 |
Hidden Geometry in Music of Bach and Schoenberg: Reflection,
Rotation, Proportion
Jonathan Saggau
Abstract
Musicians often equate the composers Bach (1685-1750)
and Schoenberg (1874-1951) with a nearly obsessive relationship
to numbers and mathematics. We think often of Schoenberg as a father
of the numerical music organization systems of atonality and dodecaphony
while there is an oft cited encoding within Bach's works the numerical
and musical representation of his surname. Analysts often limit
these composers' creativity to numerical surface details in the
case of Bach or the tabulations of 12-tone row forms in the case
of Schoenberg; this paper illuminates elegant architectural structures
so prevalent behind such musical edifices. It comprises an architectural,
geometrical, and statistical walking tour of two beautiful constructions,
namely Schoenberg's Op. 19 No. 2 for piano and the Courante from
Bach's Suite for Solo Cello No. 2. Through graphic measurements
taken through modeling each piece of music onto a pitch-time two-dimensional
complex plane (after Cogan and Escot), the paper analyzes internal
architectural and geometric proportions of these works illuminating
a consistent use of arithmetic, geometric, harmonic, and golden
mean proportions amid these composers' works. It also provides graphical
illumination of various pitch and time bilaterally symmetrical structures
within the Schoenberg. Statistical contour correlations and oppositions
are also found between ordered pitch data sets obtained from equal-length
sections of the Bach using Spearman, Pearson, and Kendall data correlation
methods. Finally, the paper compares the statistical distribution
of pitch classes within the Schoenberg underlining his use of statistically
lowest total duration pitches as contextually unique information
at architecturally significant moments.
1. Bach
1.1 Architectural Proportions. The Courante from Bachs
second suite (BWV 1008) for violoncello composed in 1721 exhibits
a number of remarkable proportions. Using a two-dimensional pitch-time
complex plane (Figure 1) in which each unit along the abscissa equals
one quarter note of time duration and each vertical unit equals
one semitone from C2 to G4 the movement exhibits the following proportions:
- Sections A and B are the same length 48 quarter note
durations.
- The first iteration of the nadir pitch C2 at time point 117
is at the golden section of the movement.
- Pitch collection A2, Bb2, D3 of section A a local nadir
- is sounded at the negative golden mean of section A.
- The Apex of section A is at its golden mean.
- Pitch Bb3 at time point 24 of section A is at the golden mean
of local apexes at time points 14 and 30; the same Bb3 also bisects
section A.
- Pitch C2 at the nadir of section B is at the geometric mean
at time point 69 (time points in this case counted without the
repeats) of time points 48 and 96 quarter notes.
- Pitch E2 at the nadir of section A is at the geometric mean
at time point 69 (repeat included) of time points 48 and 96.
- Notice that the geometric mean proportions meet at the midpoint
of the piece.
- Local apex D4 at time point 15 of section A bisects the onset
and apex (time point 30) of section A.
- Pitch G4 at the apex (time point 77) and pitch C2 the nadir
(time point 117) are equidistant from the center of the movement.
- Pitch G4 at the apex of section B is at the reflection over
the midpoint of section B of the harmonic mean. (i.e. Negative
harmonic mean)
- (x x 96) ÷ (48+96) = 6448, 642
reflected over the midpoint (72) of section B = 80
The onset of the repeat of section B is at time point 144, a Fibonacci
number.
- Pitch D4 at time point 66 is at the negative golden mean of
section B.
- Pitch E4 at time point 77 is at the golden mean of section B.

Figure 1: Pitch-time complex plane
+ ZOOM
IMAGE +

Figure 2: Courante Bach suite for solo violoncello No.
2
+ ZOOM
IMAGE +
Notice from the above that sections A and B are of equal length
and repeated. The graph of Figure 1 shows the actual physical and
temporal development as the piece is performed, including repeats.
Notice also from both Figures 1 and 2 that each half (sections A
and B) seems to exhibit similar contour.
1.2 Pitch Data Correlations. When the above graph is converted
into a data set in which each available pitch is given a unique
integer (i.e. the bottom pitch on the graph is given a one, the
next higher a two, etc.) and the collection of pitches is ordered
from the first to the last such that each equal time duration is
given its corresponding pitch integer, common statistical data correlation
methods can determine similarities of pitch contour between equal-length
sections of the construction. When the above graph is similarly
converted into a data set in which, rather than mapping the collection
of pitches in register, pitch-classes are mapped (i.e. each C-natural
is given a unique integer, etc.), the same statistical data correlation
methods can determine similarities of pitch-class content between
sections A and B (Figure 3). The notion of pitch-class stems from
the assumption that any group of pitches at octave-divisible distances,
all A-naturals for example, sound equivalent to a large degree regardless
of register.

