Yahoo Tuning Groups Ultimate Backup tuning Why we all disagree on categorical perception; great study that might explain a lot of things

This is in response to what Carl wrote - much of which I now agree
with, because I've learned a bit more about categorical perception
after making this post. Credit goes to Keenan for finding this study,
which is newer and conflicts with some older studies I've read (most
notably Locke and Kellar).

On Tue, Jan 24, 2012 at 11:37 PM, Carl Lumma <carl@...> wrote:
>
> Mike Battaglia <battaglia01@...> wrote:
>
> > You -DO- have a diatonic-based categorical perception. The fact
> > that yours isn't as strong as someone with AP and 22 years of
> > piano training doesn't mean you don't have one. The way you
> > heard the Bach retunings is evidence of that, as you agreed
> > the other night.
>
> I already pointed out that the Bach retunings share a common
> rhythm, which makes them sound alike. And I think Gene said
> that more extreme transformations are possible which would
> still preserve much of the character of the piece -- such as
> to the Hungarian minor. In the '90s when I first started
> fooling with tunings, I would retune familiar MIDI files to
> totally crazy 12-note scales. There's only so much damage
> you can do that way.

Part of the confusion is because I've been using the term "categorical
perception" as an umbrella term for a number of scale-internalization
effects that are assumed to come from acculturation, but after reading
this study I'm going to change the terminology I use. Specifically,
"categorical perception" is defined in the psychology literature as
when a subject perceives stimuli along a sensory continuum in the
following manner:

1. The labeling functions are "sharp," that is, stimuli are
consistently identified within the categories and there is a steep
gradient between categories;
2. There are peaks of discrimination at category boundaries while
discrimination within categories is relatively poor; and
3. There is congruence between obtained discrimination levels and
those predicted on the basis of the listener's labeling function.

Note that by this definition, the whole point of categorical
perception is that listeners have more difficulty discriminating
between intra-category stimuli as they do with inter-category stimuli,
and that it overall refers to a sort of perceptual warping effect. In
other words, someone with a 12-based categorical perception would have
a more difficult time differentiating between "subminor thirds" and
"minor thirds" than between "minor thirds" and "major thirds." So in
the future, I'll simply use the term "categorical perception" to
denote precisely this effect, which is the way it's actually defined
in the psychology literature.

There are a lot of studies showing that most western listeners exhibit
behaviors that would seem to derive from "categorical perception" to
some degree (Krumhansl, Locke & Kellar). The newer study that Keenan
found, linked to below, was designed to improve on some of the
shortcomings in the Locke & Kellar study, and ended up showing that
most non-musicians don't exhibit any evidence for categorical
perception -at all.- They tested for the existence of these criteria
directly and showed how it changes with respect to training. Here's
the link:

http://www.jstor.org/pss/40285607

And what they found was that the extent to which perceptual warping
occurs is proportional to how well listeners perform on an interval
identification test, which they chose to measure something like the
"musicality" (unfortunate word choice) of the listeners. The listeners
who performed best on this test exhibited the greatest degree of
categorization. What I view as a fundamentally important result is
that there was no evidence found at all for categorical perception in
listeners in the group with the lowest ear training test scores.

Page "211" in this study, which is really page 6 or so in the PDF,
says it all. You can see the characteristic plateau + steep gradient
behavior of true categorical perception is demonstrated most for
listeners who performed best on the test, and that the extent to which
the behavior is exhibited increases with test scores. Further along in
the study, you can also see that "same/different" discriminations had
a peak near the major/minor category boundary only for the
"Hi-musical" listeners, and that the "lo-musical" listeners had no
peak at all.

I've seen some of Paul's posts in the archives that assume that all
western listeners exhibit the behavior to some degree:

/tuning/topicId_27499.html#27525

So I assume he's going off of the older Locke and Kellar study, and
not this one. This would also seem to improve on Paul's understanding
of categorical perception (whose writings in tuning and tuning-math
are an excellent research review up to this point), as I'll discuss
below. I'll pass this onto him and see what he says.

