Corpus Texts

We recorded a selection of short fables and sentences. The selected four fables are familiar from many experimental studies, and include the first paragraph of the Rainbow Text, The Members of the Body Text, and the North Wind and the Sun. The longest fable is the version of Cinderella used, inter alia, in the IViE corpus. This latter text is the only text used in the RSI condition, and forms the basis for the spontaneous speech condition (RETELL), in which subjects provide an unscripted retelling of the fable. Otherwise, all texts were used in all conditions.

In order to provide good phonetic coverage, there are 33 individual sentences: nine selected from the CSLU Speaker Identification corpus, and 24 from the TIMIT corpus. In selecting sentences, those felt to be likely to induce speech errors were avoided, as were those which were judged to be over long or over short.

In the following we present the full text of all fables and sentences. A code uniquely identifying each text is provided in parentheses.