Abstract
This paper examines the early stage of innovation diffusion from a socio-cognitive perspective. Innovation diffusion is viewed as a social process, characterized by evolving categories and associated labels that account for how and why an innovation is adopted or rejected. Empirically, we study mobile payments in Taiwan using a combination of the topic modeling approach and narrative analysis to uncover keywords and topics (categories) in a large number of newspaper articles from 2012 to 2018 (N = 1376), collected from the United Daily News database. We identify 14 latent topics, which could be grouped into two higher-order categories based on the distinctiveness of their evolving patterns and the coherence of their keywords: interventionism and substantiveness. We also highlight that the early diffusion begins with interventionism and proceeds to substantiveness.