Open round | 10 points | 56.76% | Problem statement | Official solution | Tags: Phrase translation
1–5 all contain "ka", so it's probably some functional particle (tense, aspect, etc.; all sentences here are present progressive). 1–3 all begin with "Maliki", so it means "The dog". This means "Marlu" = "The kangaroo" and "Karnta" = "The woman". The third word also begins with "marlu", "karnta", etc., so it's the object. However it has an extra suffix, which is probably case agreement, because each object has a different kind of suffix (-ngka for "marlu" and "karnta"; -rla for "nantuwu" and "kuurlu"). The last word is therefore the verb: "karrimi" = "stand"; "nyinami" = "sit"; "ngunami" = "lie". To recap, here's the sentence structure:
[Subject] ka [object][suffix] [verb]
Therefore:
a. Maliki ka nantuwurla ngunami. = The dog is lying on the horse.
b. The kangaroo is sitting in the school. = Marlu ka kuurlurla nyinami.
Applying this structure to the new sentences, we have:
Therefore "mardukuja" and "karnta" both mean "woman".
Finally we figure out whether to use "-rla" or "-ngka".
It cannot be based on anything semantics, because "mardukuja" and "karnta" both mean "woman" but use different suffixes. One plausible explanation is that it's syllable-based: "-rla" are all 3-syllable or longer, while "-ngka" are all 2-syllable (long vowels are counted as 2 syllable). Therefore in B3, "Wati", "Warlu", "Ngurra", and "Yama" are short and use "-ngka", while "Jaaji" and "Wulpayi" are long and use "-rla".