PSYC56 Music Cognition
Lecture 5 Music Scales and Tonality
Musical Key Finding
Music Scales and Tonality
The Probe Tone Method
Krumhansl & Shepard played listeners a musical context, and this context was intended to unambiguously set
up the tonality. They followed that context with a probe tone and asked listeners how well that probe tone fit
in with the musical passage.
On a scale of 1-7, how well does the probe tone fit in?
The first probe tone was 4.
The second probe tone was 2.
The third probe tone was 6.
Major and Minor Key Profiles
Krumhansl & Kessl decided let’s try scales, single chords, and did a bunch of musical contexts. They also tried
Replicated the original result, suggesting theoretical hierarchy of stability is reflected in listener’s perception
This tonal hierarchy pattern seems to hold for major and minor keys.
It transposes to other keys.
They tried different tonalities and they tried a musical context built on the note F#.
They found the profile for F# major was exactly the same for c# major just shifted over.
C and F# Major Key profiles
They argued tonal hierarchy is abstract representation of the tones and doesn’t depend on any particular key.
It can be produced using different tonal context.
Grad student gave participants a tone and asked to imagine a major tonality based on that tone then asked to
imagine a major tonality taking that tone to a 2 ndtone degree, 3 tone degree, 4 tone degree, and so far. And
she did this for both major and minor.
A number of studies have taken this tonal hierarchy and if this is a fundamental psychological principal, then
we should find some sort of abstract presentation for musical structure if we look at other cultures.
The Petroushka Chord
Krumhansl & Schmuckler were interested in looking at the perception of tonality in bitonality music, which
involves the simultaneous sounding of two different tonalities.
C# and F# major are as far as you can get in tones.
The character of the bitonal is much different than the individual tones.
How do listeners perceive the simultaneously sounding of two different tonalities?
We played the C major line to listeners, the F major line to listeners, and then played the bitonality.
The F sharp major ratings look like the F sharp major tonal hierarchy.
Bitonal-The solid line are the ratings that we observed and it looks like C sharp and F sharp major. It is like
both tonalities are in there. The dotted lines are regression equation predicting bitonality from combination
of C and F sharp major
The listeners perceive bitonality as combination of C and F sharp major. When you hear bitonal passage, do you hear two independently separate tonalities, or is this some
combination where the two tones are mushed and cannot be pulled apart.
Used selective task, would tell participants listen to the left ear and ignore the right ear, or divided attention
you would play stereo in one ear and the protone would come up in one of the ears, and told participants to
rate the protone which one came in one ear.
Listeners were horrible at this task, they were unable to selective attend or show divided attention, no matter
what they were given it looked like c major, even if given F major. They weren’t able to attend to one of two
ears but could only listen to the higher pitch.
Thought listeners were just not trained enough so brought in Cornell university orchestra.
The Cornell university orchestra who played the petroushka chord were brought in and put through all the
above experiments and got the same results, so expertise had nothing to do with the results.
The diminished scale is an 8 note scale rather than a 7 note scale and includes all the noes of the petroushka
chord that you get plus a couple of other notes. The passage of pertrouscha was a primary example of
Bottom line- this experiment tested an extension to the idea of tonality and found that the notion of an
abstract representation of the music does hold.
What happens if we look at listeners’ perceptions of music that is written outside of the tonal system?
They took two famous examples of what is called tone rows, the first was from scoenbersgs wind winet and
the second was the string quartet. Schoenberg composed this music to get rid of tonality. You have to play all
of the twelve tones in that order, of the wind quartet. By making sure all twelve tones always occur, you are
going to get rid of tonality.
They took this tone row and played it to listeners, they played excerpts of pieces. They played this tone note
to listeners. The idea is that if listeners can hear the structure of this thing, we should get a radically different
pattern then with the notion of tonality.
The probe tone ratings for group 1 on the left is wind quartet and right is
First group of listeners were listeners who were musically trained and knew 20 century music very well.
The second group were listeners with musical training but didn’t know anything with 20 century music
Get clean data when deal with musically trained listeners, when you deal with listeners that are not trained?
There hasn’t been any good work that has tried to go at that
You get the same patterned when looking at trained vs. untrained but trained tend you give you less noise in
data, but there are exceptions.
Group 2 are listeners who are musically trained but do not have experience with schoulchers 20 century th
Classical Indian Music
In this experiment, they wanted to see does abstracts structural representation, is it true regardless of
musical demonstration? They took classical Indian rock music. There are a number of different musical scales
that this music can be based on. They got samples based of these structures and brought in a group of Indian
listeners and a group of western listeners. And they used the probe tone method, played in context and gave
individual notes and asked how notes fit in with the structure.
Essentially what you find is the tones that are considered theoretically important to those different rock
forms receive high ratings in listening probe tone ratings, and tones considered less important receive lower
ratings. The listeners are showing a classical Indian tonal rating. What was more interesting that this pattern
of results held for both the Indian listeners and the western listeners? The western listeners were also able to
pick up on the structure. There were a sum of differences between Indian and western listeners but nevertheless western listeners got the structure too. How were westerners able to pick up on structure
prevalent in Indian music?
One important result we have seen is when you play a hierarchy structure stimulus to people be it western
tonality, bitonality, atonal music, listeners pick up on that tonality such that tones that are theoretically
important are perceived by listeners as being fundamentally good, these are the good tones, the tones that fit
in well with the tonality. Tones that theoretically unimportant are perceived by listeners as being a bad fit for
the tonality. Listeners’ ratings mimic the theoretical hierarchy and this hierarchy can vary depending on the
tonal system you are using. For instance we saw listeners when you play a passage of C major respond to C
major and when you play F sharp major their ratings shift to what you expect in F sharp major.
Musical Key Finding
If we have multiple tonalities, we can raise questions, with the first question being given there are multiple
tonalities, what is the relation go one tonality to the other? Can we measure the relationship between these
two tonalities? Then if we can measure the relationship, how can we represent this? We can ask questions
that have to do with our perceptions. How do we shift our perception when we have different representation
of keys? Do we have any models for how we can perceive?
Theoretical Maps of Interkey Distances
We have this idea of multiple tonalities, how can we represent this notion of multiple tonalities? This idea of
key distance is important because the musical practice of key modulation. Key modulation is when we hear a
piece of music; this much does not stay in the