Abstract
The present paper is a preliminary study to a deeper understanding of Bartók's compositional process: how he filled the music paper even if the notation appears continuous. He did not always write the draft from the top-left corner to the bottom-right corner but occasionally skipped some measures or phrases that he intended to write down later. Thus, the continuity draft is not necessarily regarded as a continuity; it might have consisted of fragments but they have been eventually concealed in the final form of the draft. This may directly affect our interpretation of Bartók's creative process. The present research takes the visual appearance of a draft as a clue to examine its hypothetical fragmentariness. Beginning with the examination of some actual blank spaces that were never filled, the paper deals with cases where the original existence of fragments is suggested by the extraordinary appearance of the draft, such as a blank space at the end of a system, notations in the margin, re-organization of bar-lines, etc. Besides systematic approaches to fragmentariness, the author offers some interpretation concerning the reasons why Bartók did not write the draft in an ordinary way. There might have been different explanations for the interruption of the linear compositional process. One of the important findings of this paper is that Bartók's bar-lines do not always mark metric accents but sometimes simply facilitate the performers' orientation. Even though the occasional special function of some of his bar-lines as mere markers of orientation is explained by Bartók in a letter to Rudolf Kolisch, the existence of philological evidence to this effect may further underline its validity and importance.
1 Introduction
The micro-chronological examination of Bartók's autographs is not entirely a new theme in Bartók scholarship. A study of the compositional process may necessarily involve the examination of the actual written order not only within a group of autographs but also within a single page. 1 So far, however, the precise reconstruction of the exact order of writing has only played a subordinate role in comparison with the interpretation of the creative process. This is indeed a natural tendency, as the former rather belongs to a purely technical aspect of the writing activity that does not seem to deserve scholarly attention. However, identifying irregular order in the actual writing may help us to recognize and then examine possible compositional problems that Bartók had to face and that might remain unnoticed otherwise. It seems evident that a composer may skip a problematic section and later supplement it; however, it is impossible to discuss it if a scholar does not notice it.
The present paper has evolved primarily from my personal experience gained from the meticulous study of Bartók's autographs, mainly sketches and drafts from which I prepared many transcriptions for the string quartet volume of the Béla Bartók Complete Critical Edition. 2 This is the primary reason why the scope of the paper is limited to Bartók's string quartets. Nevertheless, the phenomenon discussed in the present paper may also be applied to other works by Bartók, and may help future scholars to better interpret some irregular notational elements in his autographs. 3
2 Classification of the phenomena of fragmentariness
In the present paper, the phenomena of fragmentariness are grouped into three categories:
- (A) existing blank systems in the autograph (if there is some blank space longer than a system that interrupts the notation in the autograph)
- (B) unusual blank space in the autograph (if the notation is apparently continuous but there is an unusual, primarily short space that disturbs the flow of notation)
- (C) concealed fragmentariness in the autograph (the notation is continuous, but the context suggests that the given section was written in an irregular order)
(Each category contains some cases. In the following, the cases are referred to by the category's label, A-1, A-2, B-1, and so on. The order of this classification coincides with the degree of clarity.)
It should be noted, however, that the category merely deals with the appearance of the draft and is not necessarily relevant to the musical importance. Some cases (e.g., C-2 and C-3) are challenging to discover, but the most straightforward cases (e.g., A-1 and A-2) offer some insights into the composer's workshop that are no less important than the more complex cases. The discussion of each case is, whenever it seems possible, accompanied by a short analysis and interpretation concerning the cause of the fragmentariness. Some cases deserve a further discussion that also considers the overall form of the movement and the compositional strategy (especially A-2 and C-4). Due to lack of space, the present paper only deals with the most essential aspects. 4 It should also be noted that the cases below are selected not as a matter of curiosity but based on my judgement of their importance from the viewpoint of the musical and compositional technique.
3 Category a: existing blank systems in the autograph
A-1: Quotation from Pelléas and Mélisande (String Quartet no. 2, mov. III)
One of the most prominent examples concerning blank space is the last movement of the Second String Quartet (BB 75, 1914–1917). The original autograph sketch of this movement, preserved at the Budapest Bartók Archives, contains a short blank space on the second page. 5 On this sketch page, basically written in ink, it is discernible that Bartók stopped composing near the end of system 4, then continued writing in pencil (see Facsimile 1). Concerning the rest of system 4 and the entire system 5, however, Bartók drew bar-lines to mark the space of seven measures, yet did not write anything but the first note in the viola and cello parts (E/A♯1). 6 This dyad may indicate the first note of the omitted section and thus serves as a kind of reminder. 7 Then he moved to the beginning of system 6; however, interestingly, he sketched the continuation of the (possibly) imagined previous measures that had not been fixed on the paper yet.
