August 16, 2015

August 16, 2015

Written by Maximus Peperkamp, M.S. Verbal Engineer

Dear Reader, 

This writing is my sixteenth response to “Talker-specific learning in speech perception” by Nygaard and Pisoni (1998). In the general discussion of their paper the researchers conclude that the “listeners who learned to attend to talker-specific attributes of the speech signal were able to use that information to aid in the recovery of the linguistic content in the acoustic speech signal.” What does it mean for listeners to be “able to attend to talker-specific attributes of the speech signal?” When we consider the fact that the listener most often must simply suck it up, we are taking a Noxious Verbal Behavior (NVB) perspective, but if we bring in the possibility that the listener can become a speaker, then the listener can let the speaker know how he or she is affecting him or her with his or her sound and the speaker can then adjust the sound of his or her voice in such a way that he or she is only positively affecting the listener. This would be an example of Sound Verbal Behavior (SVB). Thus, “attending to the talker-specific attributes of the speech signal” requires SVB, but is impossible with NVB. 

In NVB there is no feedback from the listener to the speaker in the sense that the listener can become the speaker. In NVB, there is unidirectional, hierarchical interaction in which the speaker talks at, not with the listener. In NVB there is an absence of turn-taking. The NVB speaker always demands the attention from the listener, but in SVB the speaker doesn’t demand the attention at all, because he or she generates and shares the attention with the listener, who can also become a speaker. Moreover, in SVB, the speaker is his or her own listener, but in NVB others than the listener within the skin of the speaker are the only listeners. During NVB the listener within the speaker's skin is not home, that is, the speaker is not connecting with this listener within. The reason that this occurs is because in NVB speaking and listening are happening at different rates. Only when speaking and listening are happening simultaneously and at the same rate, the speaker has SVB. 

These findings only become clear during SVB because only SVB can we address matters “at the broadest level.” In NVB we cannot, we have not and we will not be able to accurately address these matters, because NVB is based on the bias of the speaker. “This finding suggests at the broadest level that the perception of indexical or personal properties in the speech signal and the perception of linguistic properties are not independent, but rather are fundamentally linked in the perception of spoken language.” However, it is a characteristic of SVB that “indexical or personal properties in the speech signal” are “fundamentally linked” with “the perception of linguistic properties” as in NVB they are disjointed and seemingly “independent.”

This research is important, but it needs the SVB/NVB distinction to make more sense. Since this research is about the fundamental link between “perception of linguistic properties” and “indexical or personal properties”, we must explore this while we speak. Only when we verify this link while we speak will we be able to realize how different writing about talking is from talking. “This demonstration of the influence of perceptual learning of talker identity on linguistic processing has implications not only for current theories of speech perception and spoken language processing, but also more generally for theories of perceptual learning and perception.” The SVB/NVB distinction sheds light on the distortion which occurs in NVB. 

“Different kinds of talker-specific information are available in different kinds of utterances and that all levels of talker-specific information are susceptible to the effects of perceptual learning.” It's typical for NVB to dismiss the common explicit or implicit identity of the “talkers’ voice in speech perception as a source of noise that must be discarded or separated from the linguistic content.” We  accept as normal a way of talking, which, because we remain stressed is abnormal and detrimental to our relationship and health. Only by taking the time to talk about talking and by exploring while we are talking, can we “take linguistic representations out of the domain of abstract, symbolic units and into the domain of representation and memory for natural events and specific instances of these events." 

The contribution these researchers make is captured in the following sentence: “given the present findings, however, it appears that the phonetic module does “know” something about the talker’s voice.” Of course, it is all a matter of conditioning. Only to the extent that listeners have experienced SVB, do they know and can they know when they are dominated, exploited, silenced, ignored and marginalized by NVB speakers. However, the sad fact is that, by and large, people don’t receive enough SVB reinforcement to be able to withdraw from NVB. In other words, they keep being engaged in NVB and are consequently negatively affected by it.  

“Talker-specific perceptual operations are retained or developed during the course of training, and listeners find speech from familiar talkers to be more intelligible than speech from unfamiliar talkers because they are better able to disentangle talker from linguistic information." This seems to reflect what is happening in the normal course of development: “The perceptual operations that are specifically associated with unraveling the variations introduced by particular talkers could be modified to become more efficient.” Ideally, those who raise us don’t require us to “disentangle talker from linguistic information.” Ideally, we are brought up with mostly SVB.