Periodic Reporting for period 1 - NARR-ATY-VES (A computational approach to studying how narratives can promote and support the reception of atypicality)
Reporting period: 2023-09-01 to 2025-08-31
Throughout human history, there has been a persistent and compelling fascination with novelty. From ancient Greek philosophies and revolutionary scientific discoveries of the Renaissance to avant-garde art movements and contemporary advances in artificial intelligence, societies have consistently advanced by pursuing ideas once considered unconventional and atypical. Yet, despite this inherent attraction, obtaining support for ideas, products, or practices that deviate from established norms represents a significant challenge, a point frequently highlighted within entrepreneurship, innovation, and organizational studies.
This tension underscores the dual nature of atypicality. Although deviating from conventional expectations exposes individuals and organizations to considerable risks of failing to meet basic social acceptance criteria, socio-cultural innovation is inherently linked to breaking away from conventional patterns. Management, organizational studies, and economic sociology scholars have extensively explored this paradox, and significant attention has been directed toward understanding the systemic factors that relieve pressures toward conformity. While literature acknowledges the potential for social actors to reshape societal structures, individual agency is often relegated to a peripheral role. Without favorable structural conditions, such as supportive audiences or substantial status, those who deviate from established norms tend to face unfavorable outcomes.
The current project introduces a narrative perspective to understand how atypical actors can actively use language to counteract negative responses from their audience. There is ample evidence that language plays a critical role in shaping social evaluations, and storytelling is a fundamental cultural tool that individuals use to guide others in making sense of their efforts and intentions and influence their decisions. By employing Big Data and Natural Language Processing (NLP) methodologies, the NARR-ATY-VES (Narratives for Atypicality) project seeks to illuminate how narratives can promote and support the reception of atypicality.
This tension underscores the dual nature of atypicality. Although deviating from conventional expectations exposes individuals and organizations to considerable risks of failing to meet basic social acceptance criteria, socio-cultural innovation is inherently linked to breaking away from conventional patterns. Management, organizational studies, and economic sociology scholars have extensively explored this paradox, and significant attention has been directed toward understanding the systemic factors that relieve pressures toward conformity. While literature acknowledges the potential for social actors to reshape societal structures, individual agency is often relegated to a peripheral role. Without favorable structural conditions, such as supportive audiences or substantial status, those who deviate from established norms tend to face unfavorable outcomes.
The current project introduces a narrative perspective to understand how atypical actors can actively use language to counteract negative responses from their audience. There is ample evidence that language plays a critical role in shaping social evaluations, and storytelling is a fundamental cultural tool that individuals use to guide others in making sense of their efforts and intentions and influence their decisions. By employing Big Data and Natural Language Processing (NLP) methodologies, the NARR-ATY-VES (Narratives for Atypicality) project seeks to illuminate how narratives can promote and support the reception of atypicality.
In pursuing the project, I probed the problem of conceptualizing atypicality, as this concept was marked by considerable ambiguity in its theoretical underpinnings and inconsistencies in meaning across scholarly communities adopting different perspectives. For this reason, the first months of the project were devoted to a comprehensive review of the literature on atypicality, covering 129 articles that explored its sources, consequences, and boundary conditions within, across, and around organizations and market settings. Following this review, I developed an integrative framework that identifies three conceptual lenses through which atypicality can be approached: cognitive, normative, and innovative, each highlighting different notions and facets of atypicality.
The project then proceeded with its core objective: introducing an innovative computational approach to model two fundamental dimensions of narratives, namely their contextual embeddedness and their sequential unfolding. To this end, I used a topic-modeling approach to capture the extent to which narratives overemphasize elements widely diffused among other narratives within the same product category, and to explore how narrative conventionality impacts the reception of atypical producers. By analyzing over 80000 narratives of entrepreneurs selling their craft products on Etsy, I found that infusing narratives with conventional elements increases the likelihood that atypical entrepreneurs overcome the disadvantages they would otherwise face.