To me, this explains why folks like Paul and Keenan have always just
immediately understood when I say things about categorical perception,
and why Gene and Ryan have never really understood it, and why you've
always sort of understood half of the things I say about it, and the
other half you disagree with. It's either something you experience or
you don't, and if you don't then it sounds like the things someone
from the other camp says aren't making any sense to you. End result:
Paul thinks that Gene's tiny comma shifts sound "seasick" to him,
whereas when I say that I hear 350 cents as multistable between a
major and minor third, Gene thinks I've lost my mind. All there in
nice pretty graphs in the study linked to above.

-Mike

🔗Carl Lumma <carl@...>

--- In tuning@yahoogroups.com, Mike Battaglia <battaglia01@...> wrote:
>
> This is in response to what Carl wrote

Really? It doesn't seem to address anything I wrote.

> To me, this explains why folks like Paul and Keenan have
> always just immediately understood when I say things about
> categorical perception, and why Gene and Ryan have never
> really understood it, and why you've always sort of
> understood half of the things I say about it, and the other
> half you disagree with. It's either something you experience
> or you don't, and if you don't then it sounds like the
> things someone from the other camp says aren't making any
> sense to you. End result: Paul thinks that Gene's tiny comma
> shifts sound "seasick" to him, whereas when I say that I
> hear 350 cents as multistable between a major and minor third,
> Gene thinks I've lost my mind. All there in nice pretty
> graphs in the study linked to above.

? Thanks for the link to the study, but I am baffled as
to how you think it supports any of the claims in this
paragraph.

-Carl

🔗Mike Battaglia <battaglia01@...>

🔗clamengh <clamengh@...>

🔗Mike Battaglia <battaglia01@...>

🔗lobawad <lobawad@...>

--- In tuning@yahoogroups.com, Mike Battaglia <battaglia01@...> wrote:

>
> To me, this explains why folks like Paul and Keenan have always just
> immediately understood when I say things about categorical perception,
> and why Gene and Ryan have never really understood it, and why you've
> always sort of understood half of the things I say about it, and the
> other half you disagree with. It's either something you experience or
> you don't, and if you don't then it sounds like the things someone
> from the other camp says aren't making any sense to you. End result:
> Paul thinks that Gene's tiny comma shifts sound "seasick" to him,
> whereas when I say that I hear 350 cents as multistable between a
> major and minor third, Gene thinks I've lost my mind. All there in
> nice pretty graphs in the study linked to above.
>
> -Mike
>

Hearing a middle third as "multistable between a major and minor
third" is just collateral damage inflicted by conditioning, if you are unable to assign the interval a musical category of its own and insist on hearing it as either major or minor.

Same goes for any interval, within reason. I think the ancient Greeks and India have it right in putting the "reasonable" smallest step size, i.e. category, at roughly a quartertone. (cf. enharmonic genus, shruti)

I think it is foolish to put a limit on color or feeling perception within that reasonable limit to distinct step size, though. And I would be wary about assuming there is anything near symmetry and consistency in either categorical or fine perceptions. I am sure many guitarists have coloristic perception when it comes to variation in bent blue notes so extremely fine that it is truly microtonal categorical perception, yet are oblivious to the singer so far off unisons as to be bitonal.

🔗Mike Battaglia <battaglia01@...>

Hi Cameron,

On Wed, Jan 25, 2012 at 10:12 PM, lobawad <lobawad@...> wrote:
>
> Hearing a middle third as "multistable between a major and minor
> third" is just collateral damage inflicted by conditioning, if you are unable to assign the interval a musical category of its own and insist on hearing it as either major or minor.

