THE MATHEMATICAL ARCHITECTURE OF BACH'S "THE ART OF FUGUE"

Author(s): Loïc Sylvestre and Marco Costa

Source: Il Saggiatore musicale , 2010, Vol. 17, No. 2 (2010), pp. 175-195
Published by: Casa Editrice Leo S. Olschki s.r.l.

Stable URL: https://www.jstor.org/stable/43030058

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

Casa Editrice Leo S. Olschki s.r.l. is collaborating with JSTOR to digitize, preserve and extend
access to Il Saggiatore musicale

This content downloaded from

82.49.44.75 on Sun, 28 Feb 2021 13:22:24 UTC
All use subject to https://about.jstor.org/terms
 Loïc Sylvestre - Marco Costa
Les Villards-sur-Thônes - Bologna

THE MATHEMATICAL ARCHITECTURE

OF BACH'S THE ART OF FUGUE

Introduction

The Art of Fugue BWV 1080 is one of Johann Sebastian Bach's finest
masterpieces, a true testament to his achievement as a composer. The Ar
of Fugue is the last of Bach's great monothematic cycles, after the Musical Of
fering BWV 1079, the Goldberg Variations BWV 988, and the Canonic Varia-
tions on «Vom Himmel hoch da komm' ich her» BWV 769.
Handwriting analysis and a study of watermarks in the autograph manu-
script suggest that the work was probably started in the early 1740's.1 The
first known surviving autograph version was copied by the composer in
1745 and is kept at the Staatsbibliothek, Berlin, under the manuscript refer
ence Mus. ms. Bach P200.2 At the time of his death, Bach was supervisin
the publication of The Art of Fugue. Publication had proceeded to the point
where engraver's plates had been produced but in no definitive order. Car
Philipp Emanuel Bach probably ordered the engraver's copies into what
now known as the 1751 printed sqcore. The index of both sources
summarized in Table 1 (here on p. 176). The fugues except the final on
use the same subject in D minor.
In the first published version, in comparison to the earlier manuscript, two
canons (BWV 1080/16-17), one complete fugue (BWV 1080/4), and one frag-

1 See Chr. Wolff, Zur Entstehungsgeschichte von Bachs " Kunst der Fuge", Ansbach, Bachwoche
Ansbach, 1981, pp. 77-88; and Id., Sulla genesi dell' " Arte della fuga " (1983), in Musica Noètica. Johann
Sebastian Bach e la tradizione europea , ed. by M. T. Giannelli, Genoa, ECIG, 1986, pp. 413-427.
2 See J. S. Bach, Die Kunst der Fuge BWV 1080: Autograph, Originaldruck , ed. by H. G. Hoke
Leipzig, VEB Deutscher Verlag für Musik, 1979; and Id., Die Kunst der Fuge BWV 1080, ed. by S. Var-
tolo, Florence, SPES, 2008. See also Id., Die Kunst der Fuge BWV 1080, ed. by D. Moroney, Munich
Henle, 1989.

This content downloaded from

82.49.44.75 on Sun, 28 Feb 2021 13:22:24 UTC
All use subject to https://about.jstor.org/terms
176 LOÏC SYLVESTRE - MARCO COSTA

Table 1 - Order of pieces, shown in BW

including the three appendices) and in th

manuscript P200 1, 3, 2, 5, 9, 10¿, 6, 7

(aligned), 13,1-13,2 (aligned),
printer's copy), 18,1 (App. 2),
first printed edition (1751) 1, 2, 3, 4, 5, 6
15, 16, 17, 18,1, 18,2, 19, BW

mentary fugue (BWV 1080/19 include

pendix) were added. Also included we
(BWV 1080/1 0^, which corresponds t
Counterpoint 10 BWV 1080/10 is an e
the chorale fantasia for organ «We
668^), and an arrangement for two
1 and 2, however comprised in the se
early version of one of the mirror f
included by the editor as a compleme
(BWV 1080/19), but it was not intend
as it is in a completely different sty
The fugues - each of which is design
voice on a separate stave and most are
One of the most intriguing problems c
ordering of its various compositions,
problem.4 The assignment of fugue n
numbers is in fact posthumous.5 The
175 1 edition, and in all subsequent ed

3 See J. S. Bach , ed. by M. Boyd, Oxford - N

J. Chailley, L'art de la fugue de J. -S. Bach: étude
de l'œuvre , Paris, Leduc, 1971; H. H. Eggebrecht,
München-Zürich, Piper, 1984; P. Schleuning, J
Entstehung, Analyse, Kassel, Bärenreiter, 1993.
4 See F. Szymicho WSKI , Zu den Neuordnungen v
ser und Hans Theodor David, «Zeitschrift für Mu
MANN, Die " Kunst der Fugue " als Klavierwerk.
1938, pp. 1-61; J. Chailley, L'ordre des morceau
LIII, 1967, pp. 110-136; G. G. Butler, Ordering
«Musical Quarterly», LXIX, 1983, pp. 44-61; K
Kunst der Fuge, Kassel, Bärenreiter, 1996 («Ne
E. Livingston, Notes on the "Art of Fugue": A
Butler, Scribes, Engravers, and Notational Styles:
Bach, ed. by G. G. Buder, G. B. Stauffer and M. D
2008, pp. 111-123.
5 See H. G. Hoke, Zur Handschrift Mus. ms. aut
lin, in Bach, Die Kunst der Fuge cit. (see here in.

