Axis Tonality and Golden Section Theory Reconsidered

Author(s): Jnos Krpti

Source: Studia Musicologica Academiae Scientiarum Hungaricae, T. 47, Fasc. 3/4, Bartk's
Orbit. The Context and Sphere of Influence of His Work. Proceedings of the International
Conference Held by the Bartk Archives, Budapest (22-24 March 2006). Part I. (Sep., 2006), pp.
417-426
Published by: Akadmiai Kiad
Stable URL: http://www.jstor.org/stable/25598271 .
Accessed: 03/03/2014 11:43
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
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 [email protected].

Akadmiai Kiad is collaborating with JSTOR to digitize, preserve and extend access to Studia Musicologica
Academiae Scientiarum Hungaricae.

http://www.jstor.org

This content downloaded from 130.64.196.119 on Mon, 3 Mar 2014 11:43:44 AM

All use subject to JSTOR Terms and Conditions

Axis Tonality andGolden Section Theory

Reconsidered

JanosKarpAti
Liszt Ferenc University ofMusic

Liszt Ferenc

E-mail:

(Received:
Abstract:

After

three decades

ter 8, H-1061

Budapest, Hungary

[email protected]

25 March

2006; accepted:

25 May

2006)

of our personal,

conducted
with
discussions
publicly
to his theories at a
related
again my objections
of Bence
Szabolcsi.
Since my lecture was given in
to present
version was
in Hungarian,
I feel it necessary

in 1999 I raised
Lendvai,
inmemory
conference
organized

Erno

and its published

Hungarian,
some of my objections
at an international forum as well, with particular reference to the
in spite of serious criticism
literature
fact that in the Bartok
(Petersen, Gillies)
several analysts today still employ Lendvai's
theories in a servile way. My objections
of Riemann's
four points. (1) The extension
three-function
theory to the
system is theoretically arbitrary and an impasse. (2) The axis interpretation
of tonalities
is in flat contradiction with Bartok's
by identification of polar keys
tonal thinking. (3) The pentatonic
scale interpreted as a golden section system is very
in doubt according
to the basic findings of ethnomusicology.
much
(4) The typical
focus upon
twelve-tone

as a, p etc. are phenomenologically

chord structures named by Lendvai
numbers
is arbitrary, because
the actual
correct, but their interpretation using Fibonacci
intervals represent other ratios.

Bartokian

Keywords:

Bela

Bartok,

Erno Lendvai,

theory of music,

golden

section

After three decades

of personal, publicly conducted discussions with

in
I
1999
raised
Lendvai,
again my objections related to his theories at
tional conference. At that timemy lecture was given, and later published,
in Hungarian,1 and I feel it necessary to present my objections at an
national forum as well. Many

ofmy colleagues may

Erno
a na
only
inter

think I am an old man who

1
kerdesei: Meg egyszer Lendvai Erno elmeleterol" [Questions of the Bartok
?A Bartok-analitika
In: Bartok-analitika.
Analysis: Erno Lendvai's Theory Reconsidered], Muzsika XLIII, 2000, 4:11-16.
Vdlogatott tanulmdnyok [Bartok Analysis, Selected Writings] Budapest: Rozsavolgyi, 2003,156-167.
StudiaMusicologica Academiae ScientiarumHungaricae 47/3-4, 2006, pp. 417-426
D0I:
l0.1556/SMus.47.2006.3-4.15
0039-3266/$ 20.00 ? 2006 Akademiai Kiado, Budapest

This content downloaded from 130.64.196.119 on Mon, 3 Mar 2014 11:43:44 AM

All use subject to JSTOR Terms and Conditions

418 Janos Kdrpdti

to repeating

riding his hobby-horse, but itworth

greater number of Lendvai fans.

