pet"
• t – like English "top"
• k – like English "kit"
• f – like English "fun"
• m – like English "man"
• n – like English "no"
• s – like English "see"; th (lisped s) is also fully correct
• w – pronounced as in English "wise" [w] or as in German/Dutch/Afrikaans "Victor" [v]; both are fully correct
• y – like English "yes"
• r – may be pronounced [r], [ɾ], or [l] depending on the speaker

Note on s, w and r: These are not allophonic rules based on environment, and not dialect clusters. All variants are fully correct, do not change meaning, and belong to the single standard form of Sorantoki. Speakers choose whichever variant is natural for them. Communication always supersedes precision.

1.2 Vowels

a, e, i, o, u

1.3 Diphthongs

ai, au, ei, oi, ou

Only these five vowel-pairs count as single-nucleus diphthongs. Any other vowel-vowel sequence (ea, eo, oe, oa, ua, ui, etc.) is not a legal nucleus and cannot occur within a syllable or across a root boundary — see §1.5 and §1.7.

1.4 Syllable Structure

(C)V|D(N) where N = m or n, D = Diphthong

• Allowed codas: m, n only
• No consonant clusters

1.5 Root-Level Rules

• A syllable nucleus is either one vowel or one of the five listed diphthongs — never any other vowel-vowel sequence.
• If a syllable ends in a vowel or in a nasal coda (m/n), the following syllable must begin with a consonant. This is what prevents vowel hiatus (two vowel sounds running together across a syllable boundary) anywhere in a legal root.
• A word-initial syllable is the only syllable allowed to begin with a bare vowel (e.g. oru, awani, uton), since there is no preceding syllable to create hiatus with.

Scope of these rules: §1.4–1.5 govern content roots — nouns, verbs, adjectives, and derived vocabulary. Sorantoki's small closed class of grammatical particles (es, in, no, ka, su, kon, etc.) is memorized as a unit by every learner in the first lesson regardless of shape, the way function words often are in natural languages, and is not required to fit the content-root template. Exactly one particle, es, ends in a non-nasal consonant (s); this is a deliberate, single, clearly labeled exception for a high-frequency grammatical particle, not a crack in the content vocabulary's regularity. No content root is exempted.

1.6 Stress

Stress always falls on the penultimate syllable of the whole word.

• Prefixes (noi-/noin-, pe-) are counted normally → they can shift the stress rightward.
• Suffixes (-n, -in, -ri, -ta, -pa, -ru, -ti, -ka, -ma) and the plural marker are "invisible" to stress → they never take it and never move it. (Note: -ma, the ordinal suffix, was added to this invisible list in this edition — see §17.2.)

Examples: - ká.sa → ká.san - pú.ka → pú.kan - pon → nói.pon - sero → pe.sé.ro - ú.no → ú.noma (1st; see §17.2 — no longer a stress exception)

1.7 Hiatus Resolution in Root Formation

When two vowels would otherwise land next to each other across a morpheme boundary (and the pair is not one of the five legal diphthongs), Sorantoki resolves it with an epenthetic glide rather than allowing the hiatus: w after a back vowel (o, u), y after a front vowel (e, i), and y by default after a. This is already how several native roots are built (e.g. awani "air," ponau "happy" resyllabify cleanly using this logic), and it is now applied consistently everywhere, including the seven legacy roots that used to violate §1.5 (see Revision Notes):

2. Core Principles

2.1 Word Order

Strict SVO:

1. Subject
2. (Adverbs / Time words / Place words ...)
3. [TAM Serial Verb Chain]
4. (si | kuno) ← reflexive or reciprocal, immediately after the main verb
5. Direct Object / Stative Predicate / Complement
6. (Prepositional phrases, extra objects, etc.)

A topicalized phrase (§20A) may optionally precede all of this, marked by fan.

2.2 Modifiers

Single-word modifiers (adjectives, possessives, adverbs, numbers, demonstratives) always precede the head.

Adverbs of manner, time, and place (including nau "now", supa "already", suno "soon", piko "here", weyo "there", rapiri "quickly", etc.) are head-initial modifiers and immediately precede the TAM serial-verb chain.

Clause-sized modifiers are the one systematic exception: relative clauses (§12) follow their head noun rather than preceding it. This is not arbitrary — it is the same "heavy constituents go last" pattern found in many otherwise strictly head-initial languages (including English itself, which puts adjectives before the noun but relative clauses after it). A single-word modifier precedes; a clause-sized modifier follows. This is a general, principled split, stated once here so it never needs to be treated as a surprise later.