Autograph of String Quartet no. 2, mov. III (BBA, BAN 494a–b, fol. 1v, systems 4–6)
Citation: Studia Musicologica 62, 3-4; 10.1556/6.2021.00017
Ferenc Bónis discussed Bartók's possible quotation from Debussy's Pelléas and Mélisande. 8 According to him, the quotation begins at measure 88. This is the very first measure of the missing section; thus, the Pelléas passage is missing from the sketch. This interesting coincidence is only recently discussed in the string quartet volume of the BBCCE. 9 The editors argue that the absence of this passage in the sketch underlines Bónis's finding. Even though it remains an open question to what extent this passage can be regarded as a quotation, it is still plausible that leaving a blank space is related to the fact that here Bartók referred to Pelléas. 10
A-2: Extensive blank space (String Quartet no. 4, mov. V)
A much larger blank space can be found near the end of the draft of the fifth movement of the Fourth String Quartet (BB 95, 1928). 11 The blank space can be found at the bottom of page 27 and at the top of page 28 (see Example 1). The blank space is equivalent to three blank systems (one on p. 27 and two on p. 28). 12 After that, Bartók continued the draft at the beginning of the new system (system 3) on page 28. According to the published version, the missing section corresponds to measures 331–364. 13
Content of the draft of String Quartet no. 4, mov. V (62FSS1, pp. 27–28)
(The missing section is put in a grey background)
Citation: Studia Musicologica 62, 3-4; 10.1556/6.2021.00017
The present case differs considerably from the previous one (case A-1). In the case of the Second String Quartet, Bartók marked only the first note of a new section that he eventually left blank and resumed the sketch as a continuation of the missing measures. In contrast, in the present case, he stopped the composition without reaching a marked closure, then started writing a new beginning by leaving out some systems. In this case, the blank space would be a piece of direct philological evidence that supports an otherwise frequent recognition: a composition has several important structural points, from where composers are able to write a new formal section, and to which they compose a transitional passage. This is, however, a less likely impression for a scholar to gain from an examination of Bartók's continuity draft.
It can be a common feature that in both cases, the missing section does not contain any materials that are essentially new to Bartók. In the present case, he did not quote anything but re-organized the materials he used at an earlier point of the movement: cf. measures 99–148, which essentially correspond to the missing 34 measures, but the materials appear in a different order.
It is remarkable that Bartók apparently did not randomly leave blank spaces. The three blank systems on pages 27–28 should have been enough to accommodate the still missing 34 measures. In fact, he wrote these 34 measures almost precisely in the space of three systems on pages 38–39 in the fair copy (see Example 2). 14
Content of the autograph fair copy of String Quartet no. 4, mov. V (62FSFC1, pp. 38–39) (The section corresponding to the missing section in the draft is put in a grey background)
Citation: Studia Musicologica 62, 3-4; 10.1556/6.2021.00017
There is, however, an alternative interpretation. As is mentioned above, the missing part largely corresponds to measures 99–148. These measures can be found on pages 23–24 in the draft. Bartók repeatedly revised this section: he introduced a lot of deletions and insertions to the section, and re-wrote the section containing fifth-chords (mm. 121–48, according to the published score), so it is difficult to precisely assess the area the section occupies. Nevertheless, regarding only the valid measures and excluding insertions written in the blank space of the page, the section is written in a space also equivalent to three systems (see Example 3). Thus, it is not necessary to assume that Bartók exactly imagined the content of the missing section; he might have recalled his experience of how much space he needed for the corresponding measures in an earlier part of the movement and left essentially the same space there to fill in later. 15
Content of the draft of String Quartet no. 4, mov. V (62FSS1, pp. 23–24)
(The section corresponding to the missing section in the draft is put in a grey background)
Citation: Studia Musicologica 62, 3-4; 10.1556/6.2021.00017
A-3: Blank space concealed by paste-ups (String Quartet no. 5, mov. III)
This can be considered an irregular case in several respects. First, the blank space has been present on a draft page, but it was concealed by two paste-ups containing the final version. Second, despite the apparent discontinuity of the music written on three different pieces of paper (one complete sheet of paper and two paste-ups on it), the music might have been notated more or less continuously. Nevertheless, the existence of the blank space suggests the emergence of a compositional problem Bartók had to face.
The blank space can be found at the top half of page 25 that contains measures 58–76 from the third movement of the Fifth String Quartet (BB 110, 1934). 16 There are eight blank staves, to which Bartók later pasted two fragmentary pieces of paper (see Example 4). The paste-ups contain measures 58–65, and while a smaller piece of paper contains the first two measures (i.e., mm. 58–59), a larger piece of paper has the rest (i.e., mm. 60–65). Following the paste-ups, the music continues from measure 66.
Content of the draft of String Quartet no. 5, mov. III (71FSS1, pp. 25, 53, 51)
(The sections related to the pasted-ups are put in a grey background)
Citation: Studia Musicologica 62, 3-4; 10.1556/6.2021.00017
However, these measures might have originally represented a continuity on a separate folio that was eventually cut into at least three pieces. While the largest part of the folio, approximately a half-sized sheet of paper (the lower half missing), bears stamped page numbers 51 and 53, the other two pieces of paper were pasted on page 25. It should be noted here that the paper type of this folio (24-lined music paper) is different from the main body of the continuity draft (18-lined music paper). 17 During the composition of the Fifth String Quartet, Bartók used the 24-lined music paper primarily for sketches and later insertions; indeed the current page 53 contains various side sketches to the first and third movements. 18
Bartók originally made two notational versions of measures 58ff. on the original, complete page 53. The one can be considered an incomplete preliminary sketch; this sketch remains at the bottom of the current fragmentary page 53, except for the first two measures that were cut out and used as part of the final version of the draft. The other is essentially a continuity draft that might have originally proceeded to the other side of the folio, page 51. This page actually contains an earlier form of measure 66 that follows the paste-ups on page 25.