keeps

because

of the still

at the Budapest Academy

of Music
that studied musicology
as
Lasz
and
Balint
Sarosi
such
classmates
between 1951 and 1956
including
inBartok analysis. We
witnessed directly Lendvai's
lo Somfai
appearance
had in our hands his book entitled Bartok's style published in 1955, and fur
The

class

Some
thermore, we could listen to his fascinating lectures at theAcademy.
most
his
the
and
students
considered
teachers
theory
important
colleagues
and convincing discovery in Bartok analysis, but others disputed itmainly
and extreme aesthetic implications.
because of its arbitrarymethodology
International reception of his theory was similarly divided. Enthusiastic
for example the Hungarian-born Andras Szentkiralyi
followers emerged
but objections were not lacking either. Very early, even in 1971, the year of
from Hamburg, Peter
Lendvai's English language publication, a musicologist
Petersen, in his book on Bartok's tonality, wrote this very sharp and apposite
opinion:
...the

connecting
system is nothing

of axis
tonality
else but fiction.2

with

the harmonic-functional

relation

Since dealing with my academic studies inBartok analysis I could not avoid
theories; and itwas impossible not to accept a number
encountering Lendvai's
statements and appropriate terminology. But early
of his phenomenological
inHungary had some political overtones, mainly from
criticism of Lendvai

in
aestheticians, which iswhy I and some ofmy colleagues
as
to
such
to
not
want
be
considered
did
Bartok research
attackers,
belonging
and theoretical level.
because our objections remained on a purely musical
Thus those who disputed his theories restricted themselves to criticizing only
some peripheral elements, the followers became ever stronger, and occupied
that the problem of
important places in the education. Many musicians believe
axis theory can
Bartok analysis has been completely resolved, and Lendvai's
the side ofMarxist

and taught like Rameau's

harmony, since it seems
terms
and
with
learnable
comprehensible figures.
system

be used

to be a closed

Inmy firstbook on theStringquartetspublished inHungarian in 1967 and

in 1975, and later inmy book on Bartok s Chamber Music, while

some of Lendvai's
ideas, I started to dispute some of the theses
acknowledging
forUS publication
he presented. When I revised my Bartok s Chamber Music
on the inter
in 1994, consulting the studies that had appeared meanwhile
was
some
literature
of
the
I
that
realized
national scene,
similarly
analytical
from
theory. I received for example encouragement
criticizing the Lendvai
in English

2 Peter

Petersen, Die

StudiaMusicologica

Tonalitat in Instrumentalschaffen von Bela Bartok (Hamburg,

Academiae ScientiarumHungaricae

47, 2006

This content downloaded from 130.64.196.119 on Mon, 3 Mar 2014 11:43:44 AM

All use subject to JSTOR Terms and Conditions

1971), 11-12.

419

Axis Tonality and Golden Section Theory Reconsidered

Malcolm

Gillies's

review-article

on Lendvai's

Workshop
Bartok book, which

of Bartok and Ko
includes as an Ap

daly (1986) and from Paul Wilson's

pendix a whole chapter of serious criticism.3 There is no room to quote the
main points of these, but I quote one passage from Gillies which very well
illustrates the problem:
If I have concentrated somuch on the shortcomingof thisbook it is because
Lendvai's

Workshop

is dangerous

reading,

especially

if undertaken

casually.

Itsdistortionscan fairlyeasily be identified-but onlywith therelevant scores

inhand and a generous allocation of time...These conceptswill be called upon
by others in a superficialway. Already some of thework of Tibor andMaria
Bachmann

shows

that this danger

is a real one.4

In this paper I do not strive for anything other than to focus on some elements
of Lendvai's
theories which I consider basic errors on which no system can be
built. Please do not think that I am doing this because I feel secure now that

cannot reply with his dynamic arguments, well known to his

I had disputes with him during his life both in speech and in
older colleagues.
the
way, he is still present through his disciples, and his writings
writing. By
are accessible in full invarious languages.
Erno Lendvai

My objections focus upon four points.

