el/elp/src/morphology-txb.el

// morphology-txb.el - Tocharian B morphology for the NLG engine.
//
// Implements Tocharian B verb conjugation and noun declension for the ca. 500-1000 CE
// period.  Designed as a companion to morphology.el and called by the engine when
// the language profile code is "txb".
//
// Language profile: code=txb, name=Tocharian B, morph_type=fusional, word_order=SOV,
//   question_strategy=particle, script=latin, family=tocharian.
//
// Tocharian B is an extinct Indo-European language attested in the Tarim Basin
// (modern Xinjiang, China).  Most surviving texts are Buddhist in content.
// The transliteration used here follows standard scholarly convention (Sieg, Siegling,
// Winter); subscript dots are omitted in identifiers but retained in string literals.
//
// Verb conjugation covered:
//   Tenses:   present (class I endings), imperfect not implemented
//   Persons:  first/second/third x singular/plural (slots 0-5)
//   Irregulars: käm- (to come), yä- (to go), wes-/ste (to be),
//               lyut- (to see), wak- (to speak)
//   Canonical map: "be" -> "ste" (3sg) / "wes" (other)
//
// Noun declension covered:
//   Masculine o-stem: nom/acc/gen/dat x sg/pl
//   Feminine ā-stem:  nom/acc/gen/dat x sg/pl
//   (The full 8-case paradigm is simplified to 4 cases for the engine.)
//   Number: singular, plural
//   Articles: none — Tocharian B has no article system
//
// Depends on: morphology.el (str_eq, str_len, str_slice, str_ends_with)

// ── String helpers ─────────────────────────────────────────────────────────────

fn txb_drop(s: String, n: Int) -> String {
    let len: Int = str_len(s)
    if n >= len { return "" }
    return str_slice(s, 0, len - n)
}

fn txb_ends(s: String, suf: String) -> Bool {
    return str_ends_with(s, suf)
}

// ── Person/number slot ─────────────────────────────────────────────────────────
//
// Maps person x number to a 0-based paradigm slot.
//   0 = 1st singular
//   1 = 2nd singular
//   2 = 3rd singular   (most frequently attested)
//   3 = 1st plural
//   4 = 2nd plural
//   5 = 3rd plural

fn txb_slot(person: String, number: String) -> Int {
    if str_eq(person, "first") {
        if str_eq(number, "singular") { return 0 }
        return 3
    }
    if str_eq(person, "second") {
        if str_eq(number, "singular") { return 1 }
        return 4
    }
    // third
    if str_eq(number, "singular") { return 2 }
    return 5
}

// ── Present class I endings ────────────────────────────────────────────────────
//
// Tocharian B present class I (most regular verbs).  The full paradigm is
// complex with active/medio-passive distinction; only the active series is
// implemented here.
//
// Active present class I endings applied to stem:
//   Sg: 1st -au   2nd -ät   3rd -em
//   Pl: 1st -emane  2nd -em  3rd -em
//
// Note: the plural endings show extensive syncretism in the attested corpus.

fn txb_pres1_suffix(slot: Int) -> String {
    if slot == 0 { return "au" }
    if slot == 1 { return "ät" }
    if slot == 2 { return "em" }
    if slot == 3 { return "emane" }
    if slot == 4 { return "em" }
    return "em"
}

// ── Irregular verb tables ──────────────────────────────────────────────────────

// käm- (to come) — suppletive present paradigm
//   kam, käm, käm, kamnäṃ, kamnäṃ, kamnäṃ

fn txb_kam_present(slot: Int) -> String {
    if slot == 0 { return "kam" }
    if slot == 1 { return "käm" }
    if slot == 2 { return "käm" }
    if slot == 3 { return "kamnäṃ" }
    if slot == 4 { return "kamnäṃ" }
    return "kamnäṃ"
}