Right. And the point of all this is that I think this notion of
categorical perception - using the actual, psychological definition -
gives us a concrete framework for discussing things like this
collateral damage. For example, note what I said here (this was lifted
right out of the study text):

> Specifically,
> "categorical perception" is defined in the psychology literature as
> when a subject perceives stimuli along a sensory continuum in the
> following manner:
>
> 1. The labeling functions are "sharp," that is, stimuli are
> consistently identified within the categories and there is a steep
> gradient between categories;
> 2. There are peaks of discrimination at category boundaries while
> discrimination within categories is relatively poor; and
> 3. There is congruence between obtained discrimination levels and
> those predicted on the basis of the listener's labeling function.

The key element here is #2, which defines categorical perception as
being something entirely different from just the application of a set
of discrete labels along a sensory continuum: for something to really
be considered "categorical perception," a listener must have -more
difficulty- distinguishing between intra-category stimuli than
inter-category stimuli, which indicates a sort of perceptual warping
that's occurring.

So you might take a nonmusician and tell him "this is a major second,"
and "this is a minor second." Then, you might take a set of chords
with intermediate-sized thirds, and test to see "which one" the
subject hears each one as. This by itself isn't evidence for
categorical perception, however - it's only evidence that asked the
subject to use discrete labels for a continuous percept. What WOULD be
evidence for categorical perception is if you then tested the
perceived difference between different stimuli, and found that all
other factors being equal, there was a peak of discriminability near
the categorical divide.

For example, if a highly trained western musician exhibited more
difficulty in distinguishing between a "supermajor second" and a
"major second" than between a "supermajor second" and a "subminor
third," even if you control for things like concordance, that'd be
evidence for something like categorical perception.

So this would seem to be the same thing as the "collateral damage" you
speak of. Another area in which this happens is for things like vowel
phonemes, which people tend to perceive categorically and in a manner
consistent with the vowels used in their own language - then they
switch to another language and they have trouble switching to the new
set of phonemes. Formants, as you know, are a continuous percept.

> Same goes for any interval, within reason. I think the ancient Greeks and India have it right in putting the "reasonable" smallest step size, i.e. category, at roughly a quartertone. (cf. enharmonic genus, shruti)

I don't know, I one can probably learn to easily distinguish between
intervals differing by 33 cents or so. I don't see why not.

> And I would be wary about assuming there is anything near symmetry and consistency in either categorical or fine perceptions.

What do you mean by symmetry or consistency?

> I am sure many guitarists have coloristic perception when it comes to variation in bent blue notes so extremely fine that it is truly microtonal categorical perception, yet are oblivious to the singer so far off unisons as to be bitonal.

That'd be something interesting to test, if there's variation in
categorical perception for musicians that play fixed vs variable pitch
instruments.

What I mostly care about is now looking up strategies for a listener
with one categorical perception to re-categorize things differently.
What sorts of stimuli are best to trigger recategorization? Perhaps we
can apply those same sorts of principles to scales, and then find
scales that Western listeners with a strong categorical perception are
most likely to find to be extremely trippy.

-Mike

🔗lobawad <lobawad@...>

--- In tuning@yahoogroups.com, Mike Battaglia <battaglia01@...> wrote:

> The key element here is #2, which defines categorical perception as
> being something entirely different from just the application of a set
> of discrete labels along a sensory continuum: for something to really
> be considered "categorical perception," a listener must have -more
> difficulty- distinguishing between intra-category stimuli than
> inter-category stimuli, which indicates a sort of perceptual warping
> that's occurring.
>
> So you might take a nonmusician and tell him "this is a major second,"
> and "this is a minor second." Then, you might take a set of chords
> with intermediate-sized thirds, and test to see "which one" the
> subject hears each one as. This by itself isn't evidence for
> categorical perception, however - it's only evidence that asked the
> subject to use discrete labels for a continuous percept. What WOULD be
> evidence for categorical perception is if you then tested the
> perceived difference between different stimuli, and found that all
> other factors being equal, there was a peak of discriminability near
> the categorical divide.
>
> For example, if a highly trained western musician exhibited more
> difficulty in distinguishing between a "supermajor second" and a
> "major second" than between a "supermajor second" and a "subminor
> third," even if you control for things like concordance, that'd be
> evidence for something like categorical perception.
>
> So this would seem to be the same thing as the "collateral damage" you
> speak of. Another area in which this happens is for things like vowel
> phonemes, which people tend to perceive categorically and in a manner
> consistent with the vowels used in their own language - then they
> switch to another language and they have trouble switching to the new
> set of phonemes. Formants, as you know, are a continuous percept.