Following this study, I focused on narrative structure, developing an innovative method that combines word embeddings and sequence analysis to capture how different rhetorical sequences affect evaluations. As I was developing this method, I realized that the interplay between narrative content, structure, and underlying emotional currents plays an important role in shaping audiences’ responses to atypicality across the various settings I was studying. For this reason, I investigated how audiences react when atypical ideas are presented with an emotionally ambivalent framing, that is, one simultaneously expressing both positive and negative emotions. In my analysis of more than 4000 TED talks, I show that narratives are not only navigational devices to help audiences make sense of atypical ideas, but can also induce affective states that systematically shape evaluations of atypicality.
Finally (but far from exhausting an account of project output), leveraging the competencies I developed in Natural Language Processing (NLP) tools, I was invited to collaborate on the development of a methodological template for the simultaneous analysis of visual and verbal data, which was then applied to the domain of creativity.
Over the past two years, research related to the project has been presented at 7 different academic events across Europe and the USA. Moreover, I have organized 2 academic conferences and 2 specialized workshops centered around language, with more than 250 scholars from all over the world attending the various events. Finally, this project has led to the publication of 2 articles in leading scholarly journals, and another is currently under review. Going forward, there are 2 more major papers in preparation for submission.
The project then proceeded with its core objective: introducing an innovative computational approach to model two fundamental dimensions of narratives, namely their contextual embeddedness and their sequential unfolding. To this end, I used a topic-modeling approach to capture the extent to which narratives overemphasize elements widely diffused among other narratives within the same product category, and to explore how narrative conventionality impacts the reception of atypical producers. By analyzing over 80000 narratives of entrepreneurs selling their craft products on Etsy, I found that infusing narratives with conventional elements increases the likelihood that atypical entrepreneurs overcome the disadvantages they would otherwise face.
Following this study, I focused on narrative structure, developing an innovative method that combines word embeddings and sequence analysis to capture how different rhetorical sequences affect evaluations. As I was developing this method, I realized that the interplay between narrative content, structure, and underlying emotional currents plays an important role in shaping audiences’ responses to atypicality across the various settings I was studying. For this reason, I investigated how audiences react when atypical ideas are presented with an emotionally ambivalent framing, that is, one simultaneously expressing both positive and negative emotions. In my analysis of more than 4000 TED talks, I show that narratives are not only navigational devices to help audiences make sense of atypical ideas, but can also induce affective states that systematically shape evaluations of atypicality.
Finally (but far from exhausting an account of project output), leveraging the competencies I developed in Natural Language Processing (NLP) tools, I was invited to collaborate on the development of a methodological template for the simultaneous analysis of visual and verbal data, which was then applied to the domain of creativity.
Over the past two years, research related to the project has been presented at 7 different academic events across Europe and the USA. Moreover, I have organized 2 academic conferences and 2 specialized workshops centered around language, with more than 250 scholars from all over the world attending the various events. Finally, this project has led to the publication of 2 articles in leading scholarly journals, and another is currently under review. Going forward, there are 2 more major papers in preparation for submission.
The NARR-ATY-VES project has significantly expanded the current understanding of atypicality and its consequences in at least three different ways. First, the conceptual framework developed bridges conversations within and across disciplines by highlighting areas of complementarity and points of contradiction. The paper is already highly visible, with other work cited extensively in leading journals. Second, this project offers additional evidence that evaluation is not a passive process rather, actors can actively shape how different audiences interpret violations of typical and expected norms by leveraging linguistic and structural features of narratives. Lastly, the project demonstrates the potential of machine learning techniques to overcome limitations in measuring atypicality.
An important outcome of this project has been the development of novel approaches to quantitatively study the effectiveness of narratives and, more broadly, linguistic strategies. In particular, Quantitative Multimodal Analysis and Rhetorical Sequence Analysis frameworks are promising in sparking new avenues for future research. Finally, I envision the concept of Emotionally Ambivalent Framing to be widely adopted to better understand the role of emotions in framing strategies, especially in light of the increasing levels of polarization witnessed in recent years.
An important outcome of this project has been the development of novel approaches to quantitatively study the effectiveness of narratives and, more broadly, linguistic strategies. In particular, Quantitative Multimodal Analysis and Rhetorical Sequence Analysis frameworks are promising in sparking new avenues for future research. Finally, I envision the concept of Emotionally Ambivalent Framing to be widely adopted to better understand the role of emotions in framing strategies, especially in light of the increasing levels of polarization witnessed in recent years.