// yä- (to go) — class IV verb; short stem
//   yau, yät, yäm, ymäṃ, ymäṃ, yänmäṃ

fn txb_ya_present(slot: Int) -> String {
    if slot == 0 { return "yau" }
    if slot == 1 { return "yät" }
    if slot == 2 { return "yäm" }
    if slot == 3 { return "ymäṃ" }
    if slot == 4 { return "ymäṃ" }
    return "yänmäṃ"
}

// wes- / ste (to be) — the copula/existential "be" verb
//   3sg form "ste" is by far the most attested; elsewhere "wes" is used.
//   Simplified paradigm: "ste" for 3sg, "wes" for all other slots.

fn txb_wes_present(slot: Int) -> String {
    if slot == 2 { return "ste" }
    return "wes"
}

// lyut- (to see) — class I regular stem, fully attested
//   lyutau, lyutät, lyutem, lyutemane, lyutem, lyutem

fn txb_lyut_present(slot: Int) -> String {
    if slot == 0 { return "lyutau" }
    if slot == 1 { return "lyutät" }
    if slot == 2 { return "lyutem" }
    if slot == 3 { return "lyutemane" }
    if slot == 4 { return "lyutem" }
    return "lyutem"
}

// wak- (to speak) — class I regular stem
//   wakau, wakät, wakem, wakemane, wakem, wakem

fn txb_wak_present(slot: Int) -> String {
    if slot == 0 { return "wakau" }
    if slot == 1 { return "wakät" }
    if slot == 2 { return "wakem" }
    if slot == 3 { return "wakemane" }
    if slot == 4 { return "wakem" }
    return "wakem"
}

// ── Canonical verb mapping ─────────────────────────────────────────────────────
//
// Maps English semantic labels to Tocharian B citation forms (verbal stems).

fn txb_map_canonical(verb: String) -> String {
    if str_eq(verb, "be")    { return "wes" }
    if str_eq(verb, "come")  { return "käm" }
    if str_eq(verb, "go")    { return "yä" }
    if str_eq(verb, "see")   { return "lyut" }
    if str_eq(verb, "speak") { return "wak" }
    if str_eq(verb, "say")   { return "wak" }
    return verb
}

// ── txb_conjugate: main conjugation entry point ───────────────────────────────
//
// verb:   Tocharian B stem or English canonical label
// tense:  "present" (only present is implemented; others return base form)
// person: "first" | "second" | "third"
// number: "singular" | "plural"
//
// Returns the inflected form.  For unknown verbs the stem is returned as-is —
// the corpus is small enough that most productive verbs are known to the engine.

fn txb_conjugate(verb: String, tense: String, person: String, number: String) -> String {
    let v: String = txb_map_canonical(verb)
    let slot: Int = txb_slot(person, number)

    // ── wes-/ste (to be) ──────────────────────────────────────────────────────
    if str_eq(v, "wes") {
        if str_eq(tense, "present") { return txb_wes_present(slot) }
        return v
    }

    // ── käm- (to come) ────────────────────────────────────────────────────────
    if str_eq(v, "käm") {
        if str_eq(tense, "present") { return txb_kam_present(slot) }
        return v
    }

    // ── yä- (to go) ───────────────────────────────────────────────────────────
    if str_eq(v, "yä") {
        if str_eq(tense, "present") { return txb_ya_present(slot) }
        return v
    }

    // ── lyut- (to see) ────────────────────────────────────────────────────────
    if str_eq(v, "lyut") {
        if str_eq(tense, "present") { return txb_lyut_present(slot) }
        return v
    }

    // ── wak- (to speak) ───────────────────────────────────────────────────────
    if str_eq(v, "wak") {
        if str_eq(tense, "present") { return txb_wak_present(slot) }
        return v
    }

    // ── Regular class I verb ──────────────────────────────────────────────────
    //
    // Apply present class I endings directly to the stem.
    if str_eq(tense, "present") { return v + txb_pres1_suffix(slot) }