This content downloaded from

82.49.44.75 on Sun, 28 Feb 2021 13:22:24 UTC
All use subject to https://about.jstor.org/terms
THE MATHEMATICAL ARCHITECTURE OF BACH'S THE ART OF FUGUE 111

Table 2 - List of pieces included in The Art of Fugue (BWV 1080) along
classification, typology, and total bars. The first column shows the identific
bers considered in this paper. The roman numbers in square brackets
numeration of pieces on the autograph manuscript P200. Total bars, when
cated, refer to the first printed edition (1751).

num- BWV , T 7 7
; » .f. . Name , T type JMr total 7 bars 7
ber ; classification » .f. . JMr

1 1080/1 Contrapunctus 1 [I] simple fugue 78

2 1080/2 Contrapunctus 2 [EI] simple fugue 84
3 1080/3 Contrapunctus 3 [II] simple fugue 72
4 1080/4 Contrapunctus 4 simple fugue 138
5 1080/5 Contrapunctus 5 [IV] counter-fugue 90
6 1080/6 Contrapunctus 6 a 4 in stylo francese counter-fugue 79
[VII]
7 1080/7 Contrapunctus 7 a 4 per augment, et counter-fugue 61
diminuì. [VIII]
8 1080/8 Contrapunctus Sai [X] triple fugue 188
9 1080/9 Contrapunctus 9 a 4 alla duodecima [V] double fugue 130
1 080/1 0j Contrapunctus a 4 [VI] double fugue 100
10 1080/10 Contrapunctus 10 a 4 alla decima double fugue 120
11 1080/11 Contrapunctus 11 a 4 [XL] triple fugue 184
12 1080/12,1 Contrapunctus inversus 12 a 4 [XIII] mirror fugue 56 {rectus)
and 2 and Contrapunctus inversus a 4 [XIII] 56 {inversus)
13 1080/13,1 Contrapunctus a 3 [XIV] and mirror and 71 ( rectus )
and 2 Contrapunctus inversus a 3 [XIV] counter-fugue 7 1 ( inversus )
- 1080/14 Canon per augmentationem in contrario canon 109
motu [XII; XV; App. 1]
- 1080/15 Canon alla ottava [IX] canon 103
- 1080/16 Canon alla decima contrapunto alla canon 82
terza

- 1080/17 Canon alla duodecima in contrapunto canon 78

alla quinta
- 1080/18,1 Fuga a 2 clav. [App. 2] {rectus), and fugue for two 71 {rectus)
and 2 Alio modo Fuga a 2 clav. {inversus) instruments 7 1 {inversus)
14 1080/19 Fuga a 3 soggetti [App. 3] (unfinished) 239 (ms)
triple (possibly 233 (first ed.)
quadruple) fugue

lowing an increasing level of musical complexity, the first block of four simple
fugues was placed at the beginning, followed by three stretto fugues, four
double/triple fugues, four mirror fugues, and ending with fuga a 3 soggetti
(BWV 1080/19), a fugue probably intended for four voices (seeTable2, here
above).
Gregory Buder, in particular, has been involved in detailed research into the
engraving of The Art of Fugue, dealing with all the cases in which tell-tale signs of

This content downloaded from

82.49.44.75 on Sun, 28 Feb 2021 13:22:24 UTC
All use subject to https://about.jstor.org/terms
 178 LOÏC SYLVESTRE - MARCO COSTA

erased page numbers remained in evid

example, that the augmentation can
been intended to follow the other th
and that the four canons probably f
Fugue BWV 1080/19 appears unfi
peremptorily at bar 239, while in th
233. Below bar 239, in the manuscrip
son, Carl Philipp Emanuel Bach: «Ueb
Contrasubject angebracht worden, is
generated romantic images of Bach
19 shortly before his death, but calligr
Probably the editor was striving fo
the last page of Fugue BWV 1080/19
steady hand as opposed to the erratic h
page of Fugue BWV 1080/19 was defi
death.
Many researchers have found numerical symbolism in the works of Johann
Sebastian Bach.8 These studies were particularly promoted by Friedrich Smend
who, in the introduction to his third book on Bach's Church Cantatas,9 pointed
out numerous examples of number symbolism. Much of this symbolism includes
numbers derived from the "number alphabet" in which each letter is associated
with the number of its ranking position in the alphabet. Early on in his study he
drew attention to the number 14, which has since become widely known as "the
Bach number", being derived from B+A+C+H = 2+1+3+8 = 14. These tech-
niques of gematria were well known in Bach's days. Other numerical symbolisms
were associated with theological numerology.
Some of these studies have been criticized for focusing on numeric sym-
bolism, loose logic, and lack of systematic analysis - especially by Ruth Tadow. 10

6 See Butler, Ordering Problems cit., and Id., Scribes, Engravers, and Notational Styles cit.
7 See Wolff, Zur Entstehungsgeschichte von Bachs " Kunst der Fuge" cit.; Id., Sulla genesi cit.; Id.,
Johann Sebastian Bach. La scienza della musica (2000), Milan, Bompiani, 2003.
8 See e.g. H. Norden, Proportions in Music , «Fibonacci Quarterly», II, 1964, pp. 219-222; J.-J.
Duparcq, Contribution à l'étude des proportions numériques dans l'œuvre de ]ean Sébastien Bach , «Re-
vue musicale», n. 301-302, 1977, pp. 1-59; H. A. Kellner, Was Bach a Mathematician? , «English
Harpsichord Magazine», II, 1978, pp. 32-36; Nombre d'or et musique , ed. by J.-B. Condat, Frankfurt
a.M., Lang, 1988; K. van Houten - M. Kasbergen, Bach et le nombre , Liège, Mardaga, 1992.
9 See J. S. Bach, Kirchen-Kantaten , ed. by Fr. Smend, HI: Vom 8. Sonntag nach Trinitatis bis zum
Michaelis-Fest , Berlin-Dahlem, Christlicher Zeitschriftenverlag, 1947, pp. 5-21.
10 See R. Tatlow, Bach and the Riddle of the Number Alphabet, Cambridge, Cambridge Univer-
sity Press, 1991; Id., The Use and Abuse of Fibonacci Numbers and the Golden Section in Musicologa
Today , «Understanding Bach», I, 2006, pp. 69-85; Id. Collections, Bars and Numbers: Analytical Co-
incidence or Bach' s Design?, «Understanding Bach», II, 2007, pp. 37-58.