The simplest explanation of the existence of paste-ups would be that Bartók decided to draft measures 58–65 not on page 25 (the page belonging to the main body of the continuity draft) but on page 53, originally as a preliminary sketch, for some reason. After that, he decided not to copy the material onto page 25 but instead cut out two pieces of paper from page 53 and pasted them onto page 25 to save time, and then drafted the continuation on page 25, possibly based on a preliminary sketch on page 51. This should be regarded as a unique compositional process in Bartók's workshop, as the paste-ups are usually used for the later replacement of several measures. 19 However, the extraordinariness might be explained by the biographical fact that Bartók composed the Fifth String Quartet in haste in a single month, quite a short period, in order to meet the deadline. 20 According to this simple explanation, the blank space is somewhat deceptive, as it does not necessarily mean that Bartók skipped some systems at the time of composition. Nevertheless, the blank space still suggests that the composition of measures 58ff. required some special attention from Bartók, as he decided to write the section on a separate page.
There could be several possible reasons why he did so. The most likely one is that Bartók might have felt the necessity to make a preliminary sketch for this passage. From a compositional point of view, this passage can be worked out easily and mechanically as a chromatic double-canon in contrary motion, where two upper and lower parts go upwards and downwards, respectively. 21 However, from a practical point of view, writing a voice in contrary motion could have been a problem. 22 In addition to this, the use of a lot of accidentals might have caused some additional difficulty, for instance in aligning the notes of the four parts vertically. 23
It is, however, still possible to assume a hypothetical interruption between measures 58–65 and 66ff., because there are two unusual issues on the paste-up on page 25. First, the last measure of the second system is partially written on a hand-written staff on the right margin (see Facsimile 2), as if there had already been other material at the beginning of the next system (i.e., in this case, the beginning of p. 51). 24
Draft of String Quartet no. 5, mov. III (71FSS1, p. 25, systems 2–3)
Citation: Studia Musicologica 62, 3-4; 10.1556/6.2021.00017
Second, the music was originally not connected smoothly from measure 65 to measure 66, either in the case of the preliminary version on page 51 or the case of the original layer on page 25 (see Example 5a). The first violin part ascends more than two octaves in measures 58–66, from C♯1 to E3, but E3 as the destination and the highest note of the passage originally lasted only an eighth note and immediately descended to B1. This sounds too abrupt and dissatisfying from a melodic point of view. Bartók improved this connection by exchanging the two violin parts and letting the first violin hold E3 (see Example 5b). 25 This revision suggests that measure 66 was written earlier than the previous measures, and measure 66 might have served as a kind of composition “bookmark” where the previous measures should arrive. At that time, however, Bartók had not yet considered the inner melodic logic of the voices.
Transcription from the draft of String Quartet no. 5, mov. III (from 71FSS1, pp. 25 and 51, original form), above: pp. 25 and 51; below: p. 25
Citation: Studia Musicologica 62, 3-4; 10.1556/6.2021.00017
Transcription from the draft of String Quartet no. 5, mov. III (from 71FSS1, p. 25, final form)
Citation: Studia Musicologica 62, 3-4; 10.1556/6.2021.00017
4 Category B: unusual blank space in the autograph
B-0: Superfluous blank space at the end of an insertion (String Quartet no. 5, mov. III)
While some of the previous cases (cases A-1 and A-2) demonstrated that Bartók was able to almost precisely estimate the necessary space when he skipped some systems, sometimes he seems to have left more space than he eventually needed; consequently, some unusual blank spaces remain on certain pages of his autographs. While it is not always easy to judge whether a part of the measure written on hand-ruled staff in the margin was caused by the lack of space at the beginning of the next system (cf. case A-3) or was just done so, the existing space can easily be observed. 26
Before discussing the unusual blank space, we shall briefly examine the space at the end of the material that replaced a part of the continuity draft at a late stage of composition. The replacement can be considered a phenomenon analogous to the filling-out of the blank space, as both were written later than the surrounding part of the continuity draft. A possible difference is that Bartók wrote the replacement in a place independent from the main body of the continuity draft (usually on a separate folio); consequently, he was not bothered by the available space.
The draft of the Fifth String Quartet contains several insertions, probably due to its hasty compositional process. 27 Page 23 can be one such example. This page contains a considerably later replacement of the last measure of page 22 and a substantial part of page 24 (see Example 6). 28 Except for the last system, the material on page 24 was replaced by that on page 23. 29 It can be observed that Bartók did not use all the space on page 23; he left about a one-third blank space in the last system. 30
Content of the draft of String Quartet no. 5, mov. III (71FSS1, pp. 22, 24, and 23)
(The sections related to the replacement are put in a grey background)
Citation: Studia Musicologica 62, 3-4; 10.1556/6.2021.00017
B-1: The “soft part” and the problem of repetition (String Quartet, no. 5, mov. III) 31
Some of the unusual short blank spaces within the continuity draft are apparently related to the so-called “soft part,” one of the characteristic features in Bartók's music. 32 It has been observed that he did not always strictly deal with repeated measures, either during the compositional process or in performances of his piano music. 33 Such repeated measures are usually accompanied by a change of tempo and dynamics. Their function is to create the necessary space between two more important sections or phrases, and for this purpose the number of the measures may elastically change to create a smooth transition. Thus, the precise number of repeated measures might have been of secondary importance. 34
A draft page from the Fifth String Quartet contains a space related to a “soft part.” The space can be found at the beginning of the Trio of the third movement. On page 19, the last one-third of the first system remains empty (see Facsimile 3). This should be regarded as unusual within the continuity draft, as such a blank space usually appears at the end of the replacement or insertion (cf. case B-0). 35 In this case, however, the existence of a blank space might have been caused by the fact that Bartók did not determine the precise number of measures when he started to write the first system.
