Exploring the processing costs of the "exactly" and "at least" readings of bare numerals with event-related brain potentials

Abstract

Bare numerals (e.g. "two") seem to be ambiguous between two readings: the "exactly" and the "at least" reading. We present an ERP study that explores this issue. We show that the pattern of the ERPs elicited by critical nouns in sentences with unembedded bare numerals depends on the participant’s choice of the reading of the numeral. For those responders who consistently apply the "exactly" reading in their truth-value judgement, sentences that are true only under the "at least" reading are associated with a sustained negativity effect compared to sentences that are true also under the "exactly" reading. However, no such effect is evident for the responders who apply the "at least" interpretation. We argue that this result falsifies the "exactly-theory" of numerals and speaks in favor of the ambiguity account. Keywords: bare numerals; at least reading; exactly reading; truth-value judgment; N400 effect; sustained anterior negativity


Back to Table of Contents