2.3 Word-Class Flexibility

• Any adjective → verb automatically ("be Adj")
• Any noun → modifier ("N-like")
• Light verb turan = causative / "make X"

2.4 Copula

Predicates are zero-marked by default. es is Sorantoki's copula particle. It is optional when the predicate is untensed and unambiguous (plain present-tense zero copula), and it is required whenever the predicate needs to carry tense, aspect, or negation, because those markers need a verb slot to attach to. See §8 and §8B for the full picture — this is one unified system, not three separate ones.

• mi tokoto → "I am a doctor" (zero copula, plain present)
• mi es tokoto! → "I really am a doctor!" (emphatic use)
• mi ya es tokoto → "I was a doctor" (tensed — es required here, see §8B)

In Sorantoki, any root placed directly after a subject with no tense/aspect particle automatically means "be [root]." Therefore tai = "big / be big", pon = "good / be good", tokoto = "doctor / be a doctor", etc.

3. Pronouns and Plurals

Pronouns pluralize with the simple suffix -n added directly to the stem, regardless of what sound the stem ends in. This is not an irregularity in the noun-plural system of §6 — it is a separate, entirely regular rule that belongs to the closed pronoun paradigm. Pronouns are always definite by nature (they always pick out specific referents), so the generic-vs-definite plural distinction described in §6 for common nouns simply doesn't apply to them; they never needed the -in allomorph in the first place.

4. Noun Phrases

4.1 Structure

(Demonstrative) + (Number) + (Modifiers...) + NOUN + (Plural -n, if specific)

4.2 Demonstratives

• ko = this / these (near speaker)
• so = that / those (far or previous)

4.3 Examples

• ko tai kasa = "this big house"
• so pukan = "those books"
• tai tinin = "the big ones" (plural -n or -in attaches only to pronouns or to the generic noun tin "thing"; not directly to adjectives except toun "all")

5. Possession

Possession is marked with te:

Possessor + te + Possessed

• mi te kasa = "my house"
• tan te mon = "their money"
• yan te tai korem = "a person with a big heart"

5.1 Colloquial Spoken Variant

In casual everyday speech, te may be dropped when the possessor is a pronoun or with body parts and kinship terms:

• mi kasa = "my house" (colloquial)
• tu puka = "your book" (colloquial)
• mi moran = "my mouth" (very common colloquial)
• ta peman = "her family" (very common colloquial)

However, the forms with te (mi te kasa, mi te moran, ta te peman) remain the unambiguous standard and are recommended for learners.

5.2 Chained Possession (Deep Possessives)

A chain A te B te C is read strictly left-to-right and left-branching: the leftmost two elements bind first, and that whole phrase becomes the possessor of the next element, and so on. There is exactly one legal parse — this was never actually ambiguous, only undocumented.

• mi te kasa = "my house"
• (mi te kasa) te puka = mi te kasa te puka* = "my house's book" (i.e., "the book of [my house]")

For chains longer than two links, most speakers find a relative clause easier to follow than a long te-chain:

• puka su ta te kasa te mi → "the book, [which is] of the house of mine" (equivalent alternative phrasing using su, §12)

Either strategy is correct; the te-chain is more compact, the su-clause is easier to parse aloud for very deep nesting.

6. The Definite Plural Marker -n / -in

Sorantoki has no indefinite plural. Bare nouns are used for generics and indefinite plurals.

• kasa = "house / houses" (in general)
• mi peron kasa = "I like houses" (any houses, houses in general)
• ruwa in niran = "Rain is cold" (generic)

The suffix -n (after consonant) or -in (after vowel or stem-final n) is added only when the plural referent is definite or specific:

• ko tai kasan = "these big houses" (specific ones)
• so pukan = "those books" (specific books we're talking about)
• mi wita tai tinin = "I saw the big ones" (specific big things, using generic tin + -in)

When the head noun is omitted, -in attaches to tin "thing" or (very informally) directly to the adjective:

• tai tinin = "the big ones"
• pon tinin = "the good ones"

7. The TAM Serial Verb Template

The verb phrase uses exactly six ordered slots. Only Slot 4 may contain 1 or 2 verbs, or nothing. All other slots accept exactly one particle or nothing.

7.1 Quick Reference Line

1 Tense (ya / wi / ∅) • 2 Aspect (in / pa / sa / ∅) • 3 Neg (no / ∅) • 4 Modal/Phase, max 2, left word scopes over right word (kan muso wan fowa sita suto pin na) • 5 Direction/Purpose (to ku ki po / ∅) • 6 Main verb (or es)

7.2 Examples

NOTE: The negation particle no negates predicates, not nouns. It is part of the serial-verb system and never functions as a derivational morpheme.