Draft of String Quartet no. 5, mov. III (71FSS1, p. 19, systems 1–2)
Citation: Studia Musicologica 62, 3-4; 10.1556/6.2021.00017
The first system currently contains measures 1–8, and the following measures are written from the beginning of the next system. These eight measures serve as a transition between the main parts of the Scherzo and the Trio. The eight measures can be divided into two sections: the first half (mm. 1–4) consists of two voices, a rapid ostinato by the first violin, and a prolonged dyad by the viola; in the second half (mm. 5–8), the second violin joins them and emphasises the meter of the section 
by doubling the note at the beginning of each beat played by the first violin.
Considering the fact that the first system only contains seven written-out measures, and its third measure is repeated by a verbal remark “bis,” it is likely that the first system originally had only 3+3 measures for the transition (i.e., the last, seventh measure might have been a later addition). The lengthening of the transition passage can be further traced back to an early version of the transition on page 18, in the second half of the third system (see Facsimile 4). As this version belongs to the first version of the Trio, there are several differences, but its very function as a transition is the same as in the final version.
36
Apparently, this earlier version contains six measures, but it is certain that this version originally contained only five measures (a simile symbol 
is added in the right margin, and the necessary clef-change in the viola part is written earlier than that).
Draft of String Quartet no. 5, mov. III (71FSS1, p. 18, systems 3–4)
Citation: Studia Musicologica 62, 3-4; 10.1556/6.2021.00017
The question as to why Bartók gradually increased the number of measures can be answered by examining the difference in tempo between the Scherzo and the Trio (in the published score, the MM numbers per eighth-note are 414 and 600, respectively).
37
Even though the draft does not contain any tempo markings, it is very likely that Bartók already intended an accelerando due to the change of the basic note-value. This change cannot be directly observed in the actual time signature 
and 
because both merely represent how the eighth notes, the base note-value of the measure, are grouped into beats. However, based on the thematic material, it is clear that the eighth notes have a melodic quality in the main part of the Scherzo; in the Trio, the eighth notes appear only in the rapid ostinato figures, and (dotted) quarter notes possess a melodic quality instead. Consequently, the Trio should be intended to be considerably faster than the main part of the Scherzo.
During the drafting process, however, Bartók probably did not yet decide the precise speed ratio between the Scherzo and the Trio. This might have caused the problem concerning how many measures the performers would need in order to achieve the accelerando in a natural manner. He might have first imagined fewer measures (five measures, according to the original layer of the first version), 38 then gradually increased the number of measures to eight (as in the final version). 39 It is possible that five measures were enough for him if he played a similar passage on the piano; however, he had to adjust to an ensemble involving four string players that might require more measures in such a passage. Due to this problem, Bartók might have decided not to fix the number of transition measures immediately; instead, he started writing the main section of the Trio in the next system so that he could modify the number of measures later. 40
B-2: The “soft part” within a quasi-Cadenza (String Quartet no. 6, mov. II)
The draft of the Sixth String Quartet (BB 119, 1939) also has a short blank space related to the “soft part,” but its context differs considerably from the previous case. 41 The short blank space (ca. one-quarter of a system; see Example 7) can be found on page 10, between the end of the trio section (mm. 114ff.) and the return of the main section (m. 122). Before the break, all four parts repeat a four-note figure in parallel motion. Although this section from the draft does not contain any verbal performing instructions, the character is obviously quasi cadenza as is indicated in the published score (at the beginning of m. 115). The use of a long measure that contains ten quarter notes (i.e., the second measure of the system) typically refers to a “cadenza” passage.
Transcription from the draft of String Quartet no. 6, mov. II (79FSS1, p. 10, excerpt)
Citation: Studia Musicologica 62, 3-4; 10.1556/6.2021.00017
In this case, however, not only the number of repeats but also the transition meant a problem for Bartók, i.e., to smoothly introduce the return of the main section. He might have needed to compose the beginning of the main section first in order to better plan the transition (cf. case B-1).
B-3: Reversed order of composition in a replacement (String Quartet no. 5, mov. III)
A different kind of blank space can be found in the draft of the third movement of the Fifth String Quartet. On page 23, there is a very short blank space in the third measure of system 2 (see Facsimile 5; according to the published score, the blank space is between the first two notes in m. 48). The published score has a comma (’) in the corresponding place, but it is unlikely that the short space might have meant a short break between the first two notes.
Draft of String Quartet no. 5, mov. III (71FSS1, p. 23)
Citation: Studia Musicologica 62, 3-4; 10.1556/6.2021.00017
On the one hand, the comma seems to be a second thought, having been added at the very last moment of the composition (i.e., it is still missing from the fair copy on transparent tissue). 42 On the other hand, the short space was most likely caused by the fact that Bartók wrote this system in an irregular order. He might have started drafting the phrases following the short blank space before writing the previous notes.