Figure 3: Data Correlations
+ ZOOM
IMAGE +
Observe that there exists great statistical pitch contour similarity
between sections A and B as noted when examining the complex plane
graph (Figure 1) above. Note also that contour similarity diminishes
when comparing subsections by further halving sections A and B to
obtain divisions of 25% of the total duration of the piece. The
above outlines also striking internal oppositions of pitch and pitch-class
structure. Namely, the first halves of sections A and B (labeled
1st and 3rd 25% above) share great similarity in both pitch and
pitch class domains whereas halves within sections (1st compared
to 2nd 25% and 3rd compared to 4th 25%) share considerably less
contour and pitch-class information. Also note that the correlation
between the 2nd and 3rd 25% (spanning the border between sections
A and B) exhibits the very least similarity among adjacent divisions
in both contour and pitch-class domains. This recursive similar/dissimilar
structure begins to illuminate an interesting phenomenon found in
many AABB pieces of Bach [1].
Most listeners will experience the borders between repeats of sections
A and B and between sections A and B themselves particularly jarringly
due to almost complete change of musical context at those points
as illustrated by the correlations above. Bach places the first
25% of the piece (first 50% of section A) squarely in the key of
D minor and gradually modulates to the dominant key of A major (and
for a short while, A minor) within the second 25%. The listener
is then jolted suddenly back to the beginning of the piece (D minor)
hearing section A again. The same process is repeated within section
B beginning with A major, gradually modulating to D minor and suddenly,
once again, jerking the listener back to A major, this time with
a gigantic leap of more than two octaves. We can see from bar graphs
(Figure 4) below the magnitude of oppositions of pitch and pitch-class
usage between sections. Notice also that the first 25% and the final
25% exhibit great statistical similarity, indicating a return to
opening materials (D minor key).

Figure 4: Pitch Content
+ ZOOM
IMAGE +
2. Schoenberg
2.2 Architecture. Schoenbergs Opus 19 No. 2 for solo
piano composed in 1911 (nearly two hundred years after the Bach
piece) exhibits remarkably similar architectural proportions and
processes. Using a two-dimensional pitch-time complex plane as above,
the music exhibits the following proportions:
- Range (semitones) represents approximate golden mean (GS) proportion
relative to time (eighth note). Ratio of the length of X Axis
in durational values to the length of the Y Axis in semitones
= Golden mean -- 46 semitones / 72 eight notes)
- Significant events structurally are represented in similar relative
positions in the pitch and time domains (+GS and GS, .44
and .54 lines represent the same positions relative to
minima and maxima on both the pitch and time axes).
- Golden Sections (+,-) in both pitch and time domains mark significant
events.
- When the 14 Eighth note introduction is subtracted from the
piece (See dark lines):
- The remaining 57 eighth notes are bisected by the positive golden
mean of the piece.
- The golden mean of the remaining 57 eighth notes is found at the
end of the chord containing the lowest pitch of the piece. (Notice
that the golden section (labeled GM) of the piece as a whole is
found at the beginning of this chord.)

Figure 5: Pitch-time complex plane
+ ZOOM
IMAGE +
2.3 Rotation. Note that structurally significant points
within each axis, pitch and time, occur at similar relative positions.
For instance, the higher of the two ostinato pitches (the opening
pitches repeated throughout the piece) occurs at the positive golden
section of the pitch (Y) axis while the chord at the nadir of the
pitch axis occurs at the positive golden section of the time (X)
axis. Notice also the chain of thirds just prior to the above chord
commences at .54 of the total time duration of the pitch while the
bottom pitch of the above ostinato occurs at .54 of the total pitch
span. There exist other such proportional rotations of the pitch
axis onto the time axis as illustrated in Figure 6 below.

