Alternative Bar Structures in Bartók's Music

Author(s): János Kárpáti

Source: Studia Musicologica Academiae Scientiarum Hungaricae, T. 47, Fasc. 2 (Jun., 2006),
pp. 127-140
Published by: Akadémiai Kiadó
Stable URL: http://www.jstor.org/stable/25598247
Accessed: 06-04-2016 17:51 UTC

REFERENCES
Linked references are available on JSTOR for this article:
http://www.jstor.org/stable/25598247?seq=1&cid=pdf-reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.

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

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].

Akadémiai Kiadó is collaborating with JSTOR to digitize, preserve and extend access to Studia
Musicologica Academiae Scientiarum Hungaricae

This content downloaded from 200.130.19.173 on Wed, 06 Apr 2016 17:51:57 UTC
All use subject to http://about.jstor.org/terms
 Alternative Bar Stuctures
in Bartok's Music
Janos KArpAti
Budapest
E-mail: hi [email protected]

(Received: 1 February 2006; accepted: 25 March 2006)

Abstract: Already in some of Bartok's juvenile compositions, a definite attraction is

manifested for the irregular rhythmics. This is an early appearance of additive metrics,
opposed to the conventional divisive one. An extremely rich selection for irregular
meters was offered to Bartok by Rumanian folk song material, particularly in the "colin
das". Bartok used the term "Bulgarian rhythm", although knew well this was not an
exclusively Bulgarian peculiarity. The phenomenon shows clearly the existing form of
the additive-substractive metrics, because the quantitatively various bar structures
alternate by augmenting or diminishing by one quaver (as the common denominator), a
kind of "mistuning in the time". This thinking has been well documented in a series of
various examples, but the Sonata for two pianos and percussion (1937) provides an
extremely rich selection of rhythm combination, in which the inner overstructurating of
the bar, resp. the mistuning of the meter plays the principal role.

Keywords: Bela Bartok, divisive and additive metrics, Bulgarian rhythm

In the Bartok literature there are relatively few works dealing with his very
exciting rhythm, at least not in a profound way. Our excellent teacher, pro
fessor Lajos Bardos, from whom we learned very much about analytical
thinking, promised in his study Folk Rhythms in Bartok s Music more than he
realized, because he gave a dictionary-type list of the simpler patterns, but did
not go in depth into treating the irregular and asymmetrical elements.' How
ever, in this regard a lecture by Janos Breuer given in honour of Lajos Bardos
with the title Colinda Rhythm in the Music of Bartok can be considered a

1 Harminc irds. Budapest: Zcncmiikiado, 1969, pp. 9-40.

Studia Musicologica Academiae Scientiorum Hungaricae 47/2, 2006, pp. 127-140

DOI: 10.1556/SMus.47.2006.2.1
0039-3266/$ 20.00 > 2006 Akademiai Kiado, Budapest

This content downloaded from 200.130.19.173 on Wed, 06 Apr 2016 17:51:57 UTC
All use subject to http://about.jstor.org/terms
 128 Jdnos Kdrpdti

pioneering study which directed our attention to Rumanian origin of the

frequently and surprisingly metric changes of Bartok's music.2 Another
important contribution to our understanding of Bartokian rhythm was Laszlo
Somfai's study entitled Analytical Notes on the Piano Year 1926, particularly
chapter VI about" metre-breaking rhythmical practice".3
In one of my previous studies I attempted to analyze Bartok's melody,
chords, counterpoint and tonal relations via a dualism of perfect and mistuned
structures.4 There I raised the idea of "rhythmic mistuning". At first sight the
word mistuning seems inappropriate to characterize a process in time.
Nevertheless I think it could be appropriate with certain abstraction, since
mistuning - in the last analysis - is simply the +/- modification of a given per
fect structure by the smallest unit of the system. Accordingly, if the structure
moves in "musical space", the smallest unit is the semitone, and if it moves in
"musical time", the smallest unit is the denominator of the fraction marking the
metre (usually the quaver or semiquaver). The present paper aims to give a
more detailed analysis of this idea.
First of all it seems necessary to clarify our terminology. In everyday prac
tice there frequently occurs a mixed, sometimes inappropriate use of the
words rhythm and metre. Relying on lexical definition, I am recommending
some simpler but still appropriate formulations as follows.
The rhythm is the frequency and the proportion to each other in time of the
notes. The metre is simply the organized grouping of the rhythm by accen
tuation, prosody or other factors. Although Bartok mostly uses the expression
"Bulgarian rhythm", when he marks the bar structure at the beginning of such a
piece with fractions (e.g. 2+2+3/8), he in fact defines the grouping of the notes,
that is the metre. Our suspicion regarding the incorrectness of Bartok's wording
is confirmed by Timothy Rice, the eminent expert on Bulgarian folk music,
who, in his lecture given at the Bartok symposium held in Los Angeles in 1995,
proposed putting "Bulgarian metre" in place of "Bulgarian rhythm".5
Another question of terminology concerns the inner structure of the metres
(bars) and how they are connected. In European art music - let us say from the
mensural thinking of the Ars nova onwards - it was the division of identical time
units which dominated. This divisional principle and thinking still dominated
even when unusual kinds of bar cropped up, e.g. 5/4 in Chopin and Tchaikow

