Articulate the speech, I pray you

[From Rick Marken (980403.1010)]

Bruce Nevin (980403.1207 EST)

I still have this problem with that story:

When one's mouth is numbed with novocaine, one's speech is *audibly*
affected.... When one's hearing is prevented, one's speech is not
noticeably affected.

But when one hears a delayed or distorted version of one's own
speech, the speech IS noticeably affected. I think there's good
evidence that, in speech, people control acoustical perceptions by
controlling kinesthetic (articulatory) perceptions.

Why in the world do you have a problem with the idea that one of
the perceptions we control when we talk is what we hear? The delayed
feedback results clearly rule out the notion that speech involves
the control of only kinesthetic perceptions. What's the problem?

Best

Rick

···

--
Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates e-mail: rmarken@earthlink.net
http://home.earthlink.net/~rmarken

[From Bruce Nevin (980403.1207 EST)]

Rick Marken (980403.1010)--

I still have this problem with that story:

When one's mouth is numbed with novocaine, one's speech is *audibly*
affected.... When one's hearing is prevented, one's speech is not
noticeably affected.

But when one hears a delayed or distorted version of one's own
speech, the speech IS noticeably affected. I think there's good
evidence that, in speech, people control acoustical perceptions by
controlling kinesthetic (articulatory) perceptions.

Why in the world do you have a problem with the idea that one of
the perceptions we control when we talk is what we hear? The delayed
feedback results clearly rule out the notion that speech involves
the control of only kinesthetic perceptions. What's the problem?

Rick, I know you're not stupid. I don't know if you're just careless or if
you're controlling a perception of being in a world full of stupid people.
It appears to me that you are scanning messages quickly (and carelessly)
for something that disturbs a perception that you are controlling. Whatever
the reason, you're understanding what I'm saying way too fast.

I say *again*, we control both acoustic and articulatory perceptions of
speech, and we control the acoustic perceptions by means of controlling the
articulatory perceptions.

The "deaf-folding" experiment shows that acoustic perception is not
*necessary* and that articulatory perception is *sufficient* for speaking
over time spans of an hour at least (probably much longer). The point of
the statement that you quote above, contrasting anaesthetic and
"deaf-folding", is that articulatory perception is *necessary* for speaking
in very short (immediate) time spans. Bill thinks that when we can't hear
ourselves talking we control imagined acoustic perceptions. That might be
true, but it does not necessarily follow, since articulatory perception is
both necessary and sufficient (as input to configuration and sequence
detectors) for pronouncing words.

Before you start composing a response, please read on.

We *also* control perceptions of what our speech sounds like. Even though
auditory perception is neither necessary (over relatively short time spans
--how long does it take for a deafened person's speech to start to sound
strange?) nor sufficient for pronouncing the words that we hear, acoustic
perception is both necessary and sufficient for verifying that we
pronounced the right word, and it is both necessary and sufficient for
setting and adjusting the reference values for our control of the
articulation of words. We do this not during the course of pronouncing a
word, as you seem to believe, but over numerous repetitions of words
containing the phonemes whose reference values for articulation are
changing. Please refer to Bill's comparison with throwing a ball.

Except for prolonged vowel sounds, the control of acoustic perceptions is
too slow to be involved in ongoing control of the articulations that result
in them, and in the case of stop consonants the control of acoustic
perceptions is too delayed (you can't hear and recognize that 15ms-45ms
burst-plus-transition until it's too late to modify the articulations that
produce it). Except for prolonged vowel sounds, the way our control of
acoustic perceptions controls our pronunciation is (a) by our re-saying a
word when we perceive that we said a different word by mistake (this is
what gets discombobulated with delayed feedback), and (b) by adjusting the
reference perceptions for articulating the phonemes in the word. This last
happens when we learn to speak, when we adapt to speaking a different
dialect or "accent", and evidently when we adapt to a disturbance
introduced into the acoustic signal fed back to headphones, though I still
haven't seen the article in Science or heard from Houde.

  Bruce Nevin

[From Rick Marken (980403.1600)]

Bruce Nevin (980403.1207 EST)

I say *again*, we control both acoustic and articulatory
perceptions of speech, and we control the acoustic perceptions
by means of controlling the articulatory perceptions.

Cool! I agree.

The "deaf-folding" experiment shows that acoustic perception is
not *necessary* and that articulatory perception is *sufficient*
for speaking

Ok. I would say that people are able to control the kinesthetic
perceptions of articulation so that an undisturbed side effect of
those articulations is pretty good acoustic speech sounds.

The point of the statement that you quote above, contrasting
anaesthetic and "deaf-folding", is that articulatory perception
is *necessary* for speaking in very short (immediate) time spans.

Well, for speaking clearly, yes. But I think that even if we cannot
perceive (and, thus, control) the articulations (as is the case with
novacaine) we can "throw" the articulations around well enough
to control the acoustic results of these articulations.

We *also* control perceptions of what our speech sounds like.

Ah. That's nice to hear;-) I agree!

Except for prolonged vowel sounds, the control of acoustic
perceptions is too slow to be involved in ongoing control of
the articulations that result in them

I think it would depend on which perceptual aspect of the acoustic
signal you are talking about. I think we continuously adjust
articulations for real time control of intensity, sensation
(voicing), configuration (spectrum) and probably even transitions
in the acoustic signal.

Anyway, at least you agree that some of the most important
perceptual variables that we control during speach are acoustic
variables.

Best

Rick

···

--
Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates e-mail: rmarken@earthlink.net
http://home.earthlink.net/~rmarken

[From Bruce Nevin (980403.1933 EST)]

Rick Marken (980403.1600)--

[...] articulatory perception
is *necessary* for speaking in very short (immediate) time spans.

Well, for speaking clearly, yes. But I think that even if we cannot
perceive (and, thus, control) the articulations (as is the case with
novacaine) we can "throw" the articulations around well enough
to control the acoustic results of these articulations.

Novocaine generally does not deaden pressure/touch sensations in all places
of the tongue, gingivae, and palate that are involved in articulation, so
remembered experiences of walking out of the dentist's office and saying
"phthankths" are not a good test. The topical anaesthetic would be a better
test if it could be made to cover all the bases, so to speak.

It may be that what is perceived as "throwing" articulations around the
oral cavity is kinesthetic perceptions of tongue contour, which feel less
finely differentiated than pressure/touch at different points. My
impression is that they are less finely controllable than pressure and
touch sensations. I don't know of a way that kinesthetic sensations could
be deadened separately from pressure/touch.

  Bruce