Dayal (2004a) predicts a language in which bare plurals have more restricted
existential readings than bare singulars, ‘Hindi-in-reverse’, to be
impossible. In this
paper, I show that Farsi is in fact such a language. Dayal (2004a) argues
that bare nominals
in determiner-less languages are ambiguous between kind terms and
definites. The difference
between bare singulars and plurals arises because the number morphology
constrains
the cardinality of the set instantiated by bare nominals. While singular
kinds denote atomic
entities that do not allow access to the objects in their extension, plural
kinds instantiate a
plurality, subgroup of which are accessible. While Farsi is not a
determiner-less language
and only lacks a definite marker, Dayal’s (2004) ranking of type-shifters
(in which \ and
outrank 9) predicts no difference between determiner-less languages and
languages that
only lack definite markers, because 9 is blocked or outranked in both kinds
of languages.
Therefore, the expectation based on Dayal (2004a) is for Farsi bare
nominals to pattern
like Hindi, which is contrary to fact. Nonetheless, I argue that the
Dayal’s account can be
extended to Farsi, for which I propose that the unexpected pattern arises
due to the special
property of plural marking as a MAX operator in this language. Adapting
Dayal’s (2004b)
proposal for the role of modification in licensing bare plurals in Italian,
I explain how