2 Studia Musicologica 17 (1975), pp. 39-58.

3 Tizeimyolc Bartok-tanulmdny. Budapest: Zencmiikiado, 1981, pp. 171 -176.
4 ^'Perfect and mistimed structures in Bartok's music*', Studia Musicologica 36 (1995), pp. 365-380.
5 "Bcla Bartok and Bulgarian Rhythm", in Bartok Perspectives. London New York, Oxford Univ.
Press, 2000, pp. 196-212. This distinction by Rice is based on the study Mieczyslaw Kolinsky: "A Cross
cultural Approach to Mctrorhythmic Patterns", Ethnomusicology 17(1973), pp. 494-506.

Studio Musicologica Academiae Scientiarum Hungaricae 47, 2006

This content downloaded from 200.130.19.173 on Wed, 06 Apr 2016 17:51:57 UTC
All use subject to http://about.jstor.org/terms
 Alternative Bar Stuctures in Bartok's Music 129

sky.6 In these cases the 5/4 cannot be divided, as the triplet, quintuplet or
septuplet falls on a single time unit. We find such a typical divisional case in
Bartok too, e.g. in the 1st movement of the 2nd Sonata for violin and piano
(Example 1).

W *** ?s= -=
Puco piu andante, J* - 108

Example I
Accordingly divisional metre means that there are identical time units in
use, independently of whether their inner division is duple or not. And it has to
be added that divisional metre can be interpreted in principle inversely, i.e. in
multiplicative form; this means that for example 6/8,9/8 or 12/8 are simply the
double, triple or quadruple product of 3/8.
A principle opposed to divisional metre is another musical practice - mainly
in traditional and modern music, additive metre. This is already present when
the inner structure of a 5/4 or 5/8 metre is divided unequivocally into 2+3 or
3+2 units. This is the basic case of additive metre, already known in the ancient
Greek theory as hemiolia, i.e. "one and a half (in the Middle Ages sesquialte
ra), since to the basic unit half of it is added. It was practiced already in
divisional metre as hemiola, meaning that a triple divided metre (e.g. 3/4)
changes to double (6/8). In this case the metric pulse does not change (3x2/8 =
2x3/8). The change of the inner structure of the bar refreshes the metre and the
musical process becomes more lively and interesting. It is a frequent phenome
non in the Italian, Spanish and Arabic folk music of the Mediterranean Region,
as I found in my earlier investigations, and it occurs also in Italian art music;
e.g. Monteverdi uses the anacreontic metre in this hemiolic way in scenes in
Orfeo, and Incoronazione di Poppea, alternating 6/8 and 3/4: "Vi ricord[a] o
bosch[i] ombrosi" or "Questa vit[a] e dolce troppo".7 Much harder in effect is
the hemiola in the Scherzo of Beethoven's Symphony III, when the fast 3/4
metre changes only for four bars to Alia breve 4/4. This is a real additive pro

6 Cf. Bartok's lecture "Az ugyncvezctt bolgar ritmus" held in Budapest April 6, 1938. English version:
"The So-called Bulgarian Rhythm", in Beta Bartok Essavs. Selected and edited by Benjamin Suchoff.
London: Fabcr & Fabcr, 1976, pp. 40 49.
7 "Melodic, vers et structure strophique dans la musique berberc (imazighen) du Maroc Central", Studia
Musicologica 1 (1961), pp. 451 -473.