> -Mike
>

That kind of categorical perception, where everything within each region is "the same", I would call "simply not listening to the physical reality of the thing". It's a common problem among trained musicians, and is deeply unmusical, unless you conceive of music only as a language within a culture and ignore the corporeal element.

My own perceptions are a mix (or fustercluck for all I know) of different kinds, physical, categorical, etc. I did have to deliberately break some learned things, but other than this I highly doubt I'm any different than the majority of folk in perceiving things in a variety of different ways, even simultaneously.

🔗Mike Battaglia <battaglia01@...>

On Thu, Jan 26, 2012 at 8:47 AM, lobawad <lobawad@...> wrote:
>
> That kind of categorical perception, where everything within each region is "the same", I would call "simply not listening to the physical reality of the thing". It's a common problem among trained musicians, and is deeply unmusical, unless you conceive of music only as a language within a culture and ignore the corporeal element.

How do you know that it can't lead to some "musical" effects as well,
though? I can pick up a lot of "implications" when I hear something
like really fast, harmonically very free jazz that I'm not sure that
nonmusicians can. I don't consider it "good" or "bad" but rather just
an adaptation to being habitually exposed to a certain type of
stimulus. The problem is when you can't re-adapt to new stimuli, I'm
afraid.

You said that it's basically just not listening to physical reality.
Yes, I agree on some level. But, it can be very difficult, sometimes,
to listen to physical reality. Consider someone learning a new
language: they hear a set of phonemes which seem to be weird,
intermediate versions of phonemes that they know, existing between
categories and so on. So, this person might tend to round off to the
nearest phoneme. Then, when he or she talks to a native speaker in the
language being learned, and rounds the phonemes off to those of the
original language - it's the native speaker who suddenly hears weird,
intermediate phonemes! Our language learner now has "an accent." And
learning to drop your accent is difficult at first, although not
impossible. Jason Conklin earlier suggested that the L2 second
language literature might have a lot of good work on this, which might
be applicable to categorical reorientation at large.

Well, to a listener with a strong 12-based categorical perception, a
lot of microtonal music also has an accent. I don't think there's much
problem in readjusting - you just have to listen for other things than
you already are.

But, now that we're past this perceptual warping thing, I'm more
interested in figuring out what all of the different things are that
I'd associate with the term "major." Concordance is probably one of
them, but... what else? What other logical, "implicational" things
might there be? What sorts of things does one throw into the "major"
box as one acquires more and more common practice training? The whole
"categorical perception" thing seems to have been a slight red
herring, and now we're back to this again.

-Mike

🔗lobawad <lobawad@...>

--- In tuning@yahoogroups.com, Mike Battaglia <battaglia01@...> wrote:
The problem is when you can't re-adapt to new stimuli, I'm
> afraid.

That's my point. Obviously you can get art out of aspects of music which require immense conditioning- people can tell different kinds of "house" music apart, for example. I am sure that's a more subtle and cultured accomplishment than getting implications in free jazz, for even though I almost never listen to jazz, I bet you we could sit down and listen together and I could describe a whole slew of things for which I do not have a jazz-educated vocabulary, but you would recognize as subtle implications. Whereas if you put a gun to my head, I doubt I could reliably differentiate between happy house and intelligent diaper house, or whatever.

>
> You said that it's basically just not listening to physical reality.
> Yes, I agree on some level. But, it can be very difficult, >sometimes,
> to listen to physical reality.

Yes, and there are many complex and thorny issues here. My opinion is that the richest listening incorporates both conditioned and raw perceptions. Most music works against this, however- if every. single. damn. song. is in 4/4, clearly physical, animal, sexual, perception has been brutally maimed.