Before discussing the irregular compositional process, it is necessary to note that this page 23 is prepared as a replacement of the last measure of page 22 and a considerable part of page 24 as mentioned above (cf. case B-0). Bartók did not compose a totally new section but slightly elaborated the music he had already written on pages 22 and 24. There are many differences between the content of the original form (on pp. 22 and 24) and the revised form (on p. 23), but the most important one is that the music on pages 22 and 24 is written throughout in the same meter 
,
43
despite the fact that the section consists of phrases of irregular length (see Example 8; e.g., the first violin plays phrases consisting of 13 and 6 eighth notes). In this section, the phrases are independent of the actual meter.
44
Transcription from the draft of String Quartet no. 5, mov. III (71FSS1, pp. 22 and 24, excerpt)
Citation: Studia Musicologica 62, 3-4; 10.1556/6.2021.00017
On page 23, however, changes of time signature are introduced in measures 42 and 47 so that at least the longest phrases consisting of 13 notes (beginning at measures 41, 44, and 48, played by three different instruments) should fall into the same metric scheme (see Example 9). 45 This might have been done for the sake of performers; it should have been easier to orientate if the same phrases had the same metric accents, especially in the case of a movement that applies an asymmetric meter in quite a fast tempo.
Transcription from the draft of String Quartet no. 5, mov. III (71FSS1, p. 23, leading parts only)
Citation: Studia Musicologica 62, 3-4; 10.1556/6.2021.00017
Concerning the phrase after the blank space, however, Bartók originally used a different metric scheme (see Example 10). This may underline the hypothesis that Bartók had written the phrase following the blank space earlier than the previous phrases, and he seems not to have decided yet to facilitate the performance. 46
Reconstruction of the original intention, based on the draft of String Quartet no. 5, mov. III (71FSS1, p. 23, leading parts only)
Citation: Studia Musicologica 62, 3-4; 10.1556/6.2021.00017
The place of the original bar-lines deserves further attention. From measure 48, the original bar-lines are drawn three eighth notes earlier as compared to the final bar-lines. These bar-lines would be in the right places if Bartók had not introduced the change of time signatures in the previous measures. If this is not a mere coincidence, then he had already counted the exact number of eighth notes in the skipped measures. 47 In this case, the reversed order of composition might have been affected by the fact that the composition of the smooth connection to the existing material (i.e., the last system on p. 24; see Example 6) was Bartók's primary concern. 48
5 Category C: concealed fragmentariness in the autograph
C-1: Drafting a new passage in a new system (String Quartet no. 5, mov. V)
The knowledge of the fact that Bartók occasionally skipped a few measures may encourage us to assume that apparently continuously written draft pages might have originally contained blank spaces. It is, however, necessary to rely on some irregular elements of the notation to persuasively discuss the existence of such blank spaces at an early stage.
A good example could be found in the draft of the fifth movement of the Fifth String Quartet. On page 32, near the beginning of the movement, the last system contains a four-part octave canon at the distance of a quarter note (mm. 43ff. in the published version; see Facsimile 6). An irregular element of this passage is that the two lower parts (the viola and cello) originally began with a rest, but later an eighth note was written at the beginning. 49
Draft of String Quartet no. 5, mov. V (71FSS1, p. 32)
Citation: Studia Musicologica 62, 3-4; 10.1556/6.2021.00017
It frequently happens that the modification introduced in a measure is related to that in the surrounding measures. In this case, the revision at the beginning of measure 43 suggests that Bartók drafted the given measure earlier than the previous ones. The direct proof is that in the previous phrase, the two upper and lower parts are paired, 50 and these pairs play an octave canon at the distance of a quarter note (mm. 35–42). If Bartók had notated these measures in the previous system, he would not have originally written rests at the beginning of measure 43 in the two lower parts.
As there is no apparent break in the first violin part, it seems possible that Bartók drafted only the first violin part throughout the last two systems and then elaborated the lower three parts in a reversed order. This might already be an interesting finding, as it reveals that an apparently continuous passage originally consisted of several blocks that were elaborated independently.
However, two additional issues deserve attention and may slightly modify the above hypothesis. First, the previous measure (m. 42) is drafted on a hand-ruled staff in the right margin. Bartók usually left the margin blank for possible later revision, and he rarely wrote an entire measure in the margin unless he did not have enough space on the page. 51 He might have had to use the right margin in this case because he had already started notating the following system. Second, the last system begins with a new phrase. This may happen merely by accident; however, as is discussed, in some of the abovementioned cases (especially A-2 and B-1), Bartók indeed used a new system to write new material after skipping a section. Consequently, wherever the beginning of a system coincides with that of a phrase, there is some possibility that Bartók started drafting before completing the previous system. In this case, it is notable that the passage written as a four-part canon seems to have required special attention in order that the intended contrapuntal texture might work well. Thus, he might have worked out the music in advance, independently of the previous measures. 52
On the same page, there is an additional proof that Bartók intentionally started drafting a new system with new material. System 2 on the same page, albeit partially crossed-out, begins with a new passage beginning on E2. In fact, this is not an independent passage but a part of a lengthy thematic statement corresponding to measures 14–26 in the final version (systems 1–2 on page 32). Based on the published score, this phrase might have been intended as a transposition of the preceding (also partially crossed-out) phrase on B♭ 2 a diminished fifth lower. In the draft, however, these phrases were not written continuously. At the end of system 1, the second violin goes down to F1 but suddenly leaps a major seventh higher to E2, the next note in the next system. This strongly suggests that system 2 had already been written before system 1 was completed. Thus, the end of system 1 was not followed by the beginning of system 2. Concerning this hypothetical chronology between systems 1 and 2, it is important to note that the end of system 1 is written on a hand-ruled staff in the right margin.
