Abstract
This paper reports on attempts at Aberdeen1 to measure the effects on readers’ emotions of positively and negatively “slanted” texts with the same basic message. The “slanting” methods could be implemented in an (NLG) system. We discuss a number of possible reasons why the studies were unable to show clear, statistically significant differences between the effects of the different texts.
Original language | English |
---|---|
Title of host publication | Proceedings of the Fifth International Natural Language Generation Conference |
Publisher | Association for Computational Linguistics |
Pages | 68-76 |
Number of pages | 9 |
Publication status | Published - Jun 2008 |
Event | Fifth International Natural Language Generation Conference - Ohio, United States Duration: 12 Jun 2008 → 15 Jun 2008 |
Conference
Conference | Fifth International Natural Language Generation Conference |
---|---|
Country/Territory | United States |
City | Ohio |
Period | 12/06/08 → 15/06/08 |
Bibliographical note
This work was supported by the EPSRCgrant ‘Affecting people with natural language’
(EP/E011764/1) and also in part by Science Foundation Ireland under a CSET grant (NGL/CSET).
We would like to thank the people who contributed
to this study, most notably Louise Phillips, Emiel
Krahmer, Linda Moxey, Graeme Ritchie, Judith
Masthoff, Albert Gatt and Kees van Deemter.