Do all languages make countability distinctions? Evidence from Nez Perce.

At first glance, Nez Perce looks like a language lacking any correlate of the traditional mass-count distinction. All Nez Perce nouns behave like canonical count nouns in three ways: all nouns combine with numerals without an overt measure phrase, all NPs may host plural features, and all NPs may host adjectives like big and small. I show that Nez Perce nevertheless makes two countability distinctions in noun semantics. A sums-based (cumulativity) distinction is revealed in the interaction of quantifiers with plural; a parts-based (divisiveness) distinction is revealed in certain quantity judgments. Both types of evidence involve complex structures to which language learners likely have little to no actual exposure. I suggest that Nez Perce furnishes a poverty of the stimulus argument in favor of semantic countability distinctions as a language universal