Pythagoras, Consonance and Dissonance - research comparing Western and Javanese instruments

[Bee_Abney]

Article:

Pythagoras was wrong: there are no universal musical harmonies, study finds

The tone and tuning of musical instruments has the power to manipulate our appreciation of harmony, new research shows. The findings challenge centuries of

www.cam.ac.uk

Research paper:

Timbral effects on consonance disentangle psychoacoustic mechanisms and suggest perceptual origins for musical scales - Nature Communications

Consonance is crucial to diverse musical styles and is traditionally attributed to simple frequency ratios between tones. Here, the authors show timbral effects on consonance that challenge this view and suggest perceptual origins for musical scales.

www.nature.com

Click to read more...

 

[AnhrithmonGelasma]

The headline seems misleading.

The study found that harmony was not "universal" in the sense that perceptions of which intervals were consonant varied with the instrument type: "“When we use instruments like the bonang, Pythagoras's special numbers go out the window and we encounter entirely new patterns of consonance and dissonance,” ... “Western research has focused so much on familiar orchestral instruments, but other musical cultures use instruments that, because of their shape and physics, are what we would call ‘inharmonic’."

But it also apparently found that perception of consonance wasn't primarily determined by the participant's culture or prior musical experience, but by which instruments were being used: "the study suggests that its participants – not trained musicians and unfamiliar with Javanese music – were able to appreciate the new consonances of the bonang’s tones instinctively. ... “Our findings suggest that if you use different instruments, you can unlock a whole new harmonic language that people intuitively appreciate, they don’t need to study it to appreciate it.”"

The article also fails to mention that inharmonic instruments have less of a clearly defined pitch, like a bell....

 

[Bee_Abney]

The headline seems misleading.

The study found that harmony was not "universal" in the sense that perceptions of which intervals were consonant varied with the instrument type: "“When we use instruments like the bonang, Pythagoras's special numbers go out the window and we encounter entirely new patterns of consonance and dissonance,” ... “Western research has focused so much on familiar orchestral instruments, but other musical cultures use instruments that, because of their shape and physics, are what we would call ‘inharmonic’."

But it also apparently found that perception of consonance wasn't primarily determined by the participant's culture or prior musical experience, but by which instruments were being used: "the study suggests that its participants – not trained musicians and unfamiliar with Javanese music – were able to appreciate the new consonances of the bonang’s tones instinctively. ... “Our findings suggest that if you use different instruments, you can unlock a whole new harmonic language that people intuitively appreciate, they don’t need to study it to appreciate it.”"

The article also fails to mention that inharmonic instruments have less of a clearly defined pitch, like a bell....

I agree, it's part of why I didn't quote it in my post or the title of the post. The research is very incomplete, but that's the case with each piece of research. It will be interesting to see the broader context. But it is suggestive, both that Pythagoras's specific universal claims are false and pointing in interesting directions. And there's lots of nice ideas to explore further. The original paper might be much clearer and more precise than the journalistic article; but I haven't read it yet.

 

[PeterN]

Article:

Pythagoras was wrong: there are no universal musical harmonies, study finds

The tone and tuning of musical instruments has the power to manipulate our appreciation of harmony, new research shows. The findings challenge centuries of

www.cam.ac.uk

Research paper:

Timbral effects on consonance disentangle psychoacoustic mechanisms and suggest perceptual origins for musical scales - Nature Communications

Consonance is crucial to diverse musical styles and is traditionally attributed to simple frequency ratios between tones. Here, the authors show timbral effects on consonance that challenge this view and suggest perceptual origins for musical scales.

www.nature.com

Someone should make same study to challenge it, bcs most 'experts' nowadays cannot be taken seriously anymore. The chance of Pythagoras being right instead of some woke researcher at Cambridge, is pretty damn solid evidence already. heh

Anyway, Burmese music in particular, (similar to Javanese) is quite a mess, but a rhythm is the tonality. Basically. (except that Burmese have pretty much fu--ed up that too, but....oh well this is getting complicated). Theres a musical chaos that can be liked, but it would not mean Pythagoras is wrong.

