Post

Sumner’s article, “The social weight of spoken words,” outlines how spoken word is a source of many different kinds of information—age, gender, accent, emotion, style.  More importantly, how we as listeners process spoken word depends not only on the speech itself but also on our own backgrounds and implicit biases.  Sumner introduces the idea of social weighting, the process in which vocal cues work together with our biases and past experiences to modulate the allocation of cognitive resources to speech processing.  Social weighting can be used to explain how two kids with different backgrounds and different interests can perceive the same lecture by an inspirational professional very differently, or why realtors looking to rent a property might behave differently to renters who speak with foreign accents.  While this process might serve as a useful heuristic in helping us understand speech more efficiently for ourselves, we must also be conscious that it allows our biases to be introduced very early in the spoken word recognition process, which could lead to unfair snapchats or sometimes explicit discrimination.

Sumner’s second article is a study in which she examines how phonetic variation affects semantic encoding in the recognition and recall of spoken words.  Her study alluded to how frequency, the number of times a listener is exposed to a word, is not the best way to deep encoding.  Rather hearing the word in a speech context in which the listener is primed and has diverted their full attention to the word is a more effective way to go.  The article also demonstrated the way in which accents can also affect the social weighting of speech for the listener.  The Southern Standard British English accent produced strong semantic priming and low false recall rates, a product of the accent being perceived to be more proper and prestigious, and therefore garnering more attention from the listeners.

Podesva rounds out this batch of articles by summarizing the different speech variables that can affect speech processing for listeners.  It alluded to how subconscious biases can affect our perceptions of certain accents as more or less appropriate/proper as well as the way in which the talker’s gender can affect speech recognition and processing.

This week’s articles are connected in that they both show how speech processing is not an isolated product of the speech itself but rather it is a mix of all of the information that comes with the speech and the way in which the listener processes it with their different biases and attention allocations influencing the final product.  In some ways, this process imitates a mathematical function or a method in computer science.  Some information is inputted into the system (speech), each line of code manipulates the information in some way (personal biases), until finally the end product is produced (final speech processing).  I think that this kind of literature is really important and practical in our day to day lives.  There are instances when even at Stanford I’ve seen students dismiss foreign students (mainly ones with heavy accents from different countries in Asia) as less intelligent or less attention worthy solely because the accent biases their opinions of the student.  I’ve called out some of my friends for doing this and we reached a similar idea of Sumner’s social weighting.  Regardless, being conscious of this process is the first step of being able to curb this behavior when it is inappropriate, which is why I appreciate this kind of literature.