As a conclusion of the present case, two clues are established concerning the detection of an irregular order of composition: (1) the existence of a hand-written staff in the margin and (2) the beginning of a new phrase at the beginning of a system. These clues, especially the second one, may inspire an interpretation concerning the fragmentariness concealed in the final form of the continuity draft.
C-2: Deceptive insertion (String Quartet no. 5, mov. V)
In the continuation of the draft page discussed in the previous case, there is an unusual verbal remark by Bartók: “A. külön lapról ide” [A. from a separate page] (see Facsimile 7). This remark can be found on page 33, between the first two systems. Compared with the published score, the first two systems contain measures 52–62 and 113ff.; thus, 50 measures are missing from this page. The letter “A” refers to page 34 that has the same letter in the left margin, with a colossal brace marking all the system on the page. This means that the content of page 34 should be inserted between the first two systems on page 33. Indeed, page 34 contains the missing 50 measures (mm. 63–112).
Draft of String Quartet no. 5, mov. V (71FSS1, p. 33, systems 1–2)
Citation: Studia Musicologica 62, 3-4; 10.1556/6.2021.00017
The fact that the paper type of page 34 (24-lined music paper) is different from the rest of the continuity draft (18-lined music part) is less essential than it may appear. 53 In general, the paper type offers essential information concerning the (micro-)chronology, and the use of different kinds of paper within a single source suggests that each of them was used at different moments. 54 Thus, page 34 seems to be a considerably later insertion.
This general theory cannot, however, be applied to the present case. From the viewpoint of paper study, it should be noted that Bartók occasionally used the 24-lined paper from the very beginning of the composition process of the Fifth String Quartet to make preliminary sketches on it. 55 He used two types of music paper side by side: the 18-lined paper primarily for the continuity draft and the other, 24-lined paper primarily for sketches. The fact that he nevertheless wrote a part of the continuity draft on a page of 24-lined paper (= page 34) can be explained by the fact that Bartók wanted to notate the insertion on a single page. Judging from the fact that Bartók filled page 34, a page of 24-lined paper, with notation to its full capacity, it must have been impossible for a single page of 18-lined paper to accommodate the intended insertion. 56
The content of page 34 differs considerably from the rest of the continuity draft due to the existence of dynamics that is exceptional in the continuity draft in general. This extraordinary fact does not mean that page 34 was written much later as the insertion. Conversely, it seems more likely that page 34 was written largely together with page 33.
57
The actual reason is that in the first two systems of page 33, the content of the upper two parts is not continuous. Instead, the end of system 1 of page 33 continues to the beginning of page 34, and the end of the page continues to system 2 of page 33. In fact, the end of system 1 of page 33 contains several layers, but none of them can be repeated by two-measure simile symbols 
at the beginning of system 2. It is improbable that the substantial thematic material from measure 113 played by the two lower parts ‒ the inverted form of the main theme from measure 14 ‒ was intended to be accompanied by the figure in markedly light character as used in measures 56–62.
Thus, there arises the question why Bartók did not make the draft in a straightforward way, i.e., from top to bottom on page 33 but, instead, inserted a full draft page (p. 34) between systems 1–2. A possible explanation is that he first drafted only the lower two parts on page 33. Somewhat later, when he added the two upper parts, he decided to insert 50 measures on a separate page. 58
Judging from the fact that system 2 on page 33 (mm. 113ff.) begins with a new formal section, it is also possible that Bartók started to write the second system before he had filled the first system (cf. case C-1, although here the end of the previous system does not extend into the margin). If this is the case, page 33 (and also p. 32) originally comprised short sketches containing the beginning of new phrases that were later elaborated as part of the continuity draft. While Bartók was able to broadly assess the necessary blank space on page 32, he left too little space on page 33. Nevertheless, it is also possible that he originally planned a different formal structure in which he might have introduced the original and the inverse form of the main theme directly one after another. 59
C-3: “Bookmarking” structural points (String Quartet no. 5, mov. I)
There are further examples in the draft of the first movement of the Fifth String Quartet where the beginning of a system and a new phrase coincide. This movement begins with a characteristic theme consisting of repeated notes (B♭) and a fluctuation of a major second (B♭‒C‒B♭; see Example 11). Later in the movement, the initial theme returns several times on different pitches, signaling the beginning of a new section. The theme appears precisely seven times in the first movement, and except for the last two in the Coda, the first five occurrences of the theme are played in octave unison, emphasising the tonality of the ensuing section.