> But, now that we're past this perceptual warping thing, I'm more
> interested in figuring out what all of the different things are that
> I'd associate with the term "major." Concordance is probably one of
> them, but... what else? What other logical, "implicational" things
> might there be? What sorts of things does one throw into the "major"
> box as one acquires more and more common practice training?

Most obviously the contrast with minor. Male/female, black/white, good/bad. Middle thirds are the very first thing "xenharmonics" should do.

🔗Jason Conklin <jason.conklin@...>

On Thu, Jan 26, 2012 at 07:58, Mike Battaglia <battaglia01@...> wrote:

> **
>
>
> ... Our language learner now has "an accent." And
> learning to drop your accent is difficult at first, although not
> impossible. Jason Conklin earlier suggested that the L2 second
> language literature might have a lot of good work on this, which might
> be applicable to categorical reorientation at large.
>

I've started looking at the language acquisition and general phonology
literature again, to try & find some parallels and possibly inform an
experimental approach to musical categories. Will report when I can piece
together a coherent take on it.

My first impression of the past couple emails is that there's a bit of
danger in the "simply not listening to the physical reality" idea -- though
perhaps I ultimately agree with it. Mike, correct me if you have a much
different understanding here, but in the psychological notion (and
potential linguistic analogs) of categorical perception, there is an
assumption that we're talking about a *preconscious mode* of organizing
stimuli -- something our brains do to make sense of the stimuli that we
can't easily "get around". Granted, it is a form of conditioning, but we're
interested in the conditioning *mechanisms* as much as the effects. To say
it's "simply not listening" implies a kind of laziness that may not be
real, or even possible.

If that assumption stands, the ability to "crack" conditioned responses --
to hear new, category-breaking intervals and progressions as carrying
information -- is only possible through *further* conditioning, in a sense
-- unlearning one set of categories to adopt others, and only then being
able to transcend each and see more of "the physical reality" for what it
is. I doubt it ever happens quite so purely, but that gets across some of
what I'm thinking.

Note that for many second-language (L2) learners -- most? the vast
majority? -- completely losing one's accent is close to impossible. Not
talking about bilingualism here: This assumes that language study begins
after the "critical period" for language acquisition, which ends sometime
around puberty at the latest. (I make no claim that it is the same for
music or tuning!)

/jc

🔗Mike Battaglia <battaglia01@...>

On Fri, Jan 27, 2012 at 5:56 AM, lobawad <lobawad@...> wrote:
>
> --- In tuning@yahoogroups.com, Mike Battaglia <battaglia01@...> wrote:
> The problem is when you can't re-adapt to new stimuli, I'm
> > afraid.
>
> That's my point. Obviously you can get art out of aspects of music which require immense conditioning- people can tell different kinds of "house" music apart, for example. I am sure that's a more subtle and cultured accomplishment than getting implications in free jazz, for even though I almost never listen to jazz, I bet you we could sit down and listen together and I could describe a whole slew of things for which I do not have a jazz-educated vocabulary, but you would recognize as subtle implications. Whereas if you put a gun to my head, I doubt I could reliably differentiate between happy house and intelligent diaper house, or whatever.

I doubt distinguishing subgenres of electronica is as difficult as
keeping the 4/4 straight in your head while the drummer is playing a
phat backbeat with the dotted eighth as the new faux-pulse, and coming
back in on the 1 without messing it up.

> Yes, and there are many complex and thorny issues here. My opinion is that the richest listening incorporates both conditioned and raw perceptions. Most music works against this, however- if every. single. damn. song. is in 4/4, clearly physical, animal, sexual, perception has been brutally maimed.

http://www.youtube.com/watch?v=mcQ83tOZ4Wk

> > But, now that we're past this perceptual warping thing, I'm more
> > interested in figuring out what all of the different things are that
> > I'd associate with the term "major." Concordance is probably one of
> > them, but... what else? What other logical, "implicational" things
> > might there be? What sorts of things does one throw into the "major"
> > box as one acquires more and more common practice training?
>
> Most obviously the contrast with minor. Male/female, black/white, good/bad. Middle thirds are the very first thing "xenharmonics" should do.

