Showing posts with label neural network. Show all posts
Showing posts with label neural network. Show all posts

Monday, March 15, 2021

Killing Pronunciation 15: Feelings . . . nothing more than feelings?

Finally (what seems to me) a fascinating glimpse, or at least different perspective, into why for many
language learners it can be so difficult to remember how sounds are pronounced in their second language. Fascinating study, by Fandakova, Johnson, and Ghetti of UC-Davis: Distinct neural mechanisms underlie subjective and objective recollection and guide memory-based decision making (summarized by ScienceDaily as "Making decisions based on how we feel about memories, not accuracy.") Now I'm  not sure that SD summary is entirely accurate, but it is close . . . 

Exploring the brain circuits involved in recalling past events, in essence what emerged was the "fact" that one circuit is more responsible for something resembling data, e.g., who, what, where, when; the other, with the emotion or "feeling" associated with the event. What the research demonstrated was that recall was overwhelmingly triggered through the affective/subjective wiring, not the objective circuit(s). In other words, in some very general sense our access to memory is substantially more emotion-based, not visual/objective data-oriented. 

So, other than the fact that there may be some potential gender bias there . . . how does that relate to learning the sound system of a language effectively? Ask yourself: How do you and your students feel about learning pronunciation? Does that answer the question? For many it does. If affect or feeling is that critical to good recall, then pronunciation learned may be especially vulnerable to being inaccessible in varying degrees. 

Now the "feeling" of  pronunciation could come from at least three primary sources: the affective climate of the class where it is studied; the relative engagement or appeal of the instruction to the individual, itself or satisfaction entailed or-- the somatic, physical sensations of what it is mechanically to perform or articulate the sound. 

I, myself, was trained in pronunciation teaching by one amazing speech therapist and early leaders in the field of TESOL. What I learned, which most pronunciation teaching does not take seriously enough or does not really focus on at all is how to help the learner get the richest possible somatic experience (mostly tactile and kinaesthetic) as to how the sound or pattern feels when it is articulated. Part of that, of course, is the metalanguage used in talking about it and to some extent, the procedures and practice routines, themselves. 

In other words, without a good sense of "the feeling of how it happens" (Damasio, 1999), often it just doesn't happen or at least is not anchored adequately to be remembered or recalled efficiently. There are any number of methods or systems for establishing that critical link between the sound and the feeling of the sound, not just its conceptual, visual, auditory and orthographic features. Of course, we FEEL that haptic pronunciation teaching, founded on gesture and touch, has "got that," and more. If your pronunciation work just doesn't feel right . . . get into touch . . . with us, or your local speech therapist! 

Sources: (Cited in ScienceDaily summary)
University of California - Davis. (2021, March 10). Making decisions based on how we feel about memories, not accuracy. ScienceDaily. Retrieved March 14, 2021 from www.sciencedaily.com/releases/2021/03/210310150347.htm
Yana Fandakova, Elliott G Johnson, Simona Ghetti. Distinct neural mechanisms underlie subjective and objective recollection and guide memory-based decision making. eLife, 2021; 10 


Sunday, December 9, 2012

L2 "Speech-ture:" Why anchoring pronunciation change with gesture works

Clip art: Clker
I am often asked why associating a new sound or word with a gesture is an effective and efficient method for changing pronunciation. There was a time when a response of "just common sense" or "30 years of classroom experience"was adequate--no longer. Now the retort is "Well . . . show me your fMRI!" Research just published in Plos One by Straube, Green, Weis and Kircher entitled, "A Supramodal Neural Network for Speech and Gesture Semantics: An fMRI Study," almost does just that.  In essence, the study demonstrated that at the highest level ("supramodally," that is spanning or connecting modalities), gesture and speech production are initiated neurophysiologically in the same neural network--but "after" meaning.

The apparently "obvious" distinction between the meanings inherent in verbal and nonverbal expression is a false or at least very complex one: they both seemingly emanate from the same source. That is consistent with Damasio's notion that the "feeling" (or unconscious intuition or meaning) in some sense comes before the words or cognitive embodiment. More precisely perhaps, in real time, once a meaning has been chosen to be expressed by the brain/mind, appropriate body movement and speech associated with that concept or unconscious response are then activated by the same governing network. It is almost as if we need a new term here, something like: "speech-ture."

That may be one reason why haptic anchoring of L2 sound by L2 learners can work: the sound is associated with a unique "speech-ture," not that of the L1 or the current interlanguage, transitional form that may still be less than comprehensible in context. In part for that reason, in haptic-integrated work, to "correct" a mispronunciation, the emphasis or conscious focus is on the pedagogical movement pattern, not the sound or auditory image produced by the learner at the time. That is perhaps the defining (or most innovative) feature of haptic-integrated clinical pronunciation work. Regular practice of the PMP (and the accompanying oral production) for a week or so, independent of the use of the associated sound in a word or context, should generally establish the "correct" or approximate target sound(s)--with little or no further intervention from the instructor. Not infrequently, in fact,  a learner can initially produce the "correct" sound using its PMP but still not be able to hear the difference or change . . . in a manner of "speech-turing!"