7.3 Note on Directionals and Phase Markers

to + Verb: → andative ("go and do") when real physical motion to a location is implied - Example: mi to kasa = "I go to the house" (physical motion) - Example: mi to monya = "I go (there) and eat" (physical motion)

na + Verb: → imminent → pure phase marker in verb chains (depending on sentence tense, it can indicate imminent past, future or present)

Examples: - mi na monya = "I'm going to eat / I'm about to eat" (no motion) - tan pin na monya = "They're about to finish eating" - ta pin na nesi = "It's about to finish sleeping"

ku + Verb: → venitive ("come and do") when motion toward the speaker/deictic centre is implied

7.4 Additional Examples

• mi ya wan sero → "I wanted to learn"
• tu wi no kan ku → "You won't be able to come"
• tan pin na monya → "They are about to finish eating"
• mi wan ki puka ta → "I want to give the book to her"
• kasa ya sita kai → "The house started to fall"
• mi wan rapiri pesero po kan toki kon mutan → "I want to be taught quickly in order to speak with everyone"
• ta ya ku wita min → "She came to see us"

7.5 Disambiguation Notes

ki in slot 5 (benefactive "to/for") is always inside the verb chain. Interrogative ki "which" is an argument and appears in noun-phrase position.

• tu wan ki puka ta? = "Do you want to give the book to her?"
• ki puka tu wan ta? = "Which book do you want to give her?"

na root meaning imminent/prospective (modal/phase verb → slot 4) marks the entire predicate. Not to be confused with the separate root nau "now / right now" (used as a regular time adverb before the TAM chain; its adverb form is nauri "imminently").

• mi na monya. → "I'm on the verge of eating" (imminent).
• mi nau monya. → "I eat now / right now."
• mi nau na monya. → "I'm on the verge of eating right now."

7.6 Modal–Phase Scope Ordering (Revised)

Slot 4 holds at most two items back to back: any combination of a modal (kan, muso, wan, fowa) and a phase verb (sita, suto, pin, na), in either order. Sorantoki already uses left-to-right position to mean scope everywhere else (negation no scopes over everything to its right; TAM slots stack outward-to-inward) — Slot 4 now follows the exact same principle instead of an arbitrary fixed order. The leftmost item takes wide scope over the rightmost item.

• mi ya kan sita wita → "I was able to start seeing" (modal scopes over phase: my ABILITY was to start)
• mi ya sita kan wita → "I started to be able to see" (phase scopes over modal: the STARTING was of an ability)

Two modals or two phase verbs still cannot combine (kan wan, sita pin, etc. remain illegal) — the restriction was never about ordering direction, only about not stacking two members of the same subclass, which stays in place.

7.7 Conditional / Irrealis Mood

Sorantoki marks a hypothetical consequence with the modal fowa ("would") in Slot 4, paired with the existing clause-initial conditional introducer karo ("if") for the condition clause:

• karo mi es mon, mi wi fowa orei kasa → "If I had money, I would buy a house."
• karo tu ya ku, mi fowa ponau → "If you had come, I would have been happy." (fowa + zero-marked stative predicate)
• mi fowa pa orei kasa → "I would have bought a house" (fowa + Slot-2 perfect pa gives the past-conditional reading compositionally, with no extra machinery needed)

8. Stative / Adjectival Predicates

8.1 Zero Copula (default and preferred)

• kasa tai → "The house is big"
• mi tokoto → "I am a doctor"
• mon te mi → "I have money" (colloquial "my money")

8.2 Optional Emphasis or Clarity: es

• mi es tokoto! → "I really am a doctor!"
• kasa es tai → "The house really is big"

es is especially helpful when the predicate could be mis-parsed as a modifier or possessive chain:

• puka es in kasa → "The book is in the house" (without es, puka in kasa could briefly be read as "the in-house book")
• puka es ata te mesa → "The book is on top of the table" (without es, it might feel like "the on-top-of-table book")
• mon es te mi → "I have money" (without es, mon te mi is the normal colloquial possessive "my money")

In everyday speech most people simply omit es and rely on context or word order, but adding es never sounds wrong and is warmly welcomed when clarity matters.

8.3 Becoming / Turning X

sita + stative/adjective = "become X, get X"

• kasa sita tai → "The house becomes big"
• ta sita kouren → "She gets angry"

8A. Possession, Existence, and the Verb "Have"

Sorantoki has no dedicated verb meaning "have". Possession and existence are expressed with the same structures used for location and simple predication.