String Quartet no. 5, mov. I, mm. 1–5
Citation: Studia Musicologica 62, 3-4; 10.1556/6.2021.00017
These five occurrences of the theme are the following (note that the measure numbers given are based on the published score):
- - B♭ (m. 1 on p. 1, at the beginning of the “first theme” in the “exposition”)
- - C (m. 37 on p. 3, at the beginning of the “allusion to the first theme” in the “exposition,” leading to the “second theme”)
- - E (m. 59 on p. 4, at the beginning of the “development”)
- - F (m. 126 on p. 8, at the beginning of the “first theme in inversion” in the “development,” also in inversion and with an E upper pedal note)
- - B♭ (m. 159 on p. 10, at the beginning of the “first theme in inversion” in the “recapitulation,” also in inversion) 60
It should be noted that none of the occurrences is identical. While the first measure of the original occurrence (m. 1) contains a characteristic fluctuation motif of a major second, in the case of later occurrences, this motif is usually preceded by repeated notes of a varying number. Remarkably, concerning the second to the fourth occurrences, the beginning of the motif falls at the beginning of the new system in the draft (see Example 12a, 12b, 12c; a dashed line marks the system break on the draft pages). 61 If this is not a mere coincidence, Bartók might have first written the motif at the beginning of some systems, as a kind of “bookmark” for the beginning of the new section. For instance, concerning the second occurrence, it is likely that he had already started writing on page 3, and the lack of space obliged him to make such an unusually dense notation on page 2. Even though there are no irregular elements in the surrounding sections (differently from the previous cases such as C-1 and C-2), it is remarkable that Bartók's notation is extremely dense in one part of page 2: systems 2–3 on page 2 contain 11 measures (from the second half of m. 21 to the first half of m. 32). In a later part of the draft, the music is more spaciously notated: for instance, the bottom two systems on page 9 contain only nine measures (mm. 147–55) whose content is similar to that on page 2.
String Quartet no. 5, mov. I, mm. 37–41
Citation: Studia Musicologica 62, 3-4; 10.1556/6.2021.00017
String Quartet no. 5, mov. I, mm. 59–63
Citation: Studia Musicologica 62, 3-4; 10.1556/6.2021.00017
String Quartet no. 5, mov. I, mm. 125–129
Citation: Studia Musicologica 62, 3-4; 10.1556/6.2021.00017
Later, he may have added the preceding, repeated notes that can be considered the so-called “soft part” to prepare the return of the characteristic motif. 62 At least in a single case, it can be established that Bartók increased the number of repeated notes by later inserting them in the margin: concerning the second occurrence ‒ different from the published score ‒ in the draft, the initial motif began directly after the bar-line on page 3. 63 However, the bar-line was later shifted by a half measure due to the insertion of the repeated notes at the end of the previous page, on the hand-ruled staves (corresponding to the first half of m. 38; see Example 13).
Transcription from the draft of String Quartet no. 5, mov. I (71FSS1, pp. 2–3, excerpt), before the insertion and the shift of bar-lines
Citation: Studia Musicologica 62, 3-4; 10.1556/6.2021.00017
Transcription from the draft of String Quartet no. 5, mov. I (71FSS1, pp. 2–3, excerpt), after the insertion and the shift of bar-lines
Citation: Studia Musicologica 62, 3-4; 10.1556/6.2021.00017
Concerning the fourth occurrence on page 8, however, it is highly likely that the draft was not written continuously, and there is evidence other than the notation itself. When writing the draft on page 6, one of the previous pages, Bartók immediately prepared two series of insertions on the following page 7 (see Example 14). As he wrote the continuation of page 6 directly following the insertion on page 7, the revision must have taken place before he had finished page 6 at the latest. Such immediate revision should be considered rather normal, even if it is not too frequent. 64 At this point, it is noteworthy that Bartók used the pages in an irregular order: the content of a verso page (p. 6) continues to another verso page (p. 7); the other side of the latter page is page 8. A possible explanation is that he used a blank, separate single folio to write the insertion. However, this is unlikely to have happened, as he usually wrote such immediate revisions either on the same page or on the following page, which belongs to the main part of the continuity draft. 65
Content of the draft of String Quartet no. 5, mov. I (71FSS1, pp. 6–7)
(The sections related to the insertion are put in a grey background).
Citation: Studia Musicologica 62, 3-4; 10.1556/6.2021.00017
A more likely interpretation is that page 8 originally followed page 6, and Bartók had already entered the return of the initial motif at the beginning of page 8, possibly before he started writing on page 6. Note that in the last two systems on page 6, Bartók has already used the right margin. This might have been caused by the lack of space for writing the music leading to the beginning of page 8. The problem might have been solved by reversing the folio and writing the continuation of page 6 on page 7, instead of page 8.
The above hypotheses suggest that similarly to the fifth movement of the Fifth String Quartet (cf. cases C-1 and C-2), the draft might have originally consisted of shorter blocks that Bartók started writing independently from each other, then he later linked them to create a continuous notation. It cannot be established, however, how much space existed between the blocks and whether these blocks existed simultaneously (and scattered over the continuity draft). It is possible that he skipped only a few measures (cf. case A-1); yet it cannot be excluded that he left a significantly larger free space comprising several systems (cf. case A-2).