 

[Bee_Abney]

Someone should make same study to challenge it, bcs most 'experts' nowadays cannot be taken seriously anymore. The chance of Pythagoras being right instead of some woke researcher at Cambridge, is pretty damn solid evidence already. heh

Anyway, Burmese music in particular, (similar to Javanese) is quite a mess, but a rhythm is the tonality. Basically. (except that Burmese have pretty much fu--ed up that too, but....oh well this is getting complicated). Theres a musical chaos that can be liked, but it would not mean Pythagoras is wrong.

Well, researchers need to attract funding; and articles need to attract clicks. The chances of Pythagoras being correct? Well, as the philosopher Saul Kripke once said, to parphrase - what I said isn't a theory, at least, I hope it isn't a theory; because if it is a theory I'm pretty sure that it is false.

('[W]oke'? At Cambridge? Who knows?!)

 

[PeterN]

Well, researchers need to attract funding; and articles need to attract clicks. The chances of Pythagoras being correct? Well, as the philosopher Saul Kripke once said, to parphrase - what I said isn't a theory, at least, I hope it isn't a theory; because if it is a theory I'm pretty sure that it is false.

('[W]oke'? At Cambridge? Who knows?!)

Im wading atm through the academic article, and Im guessing (at this, around mid article) the summary and result will be that the harmonious chord is some kind of power structure. If its not said loud, its probably hidden there.

 

[Bee_Abney]

Im wading atm through the academic article, and Im guessing (at this, around mid article, point) the summary and result will be that the harmonious chord is some kind of power structure. If its not said loud, its probably hidden there.

If so, they could be some variety of conservative. The more hardcore of the left and right have a lot of common ground - just differing in their response to it! But well done for reading the real paper. I had a look at it; but I'm worried that I will struggle with it - in one way or a another.

 

[gamma-ut]

University press offices are really good at mangling the conclusions of research. The paper mentions the work of William Sethares, which the experiment seems to back up, who theorised that the harmonics of instruments are major influences on consonance. The press release goes nowhere near this and pretends that the idea is brand new (though the experiment itself is).

The core of Sethares' argument is that most music systems tend to align with the Pythagorean system because they use instruments dominated by resonating chambers, which have strong harmonics built on simple ratios like fifths.

Gamelan is unusual because it's metallophone-dominated, so the harmonics are completely different to wind and string instruments, and points to a different scale, like slendro. The big obstacle for research like this is that there are so few systems that don't inevitably wind up being based on pentatonic/Pythagorean scales.

Sethares' book by the way is Tuning, Timbre, Spectrum, Scale.

 

[quietmind]

Interesting to see this piece of research.

I did research on a similar thing in 1979. I was wondering why the gender wayang (photo of instruments attached) that I brought home from Bali and sounded so great there, turned out to sound so bad and out of tune after a few weeks back in California. I wondered... did the difference in humidity screw things up with the bamboo? Nah... these are metal bars, right?

Another question I had was around tuning. Gamelans are not tuned often. Genders, for instance, maybe every hundred years? That might have been a joke, because the Balinese sense of time is very different... but probably not more than every 20 years. Certainly not like we obsessively tune western instruments.

And in looking at primary intervals like octaves, those could range between a major 7th and a minor 9th in practice. So what about that?

Thus I was inspired to do a research project on it.

Incidentally, a gender is way brighter than a bonang, so I think there would be even more impact of the inharmonicity.

Anyway...

I sampled the 2 larger gender wayang and then did DSP analysis to break out the individual overtones (which of course were inharmonic). I could take that data and from it resynthesize (additively) new gender tones and they would be identical to the original recording. I could also add in another overtone that was not there. This I did at an octave above the fundamental frequency. Turned out there was a hole in the original set of partials at the octave. Also, the original gender tuning had highly stretched octaves.

Then I had my resynthesized instruments play a familiar gender wayang piece. The tune was created in four conditions:

1. Original gender scale tuning, original series of overtones.
2. Original gender scale tuning, overtones included my added one at the octave.
3. Modified gender scale tuning (octave set to an octave), original series of overtones.
4. Modified gender scale tuning (octave set to an octave), overtones included my added one at the octave.

I took this back to Bali and played it for professional gender players as well as for the 3 gong makers on the island who tuned instruments. And asked (with the proper scientific protocol) which they preferred when the various of the four conditions were paired off. It seems that when the octave-added overtone series was used, they preferred the in-tune octave scales. When the octave overtone was not there, as in the original series of overtones, they liked the original tuning.