1. Simple possessive (most common in everyday speech)

[possessed] te [possessor] - mon te mi → "my money" - kasa te tu → "your house"

2. Clear existential / possessive "have" (recommended whenever context could be ambiguous)

[thing] es te [possessor] - mon es te mi → "I have money" (lit. "money exists of me") - puka es te ta → "She has a book"

3. Negative possession

[thing] no es te [possessor] or simply no [thing] te [possessor] - mon no es te mi / no mon te mi → "I have no money"

4. Locative and general existential (always uses es for maximum clarity)

[thing] es [location] - puka es in kasa → "There is a book in the house / The book is in the house" - yan es we? → "Where is the person?" - mutan es piko → "There are people here"

Key learner rule: When you are talking about possession, existence, or location (i.e., when you would use "have", "there is", or "be located" in English), always add es. Adding es is never wrong and is the safest, most understandable choice for learners and for anyone who wants to avoid even momentary confusion.

Summary table

8B. Es as the Slot-6 Copula Verb

Every example above with tense, aspect, or negation already puts es right where Slot 6 (the main-verb slot) would go. This edition makes that explicit as its own rule, because it closes a real gap: a bare stative predicate (adjective, location, existence, possession) has nothing for Slot 1–5 particles to attach to unless es fills Slot 6.

• Untensed / unmarked → zero copula is fine: kasa tai ("the house is big"), puka in kasa (colloquial, "the book's in the house")
• Tensed, aspectual, negated, or modal → es fills Slot 6 and the TAM chain attaches normally in front of it:
• kasa ya es tai → "The house was big"
• puka wi es in kasa → "The book will be in the house"
• mi in es in kasa → "I am (right now) at the house" (progressive locative)
• mon no es te mi → "I have no money" (this was already the correct pattern in §8A — it is simply Slot 3 no + Slot 6 es + complement te mi, now stated as a general rule rather than a special case)

This means Sorantoki has, in effect, exactly one copula verb, and it is optional only when no TAM marking is needed on the predicate. There is no separate "existential es," "locative es," and "emphatic es" — it is one particle with one job (heading a stative predicate), used with or without emphasis, with or without TAM marking.

On the in / in overlap: the aspect marker in (Slot 2, progressive) and the preposition in (location) are homophones, but they never collide in practice, because aspectual in is always immediately followed by a real Slot-6 verb (now including es), while prepositional in is always immediately followed by a noun phrase and never by a verb. mi in kasa (no verb after in) can only be the zero-copula locative "I'm in the house"; it is not a well-formed use of aspectual in, because aspectual in with nothing in Slot 6 is not a complete predicate. When in doubt, adding es (§8.2) removes any hesitation either way.

9. Passive / Resultative

9.1 Prefix pe- on Main Verb

The pe- passive is formed only from transitive verbs — verbs that take a direct object in their active form. Sorantoki does not passivize intransitive verbs (motion verbs like kai "fall", ku "come", to "go", or states like nesi "sleep"); there is no patient argument for the passive to promote, so forms like kasa pekai* or ta peku* are not grammatical.

• puka pewita → "The book was seen"
• mon peki ki min → "Money was given to us"
• ta peturan pon → "It was made good / repaired"

9.2 Passive Agent Marking

Passive agents are marked by kon:

• puka pewita kon mi = "the book was seen by me"
• mon peki kon ta = "money was given by her"
• kasa peturan kon seroru = "the house was built by the teacher"

The particle kon, already used for instrument/comitative, extends naturally to passive agents.

Examples: - puka pewita kon mi → "the book was seen by me" - mon pewita kon mi → "money was seen by me" - mon peki kon mi → "money was given by me" - kasa peturan kon tan → "the house was made by them"

10. Questions

10.1 Yes/No

Sentence-final ka?

Yes/No answers normally repeat the verb: Pon ka? – Pon. / No pon.

10.2 Wh-words (in situ)

10.3 Wh-words at Any Depth (no pied-piping needed)

Because Sorantoki wh-words never move from their base position, there is nothing to "pied-pipe" — a wh-word stays exactly where the corresponding answer would go, no matter how deeply it is embedded (inside a possessive, inside a prepositional phrase, inside a relative clause). This resolves questions that look complex in languages with wh-movement:

• tu ya to kasa te ye? → "Whose house did you go to? / To whose house did you go?" (ye sits inside the possessive te-phrase, exactly where "who" would sit in a statement)
• tu toki kon ye? → "Who did you speak with?"
• ta wita puka su ye peturan? → "Did she see the book that who built?" → naturally, "Who did she see the book that [he/she] built?" (ye inside the relative clause)

11. Comparison

11.1 More Than

maso ... pase ...

