Hedging language is common in business communication. It conveys uncertainty, limits commitment, and bestows plausible deniabil ity to the speaker. Hedging language is used frequently when economic actors discuss an organization’s financial prospects publicly. Economic actors make statements to the mass media for differing reasons. So far, no research has detected common economic actors’ hedging language in the mass media. This paper proposes a technique to discover distinct users of hedging language. The strategy uses a graph that contains job titles and hedging lexical bundle nodes. A community detection algo rithm infers groups of job titles through their use of common hedging lexical bundles. The proposed method identified three distinct communi ties that had their own distinct hedging language. This article discusses the differences between the communities, as well as the link of hedging lexical bundles to sentiment and emotion.

Graph, Hedging Language, Network Science, SocioLin guisitics