So I realized that after I had returned to California and was again hearing a bunch of western tuning (I played piano, and yes, stretched octaves there too, but not nearly like in the gender)... it was not the instruments that changed. It was my brain.

 

[Snarf]

University press offices are really good at mangling the conclusions of research. The paper mentions the work of William Sethares, which the experiment seems to back up, who theorised that the harmonics of instruments are major influences on consonance. The press release goes nowhere near this and pretends that the idea is brand new (though the experiment itself is).

The core of Sethares' argument is that most music systems tend to align with the Pythagorean system because they use instruments dominated by resonating chambers, which have strong harmonics built on simple ratios like fifths.

Gamelan is unusual because it's metallophone-dominated, so the harmonics are completely different to wind and string instruments, and points to a different scale, like slendro. The big obstacle for research like this is that there are so few systems that don't inevitably wind up being based on pentatonic/Pythagorean scales.

Sethares' book by the way is Tuning, Timbre, Spectrum, Scale.

Thanks for this context. The study looks interesting. The sample sizes of the some of the individual experiments/studies are a bit small, though.

 

[Locrian MS-20]

The first paragraph of phys.org belches "The findings challenge centuries of Western music theory". I can not be bothered to read on (without payment). What a ridiculous thing to say. This is a bad faith assertion with a biased subjective premise. It reminds of the journalistic tactic of asking "So then Prime Minister how are things going with you now since you have stopped beating the wife?" Stick to the physics phys.org.

 

[rgames]

I think that paper is a good example of how academia has lost its way.

Take a look at the plot in Fig 1 - the black line (which is the basis of their claim in the later plots) is the mean through the pale blue dots behind it. If you take away the black line and just look at the pale blue dots you see the reality: asking people to rate the "pleasantness" of what they hear gives you mostly random answers. To claim some revelation in behavior based on what's in that paper seems an extremely large leap. At best they can say that their claim applies to some tiny fraction of people.

If you want to know why there's a reproducibility crisis in some scientific fields (including psychology, to the extent that it is a science), take a look at that paper. It seems to have started with a desire to achieve a clickbait outcome and has crafted a story to support it.

 

[gamma-ut]

It's hardly the first paper to use this type of technique to track consonance, so I'm not sure why you've got a beef with it over reproducibility. Fig1b is not radically different to many prior studies. I'm also unsure why a clustering algorithm based on the actual numbers is worse than simply eyeballing some dots.

I very much doubt they had the press-release title in mind when doing this research and were primarily trying to either agree or disagree with Sethares' hypothesis that timbre helps determine perceived harmonic consonance.

 

[AnhrithmonGelasma]

The first paragraph of phys.org belches "The findings challenge centuries of Western music theory". I can not be bothered to read on (without payment). What a ridiculous thing to say. This is a bad faith assertion with a biased subjective premise. It reminds of the journalistic tactic of asking "So then Prime Minister how are things going with you now since you have stopped beating the wife?" Stick to the physics phys.org.

But they didn't say which centuries....

 

[Dirtgrain]

I'm curious if people of other cultures with these instruments with different harmonics don't appreciate common western chord progressions. Do they not recognize that IV moves nicely to the V which resolves nicely to the I? I'm familiar with all sorts of world music and instruments and scales. I just wonder if there are people for whom the IV-V-I or I-VI-II-V doesn't not make harmonic or melodic sense.

 

[gamma-ut]

I'm curious if people of other cultures with these instruments with different harmonics don't appreciate common western chord progressions. Do they not recognize that IV moves nicely to the V which resolves nicely to the I? I'm familiar with all sorts of world music and instruments and scales. I just wonder if there are people for whom the IV-V-I or I-VI-II-V doesn't not make harmonic or melodic sense.

David Huron looks at this in the book Sweet Anticipation. I can't remember if it was his research or if he was quoting it but IIRC they found that musicians who had no foundation in western music – I think it might have been Gamelan players – did not recognise the classic cadence or at least did not regard those moves as special. But they did pick it up quite quickly after exposure.

You get a similar effect the other way with westerners after greater exposure to middle-eastern or Indian scales.

The upshot is: enculturation is key.

 

[rgames]

It's hardly the first paper to use this type of technique to track consonance

Understood, but that's the point: there are all sorts of papers doing this sort of thing - claiming a revelation with scant evidence to support it.