    // Unknown tense: return base form
    return v
}

// ── Masculine o-stem declension ───────────────────────────────────────────────
//
// Tocharian B masculine o-stem endings (simplified 4-case system):
//   Singular: nom -e, acc -e, gen -entse, dat -ene
//   Plural:   nom -i, acc -i, gen -entwetse, dat -ene
//
// The o-stem masculine is the most common nominal class.  The nom/acc syncretism
// in the singular is a characteristic Tocharian B feature.

fn txb_decline_masc(noun: String, gram_case: String, number: String) -> String {
    if str_eq(number, "singular") {
        if str_eq(gram_case, "nominative") { return noun + "e" }
        if str_eq(gram_case, "accusative") { return noun + "e" }
        if str_eq(gram_case, "genitive")   { return noun + "entse" }
        if str_eq(gram_case, "dative")     { return noun + "ene" }
        return noun + "e"
    }
    // plural
    if str_eq(gram_case, "nominative") { return noun + "i" }
    if str_eq(gram_case, "accusative") { return noun + "i" }
    if str_eq(gram_case, "genitive")   { return noun + "entwetse" }
    if str_eq(gram_case, "dative")     { return noun + "ene" }
    return noun + "i"
}

// ── Feminine ā-stem declension ────────────────────────────────────────────────
//
// Tocharian B feminine ā-stem endings (simplified 4-case system):
//   Singular: nom -a, acc -a, gen -antse, dat -ane
//   Plural:   nom -ä, acc -ä, gen -antse, dat -ane
//
// The ā-stem feminine shows the same nom/acc syncretism as the masculine.

fn txb_decline_fem(noun: String, gram_case: String, number: String) -> String {
    if str_eq(number, "singular") {
        if str_eq(gram_case, "nominative") { return noun + "a" }
        if str_eq(gram_case, "accusative") { return noun + "a" }
        if str_eq(gram_case, "genitive")   { return noun + "antse" }
        if str_eq(gram_case, "dative")     { return noun + "ane" }
        return noun + "a"
    }
    // plural
    if str_eq(gram_case, "nominative") { return noun + "ä" }
    if str_eq(gram_case, "accusative") { return noun + "ä" }
    if str_eq(gram_case, "genitive")   { return noun + "antse" }
    if str_eq(gram_case, "dative")     { return noun + "ane" }
    return noun + "ä"
}

// ── Gender detection heuristic ─────────────────────────────────────────────────
//
// In Tocharian B the gender of a noun must ideally be supplied by the lexicon.
// Gender was originally inherited from Proto-Indo-European but many neuters
// merged into masculine.  This heuristic defaults to masculine; known feminine
// items should be supplied via the vocabulary layer.

fn txb_detect_gender(noun: String) -> String {
    // No reliable phonological gender markers in Tocharian B stems.
    // Default: masculine.
    return "masculine"
}

// ── txb_decline: main declension entry point ──────────────────────────────────
//
// noun:      Tocharian B noun stem (uninflected base form)
// gram_case: "nominative" | "accusative" | "genitive" | "dative"
// number:    "singular" | "plural"
//
// Returns the inflected form.  Falls back to the bare stem on unknown input.

fn txb_decline(noun: String, gram_case: String, number: String) -> String {
    let gender: String = txb_detect_gender(noun)

    if str_eq(gender, "feminine")  { return txb_decline_fem(noun, gram_case, number) }
    // masculine (default)
    return txb_decline_masc(noun, gram_case, number)
}

// ── txb_noun_phrase: noun phrase builder ──────────────────────────────────────
//
// Tocharian B has no article system.  Definiteness was not grammatically marked
// by a separate morpheme — context and word order served that function.
// This function therefore ignores the "definite" parameter and returns the
// declined noun form directly.
//
// noun:      Tocharian B noun stem
// gram_case: "nominative" | "accusative" | "genitive" | "dative"
// number:    "singular" | "plural"
// definite:  "true" | "false" (ignored — no articles in Tocharian B)
//
// Returns the declined noun form.

fn txb_noun_phrase(noun: String, gram_case: String, number: String, definite: String) -> String {
    return txb_decline(noun, gram_case, number)
}