Yes, that is a pretty good one. Carl wrote something similar recently,
about finding MOS's with consonant dyads which share a generic
interval class. I was thinking about it with triads a while ago, and I
came to the conclusion that the answer was to pair 4:6:7 and 6:9:11.
The latter is much more ambiguous, floaty, cloudy, yearning, whatever,
not sure what 12-EDO third it is on top maybe, yada yada... and then
4:6:7 is a total crunchfest. All of those first emotions sound like a
decent type of minor to me.

So far contenders to have this pairing are 2.3.7.11 maqamic[7] and
mavila[7], with the former probably being better, but I really do like
the latter I must say.

Any others? I feel like there's still more. Sometimes I feel acutely
aware of how a major third in meantone places as two major seconds,
and how each major second is two fifths, which means I'm also aware of
how each major third lays out on the circle of fifths, and I carry
that perception to other tunings without realizing it. I'm trying to
identify stuff like that; stuff which I'm carrying around without
being aware of it.

-Mike

🔗lobawad <lobawad@...>

--- In tuning@yahoogroups.com, Mike Battaglia <battaglia01@...> wrote:
>
> On Fri, Jan 27, 2012 at 5:56 AM, lobawad <lobawad@...> wrote:
> >
> > --- In tuning@yahoogroups.com, Mike Battaglia <battaglia01@> wrote:
> > The problem is when you can't re-adapt to new stimuli, I'm
> > > afraid.
> >
> > That's my point. Obviously you can get art out of aspects of music which require immense conditioning- people can tell different kinds of "house" music apart, for example. I am sure that's a more subtle and cultured accomplishment than getting implications in free jazz, for even though I almost never listen to jazz, I bet you we could sit down and listen together and I could describe a whole slew of things for which I do not have a jazz-educated vocabulary, but you would recognize as subtle implications. Whereas if you put a gun to my head, I doubt I could reliably differentiate between happy house and intelligent diaper house, or whatever.
>
> I doubt distinguishing subgenres of electronica is as difficult as
> keeping the 4/4 straight in your head while the drummer is playing a
> phat backbeat with the dotted eighth as the new faux-pulse, and >coming
> back in on the 1 without messing it up.

Depends on what you enjoy and feel drawn to, I guess. I love such things as rubato.

>
> > Yes, and there are many complex and thorny issues here. My opinion is that the richest listening incorporates both conditioned and raw perceptions. Most music works against this, however- if every. single. damn. song. is in 4/4, clearly physical, animal, sexual, perception has been brutally maimed.
>
> http://www.youtube.com/watch?v=mcQ83tOZ4Wk

Lovely timbres, that's the art of our times. Otherwise it's pretty trite, uptight and formalistic.

>
> > > But, now that we're past this perceptual warping thing, I'm more
> > > interested in figuring out what all of the different things are that
> > > I'd associate with the term "major." Concordance is probably one of
> > > them, but... what else? What other logical, "implicational" things
> > > might there be? What sorts of things does one throw into the "major"
> > > box as one acquires more and more common practice training?
> >
> > Most obviously the contrast with minor. Male/female, black/white, good/bad. Middle thirds are the very first thing "xenharmonics" should do.
>
> Yes, that is a pretty good one. Carl wrote something similar recently,
> about finding MOS's with consonant dyads which share a generic
> interval class. I was thinking about it with triads a while ago, and I
> came to the conclusion that the answer was to pair 4:6:7 and 6:9:11.
> The latter is much more ambiguous, floaty, cloudy, yearning, whatever,
> not sure what 12-EDO third it is on top maybe, yada yada... and then
> 4:6:7 is a total crunchfest. All of those first emotions sound like a
> decent type of minor to me.
>
> So far contenders to have this pairing are 2.3.7.11 maqamic[7] and
> mavila[7], with the former probably being better, but I really do like
> the latter I must say.