Applying scientific methods to seemingly random processes *can* reveal patterns (Galton Board), but claiming that result where the metric is a human being's response to the question "How pleasant is it?" is fraught with peril, as the data in the first plot show (note that the blue dots are omitted from all other plots).

Think of it this way: you could get their results if their claim only applies to a tiny fraction of their subjects. Let's say they tested 300 people and 5 of them supported their claim and the rest were random. You could probably get the same result because the random portion of the subject pool would average to zero effect and the few remaining people would bias it towards the results they presented.

Are 5 out of 300 people significant enough to warrant their claims? To me, no. They could very easily take their data and present it that way: what percentage of the subjects show the behavior claimed? But they did not. That kind of presentation would be much more compelling.

EDIT: Here's the bottom line: a justifiable title of the article and paper could be "For a tiny fraction of people, timbre has an impact on perceived consonance"

 

[quietmind]

I'm curious if people of other cultures with these instruments with different harmonics don't appreciate common western chord progressions. Do they not recognize that IV moves nicely to the V which resolves nicely to the I? I'm familiar with all sorts of world music and instruments and scales. I just wonder if there are people for whom the IV-V-I or I-VI-II-V doesn't not make harmonic or melodic sense.

I can speak to how it was in Bali circa 1978-9 when I lived in village, studied and lived with master musicians (and went native). This was before there was much exposure to Western music in that village (pre-electricity, pre-TV, etc.). These musicians did not really relate emotionally to the western harmonic vocabulary. It did not really interest them. On the other hand, they understood emotional nuances and meaning in the Balinese musical vocabulary on levels hard for me to grasp.

Now, of course, it's a different story. Like in so many ways, American "culture" has spread and become more dominant. Last time I was there (2010) young people were listening to a lot of Western pop and the musically inclined were playing guitar (and I almost had to inform them about the Balinese gender I played).

 

[quietmind]

Understood, but that's the point: there are all sorts of papers doing this sort of thing - claiming a revelation with scant evidence to support it.

Applying scientific methods to seemingly random processes *can* reveal patterns (Galton Board), but claiming that result where the metric is a human being's response to the question "How pleasant is it?" is fraught with peril, as the data in the first plot show (note that the blue dots are omitted from all other plots).

Think of it this way: you could get their results if their claim only applies to a tiny fraction of their subjects. Let's say they tested 300 people and 5 of them supported their claim and the rest were random. You could probably get the same result because the random portion of the subject pool would average to zero effect and the few remaining people would bias it towards the results they presented.

Are 5 out of 300 people significant enough to warrant their claims? To me, no. They could very easily take their data and present it that way: what percentage of the subjects show the behavior claimed? But they did not. That kind of presentation would be much more compelling.

EDIT: Here's the bottom line: a justifiable title of the article and paper could be "For a tiny fraction of people, timbre has an impact on perceived consonance"

With all respect (and I do respect you) I start to feel you are pushing it a tiny bit and I wonder why.

This comes from a place where I basically agree with most of what you are saying. In my early life as a trained research psychologist at Stanford, I was taught how to fudge statistics. Yipes! I found that idea repulsive. And I also saw first hand the academic pressure behind it, the desperation to publish or perish, which meant to stir up false controversies, a lot of BS, and "be relevant."

My mentor at Stanford was Roger Shepard (of that "tone"), and his approach to research and statistics was extremely different. So I never had to go down the road of traditional data analysis (and all the non-replicability that results).

Regarding the study at hand here, instead of rating pleasantness of individual stimuli, forced choice between pairs seems to me to be a more reliable approach. That's what I did in my own study in Bali in the 70s (see my previous post) seems to have replicated the study under scrutiny in this thread. But my overall take on this study is that it is a bit on the shallow side, and misses several important points. Hey, they didn't even cite my research, so who are these people anyway ??!!

* The real bottom line for me personally about my period of being a research psychoacoustician was that I was able to cofound a center at Stanford (CCRMA), partially funded on the basis of my studies on timbre perception, and out of that center came lots of technology that led to other stuff that led to more things, and ultimately probably led to a lot of the things we enjoy here on VI-C.

 

[Dirtgrain]

Noam Chomsky's theory of language acquisition holds that we have an inherited ability to learn language structures, something that evolved. I wonder if there could be something similar with music--or it could be part of the same mechanism.