• ta maso pon pase mi = "She is better than me" (lit. "She more good than me")

11.2 Less Than

nasa ... pase ...

The mirror image of §11.1, built the same way, so no separate antonym root is ever required:

• ta nasa pon pase mi = "She is less good than me"
• ta nasa rapi pase mi = "She is less fast than me" (distinct from saying she is slow — see §15.1 on noi- not being a natural antonym)

11.3 Most

se moso

• ta se moso pon = "She is the best" (lit. "She the most good")

Note: The particle se (the) optionally precedes superlatives for naturalness.

11.4 As ... as ...

sama ... sama

• ta pon sama mi sama = "She is as good as I am"
• ko kasa tai sama so kasa sama → "This house is as big as that house"

12. Relative Clauses

12.1 Marker: su

Relative clauses follow their head noun — see the note on this in §2.2.

• yan su ya wita ko puka → "The person who saw this book"
• yan su mi toki kon → "The person I speak with"

12.2 Use of Resumptive Pronouns (optional)

Zero-resumptive relatives (without a resumptive pronoun) are fully native and preferred in everyday speech. Resumptive pronouns may optionally be included for clarity in long or complex clauses:

Subject relative (zero-resumptive is standard): - yan su wita puka → "the person who saw the book"

Direct-object relative (zero-resumptive is standard): - puka su mi wita → "the book that I saw"

Oblique / possessor relative (resumptive optional for clarity): - yan su mi toki kon → "the person I'm speaking with" (standard) - yan su mi toki kon ta → "the person I'm speaking with him/her" (optional resumptive for clarity) - kasa su min wita → "the house we saw" (standard) - kasa su min wita te ta → "the house that we saw (of it)" (optional resumptive; lit. "the house we saw of it")

The resumptive may be added when the relative clause is particularly long or complex, or when disambiguation is needed.

12.3 Closing Particle wu for Nested Relative Clauses

When a relative clause itself contains another relative clause (or another su-marked phrase), it can become unclear exactly where the outer clause ends and the matrix sentence resumes. Sorantoki resolves this with an optional closing particle, wu, placed at the very end of the (possibly complex) relative clause, right before the matrix clause continues. wu is not needed for simple, single-level relatives — it exists specifically for nesting.

• yan su wita kasa su ta peturan wu ponau → "The person who saw the house that he built is happy."
• Bracketing: [yan su wita [kasa su ta peturan] wu] ponau — wu signals that the entire relative structure attached to yan is now closed, so ponau ("happy") is understood as the matrix predicate about yan, not a continuation of anything inside the relative clause.
• Without nesting, wu is simply dropped: yan su wita puka ponau → "The person who saw the book is happy" (no ambiguity, no wu needed).

13. Prepositions

Only five true prepositions exist (everything else is done by serial verbs, adverbs, or zero):

14. Additional Particles

Notes on e and o: - They are simple coordinating conjunctions, exactly like English "and" and "or". - They are never sentence-initial (except in poetic style). - No comma is required before them, though a comma may be used for extra clarity in long sentences. - They can coordinate full clauses without any additional marking.

15. Derivational Morphology

All roots remain phonologically legal (CV(N) syllables, no clusters, max one diphthong-or-vowel nucleus per syllable, no illegal hiatus — see §1.5–1.7).

15.1 Noi- Prefix (Opposite / Absence / Privative)

noi- (becomes noin- before vowel-initial roots) is a productive derivational prefix that forms the logical opposite, absence, or privative form of any root.

It applies to: - adjectives / statives - nouns - verbs (privative or reversative meaning: "un-do / lose")

Functions: - Creates the logical opposite ("un-X", "non-X", "opposite of X") - Indicates absence ("without X") - Creates privative or reversative verbs ("un-do X", "lose X") - Does not express clausal negation (that function belongs to no)

Examples: - pon → noipon (good → un-good / badish) - tai → noitai (big → un-big / smallish, not big) - sama → noisama (same → not-same / different-ish) - itam → noinitam (alive → un-alive / lifeless, "dead inside") - es → noines (exist → not-exist / nonexistent) - otem → noinotem (find → un-find / lose) - moron → noinmoron (love → un-love / loveless / lacking affection) - kasa → noikasa (house → un-house / structure that is not a house) - norem → noinorem (forget → un-forget / remember) — a clean demonstration of the prefix doing real lexical work instead of coining a brand-new root - nesi → noinesi (sleep → un-sleep / wake up) — same strategy