This content downloaded from

82.49.44.75 on Sun, 28 Feb 2021 13:22:24 UTC
All use subject to https://about.jstor.org/terms
THE MATHEMATICAL ARCHITECTURE OF BACH'S THE ART OF FUGUE 179

Systematic and rigorous studies in the use of mathematical proportions

works, however, as in Tatlow, have shown that in many cases the resul
be dismissed as arithmetical coincidence. 1 1 T atlow, for example, intro
theory of proportional parallelism in which she showed that Bach inte
manipulated the bar structure of many of his collections so that they could
to one another at different levels of their construction with simple ratios
1:1,2:1,1:2,2:3.
In this essay we report a mathematical architecture of The Art of Fug
based on bar counts, which shows that the whole work was conceived on t
basis of the Fibonacci series and the golden ratio. A proportional parallelis
is also described that shows how the same proportions were used in varying
grees of detail in the work.12

Fig. 1 - Mathematical architecture of The Art of Tugue.

An analysis of the bars of the pieces included in The Art of Fugue, reveals an
architecture strongly based on the Fibonacci sequence. A phenomenon of
similarity in the distribution of golden ratios can also be observed between m
aggregate and more detailed levels of analysis. The numbers identify the C
terpoints (see Table 2). The number of bars is given in brackets. Connector l
show the golden ratios. Level 1 is the most aggregated level. In level 2 Coun
points 1-7 and Counterpoints 8-14 are each subdivided according to the go
ratio. Level 3 shows all the golden ratios considering single or small Coun
point aggregates. The number of bars in brackets shows a good approximat
to these elements of the Fibonacci sequences: ..., 89, 144, 233, 377, 610, 98
1597, ...

11 See Tatlow, Collections, Bars and Numbers cit.

12 Sec íbtd.

This content downloaded from

82.49.44.75 on Sun, 28 Feb 2021 13:22:24 UTC
All use subject to https://about.jstor.org/terms
180 LOÏC SYLVESTRE - MARCO COSTA

Method and Results

In this paper we refer to Counterpoint 14 as the fugue which is classified as

BWV 1080/19 (named Fuga a 3 soggetti in the first printed edition); Counter-
points 1-13 match BWV 1080/1-13 (see Table 2). 13
The mathematical architecture of The Art of Fugue has been investigated
considering the number of bars belonging to each piece, as reported in Table
2. The analysis is reported at three levels according to a top-down scheme, start-
ing from macrostructures that aggregate the pieces at a global level (level 1), and
proceeding to a more detailed level of analysis (levels 2 and 3). This three-level
analysis is motivated by the observation that a fractal property of self-similarity
can be found in varying degrees of detail in the work.
The total bar counts that we have considered for Counterpoints 1-13 and for
the four canons are those ofthe printed edition (1751), here in Table2. Counter-
point 14 is more problematic since in the manuscript the last bar (239) is incom-
plete, and in the printed edition it ends at the beginning of bar 233 . In this paper
we have considered the length of 239 bars, as in the manuscript. BWV 1080/10«
was not considered in the analysis since it is a preliminary version of BWV 1080/
10. BWV 1080/18,1 (Fuga a 2 clav.) and 2 (Alio modo. Fuga a 2 clav.) have not
been included in the analysis since they are harpsichord versions of BWV 1080/
13,1 and BWV 1080/13,2. The chorale BWV 668« {Wenn wir in hoechsten
Noethen Canto Fermo in Canto ) has been excluded since it was an addendum
to the first printed edition that is completely extraneous to the other material
of The Art of Fugue.
BWV 1080/12 and BWV 1080/13 are present in two versions: rectus and
inversus. The two versions are strictly related since they are obtained from the
same musical material by a process of interval inversion. In the manuscript rectus
and inversus for each of these two fugues were not written separately, but
aligned, one under the other, as a unity. For this reason the bars for BWV
1080/12 and 13 were computed once and not as the sum of rectus and inversus.
Total bar numbers differ for some counterpoints from those of the manu-
script Mus. ms. Bach P200. In particular for Counterpoints 1,2,3,9,10 the bar-
ring in the manuscript is ' while in the printed edition it is thus leading to a
doubling of bar numbers.
The allocation of bar numbers in the printed edition is due to the fact
that there is much evidence that most of the works included in the manu-
script should be considered as works in progress rather than final ver-

13 See Thematisch-systematisches Verzeichnis der musikalischen Werke von Johann Sebastian Bach,
ed. by W. Schmieder, Wiesbaden, Breitkopf & Härtel, 1990, p. 800.

This content downloaded from

82.49.44.75 on Sun, 28 Feb 2021 13:22:24 UTC
All use subject to https://about.jstor.org/terms
THE MATHEMATICAL ARCHITECTURE OF BACH'S THE ART OF FUGUE 181

sions.14 For example, music is written on plain paper in which

traced by hand with multiple drawing lines. Time signatures are so
inconsistent. For example Counterpoint 1 is marked alia breve
written in In other cases note values are diminished (as in Coun
8, 11 and 12, written in |, and in Counterpoint 13, written in |
At the beginning of Counterpoint 9 (V in the manuscript), in the A
Bach has explicidy indicated the change to be made in the final versio
more the first measure has been divided in | , as an indication of the corr
be considered throughout the work. At p. 25 of the manuscript, bef
ginning of Counterpoint 8 (X in the autograph), written in half- val
possible to read «Folgendes muss also geschrieben werden»,15 follow
beginning of the same fugue with the definitive values. The preferen
minished barring system that groups the measures 2 by 2 can be p
plained as means to facilitate counting the measures.
As regards fugue lengths, Counterpoints 1, 2, 3 were slightly shor
manuscript than in the printed edition. In Counterpoint 1 the last fou
are missing (omission of the tonic pedal and of the repetition of the
Counterpoint 2 (III in the manuscript) the last six measures are missin
sion on the dominant A, without repetition of the theme); while in Co
3 (II in the manuscript) two measures are missing at the end.