I find that middle thirds contrast markedly with "minor".
>
> Any others? I feel like there's still more. Sometimes I feel acutely
> aware of how a major third in meantone places as two major seconds,
> and how each major second is two fifths, which means I'm also aware >of
> how each major third lays out on the circle of fifths, and I carry
> that perception to other tunings without realizing it. I'm trying to
> identify stuff like that; stuff which I'm carrying around without
> being aware of it.

Breaking free of meantone (M3 as ditone) and middle thirds, yeah, those are important ingredients in clearly breaking from contemporary Western conception. Ozan Yarman and I discussed exactly this when I met him a couple of years ago.

🔗Mike Battaglia <battaglia01@...>

Hi Jason, sorry for the late response...

On Fri, Jan 27, 2012 at 1:04 PM, Jason Conklin <jason.conklin@...> wrote:
>
> My first impression of the past couple emails is that there's a bit of danger in the "simply not listening to the physical reality" idea -- though perhaps I ultimately agree with it. Mike, correct me if you have a much different understanding here, but in the psychological notion (and potential linguistic analogs) of categorical perception, there is an assumption that we're talking about a *preconscious mode* of organizing stimuli -- something our brains do to make sense of the stimuli that we can't easily "get around". Granted, it is a form of conditioning, but we're interested in the conditioning *mechanisms* as much as the effects. To say it's "simply not listening" implies a kind of laziness that may not be real, or even possible.

I do agree that it's preconscious. I also think there's something to
what Cameron's saying as well though. This is just how I feel in an
intuitive, imprecise way, and may not be what science says, but I find
if I really get into a meditative state and focus in on these "subtle,
different shades," I can start to really hear things in a more direct
way, and then form new labels. I can sort of chill out and expose
myself to a new sound until it really hits me and I can feel it sink
in. The thing for me is that it's always somewhat fragile. When I go
back to play in 12-EDO, the differences become unimportant again by
necessity, and it all falls apart.

What I've found at this point is that there are these tuning-specific
adaptations which occur: if I immerse myself in 22-EDO, I get
wonderfully used to 11-limit harmony and porcupine, so that the 327
cent intervals might seem to "imply" that they're an 11/9 over a
larger chord in a certain context. And if I immerse myself in 16-EDO,
I get to the point where I don't even really notice or care that the
675 cent 3/2's are beating so much. But it all then falls apart when I
go back to 12 for a few weeks or so. I've even noticed some minor ways
in which 12-EDO falls apart as well, with my AP misfiring for a few
hours until I get reacquainted.

What I want to do is develop presets for these adaptations and just
launch into them. Like we're about to play in 22-EDO? Reconfigure my
brain into 22-EDO mode. And it's 16-EDO coming up? Get into the state
where I don't care about the beating. Back to 12? Throw it all away
and go into my super digital only 12 note AP realm. That's kind of
where I'm at. Hopefully that makes some kind of sense.

But, behind it all, I still find there are 12 things in my head -
major thirds, minor thirds, etc. I feel like I'm on the verge of this
breakthrough where I can see what these tangled balls of yarn are
really composed on - the whole munit thing so far is the trippiest
I've got - but I still can't get there. I can never tell, when I hear
9/7 "as" a major third, wtf it means, no matter how much I seem to
focus in and meditate. And this is probably the preconscious thing you
were talking about. And it sucks, but it's interesting.

> Note that for many second-language (L2) learners -- most? the vast majority? -- completely losing one's accent is close to impossible. Not talking about bilingualism here: This assumes that language study begins after the "critical period" for language acquisition, which ends sometime around puberty at the latest. (I make no claim that it is the same for music or tuning!)

Aren't there adults who work at it and eventually succeed though? Is
it a matter of actual impossibility, or just difficulty? I know there
are people who have gone into "improve your accent" courses and come
out infinitely better than when they started.