Notes: - noi- creates logical opposites, not necessarily the natural antonym (Example: noirapi = "un-fast," but "slow" has its own root; likewise §11.2's nasa gives a true "less" comparative without ever needing noi- for that job) - Learners should treat noi- as a regular, universal opposite prefix that can attach to any root without exceptions - noi- is derivational, never part of the serial-verb negation system

15.2 Derivational Suffixes

Roots lack inherent syntactic categories; position determines function (predicate, modifier, adverbial). Suffixes serve two purposes: - Lexical creation: -ta, -pa, -ru, -ti, -ka, -ma build new vocabulary that position cannot express (e.g., seroru "teacher" from sero "learn") - Stylistic clarity: -ri creates explicit manner adverbs; bare roots remain grammatical (e.g., rapi or rapiri "quickly")

Why -ri still exists even though position already marks class: most of the time, bare-root position is unambiguous. But a manner adverb and a following object can both be simple roots sitting next to the verb, and in a language with no case marking, that occasionally creates a real attachment question (is this root modifying the verb, or modifying the object noun right after it?). -ri removes that question outright by tagging the word as unambiguously adverbial, which is why it survives as an optional tool for exactly those moments rather than being redundant busywork.

⚠️ Clarity Rule: Roots should not end with derivational suffixes (-ri, -ta, -pa, -ru, -ti, -ka, -ma) or begin with noi-/noin-, to avoid morphological ambiguity. This edition brings the dictionary itself into full compliance (see Revision Notes): the former roots tinka (now tinam, "live/reside"), kima (now kimo, "six"), and senpa (now sewo, "some") have been reshaped so the rule has no silent exceptions.

16A. Reflexives (self-action)

Sorantoki uses a single invariant particle si to mark that the object (direct or indirect) is the same as the subject.

• Position: si immediately follows the main verb (slot 6) and precedes any overt object or prepositional phrase.
• It never goes inside the TAM chain.
• It can refer to any subject (singular or plural, any person).

Word-order summary: Subject → (adverbs) → TAM chain → Main verb → si → (benefactive phrase, locative, etc.)

Examples: - mi wita si → "I see myself" - tu moron si → "You love yourself" - ta puku si → "He hit himself" - min kuyi si → "We wash ourselves" - mi ya ki mon si → "I gave myself money / I helped myself to money" - tu wi no wan wita si → "You will not want to look at yourself" - ta ya turan si pon → "She made herself (become) good again (= She improved herself)"

Notes: - si is obligatory for true reflexives; simply omitting the object means normal transitive use or impersonal/indefinite object. - Body parts and clothing are normally expressed with colloquial te-drop possession, not si: mi kuyi moran (= "I wash my mouth/face"), tu pake puka (= "You put on your clothes"). Use si only when the action is emphatically directed back at oneself.

16B. Reciprocals (each other)

Sorantoki uses the invariant particle kuno ("one another") in exactly the same position as the reflexive si.

• Position: immediately after the main verb, before any other object or phrase.
• The subject must be plural or conjoined (min, tun, tan, or "mi e tu", etc.).

Examples: - min wita kuno → "We see each other" - tun moron kuno → "You (pl.) love one another" - tan ya toki kuno → "They spoke to each other" - mi e tu wi pokan kuno → "You and I will help each other" - min ya ki puka kuno → "We gave books to each other" - ta e ye peron kuno? → "She and who like each other?"

Notes: - kuno implies mutual action; the verb is still grammatically singular in form because Sorantoki has no verb agreement. - If both reflexive and reciprocal readings are possible, context decides; si forces reflexive only, kuno forces reciprocal only. - To emphasize "each one did it to the other individually", add uno uno or yan yan: tan puku kuno yan yan → "They (each person) hit the others one by one."

Quick reference table:

17. Numbers

17.1 Cardinal Numbers

0 oru, 1 uno, 2 tau, 3 sai, 4 pei, 5 niro, 6 kimo, 7 tepu, 8 naso, 9 reko, 10 senu

11–19: senu [1–9] (e.g. senu uno, senu tau, senu sai … senu reko)

20–90: [2–9] + senu (e.g. tau senu, sai senu, pei senu … reko senu)

21, 32, 45 …: [2–9] senu [1–9] (e.g. tau senu uno, sai senu niro, etc.)

100 = pasan, 1000 = miran

101–999: [1–9] pasan [rest] (e.g. uno pasan, tau pasan sai senu pei)

1000+: [1–9] miran [rest] (e.g. uno miran, sai miran tau pasan niro senu tau)

Larger numbers follow the same pattern and can be extended indefinitely.