Studia Musicologica Academiae Scientiarum Hungaricae 47, 2006

This content downloaded from 200.130.19.173 on Wed, 06 Apr 2016 17:51:57 UTC
All use subject to http://about.jstor.org/terms
 130 Janos Kdrpdti

cedure, in that different times (metres) are placed next to each other. We should
remark here too that the addition can be interpreted as its inversion, i.e.
subtraction, (e.g. when 2 follows 3, it means a unit of 3 was taken off.)
In Bartok's two juvenile compositions, in the Scherzo or Scherzando of the
String quartet F major (1898) and Piano quintet (1903-04) a definite attraction
is manifested for the irregular rhythm. In both movements, marked 3/4, the
rhythm of the theme consists of three 2/4 (dotted crotchet + quaver) and one 3/4,
i.e. in three bars the actual structure is 2+2+2+3/4, and all this in the main
thematic material, dominating the entire movement. In this period Bartok did
not arrive at notating this metric as it really is, therefore he chose pure 3/4 with
hemiola insertions. But we can reveal its inner connection with the 5th
Bulgarian dance of Mikrokosmos.
But we must not make such a big jump! In his creative period between 1908
and 1911, Bartok handles the bar structures in certain pieces for piano with
surprising freedom. It is true, however, that the frequent changes - 3/4-1/8
5/8-7/8-3/4-4/4 - in the second of the Two elegies (Op. 8b) are very well ex
plained by the direction for the interpreter Molto adagio, sempre rubato (quasi
improvisando), and refers to the fact that the slow tempo and rubato character
itself involves additive metre.8 Furthermore, along with the rubato style goes a
complete absence of barlines, which if present, are only helping to orientate the
performer. We find a similar phenomenon in piece XII of the Fourteen Ba
gatelles (op. 8), but in contrast, in piece XIV ("Ma mie qui danse") the 3/8 Valse
metre is many times broken by hemiolas, i.e. by 2/8 units.
We find real additive metre in piece VII of Esquisses (op. 9b, 1908-10),
where 9/8-6/8-6+2/8-5/8-7/8 marked bars alternate with the tempo Poco
lento, testifying that the melody initially presented in 9/8 becomes shorter in
the process by one/two/three quavers. The 4/4-2/4-5/4 changes of bar in the
second Burlesque are part of Bartok's portrayal of the character's staggering
("a bit tipsy").
The above mentioned pieces show us nothing about the change Bartok's
creative thinking underwent in precisely these years! At this period he began
systematic folk song collecting first in Bekes and Tolna counties (1906), he
discovered Csik in Transylvania (1907), then as a result he extended his
collecting area beyond the Hungarian population to Slovaks in the Uplands
(Nyitra, Hont, Gomor) and Rumanians in Transylvania (Bihar, Hunyad,
Maros-Torda, 1908-1912). At this time folk music was not so organically in
corporated into Bartok's musical language as later on, but his method of folk
music arrangement was shaping in the two Hungarian and two Slovak books
8 It refers to the original meaning of "rubato" as "robbed" rhythmic values.

Studio Musicologica Academiae Scientiarum Hungaricae 47, 2006

This content downloaded from 200.130.19.173 on Wed, 06 Apr 2016 17:51:57 UTC
All use subject to http://about.jstor.org/terms
 Alternative Bar Stuctures in Bartok's Music 131

of For Children. The greater part of the Hungarian and Slovak material did
not provide stimulating models for Bartok and irregular rhythms, he had more
surprising encounters with folk songs from the Sekler country - primarily
Csik county in Transylvania - collected by him personally and by some of his
colleagues. He himself in Kilyenfalva found a song in giusto tempo, in July
1907, whose notation in 8/8 conceals a 2+3+3/8 structure {Example 2c). In the
same region also Kodaly collected songs with very similar rhythmic structure
in 1910 {Examples 2a, 6), and Antal Molnar discovered in 1911 the frequently
cited song Az en lovam Szajko {Example 2d) with 5+3/8 metre.9
Tempo giusto J

I El - ><i-<?uL is - nv va - la Sir- pa - la - ka fe - 16,

Poco rubato, ? ? cca &J-88_

1 Lti-da-im. Iii - da-i-rw Ti-zen-ket-ten vat-tok,

2. Szabo nem jol xzab-ta, Var- r6 scm j6l varr-w,

a) i. Az ?n )o- vain Sx*j- ko. Magam pe- dig Ian- kd.

Examples 2a-d

Bartok knew that this kind of metre is very rare in Hungarian folk song
material, and just for this reason was very happy to report his discovery to his
wife, Marta Ziegler in a letter addressed to her at Veszto in Autumn 1918:
For the moment being - pour degourdir mes membres musicaux -1 have har
monized 7 songs from the ones collected in the summer, including the famous
Borbala Angoli of Roza Okros. 1 recommend you listen to it, because to hear
something like this in Hungarian, what is more in the middle of the great Plain
and furthermore in 7/8 time, is truly sensational.,0

So was born the central piece (no. 6, Ballade) of Fifteen Hungarian peasant
songs for piano, in which Bartok substitutes the strophice chain of Borbala
Angoli s sad story by a chain of variations; and in the middle the irregular metre
is "smoothed out" to 3/2 and 6/8.
But an extremely rich selection of irregular metres was provided to Bartok
by the Rumanian folk song material, particularly in the "colindas". His first
Rumanian collection was published as early as 1913,] l and it did not depend on