-Mike

🔗Steve Parker <steve@...>

I think phonetic category is separate from accent.
I've learnt languages for work and to begin with you hear words that sometimes start with an M and other times the same word starts with a B,
even if all you're doing is rewinding and listening to the exact same sound.
I find that these 'break' and then I can hear and reproduce exactly what is there.
For any combination of speaker, native language and new language I think there are a finite and not too large number of new categories to hear.
I've definitely sussed phonetic categories in a number of languages but still have a British accent in them.
Music wise my octave has a lot of categories in it.
I knew a lot if different 'E's before I knew what they were called. When I got into JI I found that the pitches I knew were usually low limit relationships to C related to A at 440.
Unfortunately the E in 12ET has never turned into its own category - I always hear it as sharp.

Steve P.

On 28 Jan 2012, at 12:42, Mike Battaglia <battaglia01@...> wrote:

> Aren't there adults who work at it and eventually succeed though? Is
> it a matter of actual impossibility, or just difficulty? I know there
> are people who have gone into "improve your accent" courses and come
> out infinitely better than when they started

🔗Jason Conklin <jason.conklin@...>

On 28 Jan 2012, at 12:42, Mike Battaglia <battaglia01@...> wrote:

> Aren't there adults who work at it and eventually succeed though? Is
> it a matter of actual impossibility, or just difficulty?

Definitely, but it's rare. Whether that's because it's actually impossible
for a subset of individuals -- or just prohibitively difficult -- I don't
know. It's kind of hard for me to guess without reading more, as I've
always had an easier time with foreign phonology than other factors (it
tends to be my limited vocabulary, lack of idiomatic usage, and slowness in
constructing utterances that gives me away as a foreign speaker).

On Sat, Jan 28, 2012 at 07:06, Steve Parker <steve@...> wrote:

> I think phonetic category is separate from accent.
> I've learnt languages for work and to begin with you hear words that
> sometimes start with an M and other times the same word starts with a B,
> even if all you're doing is rewinding and listening to the exact same
> sound.
>
Yes, definitely! It may be worthwhile here to clarify the distinction
between phonetics and phonology, as it's at the crux of what I'm thinking
about here (sorry if you're already familiar with this, or if you just
don't care). Commence blabbing:

Phonetics looks at the sound itself (and/or its manner of articulation). So
looking at a "phonetic category" like [s], articulatory phonetics defines
it as an "unvoiced alveolar sibilant" -- i.e. in terms of how it's
articulated in the mouth and pharynx. There are endless possible
refinements here -- how forceful, exactly where on the tongue or top of the
mouth, amount of overlap with other sounds, etc. Acoustic phonetics would
define [s] in terms of its spectral qualities -- noise level and frequency
components, for example (or the formants of specific vowel sounds, for a
more familiar example). In any case, we're talking here about how to
determine what you are calling "the exact same sound".

Phonology looks at sound within the context of a single language -- a
phonological category (or "phoneme") is bound up with how sounds can be
interpreted by a speaker of a specific language -- this is where we get
into issues like you point out in language learning. In phonology, the /s/
phoneme (note the different bracketing convention) may support a range of
phonetic articulations -- you and I may say /s/ a bit differently, but any
English speaker will understand it as /s/. This may also depend in part on
the phonological, semantic, and even sociological context of an utterance.

Being from New York, I tend sometimes to say the voiced version of the
English /th/ phoneme (ideally a "voiced interdental fricative") more like a
[d] sound (a "voiced alveo-dental plosive") -- but that doesn't prevent you
from knowing my "dat" is actually the word "that". When I was in France,
though, this threw a curveball (an inswing?) to people less familiar with
English.

So, it's phonological rules that govern different ways that "the exact same
sound" may be interpreted, how different "exact sounds" may be interpreted
as one phoneme, and which sounds simply carry no information (English
probably wouldn't derive or generate the word "zthbuiookl").

To point back toard music, I'm curious whether there may be some parallels
between the way we learn and interpret phonemes and the way we learn and
interpret interval categories. You might extend the parallel by saying
phonetic categories correspond to the exact frequencies and ratios.

/jc