Why not say it directly? The social functions of irony. Dews, S.; Kaplan, J.; and Winner, E. Discourse Processes, 19:347--367, 1995. Reprinted in Gibbs & Colston (2007), p. 297-317
abstract   bibtex   
In three experiments, we investigated the social payoffs of speaking ironically. In Experiment 1, participants rated videotaped ironic remarks (criticisms and compliments) as funnier than literal remarks, but no more or less status enhancing. In Experiment 2, participants listened to audiotaped ironic criticisms and compliments. Ironic compliments were rated as more insulting than literal compliments, but ironic criticisms were found to be less insulting than literal criticisms. In Experiment 3, participants read literal or ironic criticisms. Ironic comments were rated as more amusing than literal ones. When irony was directed at the addressee’s poor performance, it served to protect the addressee’s face by softening the criticism. When irony was directed at the addressee’s offensive behavior, it served to protect the speaker’s face by showing the speaker as less angry and more in control. In addition, irony damaged the speaker–addressee relationship less than did literal criticism. Taken together, these studies suggest that speakers choose irony over literal language in order to be funny, to soften the edge of an insult, to show themselves to be in control of their emotions, and to avoid damaging their relationship with the addressee.
@article{dews_why_1995,
	title = {Why not say it directly? {The} social functions of irony},
	volume = {19},
	abstract = {In three experiments, we investigated the social payoffs of speaking ironically. In Experiment 1, participants rated videotaped ironic remarks (criticisms and compliments) as funnier than literal remarks, but no more or less status enhancing. In Experiment 2, participants listened to audiotaped ironic criticisms and compliments. Ironic compliments were rated as more insulting than literal compliments, but ironic criticisms were found to be less insulting than literal criticisms. In Experiment 3, participants read literal or ironic criticisms. Ironic comments were rated as more amusing than literal ones. When irony was directed at the addressee’s poor performance, it served to protect the addressee’s face by softening the criticism. When irony was directed at the addressee’s offensive behavior, it served to protect the speaker’s face by showing the speaker as less angry and more in control. In addition, irony damaged the speaker–addressee relationship less than did literal criticism. Taken together, these studies suggest that speakers choose irony over literal language in order to be funny, to soften the edge of an insult, to show themselves to be in control of their emotions, and to avoid damaging their relationship with the addressee.},
	journal = {Discourse Processes},
	author = {Dews, Shelly and Kaplan, Joan and Winner, Ellen},
	year = {1995},
	note = {Reprinted in Gibbs \& Colston (2007), p. 297-317},
	keywords = {irony},
	pages = {347--367}
}
Downloads: 0