9 Cf. Hungarian Folk Songs: Complete Collection Vol. I, compiled by Bela Bartok. Edited by Sandor
Kovacs and Ferenc Sebo. Budapest: Akadcmiai Kiado, 1993. Nos. AI 198 and 199.
10 Letter to Marta Zicglcr in Bartok Bela csalddi levelei. Ed. Bela Bartok jr. Budapest: Zcncmiikiado,
1981, no. 387.
1' Chansonspopulaires roumaines du departement Bihar (Hongrie). Bucuresti, 1913.

Studia Musicologica Academiae Scientiarum Hungaricae 47, 2006

This content downloaded from 200.130.19.173 on Wed, 06 Apr 2016 17:51:57 UTC
All use subject to http://about.jstor.org/terms
 132 Janos Kdrpdti

him that his great colinda collection - although the manuscript was complete
already in 1926 -reached publication only in 1935.12 In the introduction of this
volume, which contains 454 melodies, the peculiar metre of the colindas is
characterized by these words:
Special mention is required for those rhythmic patterns which show a division
into sixteenth notes, and which are called "Bulgarian" because they often
appear in Bulgarian folk music of every kind... It seems desirable to
investigate the question: In these melodies, are we not dealing with merely
accidental shortenings or lengthenings of note values on the part of the
perfonner? Since this kind of rhythm is sung throughout the entire region of
Hunedioara and, occasionally, also in Mures.-Turda and Bihor, we must say
'no' to the above question.,3
Examining the colindas - although Bartok does not touch upon this - two
types are to be observed. One can be characterized by a rather irregular
alternation of the bar structures, e.g. 2/4-3/8-2/4-3/4 etc. This is the type cited
by Janos Breuer, and referred to in movement III of the Sonata for piano
(1926).14 The Hungarian folksong that lies behind the composition is initially
presented in a 3/8 "mistuned" version, and only bar 227 introduces the
"original" 2/4 form. It is very characteristic of Bartok that in his first draft he
did not use any metric marking15 (Example 3).
Allegro motto, J " i?0

(1?==^ . ^^^Iv --j^B-p" ~ "-ft- * jffM

JT^^j^Ffguji P ' j It uXa |U

Example 3
This phenomenon shows clearly the existing form of the additive
substractive metre, because the quantitatively varied bar structures alternate
by augmenting or diminishing by one quaver (as the common denominator).
This was emphasized also by Bartok in an article written on Rumanian folk
music for a Swiss newspaper:

12 Melodien der rumdnischen Colinde (Weihnachts lieder). Wien: Universal, 1935.

13 Rumanian Folk Music by Beta Bartok, Volume Four: Carols and Christmas Songs (Colinde). Ed. by
Benjamin Suchoff. The Hague: Martinus Nijhoff, 1975, p. 31.
14 Sec note 2.
15 Bartok's Workshop: Documents of the Compositional Process. Exhibition and text by Laszlo Somfai.
Budapest: MTA Zenctudomanyi Intezet, 1995, p. 7.

Studia Musicologica Academiae Scientiorum Hungaricae 47. 2006

This content downloaded from 200.130.19.173 on Wed, 06 Apr 2016 17:51:57 UTC
All use subject to http://about.jstor.org/terms
 A Iternative Bar Stuctures in Bartok's Music 13 3

Noteworthy in this particular instance is the continual change of the measure.

We must bear in mind that these and similar "complex" melodies are per
formed by illiterate people with the greatest aplomb and in the most natural
fashion. This is the most convincing proof of the extent to which certain
theoreticians en* when they hold that repeated change of time denotes an un
natural process.16
Similarly to the booklets For Children, two series of simple arrangements
for piano were composed by Bartok from the colindas in 1915, before the great
scholarly collection was published.,7 We quote the beginning of no. 4 from the
first series, identifiable with No. 10b (D-oi roaga, sa roaga) of the collection.
This is a good example of the first procedure, i.e. irregular alternation of bar
structures {Example 4).

