Invitational round | 15 points | 19.35% | Problem statement | Official solution | Tags: Morphology
Like most semantic matching problems, we are given a starting point: Wik-Mungkan = "to swallow one's words". So either "Wik" or "Mungkan" must be the word for "words". "Mungkan" is nowhere in the dataset, but "wik" is in 22. "wik kiith" and 23. "wik thayan". We have E. "English language" and M. "law" related to "words", so "wik" must be "words". Since "kiith" appears nowhere else, it's likely "English", while "thayan" is whatever that describes the words for law. However there are many words that "thayan" could be: good? strong? trustworthy? So this lead doesn't take us very far. Nevertheless, we know that this language is head-initial.
Let's start from scratch. Most words here are two-part, and they repeat a lot. Let's make a combination table:
| ek | 8. min | 20. thayan | way | Other | |
|---|---|---|---|---|---|
| ma' | 1. ma' ek | 4. ma' thayan | 2. ma' puk pi'an 3. ma' puuy | ||
| 5. mee' | 6. mee' thayan | 7. mee' weep (10. ngak mee') | |||
| 9. ngak | 11. ngak min | 12. ngak way | 10. ngak mee' | ||
| 13. ngangk | 14. ngangk ek | 15. ngangk min | 16. ngangk thayan | 17. ngangk way | |
| 18. puuy | 19. puuy ek | (3. ma' puuy) | |||
| weep | 21. weep thayan | (7. mee' weep) | |||
| wik | 23. wik thayan | 22. wik kiith |
In the data, there's one prominent group: "water" and its 3 derivatives: "fresh water", "spring (water source)", and "undrinkable water" (we also have "crab" that could descend from "water", but since we also have "crab shell", it's more likely its own root). Among the roots here, only "ngak" and "mee'" have exactly 3 derivatives (remember that the language is head initial, so we need the shared word to be the first one, although for "spring" the "water" may be either the modifier or the head). If we choose "mee'", then "fresh water" and "undrinkable water" would need to match "thayan" and "weep" because "mee'" needs to be the head and come first. However, we also have 21. "weep thayan"; what does it mean to combine a positive adjective with a negative one? Therefore, "ngak" is a better choice here.
So we pick "ngak" = "water". It's able to combine with 8. "min" and 5. "mee'". Looking around, we see "good", for which "good"+"water" = "fresh water"; and "eye", for which "eye"+"water" = "spring". Furthermore "eye" is a noun, so it can be a head as well, matching "mee'". Therefore 5. "mee'" = F. "eye", 8. "min" = I. "good", 10. "ngak mee'" = Q. "spring", 11. "ngak min" = H. "fresh water", and 12. "ngak way" = J. "undrinkable water".
If "mee'" is "eye", then we need two more "eye"-related words for "mee' thayan" and "mee' weep". Body parts are often associated with human emotions, personalities, or feelings. Here we also have "heart" which would probably take all the emotions like "happy" and "sad", but "eye" could be associated with sleepiness: A. "awake", P. "sound asleep", T. "tired". We happen to have a 3-way combination between "mee'", "thayan", and "weep": "mee' thayan", "mee' weep", and "weep thayan", so these 3 should match with each other. If "eye" describes sleepiness, then the two "mee'" words would match "tired" and "awake". The remaining word, "sound asleep" would be "weep thayan". It should have head "sleep", so "weep" = "asleep", 21. "weep thayan" = "sound asleep", 6. "mee' thayan" = "awake", and 7. "mee' weep" = "tired" ("asleep eyes"). The best meaning for 20. "thayan" is R. "strong/firm", so that "strong"+"eyes" = "awake" and "firm"+"asleep" = "sound asleep". This also solves the mystery of 23. "wik thayan": "strong words" = "law".
"ngangk" combines with both "good" and "strong". Previously, we already said that "heart" should take all the emotions, and we do have "brave" = "strong"+"heart", and "happy" = "good"+"heart". So 13. "ngangk" = L. "heart", 15. "ngangk min" = K. "happy", and 16. "ngangk thayan" = B. "brave". We have another emotion: "sad", which is presumably "bad"+"heart". Previously we also had the contrast between "fresh water" and "undrinkable water", which is also reasonable "good"+"water" vs "bad"+"water". So "way" = "bad", 17. "ngangk way" = M. "sad". We don't know what 14. "ngangk ek" is though.
| ek | 20. thayan | Other | |
|---|---|---|---|
| ma' | 1. ma' ek | 4. ma' thayan | 2. ma' puk pi'an 3. ma' puuy |
| ngangk | 14. ngangk ek | ||
| 18. puuy | 19. puuy ek | (3. ma' puuy) |
Now we only have two roots left: "ma'" and "puuy", plus "ngangk ek". We have the following words remaining:
We need four words based on "ma'". This has to be "hand"—we can add "trustworthy" there as 4. ma' thayan = "strong hand" = U. "trustworthy". The only atomic root left must be 18. "puuy" = C. "crab". Then 3. "ma' puuy" = "hand crab"; the only matching one is J. "handcuffs". 19. "puuy ek" = D. "crab shell", so "ek" = "shell". So 1. "ma' ek" = "hand shell" = G. "fingernail", and 14. "ngangk ek" = "heart shell" = O. "shoulder blade". Finally, 2. "ma' puk pi'an" = S. "thumb".
We already analyzed that "hand" = "ma'", "bad" = "way", "weep" = "asleep". "ma' puk" only appears in "ma' puk pi'an", so it's something more specific than "hand" but more general than "thumb", so it's "finger".