(1) The extension of Riemann's three-function theory to the twelvetone system.
(2) The axis interpretation of tonalities.
(3) The pentatonic scale interpreted as a golden section system.
(4) Typical Bartokian

chord structures interpreted via the Fibonacci

1. The extension ofRiemann

sequence.

s three-function theory to the twelve-tone system

It is true that the denomination of chords as tonic, dominant and subdominant

came from Rameau, but the so-called function theory - including the succes
sion and substitution theory was created by Hugo Riemann at the end of the
19th century. The theory is logical and attractive indeed, but in several aspects
it is pure simplification. Its creator was quite aware of thiswhen entitled his
work published in 1893: Vereinfachte Harmonielehre
oder die Lehre von den
tonalen Funktionen derAkkorde. Among Hungarian music pedagogues many
followers of Riemann's
theory cropped up, but the higher level theorists did
not accept it unconditionally. A very good example of their criticism was
drawn up by professor Jozsef Ujfalussy at a musicological
conference:5

3 Paul
Wilson, TheMusic ofBela Bartok (New Haven/London,
1992).
4Malcolm
Gillies, Review Article: "Erno Lendvai: The Workshop of Bartok", Music Analysis 5/1986,
2-3.
5
?Funkci6: terminologia es logika" [Function, terminolgy, and logic]. InZeneelmelet es stiluselemzes. A

Bdrdos Lajos 75. szuletesnapja alkalmdbol tartottzenetudomdnyi konferencia anyaga [Music theory and
1977,7.
styile analysis: Conference held for the 75th birthday of Lajos Bardos] (Budapest: Zenemiikiado),
StudiaMusicologica Academiae ScientiarumHungaricae

This content downloaded from 130.64.196.119 on Mon, 3 Mar 2014 11:43:44 AM

All use subject to JSTOR Terms and Conditions

47, 2006

420 Janos Kdrpdti

Our unhistorical theoryteaching isguiltyof thefact thattheRiemann concept
of the function and its threefoldsystem live as an ever existing and eternal
manifestation in the thinkingof several generations ofmusicians.
It should be kept inmind
followers of theRiemann

Europe therewere no enthusiastic

theory, inFrance and inEngland itdid not spread at
a significant opposing party emerged already from
that inWestern

all, and even inGermany

the beginning of the 20th century, in the persons of such eminent composers
and theorists as Arnold Schonberg, Paul Hindemith, Ernst Kurth, Sigfrid
continued after the
and Heinrich
Schenker.6 The discussion
Karg-Elert
Second World War, when Helmut Federhofer and Carl Dahlhaus
stepped on to

the scene.7 By such an investigation of wider aspect it became clear that,

although the succession of chords can be reduced to certain laws of attraction,
theRiemann theorywas built upon a very artificial and speculative basis.
is no place here to recall this entire discussion, but it is appropriate to
a unique argument presented by Federhofer regarding the altered II

There

quote

function theory is
degree seventh chord. The basic priniciple of Riemann's
or
can substitute
lower
triad
that the so called "secondary"
relative)
(parallel
common
share
the given primary triad, because
notes, accordingly
they

insteadof theIV degree triadinC major (F-A-C) theII degree triad(D-F-A)

can be presented, having the same subdominant function. But if the third of the
II degree triad is altered toF sharp, says Federhofer, the note F representing the
subdominant function is no longer there, in other words D major is not the

relative of F major. So the subdominant function disappears, and the subdo

relation.
relation is overwritten by the dominant-tonic
minant-dominant
a
to
valid
and
be
law, Lendvai
permanent
eternally
theory
TakingRiemann's

builthis axis systemupon it.Let us quotehis deductionofWorkshop(p. 271 ):8

In European music, functional way of thinkinghas been establishedwith the
recognition

1
Classical

of I-IV-V-I

affinition:

SUBDOMINANT

harmony

TONIC

I
already

DOMINANT

between

distinguishes

primary

and

secondary

triads- in so faras the I, IV and V degrees may be substitutedby the relative

VI, II and III degrees respectively:
1

SUBDOMINANT
F

TONIC

DOMINANT
G

6 E.
Kurth, Die Voraussetzungen der theoretischen Harmonik
(Bern, 1913), see the main works
S. Karg-Elert and H. Schenker.
7 Carl
Dahlhaus, Einjuhrung indie systematische Musikwissenschaft
(Koln, 1971).
8 Erno
Lendvai, The Workshop ojBartok and Kodaly (Budapest, 1983).
StudiaMusicologica Academiae ScientiarumHungaricae