4 Andante tJN? 105) _ 3$?SpN ^_

Example 4
In the other basic type the prescribed metric structure remains the same
during the entire song, but this structure shows an inner asymmetric arti
culation, e.g. 2+3+3/8. As this articulation falls upon identical time units, this
belongs to the case of divisional metre, it should be added, however, this is a
particular case of divisional metre, because the basic bar structure - according
to the asymmetric division - has been created by the additive principle. In
other words: this type of metric structure is a special combination of divisional
and additive thinking. The basis of the piano colinda no. 1:7 is no. 27 of the
scholarly collection (La lina fantana), but there are a further seventeen songs
listed by Bartok in the introduction to characterize the type {Example 5).
This 2+3+3/8 division in principle stands close to the metric division of the
6th Bulgarian dance of Mikrokosmos (3+3+2/8), since the sum value of the
quavers is 8, accordingly it "could be" 4/4, but in the theoretical part of the
introduction Bartok speaks about an augmentation by semiquavers. He lists a
number of examples, but we quote only two. No. 60 with 10/16 prescription

16 "Rumanische Volksmusik", Schweizerische Sdngerzeitung September 1, 1933. "Rumanian Folk

Music", in Beta Bartok Essays. Selected and edited by Benjamin SuchotT. London: Fabcr & Fabcr, 1976,
pp. 119-127.
17 There arc two excellent studies dealing with this topic. David Ycomans: "Background and Analysis of
Bartok's Rumanian Christmas Carols for Piano (1915)", in Bartok Perspectives. London New York:
Oxford University Press, 2000, pp. 185-195; Ferenc Laszlo: "Bartok es a roman kolindak", Magyar Zene
XLIII (2005), pp. 259-272.

Studia Musicologica Academiae Scientiarum Hungaricae 47, 2006

This content downloaded from 200.130.19.173 on Wed, 06 Apr 2016 17:51:57 UTC
All use subject to http://about.jstor.org/terms
 134 Janos Kdrpdti

Andante (J ?106)

Example 5
and consists of quavers and pointed quavers (2+3+3+2/16). This metre
structure would be perceived by an unpracticed listener as simply 4/8 or 6/8.
The metric prescription of no. 11 is given by Bartok as 15/16, and with dotted
lines denotes the inner structure of the bars: 5+5+5/16. It is comprehensible
that, characterizing the "Bulgarian rhythm" discovered by him, he speaks
about an augmentation of time values by semiquavers.
As we mentioned earlier, in 1938 Bartok gave a lecture about the So-called
Bulgarian Rhythm. From this we learn that at the time when he wrote his study
about the colindas - between 1924 and 1926 - he already knew the study by
Dobri Christov entitled Bases rythmiques de la musique populaire bulgare,
published in the 37th volume of the ethnographic series of the Bulgarian
Academy of Sciences, demonstrating melodies with 5/16,7/16, triple divided
8/16.18 But Bartok had become acquainted with this phenomenon even earlier,
during his field work collecting colindas between 1912-1918, and in a further
part of the lecture he gives a more detailed account of it:
When I first saw these unfamiliar rhythms, in which such fine differences are
decisive, I could hardly imagine that they really existed. But then I seemed to
remember that in my own collection of Rumanian material I had come across
similar phenomena, but at that time had not dared - if I might put it that way -
to take note of them. Among my old phonograph notations there were dance
melodies which, with a perfectly clear conscience, I had noted down in 4/4, in
steady quarter-notes (or perhaps not with a perfectly clear conscience, because
I had written on my notations: "the ends of the bars drawn out in gipsy fash
ion"). ... Since then 1 have thoroughly revised my phonograph notations, and
it turns out that about five percent of the Rumanian material is also in so-called
Bulgarian rhythm... In the Hungarian material this kind of rhythm can only be
found in traces, and at present nothing can be inferred from them. ... There is
no trace of the rhythm in Serbo-Croate publications, but this again does not
mean it is non-existent there. I believe that it must be present there, too, but
that nobody among Serbo-Croate collectors has noticed it. An important point
is that the so-called Bulgarian rhythm obviously exists in part of the folk
music of the Turkish peoples. In one of his publications, Uspensky (and Be
laiev) gives Turkestan (Turkmenian) folk music, in which there are a con

18 Sec note no. 6, page 43.

Studia Musicologica Academiae Scientiarum Hungaricae 47. 2006

This content downloaded from 200.130.19.173 on Wed, 06 Apr 2016 17:51:57 UTC
All use subject to http://about.jstor.org/terms
 Alternative Bar Stuctures in Bartok's Music 135

siderable number of melodies notated in 5/8, with eighths of about M. M. =

300. I, too, encountered rhythms of this kind in Asia Minor, although
admittedly only in dance or instrumental music which was probably not of
ancient Turkish but of foreign origin...

