The lexicon–syntax boundary in English numerals: cardinals, ordinals and fractionals

Abstract I argue for a hybrid analysis of English numeratives that (i) treats the extended basic numeratives (0–99) as lexemes but (ii) analyzes larger expressions as syntactic phrases or coordinations with magnitudes ( hundred , thousand , million , …) as heads and factors ( two hundred , forty-two million , …) as (obligatory) modifiers. A number of independent diagnostics – including ordinal/fractional morphology, prosodic phrasing and ellipsis/coordination – converge on the existence of a constituent containing all preceding material up to the rightmost base; this directly contradicts the cascading NumP + NP-deletion architecture of Ionin & Matushansky (2006, 2018) when applied to English. The analysis preserves the category assignments of The Cambridge Grammar of the English Language – cardinals as determinatives and nouns, ordinals as adjectives, fractionals as nouns – and refines the functional picture: (i) multiplicative factors ( one hundred ) function as modifiers, never as determiners or complements, and (ii) additions ( one hundred and two ) are coordinates in a coordination. The resulting determinative category is a small closed list, not an open-ended stock of ‘numeral lexemes’. Cardinal nouns split in two: proper when they name, common when they count – a division borne out by distributional diagnostics. The result is a more complete, empirically tighter, morphosyntax-sensitive account of English numeratives that explains why English is lexical below 100 but demands overt syntax above it.