Fibonacci Series and <D

Mathematical analysis of The Art of Fugue's structure shows a strict adher-

ence to the Fibonacci sequence. A sequence un+2 = «n+i + un(n= 1) (sequence
1) is called a generalized Fibonacci sequence.16 If we set U'- a and u2 - by we
generate as an illustration of the sequence:

a, b, a + b, a + 2b, 2a + 3b, 3a + 5b, ... (sequence 1)

The classical Fibonacci sequence starts with a - 1, b- 1. The first 18 terms

are:

1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89, 144, 233, 377, 610, 987, 1597, 2584, ... (sequ

14 See Chailley, L'ordre des morceaux dans " L'Art de la fugue" cit.
15 See S. Vartolo, Saggio introduttivo , in Bach, Die Kunst der Fuge BWV 1080 cit.
fn. 2), p. 56.
16 See R. A. Dunlap, The Golden Ratio and Fibonacci Numbers , New Jersey, World Scientific,
1997.

This content downloaded from

82.49.44.75 on Sun, 28 Feb 2021 13:22:24 UTC
All use subject to https://about.jstor.org/terms
 182 LOÏC SYLVESTRE - MARCO COSTA

Fibonacci numbers are intimately li

golden ratio:
$ = 1 +y^ « 1.618
2

A generalised Fibonacci sequence has that very tidy property that all pairs of
adjacent terms are in the adjacent ratio $ but in fact it cannot be realised as any
sequence of integers. Since $ is an irrational number, it cannot be realised as a
fraction of two integers, and so it is not possible to realise the ideal starting values
as a pair of integers: it is only possible to approximate it.
From an historical perspective, the golden ratio was originally treated as a
geometrical proportion under the name of Division in Extreme and Mean Ratio
(DEMR). Ancient Greek mathematicians first studied it because of its frequent
appearance in geometry, particularly in the geometry of regular pentagrams and
pentagons. The Greeks usually attributed discovery of this concept to Py-
thagoras. The regular pentagram, which has a regular pentagon inscribed within
it, was in fact the Pythagorean's symbol.
Euclid's Elements provide the first known written definition of what is now
called the golden ratio: «A straight line is said to have been cut in extreme and
mean ratio when, as the whole line is to the greater segment, so is the greater to
the less».17

Level 1 - First Subset , Second Subset, and Canons

As shown in the Method section and in Table 2, we have analyzed fourteen

fugues (BWV 1080/1-13, BWV 1080/19) and four canons (BWV 1080/14-17)
belonging to The Art of Fugue. If the corpus of fugues is divided into two groups,
and we aggregate the first (1-7) and second (8-14) groups of seven fugues,
and consider the aggregate total number of bars, a good approximation of the
golden ratio can be found between these two subsets. The first subset
(Counterpoints 1-7) encompasses a total of 602 bars, whereas the second subset
encompasses a total of 988 bars:

E Counterpoints 8,9, 10, 11, 12, 13, 14 988 .... ,

E Counterpoints 1,2,3,4,5,6,7 602

(Here E is taken to mean 'sum of.)

E Counterpoints 1,2,3,4,5,6,7,8,9,10,11,12,13,14 1590 i ^

S Counterpoints 8, 9, 10, 11, 12, 13, 14 988

17 Euclid, Elements , Book 6, Proposition 30.

This content downloaded from

82.49.44.75 on Sun, 28 Feb 2021 13:22:24 UTC
All use subject to https://about.jstor.org/terms
THE MATHEMATICAL ARCHITECTURE OF BACH'S THE ART OF FUGUE 183

With regard to the Canons, the aggregate number of bars is 372, wh

perfect golden ratio to the first subset (Counterpoints 1-7):

E Counterpoints 1,2,3,4,5,6,7 602 Wi0 ,

H Canons 372

The aggregated sum of the bars for th

the second subset (988) and for the who
numbers that closely approximate term
(Table 3).

Table 3 - Bars for Canons, Counterpoints 1-7, Counterpoints 8-14 and their closest
Fibonacci numbers.

level 1 Canons Counterpoints 1-7 Counterpoints 8-14

The Art of Fugue 372 602 988
Fibonacci numbers òli 610 987

A fractal phenomenon of self-sim

first and second subsets separately
groups of Counterpoints can be fou

Level 2 - First Subset : Simple and

The aggregate sum of bars for Co

precisely into two groups, which ar
The first group is composed of Co
the second group is composed of

E Counterpoints 1,2,3,4
E Counterpoints 5, 6, 7 23

E Counterpoints 1,2,3,4,5,6,7 602 Wi0 ,

E Counterpoints 1,2,3,4 372

The sum of Counterpoints 6 and 7 is 140 bars (79+61), whereas Counter-

point 5 is 90 bars. 90, 140, 230, 372, and 602 correspond very closely to five ele-
ments of a generalized Fibonacci sequence starting from 90, 140 (90, 140, 230,
370, 600, ...).

This content downloaded from

82.49.44.75 on Sun, 28 Feb 2021 13:22:24 UTC
All use subject to https://about.jstor.org/terms
 184 LOÏC SYLVESTRE - MARCO COSTA

Table 4 - Bars for Counterpoints 1-4, C

numbers.

level 2: Cl-7 CI C2 C3 C4 C5 C6 C7

The Art of Fugue 372 230

Fibonacci numbers 377 233

Level 2 - Second Subset: Double/

Counterpoints 8-14 can also

in the golden ratio. The aggr
(12-13), whose total number o
fugues (8, 11, 14) whose total
golden ratio (Table 5).