And here Bartok quotes a music example as follows19 (Example 6):

?**,& -*-.i-.i? *.j-J.r-.J n J----*-?|J f J y J-*-J?*J.v.J.*~

M ** -+?> iff?* y'f?*?*?p**?+XT+{ * ? yi-* CJ*'f **TT**""
Example 6
Accordingly, Bartok, although he used the term "Bulgarian", well knew this
was not an exlusively Bulgarian phenomenon, and in this he preceded by
thirteen years Constantin Brailoiu who, in the December 1951 issue of the
Revue de musicologie, recommended to use instead of "Bulgarian" the Turkish
adjective "aksak" (lame). The eminent Rumanian ethnomusicologist - by the
way a one-time esteemed colleague and good friend of Bartok - wanted to
express - without referring to any nation - the asymmetry of this metre. By this
Bartok's discovery, that this metric phenomenon was spread over the whole of
the Balkans and south-west Asia, obtained a subsequent verification.20
Besides the Rumanian folk material, Bartok obtained special rhythmic
impulses from North-African Arab folk music too. He travelled in 1913 to Al
geria, and in the company of his first wife he recorded vocal and instrumental
pieces with his Edison-type phonograph. In his study, first in Hungarian
(1917), then in German (1920), Bartok writes about Arab folk music:
The high-level development of the rhythm is in contradistinction with the
primitivism of the range and form. The syncopations met in melodies with
strict rhythm must be mentioned as the chief characteristic differing from
Eastern European folk music. These syncopations not only occur in the fonn
customary with us, but in a manner utterly unfamiliar, where a short un
accented value is tied to a long accented one... The intricacy of the melodic
rhythm is further complicated by the accompanying rhythm of the percussion
instruments... Occasionally the accompaniment is independent to the extent
that it nearly takes on a self-contained life alongside the melody. In this case -
in point of fact met exceptionally - any connection of the melody and its
accompaniment is lacking... And, finally, curious polyrhythms come into

19 Sec note no. 6, page 47.

20 It is surprising that Timothy Rice, who published an excellent study about the topic Bartok and the
Bulgarian rhythm (see note no. 5), is ignorant of Brailoiu's communication and the internationally accepted
term "aksak".

Studia Musicologica Academiae Scientiorum Hungaricae 47, 2006

This content downloaded from 200.130.19.173 on Wed, 06 Apr 2016 17:51:57 UTC
All use subject to http://about.jstor.org/terms
 136 Jdnos Kdrpdti

being when melodic and accompanying motives, although of identical length

but fully divergent as far as rhythm, are united.21

From the typical rhythm patterns of the Biskra collection, two play an
important role in Bartok: the first is a classical hemiola, and the second a
3+3+2/8 divided "Bulgarian rhythm" {Example 7).
Bitkro No. S8

1-p p ^r^"rp PT"-f?p

6mk?w No $9. 60

jj~. ^ * J> ^ j ^ {.n .i>.A?it?.

Example 7
This second type is many times incorporated in Bartok into high level art
music forms. Laszlo Somfai drew attention to the Coda of the Piano sonata's
first movement.22 In fact, the otherwise 2/4 metric percussive effects become
3/8, and even this is broken by taking off one quaver from every third unit (2/4
- 3/8 - 3+3+2/8). A similar effect is manifested in the Finale movement of the
4th String quartet. Here the entry of the theme is prepared by a monotonous
drumming of the viola, but the quavers are accentuated in a 3 + 3 + 2 propor
tion by sforzatos and broken chords of the violoncello, creating a special
polymetre between the theme and the accompaniment {Example 8).

Example 8

A very similar occurrence is found in the version for two pianos of "Osti
nato", piece no. 146 of Mikrokosmos. In this piece the second piano has the
accents in order to produce a 3+3+2/8 pulsation. A particular metric structure
characterizes the descending scale theme of the final movement of Music for
strings, percussion and celesta. The prescribed 2/2 does not express the real
inner structure of the theme, because the quavers are in a 5+3 asymmetric

2' "A Biskra-vidcki arabok ncpzcneje". Szimfonia I, September 1917, pp. 308-323. New revised edition:
Bartok Beta irdsai 3: Irdsok a nepzenerol es a nepzenekutatdsrdl. Ed. by Vera Lampert. Budapest: Editio
Musica, 1999, pp. 87 128. In English: "Arab Folk Music from the Biskra District" (translated by Benjamin
Suchoft), in Beta Bartok Studies in Ethnomusicology. Lincoln-London, 1997, pp. 29-76.
22 Sec note no. 3.

Studia Musicologica Academiae Scientiarum Hungaricae 47, 2006

This content downloaded from 200.130.19.173 on Wed, 06 Apr 2016 17:51:57 UTC
All use subject to http://about.jstor.org/terms
 Alternative Bar Stuctures in Bartok's Music 137

articulation, counterpointed by the accompaniment's chords in a 3+3+2

articulation {Example 9a). In the major part of the performances this very fine
rhythmic effect does not come out, and the bars are accentuated simply by
halving. That this is not correct is verified by the development of the theme,
because from bar 74 Bartok himself changes the accentuation {Example 9b).

