Thought-provoking Stanford talk yesterday: ”What Can Be Ground? Noun Type, Constructions, and the Universal Grinder” by Alex Djalali, Scott Grimm, David Clausen, and Beth Levin (all of Stanford).
Abstract (for the upcoming Berkeley Linguistics Society meetings):
The thought experiment known as the “Universal Grinder”, whereby count nouns within particular morphosyntactic contexts surface as mass expressions (e.g. There is apple in the salad [and the classic There is dog all over the road]), plays a central role in argumentation in the mass/count literature. Its “universality”, however, has not been systematically investigated, though its operation has been observed to be restricted. We present the results of a sentence rating task which investigates the grinding operation across five nominal types and three constructions. Grinder sentences were on average given quite low acceptability ratings (2.33/7; SD 1.81) compared to filler sentences (5.68/7; SD 1.85). Although the different constructions did not reliably influence acceptability, acceptability was affected by noun type, from worst to best:
group terms < shape, simplex artifacts, complex artifacts < animals < foodstuff.
Foodstuff and animals are likely more felicitous in grinder sentences due to the dual life of food nouns as natural entities or food stuff and of animate nouns as natural entities or their flesh.
The results argue for a more nuanced view of the semantics of the mass/count distinction than the prevalent contextualist view countenances, as on this view a noun may be realized with either mass or count morphosyntax, with the choice determined largely by context; consequently, interactions between noun type and constructions of the type examined here are unexpected. The differential felicity of grinder sentences reflects the relationship between situational context and noun type; these in turn reflect intrinsic nominal properties and world knowledge conventionally associated with the referents of nouns.
From the same research group, an LSA paper (January 2011) “Between count and mass: Furniture and other functional collectives” by Scott Grimm and Beth Levin (slides in pdf format here), about some “misfit” non-countable nouns like furniture, mail, luggage, change, jewelry, and ammunition, which they argue “show similarities with the core count and mass nouns, but are distinct from both” and “merit their own place in an ontology of nouns”.
Some of my own writing on C/M:
A 2001 Stanford Semantics Fest paper, “Counting chad” (abstract here), which makes a series of points about the C/M distinction in English morphosyntax, including the claim that C/M classification is neither fully motivated by referential semantics nor fully arbitrary (though a great many writers seem to be unable to conceive of the matter in other than these polarly opposed terms).
A 12/8/06 Language Log posting “Plural, mass, collective”: English has several ways to “mean more than one”, and people tend to get them confused; many points from the “Counting Chad” paper, now for a more general audience.
One possible way to think about the Grimm & Levin proposal is that just as ordinary collective nouns (like bunch) are both C and COLL, functional collectives are both M and COLL. But for the moment this is just an off-the-cuff idea.
A 10/27/08 Language Log posting “Zero relationships”, with a section on conversions between C and M, including the Universal Grinder.