E Counterpoints 8, 11, 14 611 «

-

E Co

E Co
E Counterpoints 8, 11, 14 611

The aggregated sum of all fugues in the second subset in level 2 is 988. The
values of 377, 611 and 988 are very close to the Fibonacci numbers 377, 610, and
987 (Table 5).

Table 5 - Bars for Counterpoints 9-13, Counterpoints 8-14 and their closest Fibonac-
ci numbers.

level 2: C9-14 C9 CIO C12 C13 C8 Cil C14

The Art of Fugue òli 611

Fibonacci numbers 377 610

In level 3 golden ratios between

ined. The approximation to the g
the previous levels. This is possibly
ratios are lower, being less aggreg
tional number, the lower the fr
mation to the number. Furtherm
the mathematical requirements h
cal requirements.

This content downloaded from

82.49.44.75 on Sun, 28 Feb 2021 13:22:24 UTC
All use subject to https://about.jstor.org/terms
THE MATHEMATICAL ARCHITECTURE OF BACH'S THE ART OF FUGUE 185

Level 3 - Counter-fugues 5-6-7

In counter-fugues a variation of the main subject is used in both reg

inverted form. Counterpoints 5, 6, and 7 are composed of 90, 79, and
respectively. The total number of bars for counter-fugues is 230. The
Counterpoints 6 and 7 is 140 (79+61). All these correspond to terms of
eralized Fibonacci sequence starting with 90, 140 (90, 140, 230).
The number of bars of Counterpoint 5 (90) is in approximately gold
with the total of Counterpoints 6 and 7 (140) (Table 6):

E Counterpoints 6, 7 140 „ ^ „ _
Counterpoint 5 90

Furthermore, the relationship between the total of Counterpoint 6 and 7 is in

approximately golden ratio with all counter-fugues:

E Counterpoints 5, 6, 7 _ 230 _ ^ ^ ^ ^
E Counterpoints 6, 7 140

Table 6 - Comparison between the bars in Counterpoint 5, Counterpoints 6-7, and

their closest Fibonacci numbers.

level 3: counter-fugues C5 C6 C7
The Art of Fugue 90 140
Fibonacci numbers 89 144

Level 3 - Triple Fugues 8, 1 1, 14

The total number of bars for C

golden ratio to the other triple f

E Counterpoints 8, 11 372 ^
-

Counterpoint 14 239

The total number of bars for Counterpoints 8, 11, 14 (611) is in approxi-

mately golden ratio to that of Counterpoints 8 and 11:

E Counterpoints 8, 11, 14
E Counterpoints 8, 11 372

The numbers 611 and 372 are very close t

the classical Fibonacci series.

This content downloaded from

82.49.44.75 on Sun, 28 Feb 2021 13:22:24 UTC
All use subject to https://about.jstor.org/terms
 186 LOÏC SYLVESTRE - MARCO COSTA

Table 7 - Bars in Counterpoint 14, Count

nacci numbers.

level 3: triple fugues C14 C8 Cll

The Art of Fugue 239 372
Fibonacci numbers 233 377

Level 3 - Simple Fugues 2, 4

The simple fugues 2 and 4 total 84

imate the golden ratio:

Counterpoint 4 _ 138 _
Counterpoint 2 84

Table 8 - Bars in Counterpoint 2, Count

level 3: simple fugues 2 and 4 C2 C4

The Art of Fugue 84
Fibonacci numbers 89 144

Symmetrical Relationships bet

The ratio:

E Counterpoints 6, 7 _
Counterpoint 5 90

that includes counterpoints in the first subset is nearly equal to the ratio:

E Counterpoints 8, 11 _ 372 _ ^
Counterpoint 14 239

It can therefore be suggested that these subsets of triple counterpoints have

a symmetrical function in the first and second subsets of seven counter-
points.
Similarly, the ratio:

E Counterpoints 5, 6, 7 _ 230 _ ^
E Counterpoints 6, 7 140

This content downloaded from

82.49.44.75 on Sun, 28 Feb 2021 13:22:24 UTC
All use subject to https://about.jstor.org/terms
THE MATHEMATICAL ARCHITECTURE OF BACH'S THE ART OF FUGUE 187

is equal to the ratio:

E Counterpoints 8, 11, 14 _ 611 _ ^ ^

E Counterpoints 8, 11 372

These symmetrical relations between the first and second subset a

marized in Table 9.

Table 9 - Symmetrical relations between the first (Counterpoints 1-7) a

(Counterpoints 8-14) subset of The Art of Fugue. The numbers in the top r
those of Counterpoints.

1-4 1-4 1-4 1-4 12/13 12/13 9/10 9/10

2 simple 2 simple 2 mirror 2 double

fugues 3 counter-fugues fugues 3

The symmetrical equivalence between the f

considering these ratios:

E Counterpoints 1,2,3,4 _ 372 _ ^

E Counterpoints 5, 6, 7 230

is nearly equal to the ratio:

E Counterpoints 8, 11, 14
E Counterpoints 9, 10, 12, 13 377

Furthermore the ratio:

E Counterpoints 1,2,3,4,5,6,7 _ 602 _ ^

E Counterpoints 1,2,3,4 372

is nearly equal to the ratio:

E Counterpoints 8, 9, 10, 11, 12, 13, 14 _ 988 _ ^

E Counterpoints 8, 11, 14 611

These ratios are summarized in Table 10 (here on p. 188), which highlights

the symmetrical relations between the first and second subsets.

This content downloaded from

82.49.44.75 on Sun, 28 Feb 2021 13:22:24 UTC
All use subject to https://about.jstor.org/terms
 188 LOÏC SYLVESTRE - MARCO COSTA

Table 10 - Symmetrical match between

terpoints marked in light grey and dark

1_4 1-4 16/7 1 5 6/7 1-4 1-4 mi2

2c simple v -ļ . f 2 simple
v 3 -ļ counter-fugues . f r mI c2 3mirror
- triple -
. i. fugues
i r 2 double
r r
rugues c fugues r ■ fugues c fugues r

A further symmetry between the first and second

terpoints 1, 3 and Counterpoints 9, 10.
The ratio:

Counterpoint 1 _ 78 _ ^ ^
Counterpoint 3 72

is perfectly equal to the ratio:

Counterpoint 9 _ 130 _ ^ ^
Counterpoint 10 120

This symmetrical relationship is illustrated in Table 11.