V^p > ,r 'r *t*.:H

' 3 ' ' 3 ' *~1 ' 3 ' 3 ' ' _"""'

,-, 4_4 4_4__ __

* ^HZj;jTr^|j~rjTT",l
h r ' r l? r * -H
Examples 9a-b

The above quoted examples showed that additive metre and combined
additive-divisional metre are not connected only with folk song arrangements.
The fact of how sensitive a point this was in Bartok's creative thinking is ex
tremely well verified by Etude HI (op. 18), composed at the same time as
Angoli Borbdla. Although it begins with rubato arpeggios free from a metric
point of view, in the Tempo giusto section 6/8-7/8-6/8-10/16-9/16-11/16
bars succeed each other, providing an absolute extreme case of additive metre.
The rhythmic mistuning technique can here well be observed, since the qua
vers and semiquavers occur in one-by-one augmenting groups {Example 10).

Example 10
The Etudes belong among Bartok's most difficult compositions - equally
difficult to play and to understand - with wide intervallic leaps, tonally tensed
chord cumulations and dissonantly mistuned structures. It is easily un
derstandable that the rhythm and metre is similarly handled. The two Sonatas
for violin and piano belong to the same stylistic area. An example of pure
divisional polymetric has been quoted above (see Example 1). Here a very
characteristic additive structure is quoted from the 2nd Sonata {Example 11).
In some cases additive thinking is manifested in a concealed form. In
movement IV of the 4th String quartet we find a place where inside the un

Studia Musicologica Academiae Scientiarum Hungaricae 47, 2006

This content downloaded from 200.130.19.173 on Wed, 06 Apr 2016 17:51:57 UTC
All use subject to http://about.jstor.org/terms
 138 Janos Kdrpdti

jT) PKHrrtt. >*I33 krrrr ?il Sol.f"r". -

Piii vivo, J*= 132 breve _--???-rrr;-'? ???^. hrtve

** , *r^^?.-iyL?^^"*
Example 11
changed 3/4 metre accent signs < denote gradually augmented blocks (5/4
6/4-7/4-8/4) in the "guitar accompaniment". The polymetre becomes then
more tense when the accents are nudged along by one quaver (Example 12).

-.j i.?!._?___._I >

8/4

Example 12

In vocal works (operas, etc.) it goes without saying that additive metre plays a
prominent role, since the rhythm of voices depends on the prosody of the text,
while in the case of pantomime (dance play, etc.) gestures dominate. This is why
the number of examples would overgrow this study. Nevertheless, I shall quote
a passage from Cantata prof ana. From bar 43 of the second part, the rhythm of
the tenor solo has an alternation as follows: 5/4-2/4^4/4-3/4-3/2- 4/4-3/2. In
this part a definite creative attitude is manifested. Namely, if the colinda rhythm
had not left such a deep impression on Bartok's thinking, the rhythm of the tenor
solo could have been fitted into the frame work of the opening 3/2 metre. But we
know that the text of Cantata profana was compiled by Bartok from colindas,
and though the music itself is completely his own, the composer could not - and
did not want to - detach himself from the free, additive rhythm of the colindas
(Example 13).
Now we have reached the last - and summing up - idea of this argument.
Namely that in the majority of cases, additive metre is not simply a changing
metre, but a definite intervention in the process of the metre when either the
inner structure of the bars changes, or the bar gets longer (addition) or shorter
(subtraction) by a small unit (quaver, semiquaver). The fact that Bartok's
creative thinking was directed by this mechanism - called mistuning - is
Studia Musicologica Academiae Scientiarum Hungaricae 47, 2006

This content downloaded from 200.130.19.173 on Wed, 06 Apr 2016 17:51:57 UTC
All use subject to http://about.jstor.org/terms
 Alternative Bar Stuctures in Bartok's Music 139

42 Agitato, poco rubato J ? 150 ^^

M'
Kc*d - ves 6 - tics a - p&tk_ Rank_ tc_ so - sc

<r - Itiz?.' Mm te - ged _ mi __ t6 ziink A_ hzm - - - vunk he -

j>2 trrsc. .. seinpre piu mosso.

i*ye u> Es- ugy_ ha - ji ~ ga - - luak T? - - ged rei - r6l ret rc:

Example 13
proved by his lectures and writings. In his already quoted lecture The So
called Bulgarian Rhythm he comments on the 7/8 metre of a Hungarian song
[Angoli Borbala] and a Rumanian colinda [D-oi roaga] with the words:
In both melodies we feel that the 7/8 metre originated in a 3 x2 metre in eighth
notes: in he first example the fourth eighth-note has expanded into a quarter
note; on the second example, the sixth one.

