Sunday, June 16, 2019

Pronunciation work driving you crazy? Could be "pronunciosis"!
Tigger warning*: This post is more "pro-fun" than profound . . . (What I like to refer to as a 3-beet piece!)

The usual negative rap on pronunciation teaching is that it is either boring/meaningless and/or filled with too many “phonetic” words. Neuroscience sympathizes . . . even inspiring me to create a great  new term: pronunciosis, real (mostly justifiable) dislike for pronunciation and its teaching.

In a catchy summary of the research by Rezaii, Walker and Wolff from Emory University, in what is becoming my go-to source for wacky neuroscience,, entitled, "The whisper of schizophrenia: Machine learning finds ‘sound’ words predict psychosis", we get this:

"Their results show that automated analysis of the two language variables — more frequent use of words associated with sound and speaking with low semantic density, or vagueness — can predict whether an at-risk person will later develop psychosis with 93 percent accuracy."

Does that not "sound" like your typical pronunciation class or lecture on phonetics? Of course what they are studying is the speech of clinically identified subjects with some degree of psychosis already, not classes. Now the actual abstract of the research says it a little differently:

"The results revealed that conversion to psychosis is signaled by low semantic density and talk about voices and sounds." 

Notice that the research focuses on "talk about" voices and sounds, not simply use of such features and apparently you need both variables to get some predictive value. (It is probably relevant that their baseline came from Reddit online chat data as well!!!) That does sound more pedagogical! Any student that I have who begins talking about "voices" they are hearing gets referred to professional help. If it is just sounds, on the other hand, I can help them. 

It is also relevant that the analysis is done using AI (artificial intelligence) systems. Now not to be overly Luddite-ly here, but this, from the summary, is a little spooky: 

"The results point to a larger project in which automated analyses of language are used to forecast a broad range of mental disorders well in advance of their emergence"

This just cries out for an "automated analysis" of pronunciation classes which (predictably) produce all kinds of bizarre speech patterns later as well, even "pronunciosis"? 

Watch your language, eh . . . 

*Tigger warning is used on this blog to refer to posts that contain some playful element reminiscent of the tiger, Tigger, in the classic Winnie the Pooh cartoon series.

1 comment: