No open-access location is known for this record yet.
No open-access location is known for this record yet.
The literature presents two perspectives on the status of nouns naming certain entities as mass or count. On one view, this choice is predetermined by the nature of the entity named. On the other, this choice is arbitrary, though there may be some regularities or at least tendencies in lexicalization as mass or count. Proponents of arbitrariness cite doublets like mail and letters, leaves and foliage, or rice and lentils, as well as the pervasiveness of Universal Grinder and Packager effects (Pelletier 1979), which suggest that mass/count status is not tied to a lexical item. However, proponents of the alternate view note that even purportedly arbitrary minimal pairs are less arbitrary than they appear. For instance, Zwicky (1997) points out that although petunias and ice plant can both be used to cover areas of ground in a garden, the former is count as it is easily divisible into individual plants, while the latter is mass, precisely because this is not the case: a single plant covers an extended area, while separate plants are hard to distinguish. In this paper we reconcile the two views: we argue that there are cognitive–perceptual and cultural factors which generally determine whether the noun naming an entity will be morphosyntactically mass or count, while recognizing that these factors apply to conceptualizations of entities, so that entities that are open to multiple conceptualizations might indeed be lexicalized as either a mass noun or a count noun. Thus, there is some apparent arbitrariness in classification, but it is constrained by precisely the factors which give rise to mass/count classification in the first place. We also argue that there is a seeming residue of arbitrariness in mass/count assignment that arises for historical reasons, for instance, due to phonological change or meaning drift, as proposed most explicitly by Wisniewski (2010). A key property of our account is recognizing that the process of determining whether a noun is mass or count morphosyntactically is not based on the properties of the real world referent of that noun, i.e. its extension, but on a conceptualization of that referent in terms of its status on a scale of individuation, which encompasses a richer ontology than simply mass vs. count, ranging from substances to aggregates to collectives to individuals. This intermediate conceptual representation of an entity, which is determined by perceptual and cognitive factors interacting with cultural ones, allows for systematicity in the assignment of a morphosyntactic classification to a noun depending on its position on the individuation scale. Available morphosyntactic classifications vary across languages and are reflected in number marking systems. Drawing on data from several languages, we argue that while English shows a two-way morphosyntactic classification (mass vs. count), some other languages make additional distinctions. These classification systems all respect the individuation scale, with a language’s morphosyntactic classes picking out contiguous portions of this scale. Languages always provide distinct morphosyntactic treatments of ‘substances’ and ‘things’—the endpoints of the scale; entities falling between the endpoints are either assimilated to one endpoint morphosyntactically or put in additional classes. The picture proposed here contrasts with previous research on the semantics of mass and count, which while proposing analyses which differ in various respects, assumes a direct relation between the extensional realm and morphosyntactic class (e.g. Quine 1960, Link 1983, Chierchia 1998, 2010, Krifka 2007). On our picture the mapping is not direct, but involves three levels, adding a level of conceptualization, which mediates between the extensional realm and the morphosyntax. We will argue that we can better account for the larger empirical realm, including some of the key properties and puzzles raised in prior work, in this way. Such a picture is not new: in his work, Manfred Bierwisch (e.g. 1983) proposes that a conceptual system mediated between language and the real-world entities it referred, though based on rather different phenomena.
Loading…
Stored only when both papers are in the corpus — 23 citations exist upstream.
Papers that cite the same references — the relational fallback while embeddings are still backfilling.
Imported from OpenAlex (CC0) — nothing here is generated.