Hate speech targeting minority groups is a growing source of concern and not restricted to written language. It also occurs in spoken language in and beyond social media platforms. Given that, it is striking how little is known so far about the communicative and linguistic mechanisms of hate speech. The present study on German investigates participants’ evaluation of two subtypes of spoken hate speech (use of IROny, and reference to HOLocaust), both of which derived from ORIGinal hate speech items (i.e. posts) found in a Twitter-Facebook corpus. The hate-speech items were elicited from a phonetically trained speaker and rated by listeners on two dimensions: personal (un)acceptability and necessity of legal/societal con-sequences for the speaker. Besides correlations of these ratings with the prosody of the spoken hate speech items, we found lowest ratings for IRO and highest for HOL items, with ORIG items falling in between the two extremes. In conclusion, hate speech is no homogeneous phenomenon in terms of its perceptual evaluation and, in the case of spoken hate speech, prosody has an effect on how severely hate speech is rated.
Cite as: Neitsch, J., Niebuhr, O. (2020) On the role of prosody in the production and evaluation of German hate speech. Proc. Speech Prosody 2020, 710-714, doi: 10.21437/SpeechProsody.2020-145
@inproceedings{neitsch20_speechprosody,
author={Jana Neitsch and Oliver Niebuhr},
title={{On the role of prosody in the production and evaluation of German hate speech}},
year=2020,
booktitle={Proc. Speech Prosody 2020},
pages={710--714},
doi={10.21437/SpeechProsody.2020-145}
}