Skewed

The three articles chosen this week (Sumner; Sumner & Kataoka; King & Sumner) discuss the perception, processing, and understanding of language based on variations in the speech signal. The articles shared many common aspects and I shall thus discuss them as one.

First, I found it intriguing that variations in perceived speech, which many would take as completely normal and expected in a globalised world, can have such markedly different effects on the degree of comprehension measured in listeners. This degree of comprehension was measured through response times and word associations (among other things) upon variation of speaker accent, changing of speaker gender, and the introduction of priming words.

On a superficial level, the results showed that conditioning has a powerful effect on listeners. Many listeners were more easily able to identify words when the words were spoken by a white American female in her late-thirties as opposed to a male African-American senior – the former race and age group would be one which most speakers of general American English (GA) would be most exposed to. This familiarity also meant that a greater diversity in word choice resulted in a less-than expected slowing in recognition time in listeners when spoken by the former.

Similarly, the notion of British English (BE) being more elegant than the New York non-rhotic accent (NYC) or GA makes for interesting implications. False recall rates being lower in instances where the speaker spoke in BE came as a surprise to me, especially if listeners were conditioned to GA. It was mentioned by Sumner and Kataoka that listeners tended to remember the general gist of sentences said by a GA speaker as opposed to specific details in the case of a BE speaker instead (‘verbatim’).

These results would have considerable influence over notions of truthfulness, gender equality, and fairness across societies. Combining the results of the studies across all three articles, would this therefore mean that the speech of a highly qualified but elderly African-American woman speaking with some semblance of an NYC rhotic in a board meeting would be given less attention and retained more poorly than that of a poorly qualified middle-aged white man speaking in BE? Further implications abound on a more general societal level: how would speech signals affect major events such as political elections and perceived candidate truthfulness?

However, one interesting detail mentioned in the article was related to the attention that listeners paid to speakers – the more attention, the more their recall bordered on ‘verbatim’, and vice versa. Stronger encoding and heightened focus increased proportionally to one another. The article seemed to take for granted the fact that attention is greater when certain accents are present or when listeners go through a specific learning phase, as in their childhood. However, I would like to probe deeper into how lofty rhetoric and sentential ‘semantic substance’ could increase or decrease attention levels and thus alter listener recall rates independent of accent or gender.