This project investigates how the mass/count distinction determines the structure and interpretation of noun phrases in Tarascan (isolate, Mexico). It focuses on four topics: the distribution and interpretation of plural marking, the semantics of classifiers, the interpretation of the indefinite article, and a description of quantifier structure and meaning. Several languages distinguish mass nouns (e.g. English 'honey') from count nouns (e.g. English 'dog'). Mass nouns do not combine directly with numerals while count nouns do. A cross-linguistic tendency is that nouns that refer to substances are mass nouns, while nouns that refer to delimited objects are count nouns. However, some nouns behave like mass nouns despite the fact that they refer to delimited objects (e.g. English 'furniture'). The grammatical distinction between mass and count nouns is thus not entirely determined by extra-linguistic factors, and to some extent it is arbitrary and subject to linguistic variation.
Tarascan is particularly revealing in the empirical investigation of the mass-count distinction since it distinguishes not two but three classes of nouns: count nouns, like 'wichu' (dog), mass nouns like 'tékwa' (honey), and a third class, which we call "count-mass" nouns (Doetjes 1997). This third class includes items like 'purhú' (pumpkin) and 'thatsïni' (bean). They share some grammatical properties with mass nouns, some with count nouns and in other respects they constitute a class of their own. The tripartite distinction in the mass/count domain calls for a revision of the semantics of some functional categories within the Tarascan noun phrase. For instance, while mass nouns reject plural morphology and count nouns referring to more than one entity need it obligatorily, the precise contribution of plural marking with count-mass nouns cannot be accounted for by canonical theories of plurality. Moreover, classifiers and plural marking in Tarascan can co-occur in a noun phrase, a fact unexpected under well-known typological generalizations. This investigation elucidates the role of classifiers in individuating the denotation of count-mass nouns.
This research is based on semantic and syntactic judgments collected on site by Co-PI Vázquez-Rojas under the supervision of Dr. Chris Collins. The data will be obtained by elicitation and in natural occurring discourse from four consultants in Michoacán, México. Co-PI Vázquez-Rojas will also be involved in training young researchers in Mexico to undertake similar investigations in several indigenous languages spoken around the country.
In order to know what are the universal properties of language, and which aspects are subject to variation, a linguist must look into the structures of as many languages as possible, coming from a wide diversity of language families. In this particular investigation, we are concerned with how languages map the reference to real world objects into different classes of nouns, and how the different ‘mappings’ shape the structure of the noun phrases within each particular language. Our attention was focused on Purepecha, a language spoken in Central Western Mexico. Purepecha is not only not related to any European language but, to our current knowledge, it has no linguistic relatives in its area either. We found out that, while some nouns are assigned to the ‘count’ class (e.g.English hen), the same referent in Purepecha (tsikata) is mapped onto the denotation of a class of nouns that is not necessarily count. A count noun like English hen requires a plural marker to denote pluralities (hen-s), and can combine directly with numerals like ‘three’ in counting contexts. In contrast, a count-mass noun like tsikata in Purépecha can denote individual hens as well as pluralities of them without requiring a plural marker. It can combine directly with numerals like tanimu, ‘three’, but it can also combine with them by means of a classifier, which ensures that reference is being made to individual hens in order to be counted. We arrived to the conclusion that, while the contribution of the plural marker in English is to form (sets of) pluralities from (sets of) individual entities, in Purepecha the plural marker achieves the same effect by ruling out individual entities from the denotation of nouns like tsikata, ‘hen’. We also concluded that the contribution of classifiers in Purepecha is to denote exactly the opposite operation: to rule out pluralities from a set that contains them among individual entities and to yield a set that contains only the later: individual units that can be counted, say, ‘one by one’. No less interesting was finding that these classifier expressions are derived from verbs that predicate existence or location in space in Purepecha, and which also have the distributive semantics that enables them to apply to single entities 'one at a time'. The fact that Purepecha has this class of nouns (which we call count-mass nouns, following previous work from Doetjes (1995) on similar phenomena in Chinese, Dutch and English), and therefore the functional elements that can operate on them (like classifiers and plural markers), entails that Noun Phrases have projections for the different functional heads that host each of the operations that are possible in the language. The resulting structure of the Purepecha Noun Phrase might have a different shape from English Noun Phrases, but the formal definition of the operators is similar from one language to the other. Thus, while English does not have verb-derived classifiers that combine directly with nouns, it does have individualizing expressions (e.g. piece in ‘three pieces of furniture’) whose semantics (although not its syntax) is the same as Purépecha classifiers: to ultimately rule out pluralities from the reference of the noun furniture and later add a plural operator that forms only pluralities of three pieces at a time -in our example. In order to investigate the interpretation of expressions and structures in a language that is not the analysts’ native, we must develop reliable tools to collect judgments from native speakers. These tools come in the form of short experiments that must be clear and replicable. One of the main objectives of this investigation was to develop these tools for the analysis of the interpretation of nouns and number markers in under-represented languages, and to make these tools available to other researchers working with other languages in the area, in order to obtain comparable results. To that aim, Co-PI Vazquez Rojas led a workshop at the National Autonomous University of Mexico (UNAM), in collaboration with the Institute of Anthropological Research (IIA), directed to graduate students and young scholars currently researching different aspects of the semantics of indigenous languages in Mexico.
