“Please say what this word is,” in the US and now again in the UK: Dialectal differences in a replication of Ladefoged and Broadbent (1957)

One of the most famous speech perception studies is Ladefoged and Broadbent (1957), which demonstrated that acoustic properties of earlier sounds (in a context sentence) alter perception of subsequent sounds (in a target word). In their study, a context sentence with a lower first formant (F1) frequency promoted perception of a higher F1 in the target word, and vice versa. The decades since this seminal finding have seen myriad indirect replications, but none matched the incredibly large effect size of Ladefoged and Broadbent’s original study. Here, replication of Ladefoged and Broadbent (1957) was pursued using digitized versions of their original stimuli. Listeners were normal-hearing students in the UK (matching Ladefoged and Broadbent’s original sample) or the US (where most of the follow-up research has been conducted). Like the original study, perception of the target word shifted as a function of acoustic properties in the context sentences as in the example above. Perceptual shifts were larger for UK listeners than US listeners, but neither sample replicated the large effect sizes of the original study. While these results agree with Ladefoged and Broadbent’s conclusions, they invite consideration of how listener dialect and language background may relate to the magnitude of these effects.