And Bartok similarly interprets the characteristic 2+2+3/16 rhythm of the

Bulgarian Ruchenitza.
The superficial listener will evaluate this rhythm either as 3/8 or as 2/4. If we
derive it from 3/8, then we must establish an augmentation by a sixteenth
note; if from 2/4, then a diminution by the same note value. I favour the former
explanation, and as the Bulgarians' own rhythmic definition suggests, they too
seem to be of this opinion. My feeling is that this extension of the note value is
no other than the translation of a dynamic stress into terms of duration.23

This thinking has been well documented in our examples so far quoted, but
the Sonata for two pianos and percussion (1937) provides an extremely rich
selection of rhythm combination, in which the internal through-structuring of
the bar, and the mistuning of the metre plays the principal role {Example 14).
The slow introduction has an unequivocally 9/8 metre {a). At the end of the
introduction, the listener witnesses an inner transformation of the 9/8 metre
(ft), and when the Allegro molto starts, maintaining the numerator 9, the inner

b) ?* r r.er."tu.l
. t-,.J.J?L-J3-,
' -t?r?r?r?^
i j i
^ -r^~r * p * ji ? '
Example 14
23 See note no. 6, page 47.

Studia Musicologica Academiae Scientiarum Hungaricae 47, 2006

This content downloaded from 200.130.19.173 on Wed, 06 Apr 2016 17:51:57 UTC
All use subject to http://about.jstor.org/terms
 140 Janos Kdrpdti

structure becomes 1+2+2+2+2/8 which, if we consider the real rhythmic

pattern of the theme, as a matter of fact is simply 1/8+4/4. This means that the
theme in a simple metre becomes special and exciting by means of a quaver
retardation.24 But the excitement is also increased by the metre of the timpani,
which plays the inversion of the pianos' rhythm: 4/4+1/8 (c). The second
theme of the movement shows another Bulgar-like metre structure, 4+2+3/8
(_/), which occurs - among other places - in the Scherzo of the 5th String quar
tet and the 1st Bulgarian dance ofMikrokosmos. The sum of the numerator
remains 9 here too, but behind this real structure a 4/4 pattern is concealed, in
which one crotchet is stretched by a quaver - not always on the same place.
When Bartok gave a lecture series at Harvard University in 1943, he had the
opportunity to look back over his whole carrier and to give a very clear
explanation of his compositional methods and ideas. Because of his wors
ening health he could only give three of the projected eight lectures, and the
fourth one - exactly about "rhythm" - only remained in an English draft.
... We may have - though not frequently - 5/8 time or 7/8 time in our melody.
The difference between them and the regular 2/4 is not essential; it is rather a
derivative difference. In fact, 5/8 can be explained as a doubling of one of the
eighths in a 2/4 bar, and 7/8 as a doubling of one of the eighths in a 3/4 bar.
These strange bars attracted me to a high degree, and their influence can be
discovered in many places in my original works [italics mine]. As for the
strangeness of these bars, it is as nothing, however, in comparison with the
rhythm patterns I am going to describe to you in the following. They belong to
the group which we call "Bulgarian' rhythm" formations.25
[- Here the draft breaks off.]

These last lectures of Bartok's life became publicly known two decades
after his death.26 In these some of his earlier ideas presented or only touched
were repeated, enriched then by further, very explanations. At the same time,
some questions concerning his creative method were raised for the first time,
and answered in a decisive way. One answer was e.g. the "polymodal chroma
ticism", contradicting the opinions of some critics, and verifying or modi
fying, those of others. Similarly, his above quoted statement about rhythm can
be considered decisive, because, although the words about the partial stretch
ing of the bar structures are repetitions, his declaration of his personal attrac
tion verifies the theory of "mistiming".

24 A similar procedure is given 25 years earlier in the Allegro barbaro, in which the theme is delayed by a
crotchet over the pulsing accompaniment.
25 ''Harvard Lectures", in Bela Bartok Essays. Selected and edited by Benjamin Suchoff. London: Faber
&Fabcr, 1976, pp. 393 398.
26 John Vinton: "Bartok on his own Music", Journal of the American Musicological Society 19 (1966),
pp.232-243.

Studio Musicologica Academiae Scientiorum Hungaricae 47, 2006

This content downloaded from 200.130.19.173 on Wed, 06 Apr 2016 17:51:57 UTC
All use subject to http://about.jstor.org/terms