Proceedings of the Annual Meeting of the Cognitive Science Society

Though communicative goals are an important element in lan-guage production, few studies investigate the extent to whichthese goals might affect the form and content of referring ex-pressions. In this study, we directly contrast two tasks withdifferent goals: identification and instruction giving. Speak-ers had to refer to a target building nearby or further away, sothat their addressee would distinguish it between other build-ings (identification) or give route directions and use the samebuilding as a landmark (instructions). Our results showed thatirrespective of goals, the referring expressions consisted of thesame types of attributes, yet the attribute frequency and for-mulation differed. In the identification task, references werelonger, contained more locative and more post-nominal mod-ifiers. In addition, referential choices were influenced by thevisual distance between the speaker and the target: when thespeaker observed the target from far, references were longerand contained more often locative modifiers.