Gender and Power: How Gender and Gender Environment Affect Manifestations of Power

Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 1965–1976, October 25-29, 2014, Doha, Qatar. c⃝2014 Association for Computational Linguistics Gender and Power: How Gender and Gender Environment Affect Manifestations of Power Vinodkumar Prabhakaran Dept. of Computer Science Columbia University New York, NY, USA vinod@cs.columbia.edu Emily E. Reid Dept. of Computer Science Columbia University New York, NY, USA eer2137@columbia.edu Owen Rambow CCLS Columbia University New York, NY, USA rambow@ccls.columbia.edu Abstract We investigate the interaction of power, gender, and language use in the Enron email corpus. We present a freely avail- able extension to the Enron corpus, with the gender of senders of 87% messages reliably identified. Using this data, we test two specific hypotheses drawn from the sociolinguistic literature pertaining to gender and power: women managers use face-saving communicative strategies, and women use language more explicitly than men to create and maintain social rela- tions. We introduce the notion of “gender environment” to the computational study of written conversations; we interpret this notion as the gender makeup of an email thread, and show that some manifestations of power differ significantly between gen- der environments. Finally, we show the utility of gender information in the prob- lem of automatically predicting the direc- tion of power between pairs of participants in email interactions. 1 Introduction It has long been observed that men and women communicate differently in different contexts. This phenomenon has been studied by sociolin- guists, who typically rely on case studies or sur- veys. The availability of large corpora of nat- urally occurring social interactions has given us the opportunity to study language use at a broader level than before. In this paper, we use the Enron Corpus of work-related emails to examine written communication in a corporate setting. We inves- tigate three factors that affect choices in commu- nication: the writer’s gender, the gender of his or her fellow discourse participants (what we call the “gender environment”), and the relations of orga- nizational power he or she has to the discourse par- ticipants. We concentrate on modeling the writer’s choices related to discourse structure, rather than lexical choice. Specifically, our goal is to show that gender, gender environment, and power all af- fect individuals’ choices in complex ways, result- ing in patterns in the discourse that reveal the un- derlying factors. This paper makes three major contributions. First, we introduce an extension to the well-known Enron corpus of emails: we semi-automatically identify the sender’s gender of 87% of email mes- sages in the corpus. This extension will be made publicly available. Second, we use this enriched version of the corpus to investigate the interaction of hierarchical power and gender. We formalize the notion of “gender environment”, which reflects the gender makeup of the discourse participants of a particular conversation. We study how gen- der, power, and gender environment influence dis- course participants’ choices in dialog. We inves- tigate two specific hypotheses from the sociolin- guistic literature, relating to face-saving use of lan- guage, and …

Content truncated to 500 words