47, 2006

This content downloaded from 130.64.196.119 on Mon, 3 Mar 2014 11:43:44 AM

All use subject to JSTOR Terms and Conditions

of

421

Axis Tonality and Golden Section Theory Reconsidered

Romantic music continues with this progression - naturalizing the upper
relatives:

SUBDOMINANT I
F

TONIC
C

D j
1Al,

DOMINANT

El,
| EBlj

The axis system isnone other than the recognitionof thefact thatthe tonicA
and E flatnot only have C as a common relative,but also theF sharporG flat
degree.

SUBDOMINANT
F

TONIC

_C_G_
DAl,

A_Et_E_

fj

DOMINANT

as
At firstglance, this deduction looks logical and attractive. Nevertheless,
we have seen inFederhofer 's argumentation, the substitution in the case of the
lower relative is already equivocal, and the simultaneity of lower and upper
For example the VII degree represents an
I
function.
quote one of professor Bardos' statements:
absolutely ambiguous
A special case is given by the third-fourth
inversionof theVII seventh chord.
When itresolves to I degree (T), if Iwant, from theone pair of voices I am
hearing aD-T relation,but ifI am focussing on theotherpair, it is clearly S-T
character [Example 7].As amatter of fact,both functionsare unified, and as if
relatives raises

there were

further questions.

a certain

three functions,

in which
there are not
of the newer music
phenomenon
but only the tonic and some other opposing
tones, as an anti

tonic.

vn3

s? t

Example 1
to extend the
That iswhy it is an arbitrary and speculative idea of Lendvai's
relative (parallel) relationships to the twelve-tone system, and to force all the
three functions into an axis, even though in the twelve-tone system (as we
three functions.
quoted above) there are not and cannot be
The theory of axis tonality is disputable not only because Riemann's
three

fold function system is not adaptable to the twelve-tone chromatic system, but
if the first relative - as we have seen from Federhofer's example - is

because

not always valid, then the second relative starts from a false premiss. IfA ma
jor is not the relative of C major, then similarly F sharpmajor is not the relative

ofA major, and even less of C major. By such a false extension of the relativity
of tonalities, Lendvai arrives at the absurdity of the identity of polar tonalities.
StudiaMusicologica Academiae ScientiarumHungaricae 47, 2006

This content downloaded from 130.64.196.119 on Mon, 3 Mar 2014 11:43:44 AM

All use subject to JSTOR Terms and Conditions

422 Janos Kdrpdti

2. The axis interpretation of tonalities.
I quote one of his main, most emphasized statements (from page 275):
A pole can be replaced by itscounterpolewithout any change in its function.
This statement is exclusively true on the level of formal logic, whereas Bartok
did not speculate, but thought of real sounds. In his music tonality played first
of all a structural role. For example,
in the Sonata for two pianos
and
percussion,

introduction starts from F

the slow

through complicated

sharp tonality, and, going

in
at the tonality of C.
arrives
tension,
high
starts
it
from
the
tonic
and
arrives at the tonic.
theory,

material

to Lendvai's
According
itnot be a better interpretation to say that the progression starts from a
Would
distant tonality and arrives at the central, actual tonality of thework?
For Bartok polar tonalities by no means represented the same function, but
were

have

rather a distant relation substituting a traditional tonal connection. We

to hand his draft analysis in French for the firstmovement of the fifth

string quartet. He wrote:

Le theme principal a deux degrees principales: Si bemol (tonique) etMi
(Jouantle role de la dominante)?

to illustrate the axis system,

circle of fifths, so often used by Lendvai
one
the
musical
of
shows
of Bartok's principal com
actually
progression
movement
ofMusic for strings, percussion
positions, in the slow introductory