Table 1 1 - The ratio between Counterpoints 1 and 3 is equal to the ratio between Coun-
terpoints 10 and 9, showing a further symmetrical match between the two subsets.

BI^H 7 5 6 2 4 12 13 8 14 11
2 rugues
simple 3 counter-fugues 'simPk ^mirror 3 tripļe 2 double
rugues rugues rugues

Pair Matching within the Ca

The group of four canons can

the same ratio. The ratio:

Canon alla ottava _

Canon alla duodecima 78

is nearly equal to the ratio:

Canon pet augmentationem in contrario motu _ 109 _ ^ ^

Canon alla decima 82

This content downloaded from

82.49.44.75 on Sun, 28 Feb 2021 13:22:24 UTC
All use subject to https://about.jstor.org/terms
THE MATHEMATICAL ARCHITECTURE OF BACH'S THE ART OF FUGUE 189

Discussion

The results show that the whole architecture of The Art of Fugue is based on
the use of the Fibonacci numbers and the golden ratios, at least on the level of
bar totals and grouping of the pieces. Since this mathematical architecture en-
compasses the whole work, with numerous occurrences, the possibility that they
would have arisen ipso facto as a consequence of aesthetic choices or mathemat-
ical coincidences can be excluded. The significance of the mathematical archi-
tecture can probably be explained by considering the role of the work as a mem-
bership contribution to the Korrespondierende Sozietät der musicalischen
Wissenschaften,18 and to the "scientific" meaning that Bach attributed to coun-
terpoint.19
In Bach's biography there is considerable evidence of a growing interest in
Pythagorean philosophy.20 Bach had been acquainted with Johann Matthias
Gesner in Weimar, and in 1730 Gesner moved as Rector to the Thomasschule,
where Bach was Kantor. Gesner taught Greek philosophy with an emphasis on
Pythagorean thought. He even changed one of the school statutes to reflect the
Pythagorean practice of repeating all one had learned during the day before re-
tiring to bed, and the Summa pythagorica by Iamblichus was included by Gesner
in the Thomasschule norms. It is interesting that three volumes of the Summa
pythagorica (III-V) were devoted to arithmetic: De communi mathematica scien-
tia liber (Common mathematical science), In Nicomachi Arithmeticam introduc-
tionem liber (Introduction to Nicomachus arithmetic), Theologoumena arithme-
ticae (Theological principles of arithmetic).21 In all three books music is
extensively treated. Bach's emphasis on numerology and numeric symbolism
could easily have been derived from the perspective outlined in these books
in which music is described in terms of mathematical ratios and relationships
that can be found in many other domains. The study of the mathematical prop-
erties of music is understood in theological terms as a way of obtaining knowl-
edge of the divine which is embedded in a cosmological system.
Gesner was a close friend of Bach and, since their rooms were near to each
other at the school, and music played an important part in Gesner's life, we can

18 See H.-E. Dentler, L' " Arte della fuga" di Johann Sebastian Bach. Un'opera pitagorica e la sua
realizzazione y Rome-Milan, Accademia nazionale di S. Cecilia - Skira, 2000; and Id., ohann Sebastian
Bachs " Kunst der Fuge". Ein pythagoreisches Werk und seine Verwirklichung, Mainz, Schott, 2003.
19 See D. Yearsley, Bach and the Meanings of Counterpoint, Cambridge, Cambridge University
Press, 2002, as well as G. B. Stauffer's review of it («Journal of the AMS», LVIII, 2005, pp. 710-717).
20 See Dentler, L' "Arte della fuga" di Johann Sebastian Bach cit., pp. 18-53.
21 See Giamblico, Summa pitagorica , ed. by F. Romano, Milan, Bompiani, 2006 (pp. 485 ff.,
passim).

This content downloaded from

82.49.44.75 on Sun, 28 Feb 2021 13:22:24 UTC
All use subject to https://about.jstor.org/terms
190 LOÏC SYLVESTRE - MARCO COSTA

assume they had many discussions on

perpetuus BWV 1075 to him. Gesner
Lorenz Christoph Mizler, who becam
Mizler went on to create the Soziet
listed Bach as having joined the Soci
cluded Handel and Telemann). The Soc
orean philosophy and the union o
science. Each member had an oil por
bute a theoretical or practical piece
the philosophical lines of Pythagoras.
ciety that we owe both the famous 1
above mentioned Canonic Variations
BWV 769.
In the first publication of the Society, the Musikalische Bibliothek , Mizler
lists Marcus Meibom's Antiquae musicae auctores septem , graece et latine. This
book, published in Amsterdam in 1652, attempted to reconstruct the music
of ancient Greece. Very probably it was known to Bach. Mizler also refers to
the collected works of John Wallis, Opera mathematica (1693-99). Wallis was
Savilian professor of geometry at Oxford University. He made many important
contributions to mathematics, and Newton admitted that his development of
the calculus owed much to Wallis. But in addition to his work on infinitesimals,
conic sections and exponents Wallis also made specific Pythagorean refer-
ences to music and harmony. References in the Musikalische Bibliothek were
also made to Leibniz, Kepler and Robert Fludďs Monochordum mundi sympho-
niacum (1622). From his writings it is also clear that Mizler was a close friend of
Bach and that they shared common interests.
The Art of Fugue was one of three pieces written for the Society, together
with the Canonic Variations BWV 769 (a set of five canons on a popular
Christmas hymn), and The Musical Offering BWV 1079. Bach probably be-
gan The Art of Fugue in 1738, and started to rework and add to it shortly be-
fore his death, or his blindness. This mingling of music and science in Bach's
work has been confirmed by Wolff, who underlines that Bach the composer
considered himself a musical scholar producing works of musical science at a
time when Newtonian thought was very influential, especially in Leipzig.22
Bach's music was affected by the spreading culture of Newtonianism and
the general spirit of discovery that prevailed following the Scientific Revolu-
tion.

