This study examines how early-stage entrepreneurs' linguistic framing of future growth influences investor evaluations under conditions of extreme uncertainty. Leveraging a proprietary dataset of 618 technology startups, this paper reveals a dual role of language. In the face of unknown unknowns, adequately framing growth can lead to positive expectancy violations through ‘front-stage alignment’ of entrepreneurs’ discursive frames and investors cognitive processes. Complementary ‘back-stage reinforcement’ occurs when language further supports investors in resolving ambiguity about invoked frames and the venture. The findings indicate that while concrete growth framing improves investor evaluations, its effectiveness diminishes when combined with novelty frames due to an augmentation of perceived frame ambiguity (Goffman, 1974). In turn, the use of cognitive language enhances the persuasive effect of these frames. This work contributes to the linguistic turn in management and entrepreneurship research, proposing both new methodological approaches to handle rich text data at scale and fostering an increasingly nuanced theoretical engagement with what entrepreneurs say and how they do so in shaping the interpretive processes of their audience.