The

and celesta, but just in the opposite way to Lendvai's

interpretation. As it is
well known, the entries of the fugue subject follow each other in fifth layers,
and downwards
namely upwards in the dominant direction (E-B-Ft-Ctt-Gt)
each other on
in the subdominant direction (D-G-C-F-Bt),
finally meeting
Dfl and B. But this enharmonic summit of the progress cannot be interpreted as
a tonic level, as here a real "super dominant" fuses with a "super subdomi
- as
Lendvai says
but are syn
nant", and they do not neutralize each-other
thetized in themost distant tonality (Example 2).
This time I do not wish to talk about the golden section proportions

time calculations regarding the hori

exhibited inBartok's forms; ifLendvai's
zontal proportions are correct, then they cannot be disputed. I shall concen
trate inmy paper on the question whether golden section calculations are valid
In
also formelodic and chordal structures, i.e. in the "vertical dimension"?
golden section proportions, Lendvai uses the so called
numbers, "the simplest golden section sequence which can be
inwhole numbers (2-3-5-8-13)".
By thismeans he tries to demon

order to demonstrate
Fibonacci

expressed
strate thatBartok's melodic
equivalent

to 2-3-5-8-13

and chordal structures use simply those intervals

as semitones.

9 Bartok Bela Irdsai/1

[Bela Bartok'Writings

1]. Ed. Tibor Tallian

(Budapest,

1989), 217-220.

StudiaMusicologica Academiae ScientiarumHungaricae 47, 2006

This content downloaded from 130.64.196.119 on Mon, 3 Mar 2014 11:43:44 AM

All use subject to JSTOR Terms and Conditions

Axis Tonality and Golden Section Theory Reconsidered

423

gi_
ct_
Fl_
B_
E_
A
D_
G_
C_
F_
BL
_EL
Example 2
3. Let usfirst look at the question of thepentatonic
I quote from Lendvai's

scale!

Bartok's

dramaturgy:
That thegolden section is one of the innate laws ofmusic, and not an outer
stipulation, is effectively evidenced by themost ancient tone system of
mankind, thepentatonic scale. It can be considered thepurestmusical mani
festationof thegolden section.The five tone scale, and especially its"la" and
"re" modes are based upon a melodic outline consisting primarily ofmajor
second (2),minor third(3) and fourth(5) intervals.[Example 3]

His evidence
tone scale. But

Example 3
is limited to a certain, arbitrarily selected segment of the five
looking at the full octave, the mentioned pentatonic modes

contain the fifth(7) and theminor seventh(9), figuresnot fittinginto the

sequence. And it isworth mentioning that the Chinese pentatonic
scale, actually themost ancient tone system of mankind, uses the "do" and

Fibonacci

"so"modes with themajor third(4) andperfectfifth

(7) (Example4).

Example 4
StudiaMusicologica Academiae ScientiarumHungaricae

This content downloaded from 130.64.196.119 on Mon, 3 Mar 2014 11:43:44 AM

All use subject to JSTOR Terms and Conditions

47, 2006

424 Janos Kdrpdti

4. The typical Bartokian

chord structures interpreted byFibonacci

sequence

it iswell-known
that ancient tone sys
By theway, among ethnomusicologists
tems exhibit very complex relations. It is a fact demonstrated by John Ellis
more than a hundred years ago, that ancient tone systems in various regions in
theworld do not consist unconditionally of diatonic intervals, and their actual
scale structures can only be measured by Cents. How could calculating by
semitones be valid in these systems, and hence the Fibonacci
sequence?
But

let us look at how Bartok

examples
Bluebeard's
"Chinese

handles the pentatonic scale. I quote two

of the instance of the pure five tone scale. At the 5th door of Duke
Castle

march"

themelody is harmonized by mixture major triads. In the

of The Miraculous
Mandarin
the five tone melody
is

accompaniedby consecutivediminishedfifths{Example5).
Larghissimo

*-mpocoaliarg.

Example 5
the pentatonic scale cannot be interpreted as the golden
Accordingly,
section system either in its origin or as itoccurs.
or contradiction has been revealed by Malcolm
A similar miscalculation

Gillies
vai's

regarding the opening scale subject of the Cantata profana. In Lend

interpretation it is a pure golden section scale containing the charac

teristic Fibonacci

sequence:

2-3-5-8.

