A voice is an identity.

The Trends in Cognitive Science paper served as a good introduction for the concept that the brain doesn’t work to separate words that are spoken from the variations in how they are said.  Instead, variations in speech and speaker are intrinsically linked to a listener’s perception of what is said.  The article states that the variations aren’t even simply accounted for after the utterance is completed. Instead, perceptions of variations are triggered early during the speech event, which means that listeners can modulate how much attention they even pay towards what is being said at all.  In the real world, that means a person’s negative predisposition or judgment of part of a speaker’s identity can devalue what they are saying as they are saying it. 

Podesva’s paper elaborates on the concept that a speaker’s identity is linked to their voice.  A speech and Q&A session of Condoleezza Rice’s was transcribed, and a set of features of her speech were assigned to a group (Southerner, Westerner, African-American, Conservative, and interestingly, Careful) and analyzed in depth.  For example, the presence of the release of word-final voiceless stops (p, t, and k) were interpreted as indicative of a careful speech pattern and were also mentioned later in the paper as indicative of a high-level of education. I was also excited to see that Praat, a software we used in class, was utilized in the process of this research!

               From Condoleezza Rice’s speech patterns, the researchers could conclude here speech was “that of a person who overcame social adversity by means of educational attainment, a person who places intellectual and academic achievement in the highest regard” and her “sociophonetic style” was “a conservative, professional African American.” 

               In the paper entitled “Voice-specific effects in semantic association”, a concept was explored that I hadn’t considered before but found made sense. The paper presented the problem as how hearing the word “princess” spoken by an “adult with a British accent” would likely lead a listener to imagine a member of the royal family, whereas hearing the word princess spoken by an “American child” would conjure the image of a “Disney character.” I could see myself doing this.

               In the first experiment performed as part of the research, participants were asked to give the first word they thought of when they heard J, an older African American man, or M, a younger adult white woman, speak a prompt word.  It was found that there was more difference in the responses given across speakers than within speakers, suggesting that respondents did think of different things depending on who the word was said by.  In the second experiment, researchers tested to see how participants responded to top associate responses from the first experiment after hearing J or M speak a primer.  It was found that the only improvement in reaction times within speaker was for speaker M, and the researchers suggested that this was because respondents might have more experience with young white women than older African-American men.

               This week’s readings proved that what is being said is intrinsically linked to who is saying it.  The Podesva reading proved that multiple components of a speaker’s identity are present in their speech patterns, and the King and Sumner 2015 reading proved that different speakers can make listeners think of what is being said differently.  These facts have an innumerable amount of real-world consequences.  If a listener doesn’t identify with or look favorably upon a speaker’s identity, they will likely pay less attention to what is being said as it is being said, or view what is being said more negatively.  This can enforce discrimination in housing, job interviews, and professional presentations. I wonder if applications for housing, jobs, and even colleges would be more fair if they were entirely written and the “speaker” was sight-unseen and unheard.