22 See Chr. Wolff, Bach' s Music and Newtonian Science: A Composer in Search of the Foundations
of His Art, «Understanding Bach», II, 2007, pp. 95-106.

This content downloaded from

82.49.44.75 on Sun, 28 Feb 2021 13:22:24 UTC
All use subject to https://about.jstor.org/terms
THE MATHEMATICAL ARCHITECTURE OF BACH'S THE ART OF FUGUE 191

T atlow has criticised the use of Fibonacci numbers and the golden
musicology. Basing her argument on the history of mathematics, she sho
it was theoretically possible for a composer in Bach's time to resort t
tive arithmetical sequence of Fibonacci numbers. However, this was unlike
that the golden section and the Fibonacci sequence receive very lim
erage in the two most authoritative sources of the "state of the art" from
at Bach's time.23 These sources were Johann Heinrich Zedler's Große
diges Universal Lexicon aller Wissenschaften und Künste , published
1732 and 1754, and the Musicalisches Lexicon by Johann Gottfried
(1732). The lack of coverage in these two sources led Tadow to concl
«there was no interest in the phenomenon, or that the new mathemat
eries had not reached Leipzig. DEMR was probably not as divine a pr
to composers in Bach's time as some musicologists would have us th
We think that Tadow's conclusion is right and logical in a genera
perspective, considering the majority of composers, but cannot be
the specific case of The Art of Fugue by Bach that we have described in t
Certainly, in the case of almost all composers, the use of Fibonacci nu
not a common and widespread practice in composition that might b
in a universal encyclopaedia or a musical lexicon. Nor was it widesprea
works, since The Art of Fugue is the only work in which Fibonacci
were used so extensively.
Tadow herself and Roger Herz-Fischler clearly show that in Bach
both the golden section as a numeric expression and Fibonacci numb
known in scientific circles.25 Bach was interested in tuning systems
building, and he would certainly have read Mersenne's Harmonie un
(1636), which gives clear instructions on the use of the geometrical p
associated with the golden ratio in discussing tuning systems and i
making.
The first known calculation of the golden ratio as a decimal was given in a
letter written in 1597 by Michael Maestlin, at the University of Tübingen, to
his former student Johannes Kepler. He gives «about 0.6180340» for the length
of the longer segment of a line of length 1 divided into the golden ratio.26 Leon-
ard Curchin and Herz-Fischler have also found an undated handwritten anno-

23 Tatlow, The Use and Abuse of Fibonacci Numbers cit., p. 81.

24 Ibid. , p. 83 .
25 See ibid. , and R. Herz-Fischler, A Mathematical History of the Golden Number , Mineóla, NY,
Dover, 1998.
26 See L. Curchin - R. Herz-Fischler, De quand date le premier rapprochement entre la suite de
Fibonacci et la division en extreme et moyenne raison ?, «Centaurus», XXVIII, 1985, pp. 129-138:
134.

This content downloaded from

82.49.44.75 on Sun, 28 Feb 2021 13:22:24 UTC
All use subject to https://about.jstor.org/terms
192 LOÏC SYLVESTRE - MARCO COSTA

tation in a sixteenth-century hand in

ments (De divina proportione) that co
bonacci series.
Peter Schreiber has also found tha
published a numerical solution DEfor
Girard also discovered the numerica
treatise published in 1634, two year
works by Kepler were published
in 1718. In a private letter to Joac
explained how he found the nume
sequence. In 161 1 he published his d
cornered snowflake). In this essay
and icosahedron, he reports that

both of these solids, and indeed the stru

without this proportion that the geometer
videa perfect example in round number
number one, the more perfect the exam
and 1 ... Add them, and the sum will be
add 2 to this, and get 5; add 3, get 8; 5
approximately, and as 8 to 13, so 13 is to

It is clear, therefore, that the attribu

tion, a connotation first introduced
community of Kepler's epoch.
Perhaps the best proof that Bach m
ries and the golden section is Johann
book Kepler reports his discoveries o
while other chapters are dedicated to
sic, to the harmonic configurations i
in relation to the motion of the pla
of harmonic proportions that can be fo
plicit references to the "divine" pro
with the Fibonacci series.
For example, when discussing i
modes or tones serve which emotion

27 See P. Schreiber, A Supplement to J. Shalli

Algorithm ", «Historia Mathematica», XXII, 1
28 Cit. in Herz-Fischler, A Mathematical His

This content downloaded from

82.49.44.75 on Sun, 28 Feb 2021 13:22:24 UTC
All use subject to https://about.jstor.org/terms
THE MATHEMATICAL ARCHITECTURE OF BACH'S THE ART OF FUGUE 193

First, then, you will remember that the hard third arose from the pentagon and
tagon uses the division in extreme and mean ratio, which forms the divine p
However, the splendid idea of generation is in this proportion. For just as a fat
a son, and his son another, each like himself, so also in that division, when the la
is added to the whole, the proportion is continued: the combined sum takes th
the whole, and what was previously the whole takes the place of the lar
Although this ratio cannot be expressed in numbers, yet some series of num
be found which continually approaches nearer to the truth; and in that series th
ence of the numbers from the genuine terms (which are not countable but ine
by a wonderful coincidence breeds males and females, distinguishable by the
which indicate sex. Thus if the larger part is in the first place 2, and the sma
whole is 3. Here plainly 1 is not 2 as 2 is to 3; for the difference is unity, by
rectangle of the extremes 1 and 3 is less than equal to the square of the mean
by adding 2 to 3 the new total becomes 5 ; and by adding 3 to 5 the total beco
The rectangle of 1 and 3 creates a female, for it falls short of the square of 2 by
rectangle of 2 and 5 a male, for it exceeds the square of 3 by unity; the rectangle
a female, for it falls short of the square of 5 by unity. Again from 5 and 13 arise
respect of the square of 8; from 8 and 21a female, in respect of the square of
on infinitely.29