17.2 Ordinal Numbers (Now Fully Regular)

Ordinal numbers are formed with the suffix -ma attached only to the last element of the number. -ma is stress-invisible, exactly like every other suffix in §1.6 — the previous edition's separate "ordinals take final stress" rule is gone, because it was the actual crack in the "no exceptions" claim, not a necessary feature. Stress simply falls on the penultimate syllable of the base word, as always.

1st únoma, 2nd táuma, 3rd sáima, 10th sénuma, 11th senu únoma, 20th tau sénuma, 100th pásanma, 1000th míranma, 21st tau senu únoma, 357th sai pasan niro senu tépuma

Examples in sentences: - ta ku táuma = "She came second" - ko tai kasa ta sai pasánma = "This big house is the 300th most expensive" - mi pin na sero rekóma tem = "I'm about to finish the ninth lesson"

17.3 Zero

oru is used exactly like other numbers: - oru yan = "zero people" - óruma = "zeroth" (rare, but grammatical)

The entire system obeys all existing phonological rules (no clusters, only allowed codas, legal nuclei only, etc.) and follows exactly the same stress rule as every other word in the language — including ordinals, now that §1.6 and §17.2 agree.

18. Punctuation & Capitalization

(Everything is optional except when it removes real ambiguity)

18.1 Capital Letters

• All words are written lowercase.
• Exception: Capitalize only proper names of specific people, places, cities, countries, or when a common noun is used as a personal name or direct address.

Examples: - Maria ku. – "Maria is coming." - Tokio es tai. – "Tokyo is big." - Seroru ku! – "Teacher is coming!" ("Teacher" used as the person's name) - mi seroru es Maria. – "My teacher is Maria" (lowercase, because it's a normal noun here)

18.2 Allowed Punctuation Marks (only five – nothing else exists)

• . full stop (end of sentence)
• , comma (short pause, lists, non-restrictive clauses, vocatives)
• ! exclamation mark
• ? question mark
• " " quotation marks (for direct speech or titles)

That's all. No semicolons, no colons, no dashes, no apostrophes, no ellipses.

18.3 How to Write Direct Speech (quotations)

Use normal quotation marks. The sentence inside the quotes follows normal word order.

• ta toki "mi moron tu" → "She said 'I love you'"
• mi wita puka "Sorantoki v6" → "I saw the book 'Sorantoki v6'"

18.4 Commas – When They Are Useful (never required)

• After vocatives: Maria, ku ko!
• In lists: mi peron puka, mon, e kasa.
• Before/after extra information: ko kasa, su ya peturan kon mi tata, es tai. ("This house, which was built by my father, is big.")
• To avoid confusion in long sentences.

18.5 Everything Is Optional

If the sentence is clear without punctuation or capitals, omit them. Correct and equally native:

• mi moron tu
• mi moron tu.
• "mi moron tu" ta toki
• mi moron tu! ta toki

The language is designed to be spoken first. Writing is only a helper.

19. Imperatives

Drop the subject pronoun (2nd person is understood):

• rapiri monya! = "Eat quickly!"
• no kai! = "Don't fall!"
• to kasa! = "Go home!"

For politeness, add the particle ko after the verb: - monya ko = "Please eat."

20. Existential / Locative Patterns

For stating that something exists or is located somewhere, the pattern es + location is commonly used:

• puka es in kasa – "There is a book in the house. / The book is in the house."
• yan es we? – "Where is the person?"
• mon no es te mi – "I have no money." (lit. "Money does not exist of me.")

For tensed, aspectual, or negated versions of these patterns, see §8B, which is where es's full range as a Slot-6 copula is laid out.

20A. Topic-Comment Structure

Strict SVO gives Sorantoki no built-in way to front a non-subject for emphasis — no object-fronting, no clefting. This edition adds one lightweight tool for exactly that: the topic particle fan.

Structure: [Topicalized phrase] fan, [ordinary SVO clause, with a resumptive pronoun filling the gap if the topic is not the subject]

• kasa te tu fan, mi peron ta → "As for your house, I like it." (ta resumes the object gap)
• puka ko fan, mi ya wita ta in kasa → "This book — I saw it in the house."
• ta fan, mi no nora ke ta wan → "As for her, I don't know what she wants." (topicalized subject — no resumptive needed, since the subject slot is simply reused)

This reuses the same resumptive-pronoun strategy already established for complex relative clauses (§12.2), rather than inventing a new mechanism — one general tool, two related jobs.

