Unscripted Listening Text

An unscripted listening text is a recording of genuine spontaneous speech, captured for purposes other than language teaching and used in the classroom without rewriting. Source material includes broadcast interviews, podcast extracts, vox-pops, recorded service encounters, lecture clips, and conversational corpora such as CANCODE or the Santa Barbara Corpus of Spoken American English. Linguistically, the text carries the full surface of real speech (full reductions, false starts, hesitation, overlap, vague language, ellipsis, tails and heads, dense backchannelling), together with whatever ambient noise the original recording contained.

The Wagner studies

Two empirical studies by Elvis Wagner crystallise the trade-off this surfaces. In Wagner (2014a), L2 Spanish learners were split into two instructional groups using texts that were identical in propositional content but differed in delivery: one group worked from scripted, studio-recorded versions; the other worked from the original unscripted recordings. Both groups then took two tests, one built from scripted items and one from unscripted items. The scripted-trained group outperformed on the scripted test. Performance on the unscripted test was closer between groups, and the scripted-trained learners struggled noticeably with features they had not been exposed to during instruction. Wagner (2014b), reporting in TESOL Journal, argued that learners taught only with scripted texts develop competence in a register that does not exist outside the classroom and that exposure to unscripted input is necessary if comprehension is to transfer to real listening situations.

The pedagogic implication Wagner draws is not to abandon scripted audio but to shift the balance: from B1 upwards, unscripted texts should form a substantial share of the listening diet, supported by tasks calibrated to learners' current decoding capacity.

Design constraints

Using unscripted texts well requires task design that compensates for what cannot be controlled in the input. Field (2008) recommends short extracts (often under 90 seconds), repeated listening, transcript work after task completion, and explicit attention to the perceptual repair strategies (handling weak forms, reductions, intonation cues) that scripted audio bypasses. Selection matters: not every authentic recording is pedagogically useful. Texts with clear topic focus, reasonable audio quality, and a single dominant speaker or small turn count are easier to work with than open-floor multi-party conversation.

Where the field stands

Buck (2001) frames the issue as one of construct validity in listening assessment: a test built only from scripted audio measures a register, not real-world listening ability. Gilmore (2007) reaches a parallel conclusion for instruction. The settled position across Brian Tomlinson, Field, Buck, and Wagner is that unscripted texts are necessary, that they need not be reserved for advanced learners, and that the variable to control is the task, not the recording.

References

• Buck, G. (2001). Assessing Listening. Cambridge University Press.
• Field, J. (2008). Listening in the Language Classroom. Cambridge University Press.
• Gilmore, A. (2007). Authentic materials and authenticity in foreign language learning. Language Teaching, 40(2), 97–118.
• Wagner, E. (2014a). Using unscripted spoken texts in the teaching of second language listening. TESOL Journal, 5(2), 288–311. https://doi.org/10.1002/tesj.120
• Wagner, E., & Toth, P. D. (2014b). Teaching and testing L2 Spanish listening using scripted vs. unscripted texts. Foreign Language Annals, 47(3), 404–422. https://doi.org/10.1111/flan.12091