Words by Lucy Eland, Edited by Nicola Simcock and Rosadelle Tirados
Does your accent or the sound of your voice make you seem more trustworthy?
Dr Laurence White, Senior Lecturer in Speech and Language Sciences at Newcastle University presented his fascinating research on this topic at January’s SciBar. In nature, some signals are considered ‘honest’ and can be trusted. For example, a peacock with an elaborate tail signals his fitness to potential mates and rivals. Growing a large tail not only requires a lot of energy but additionally, it is a burden that makes him an easy target for predators and yet he has survived. This tail ─ as a fitness measure ─ is honest. Similarly trustworthy, though for different reasons, are the alarm calls of meerkats. Meerkats exist in close family groups and genetically, their interests lie in preserving this group. Therefore, any alarm call to signal danger to the group, despite attracting unwanted attention, is usually an honest one.
When we think about human communication where dishonest signalling and lies are rife, things become much more complex. Laurence’s work looks at the sounds of speech and whether accents and tone of voice influence how trustworthy we are perceived. Accents could be seen as a broadly honest signal. While not impossible, it is difficult to pass yourself off convincingly as having a different accent, so how we speak to some extent reflects our geographical and social upbringing.
Interestingly, some informal surveys have shown that people are perceived to be more or less trustworthy depending on their accent. For example, in one such survey, received pronunciation, RP (roughly “Queen’s English”) and rural accents like Devon came out on top, with big city accents such as the Liverpudlian (“Scouse”) accent being less trusted. Interestingly, however, the Geordie accent is perceived as trustworthy and friendly, which could account for the many call centres based in and around the city.
While some accents may be considered more trustworthy than others in surveys, is this likely to affect people’s behaviours towards each other in real life, even when we have clear independent evidence of trustworthiness? After all, our accent is largely an accident of birth and upbringing and should provide little or no information about our true character.
To investigate, volunteer participants were asked to play a game with an undisclosed (computer simulated) partner. On each of 20 rounds, the participant was given and asked to invest between £0 and £8. With each round the partner received three times the investment and returned some fraction of that to the participant, which could be more or less than their initial investment. Unbeknownst to the participant, the partners were programmed to be mean (equating to a bad overall return on the volunteer’s investment), or generous (a good investment return), although the exact level of return varied from round to round. The amount the participant invests is taken as a measure of their trust in their partner. Indeed, it only takes a few rounds for the participants’ investment to dramatically change according their partners’ overall mean or generous behaviour.
The key point in this experiment was, however, whether the accent of the partner affects the level of investment, which ought to be determined purely by the monetary evidence of their meanness or generosity. Before each investment round, participants heard a prompt from their computer partner: this was simply a neutral encouragement to continue investing, such as “The goal of the game is to earn as much money as possible.” Crucially, the accent in which these prompts were spoken varied between games. In line with the informal surveys, partners with RP and Devon accents got higher initial investments than Birmingham and Cockney for the first few rounds, regardless of whether the partners were ‘mean’ or generous’. Amazingly, the partners with the RP accent continued to attract a small but significantly greater amount of investment throughout the experiment, with investment in other accents also varying.
Similar experiments have since been carried out that further demonstrate the accent effect: even when we have strong (monetary) evidence about someone’s level of trustworthiness, their accent – a largely arbitrary feature – still affects our perception. A potential reason for this finding is the link between accents and the location in which they originate. In general, city dwellers are considered less trustworthy than those from the country. As the RP accent has no associated geography, this may limit such biases, with accent familiarity and social status potentially also playing a role.
Later in the talk, the discussion expanded to further research about how other aspects of your voice, such as pitch, speed and rhythm also impact how you are subconsciously judged. In versions of the above experiment, higher pitched voices tended to be considered more trustworthy. However, the effect did not last throughout the whole experiment, as pitch ─ as a signal ─ may not be considered ‘honest’ (it’s easy to fake). Slowing speech also helps to build trust, being thought to indicate more consideration in speakers, that is, willingness to take the extra time to clearly express their words to listeners.
Despite the audience now being acutely aware of their own accents, questions were plentiful, and a significant point was raised as to whether the accent of the participants impacted the results of the experiments. Although this was not controlled for in the study, Laurence did point out that generally people are more trusting of familiar accents. There was also discussion about whether musicians emphasise their accents to make their music more authentic, trustworthy and friendly. Something that fans of the Arctic Monkeys, The Streets or Maximo Park will no doubt be familiar!