21. Full Example Paragraph

ko tai kasa te min es se moso pon. so yan su ya ku ki mon tun, maso tan in no kan ku kausa ruwa. ye su tun na wita wen kanto ka? mi wan rapiri pesero po kan toki kon mutan. karo mi es mon, mi wi fowa orei so kasa fan mi peron ta.

Translation: "This big house of ours is really the best. That person who came wanted to give money to you all, but they could not come because of rain. Who are you all going to see when [the time falls]? I want to be taught quickly in order to speak with everyone. If I had money, I would buy that house — I like it."

Appendix: Revision Notes

This section documents what changed in this edition and why. It is not required reading to use Sorantoki — skip it unless you are curious about the editorial process.

This edition responds point-by-point to an independent stress-test of the previous grammar. Each item was individually weighed rather than applied automatically:

1. Seven phonotactic violations (koyeo, oepa, oaru, rua, yua, kui, teara) — fixed by inserting the epenthetic glides Sorantoki already uses elsewhere for hiatus resolution (§1.7), rather than loosening the phonotactic rule itself. rua was additionally found to be a duplicate of uton ("outside") and was retired rather than repaired.
2. Ordinal stress contradicting general stress — fixed by adding -ma to the stress-invisible suffix list (§1.6, §17.2) instead of keeping a special-case rule. Ordinals are now fully regular.
3. Pronoun plural labeled "the exception" — reframed, not changed in form. min/tun/tan were always phonologically motivated and semantically justified (pronouns are inherently definite); the text calling this "the exception" was the actual bug (§3).
4. Slot 4 modal-before-phase restriction — loosened into a general scope rule (§7.6): either order is now grammatical, with the leftmost item taking wide scope, matching how word order already signals scope everywhere else in the grammar.
5. No conditional mood — added fowa ("would") as a fourth Slot-4 modal (§7.7), completing the karo ("if") system without a new slot.
6. "Empty slot" for tensed locative/existential predicates — fixed by formally declaring es the Slot-6 copula verb (§8B), unifying what were previously described as three separate uses of es.
7. Recursive possession called unwieldy/ambiguous — clarified, not changed: the left-branching parse was already the only grammatical one; it just wasn't stated. A relative-clause alternative is now offered for very deep chains (§5.2).
8. Unrestricted passive of intransitives — fixed: §9.1 now explicitly restricts pe- to transitive verbs.
9. No topic/focus marking — added the topic particle fan (§20A), reusing the resumptive-pronoun strategy from relative clauses.
10. No "less than" — added nasa ... pase ... (§11.2), the mirror image of the existing maso ... pase ...
11. Ambiguous maso dictionary gloss — clarified: both senses (comparative "more" and contrastive "but/however") are now listed explicitly with the positional test that tells them apart (§14).
12. Pied-piping gap — clarified, not changed: true wh-in-situ never needed pied-piping; §10.3 spells this out with a worked example.
13. Nested relative clause ambiguity — fixed by adding the optional closing particle wu (§12.3) for nested cases specifically; simple relatives are untouched.
14. -ri suffix "in tension" with position-based class — kept, with its rationale now stated (§15.2): it resolves a real, if occasional, attachment ambiguity between a manner adverb and an adjacent object.
15. Dictionary words that coincidentally look decomposable (tinka, kima, senpa) — fixed by reshaping them to tinam, kimo, sewo, bringing the dictionary into full compliance with its own clarity rule (§15.2).
16. Thin vocabulary in kinship, body parts, food, temporal deixis, mental states, and discourse markers — expanded (see the Dictionary below). Where the existing derivational system could do the job — noinorem "remember," noinesi "wake up," monyapa "food/meal," ko soran "today," yapan soran "yesterday," tepan soran "tomorrow," gendered kin terms built from existing mano/pino — new vocabulary was avoided in favor of reusing what the grammar already provides.
17. Two additional, previously unflagged inconsistencies were caught during this revision and fixed the same way: a duplicate gloss (poran was listed for both "build" and "health" — opora already covers "health" alone, so poran now means "build" only) and an unrelated triple-conflation (meyo covered "water," "drink," and "birth" — "birth" is now the separate root natu, while "water/drink" remains together as a natural, motivated extension). tesan ("wake/decide") was likewise split: it now means "decide" only, with "wake up" reassigned to the regular derivation noinesi.
18. Politeness remaining a single, monolithic level, and -ri's optional status, were reviewed and kept as deliberate design choices consistent with the language's egalitarian aims — not every observation in a stress-test is a defect to remove.