Figure 6: Rotations of pitch axis onto time axis
+ ZOOM
IMAGE +
2.4 Symmetry. Schoenberg further emphasizes the significance
of the above structural points by focusing bilaterally symmetrical
events on them wherein they become the axis of symmetry. For instance,
the apex of the piece is an axis of symmetry in the time axis of
the beginning of the piece and the final chord in the descending
thirds chain near the end of the piece (figure 7). Also, the final
chord in the ascending thirds chain just prior to the chord containing
the nadir pitch produces the axis of symmetry for the onset of the
first pitches outside of the ostinato and the decay (the end) of
the final chord. The first pitches outside of the ostinato also
find a reflection over the golden section point of duration in the
form of the first chord in the ascending third chain. Schoenberg
thus relates the first unique event to events leading to the unique
chord at the golden section of duration and at the nadir of pitch.
Both self-similar symmetry as well as symmetry similar to the
above (centered around significant structural points) is found in
the pitch domain. When the pitch content of the music is delineated
into sections with the guidance of the time proportions above, symmetry
within the pitch domain becomes evident. The pitch content of the
first 30 eighth notes of the piece exhibit two such symmetries:
The local apex and nadir are reflections bisected by the positive
golden mean of the total pitch range.
The local nadir and the highest pitch below the ostinato are bisected
by the negative golden mean of the total pitch range.
Note other symmetrical pitch structures centered on the positive
and negative golden mean of total pitch range below. Self-symmetrical
structures such as that of eighth notes 31-40 are often bounded
by structurally important (golden mean) points within the pitch
domain as illustrated below.

Figure 7: Bilateral symmetry
+ ZOOM
IMAGE +
2.5 Pitch Statistics. Schoenberg also distributes pitch-classes
to color unique events. The bar charts below illustrate the heavy
use of pitch-classes G-natural and B-natural, the ostinato pitches
as well as the statistically secondary pitch-classes C, D, D-sharp,
F, F-sharp, and A-sharp, which are used within the golden mean chord,
the first gesture outside of the ostinato, and the apex chord (but
not the apex pitch) thus linking those temporally removed events
contextually through pitch material (and extreme registral placement).
Notice from Figure 8 below placement of the statistically lowest
total duration pitch-classes, C-sharp, E, G-sharp, and A, labeled
within the music below. The highest pitches and lowest pitches of
the descending and ascending thirds chains are most unique pitches
contextually and add weight to the gestures leading into and traveling
away from the structurally important chord of the golden mean. Note
that Schoenberg places the E-natural (F-flat), the least used pitch-class,
as the highest pitch sounded and as a member of the final descending
chain of thirds chord, thus singling them out as extremely important
events. Notice also (labeled in Figure 8) that each of those pitches
is equidistant from the other and the very beginning of the piece,
obtaining yet another bilateral symmetry.

Figure 8: Pitch statistics
+ ZOOM
IMAGE +
3. Conclusions and Similarities
We see from the above analysis the architectural genius shared
by these two composers of music. Bach and Schoenberg were not merely
interested in the surface details of their constructions as often
preoccupies the music theorist but upon the self-referencing structure
and proportion within and between their multiple dimensions. Bachs
construction seems occupied with the internal temporal proportions
while Schoenbergs more modern structure exhibits proportions
similar both within and between the pitch and temporal dimensions
through rotation and symmetry. Note that Bachs construction
in the pitch domain unfolds in a very typical (for the time period)
linear fashion perhaps not requiring such rigor proportionally to
hold the dimension of pitch together as opposed to the carefully
proportioned architectural pitch-time structure that Schoenberg
requires to tie together his more or less nonlinear unfolding. Bach
places alternations of tonic (D minor) and dominant (A major) pitch-class
material to exhibit both opposition and unification within and between
sections of his structure while Schoenbergs placement of contextually
unique pitch material similarly opposes and unifies otherwise temporally
nonlinear gestures within his construction.
(The analytical tool of the pitch-time complex plane is found first
in the masterful groundbreaking theoretical work Sonic Design of
Pozzi Escot and Robert Cogan [2].)
References
1. Hofstadter, Douglas R., Godel, Escher, Bach: an Eternal Golden
Braid (2nd edition), New York: Basic Books, 1999.
2. Cogan, Robert and Escot, Sonic Design: The nature of Sound and
Music, Cambridge, MA: Publication Contact International, 1984.
Bach, J.S., Suite for Solo Cello, No. 2
Schoenberg, Arnold, Six Small Piano Pieces, Opus 19, No. 2
Jonathan Saggau
is a PhD candidate in Music Theory and Composition at New York University,
was the winner of the BMI Student Composer Award, has received commissions
from Iowa State University, and has lectured on music in America
and Europe.
Top of page
|
 |