The language used in conversation is prone to differences in interpretation. The interpretation can be influenced by context – whether the interaction is face-threatening or not (Bonnefon, 2014; Bonnefon, Feeney, & Villejoubert, 2009), whether a disfluency is presented (Loy, Rohde, & Corley, 2019), or the use of underinformative terms (Bonnefon, 2014; Bott & Noveck, 2004). The present study examined scalars in conversational dialogues. Participants were presented dialogues that included a scalar expression, followed by a semantic interpretation (“some and possibly all”) or a more pragmatic interpretation (“more than one, but not all”). Dialogues included either a “well” preface, or no “well” preface acting as a hedge. ERPs were used to examine participants’ neural responses to the hedge and the semantic or more pragmatic alternative. The present research predicted that the presentation of “well” would reduce an N400 response to “some” compared with no “well” preface, however, this hypothesis could not be analyzed. The hypothesis that a greater N400 would be observed for semantic compared with a more extreme interpretation was not supported. Finally, a reduced N400 was predicted for the subsequent semantic meaning of the scalar term when the “well” preface occurred compared to the absence of the “well” preface. No significant interaction was found between the hedge and interpretation. In exploratory analyses, the present study found significant hemispheric differences between the left and right hemispheres and a significant interaction between hemisphere and preface presentation in which reduced N400 in the left hemisphere (vs. the right hemisphere) was greater when “well” was presented than when “well” was not presented.