Further on in Book V, Chapter n ("On the relationship of the harmo

portion to the five regular figures"), Kepler writes again about Fibona
bers and their relation to the "divine" proportion:

The second degree of relationship, which is based on origin, should be conceiv

lows. First, there are some harmonic proportions of numbers which have affi
one marriage or family, that is to say the individual perfect proportions with
family. On the other hand there is a proportion which is never expressed by w
bers, and is only demonstrated in numbers by a long series of them which gra
proach it. This proportion is called "divine", insofar as it is perfect; and it reig
ferent ways through the dodecahedric marriage. Hence the following harmon
to sketch out this proportion: 1 : 2 and 2 : 3 and 3 : 5 and 5 : 8. For it is most im
in 1 : 2, and most perfecdy in 5 : 8, and would be more perfect if onto 5 and
together, making 13, we were to superimpose 8, if that were not already cea
harmonic.30

Bach could have known about these mathematical properties since Harmo-
nices mundi is all focused on music theory and could certainly have caught his
attention. It should also be considered that Mizler, the founder of the Korre-

29 J. Kepler, The Harmony of the World (1619), Engl, transi, by E. J. Aitón, A. M. Duncan and
J. V. Field, Philadelphia, Pa., American Philosophical Society, 1997, p. 241.
30 Ibid. , p. 400.

This content downloaded from

82.49.44.75 on Sun, 28 Feb 2021 13:22:24 UTC
All use subject to https://about.jstor.org/terms
 194 LOÏC SYLVESTRE - MARCO COSTA

spondierende Sozietät der musicalisc

polymath member of the scientific c
composition, law, medicine and bein
matician and member of the Erfurt
A further possible link between Ba
meister, a theorist whom Bach read
with Kepler and Bach, Werckmeiste
point, and tied double counterpoint
the planets. Moreover, he specificall
cosmological order.32 Analysing two
plicit citations of Kepler's work Harmo
degus curiosus (1687 ) and in Hypomn
(1697).33 The latter was also revised i
Mizler.34 Access to Kepler's Harmon
could therefore have been facilitated
The key role played by Fibonacci n
chitecture of The Art of Fugue, alon
room for doubt that these relationsh
mathematical coincidences. Therefore
tentionally adopted the Fibonacci num
bers are probably used in this conte
which used to convey some particul
typical of Pythagorean philosophy, r
for concealed mathematical propertie
to Bach's admission to the Sozietät d
triple enigmatic canon for six voice
mann portrait. The canon is facing th
the other three are missing. Accordi
solutions can be found for completin
Bach's use of the Fibonacci number
The Art of Fugue makes it a unique

31 See Yearsley, Bach and the Meanings of C

32 See ibid., p. 57.
33 A. Werckmeister, Hypomnemata musica un
Olms, 1970, respectively on p. 106 and p. 38 f
34 See Th. Ph. Calvisius, review of " Hypomn
Andrea Werckmeister (1697), in L. Ch. Mizler, N
n.n., 1737, pp. 52-59.
35 See Dentler, L' "Arte della fuga " di Johan
36 See Fr. Smend, Johann Sebastian Bach bei
P- 21.

This content downloaded from

82.49.44.75 on Sun, 28 Feb 2021 13:22:24 UTC
All use subject to https://about.jstor.org/terms
THE MATHEMATICAL ARCHITECTURE OF BACH'S THE ART OF FUGUE 195

sideling its aim in connection with the Sozietät der musicalischen Wis
ten. The exact numeric proportions that we have found in The Art of
not perceptible to listeners since they are to be found in macrostructures a
at the short-term level of auditory perception. Therefore it is more probab
there are ideological-philosophical reasons for its mathematical archi
Future research should test the use of mathematical relationships in th
structure of the single pieces, by investigating The Art of Fugue at a
tailed level of analysis. Further study is required in order to ascertain if th
ematical properties highlighted in this study are also mirrored by mus
erties and if the mathematical architecture can shed further light on the i
order of pieces in The Art of Fugue.

Riassunto - Tra le ultimissime opere di Johann Sebastian Bach, L'arte del

(BWV 1080) è considerata un capolavoro della musica d'arte occidentale.
esibizione dell'ingegno contrappuntistico del compositore, essa si compone d
ghe e 4 canoni, di cui tuttavia Bach non ha stabilito la sequenza: nell'edizion
ma (1751) i pezzi sono stati disposti secondo una logica didattica, in ordine d
coltà crescente. Se però esaminiamo i rapporti tra l'estensione dei pezzi, espr
numero di battute, osserviamo che l'intero ciclo risponde a un'architettu
matematico fondata sulla serie numerica di Fibonacci; osserviamo inoltre un
lismo proporzionale dei rapporti matematici fra differenti livelli di analisi. T
diosa concezione architettonica si lascia forse ricondurre al fatto che nei suoi
anni Bach aderì alla Sozietät der musicalischen Wissenschaften, un sodalizio
tivava la filosofia pitagorica e propugnava l'intimo collegamento tra musica e
tica. In un certo senso, la perfetta architettura cui risponde l'opera si può p
siderare un sia pur criptico proclama di filosofia pitagorica; il che ve
a confermare le ipotesi formulate da Hans-Eberhard Dender nel 2000.

This content downloaded from

82.49.44.75 on Sun, 28 Feb 2021 13:22:24 UTC
All use subject to https://about.jstor.org/terms