But in reality there is also a diminished

fifth(6) and aminor seventh(9). Is a statement

valid,when of thesix intervals
two are outside the system?

With thiswe have arrivedat our fourthsubject:The typicalBartokianchord

structures interpretated via the Fibonacci

sequence.

StudiaMusicologica Academiae ScientiarumHungaricae 47, 2006

This content downloaded from 130.64.196.119 on Mon, 3 Mar 2014 11:43:44 AM

All use subject to JSTOR Terms and Conditions

Axis Tonality and Golden Section Theory Reconsidered

425

to one of Lendvai's main ideas, there is a very typical Bartokian

According
chord structure containing within itboth themajor and minor character. The

existence and frequency of such a chord type can be accepted without any
doubt and from the phenomenological
aspect it is true. Its denomination as an
chord
is
used
frequently in the Bartok literature. My
accepted and
alpha
objection focuses exclusively on its interpretation as a golden section struc
ture- but this is not a secondary question. I quote from page 13 of Workshop:
As a consequence of thisduality [i.e. of classical harmony], in theovertone
system

every

note

relates

to a fundamental

note

(root), whereas

in the pen

tatonic system (?) every note bears a relation to every note. In the typemost
characteristicofBartok andKodaly's music - theso-called alpha harmonies thisprinciple has virtuallybecome a rule: thepentatonic 8,5,3,2 intervals(or
theiroctaves) intertwine
with everynote of thechord. [Example 6\

I
V

?-minor sixth (8)

- fourth

(5)

\/*
L-1

minorthird (3)
major second

(2)

Example 6
This really would correspond to the Fibonacci
sequence, but the relations
- are
intervals
selected arbitrarily. In the frequently used seg
demonstrated
ments of this so called alpha chord consisting of five notes (called beta, gam
- we
indeed have thementioned 2,3,5,8
ma, delta)
intervals, but with as much

we also have thediminishedfifth(6), diminishedseventh(9) and

frequency

(11), while the next member of the Fibonacci sequence, the

number 13, is completely absent (Example 7).
If there is a common characteristic of these chords, it is the presence of the
diminished

octave

minor third(3) and diminishedoctave (11), and theabsence of themajor third

(4) and theperfectfifth(7). Accordinglywe cannot speak of golden section
proportions in typical Bartokian chord structures. In spite of this, the interval
to
analysis of these chords shows a definite tendency towards opposition

perfect relations, i.e. to themajor triad represented by the intervals 4-7-12.

This is rather the "mistuning" phenomenon, when themajor third (4) is sub

StudiaMusicologica Academiae ScientiarumHungaricae 47, 2006

This content downloaded from 130.64.196.119 on Mon, 3 Mar 2014 11:43:44 AM

All use subject to JSTOR Terms and Conditions

426 Janos Karpati

beta

ganuua.

delta

Example 7

stituted
by theminor third(3), theperfectfifth(7) by thediminishedor/and
augmentedfifth(6 or/and8), and theoctave (12) by thediminishedoctave or

major seventh (11). I have already presented this explanation in an earlier pa

per entitled "Perfect and mistuned structures inBartok's music".10

I am notunder the illusionthatthisshortpaperwill convince immediately

theory. I only want to draw attention to the basic

in
his
ideas
contradictions
which are, sometimes accepted as self-evident.
Bartok analysis should find and I think already has found other directions,
some results of Erno Lendvai's
maintaining
analysis, this time the peripheral
the followers of Lendvai's

ones, e.g. the denominations of chords as alpha, beta, gamma, delta. I am sorry
to say, however, that the above mentioned contradictions I have presented are
not peripheral elements in an otherwise solid system, but cardinal points that
question

the validity of thewhole

10Studia
Musicologica

system.

36 (1995), 365-380.

StudiaMusicologica Academiae ScientiarumHungaricae 47, 2006

This content downloaded from 130.64.196.119 on Mon, 3 Mar 2014 11:43:44 AM

All use subject to JSTOR Terms and Conditions