NACLO 2026 - Problem IUp In Flames

Open round | 5 points | N/A | Problem statement | Official solution | Tags: Phonology

Like 2024F, it's easy to see that pluralization is just the stem plus something else, and we need to understand how the affix is derived and how the root may have changed.

RootPlural
agɨtaagɨtɨka
babaababaika
isuɗamaisuɗamika
kalekalitini
keamakea
kɨnamakɨnamɨka
lɔŋɔtɔmalɔŋɔta
ɲeƙurumotiɲeƙurumotika
ɔʝaɔʝɨtɨnɨ
tsʼɛatsʼɛɨtɨnɨ

Within these few data points, there are many types of changes:

(Note: one could instead mimic the Paiwan problem and say that the affix is "ɨk", "ik", etc. added before the last vowel. The problem is that in most words the ending is "a" anyway, so preserving the last vowel doesn't preserve extra information; in ɲeƙurumoti, the last vowel becomes "a", so we don't want to preserve the "i". Listing out the different types of variation is already a form of analysis—even before we start accounting for the environment of each type!)

The selection of "ɨka" vs "ika" cannot be based on the last few sounds, because "kɨnama" and "isuɗama" both end in "-ama". However, notice that for "ɨka" and "ɨtɨnɨ", all the words contain one of the central vowels ɛ, ɨ, ɔ, and ʉ, while none of the "ika" and "itini" words do. Therefore, we hypothesize that "ɨka" and "ɨtɨnɨ" are variants of "ika" and "itini" that are used when the stem contains a central vowel anywhere (i.e., unbounded vowel harmony).

Next, consider when we use "ika" vs. "itini".

This again cannot be based on the ending, because both groups end in "a" and babaa" and "tsʼɛa" both end in "VV". The second group's words are generally shorter than the first group's, so "itini" is used when the root is 2-syllable, and "ika" is used when the root is 3-syllable or longer. "babaa" is 3-syllable because the "aa" long vowel is considered two syllables.

Finally, when we do use the 'remove "ama" or "ɔma", add "a"' pattern? Note that other "-ama" words like "isuɗama" and "kɨnama" don't do this, so it's not about the ending. However, "isuɗama" lie and "kɨnama" mushroom are both non-human while "keama" soldier and "lɔŋɔtɔma" enemy are humans, so the only rule that could explain this is that this pattern is used for human nouns.

Therefore:

1. arir-a (no central vowel, 3 syllables) → arir-ika
2. bɛf-a (central vowel, 2 syllables) → bɛf-ɨtɨnɨ
3. ɓo-o (no central vowel, 2 syllables) → ɓo-itini
4. atsʼ-a (no central vowel, 2 syllables) → atsʼ-itini
5. ʝak-ama (human) → ʝak-a
6. ʝulam-a (not human, no central vowel, 3 syllables) → ʝulam-ika
7. ɲanɨnɔ-ɔ (central vowel, 4 syllables) → ɲanɨnɔ-ɨka
8. ŋɨmɔkɔka-a (human) ← ŋɨmɔkɔka-ama (not sure whether the ɔa bigram is attested, but the aa bigram surely is, so -ama is better than -ɔma)