It is, however, not very likely that similar hypothetical blocks can be very frequently observed in other works by Bartók. The draft of the Fifth String Quartet might be considered an exception because the work had to be written in a relatively short period of time, and, especially in the first movement, where he created a clear and unique formal plan that eventually justifies his supposed work on independent blocks. 66
C-4: Shifted bar-lines by an eighth note (String Quartet no. 4, mov. I)
As a point of departure, the last case of the present paper takes an extraordinary issue markedly different from the previous three cases: the shift of bar-lines by such a small note-value as an eighth note. This issue may also impact performance practice, as it can be considered a piece of philological evidence of what Bartók wrote to Rudolf Kolisch: “in the 1st movement of the 4th Quartet bar-lines are very often only markings of orientation.” 67
The problem in question can be found on a draft page of the first movement of the Fourth String Quartet. At the top of page 4 (corresponding to m. 63 in the published score; see Facsimile 8), Bartók inserted an eighth note-value at the beginning of the first measure at a later point of the drafting process. Even though it is hardly possible to tell the difference between an ordinary quarter-note rest and the quarter-note rest transformed from an eighth-note rest in the two upper parts (especially when the reader does not suspect the existence of such a quarter-note rest), it is evident that an eighth note (or rest) was added at the beginning of the measure in the two lower parts, due to quite dense notation there.
Draft of String Quartet no. 4, mov. I (62FSS1, p. 3, last system)
Citation: Studia Musicologica 62, 3-4; 10.1556/6.2021.00017
Draft of String Quartet no. 4, mov. I (62FSS1, p. 4, systems 1–3)
Citation: Studia Musicologica 62, 3-4; 10.1556/6.2021.00017
The addition of an eighth note-value made it necessary to re-organize the following measures. Based on its visual appearance, the re-organization can be summarized as a shift of bar-lines leftwards by an eighth note. The old, invalid bar-lines were scratched out, 68 then the new bar-lines were inserted to the left of them. Where such an operation is not possible (i.e., at the beginning or the end of a system), an eighth note was transferred from the end of a system to the beginning of the next system. As there are no scratched-out bar-lines in the third system, Bartók must have stopped writing there and, before continuing, he revised the notation from the beginning of the page. 69
Interestingly, however, there is no such relationship between the first measure of page 4 and the last measure of page 3. This means that Bartók started writing page 4 before finishing page 3. After he reached the end of page 3, he might have realized that an eighth note was necessary to conclude the phrase played by the cello part. Instead of introducing an irregular measure consisting of four quarter notes and an eighth note, he inserted the necessary eighth note at the beginning of page 4. As the cello part leads the canon-like passage in the previous measures (mm. 60–62), it is unlikely that Bartók first notated the viola part throughout and only added the cello part later. 70
From a musical point of view, however, this case is strange, as the beginning of page 4 (= from the second eighth of m. 63) does not seem to be a new beginning that usually follows some skipped measures, as happened in the previous cases. 71 The viola part actually introduces an apparently new motif with a new rhythm, but this is nevertheless organically derived from the canon-like passage in the previous measures. This passage is derived from the phrase first introduced in the exposition (mm. 15ff.), where the phrase was freely dealt with both melodically and contrapuntally. In measures 60ff., however, this phrase undergoes a rather strict transformation process: the length of each phrase is gradually shortened, regarding the number of notes and the sum of the note-values (see Example 15). Thus, it seems that measures 63ff. might not have been written without planning the previous measures; at the same time, the unusual insertion of an eighth note at the beginning of page 4 suggests that Bartók had not yet imagined the previous measures precisely.
Transcription from the draft of String Quartet no. 4, mov. I (62FSS1, pp. 3–4, leading parts only)
Citation: Studia Musicologica 62, 3-4; 10.1556/6.2021.00017
A key to understanding this apparent dichotomy is a continuity sketch of the first movement found on pages 30–32 in the same source. 72 This sketch precedes the draft and essentially contains the materials of the movement up to the first beat of measure 93 (i.e., the beginning of the recapitulation), 73 although many elements are still missing. 74 Concerning the development section (mm. 49–92), the first measures (mm. 49–62) essentially correspond to the final version of the draft; thus, Bartók might have been able to copy the music simply from the sketch (see Facsimile 9).
Sketch of String Quartet no. 4, mov. I (62FSS1, p. 31, systems 2–4)
Citation: Studia Musicologica 62, 3-4; 10.1556/6.2021.00017
Concerning the following measures, the content is considerably different from the next version. If Bartók had already wanted to further develop the music from there, it is likely that he nevertheless considered the next measure (corresponding to m. 63 of the published score) to represent a “new beginning” where he might have been able to write the continuation without finalizing the preceding measures.
Indeed, the following measures are significantly extended in comparison with the sketch. Regarding the section containing the shift of bar-lines, Bartók elaborated about four measures in the sketch into eight measures in the draft (mm. 63–70). At this point, it is worth analysing what kind of compositional problem may be behind this elaboration and whether the existence of a specific problem justifies the irregular order of composition.
In the sketch, Bartók applied the transformation process as mentioned above more strictly and more mechanically by gradually reducing the number of notes from ten to two (see Example 16). The process continues until the very end of the sketch, where the first violin repeatedly plays a single note, F1. 75 Thus, the goal of this process seems to be the reduction of the music to a “single note,” the most primitive state of the music, 76 where the recapitulation starts as a new beginning. 77
Transcription from the sketch of String Quartet no. 4, mov. I (62FSS1, pp. 31–32, excerpt)
Citation: Studia Musicologica 62, 3-4; 10.1556/6.2021.00017
If this process guaranteed the framework of the development section, then Bartók might have been able to exercise his musical fantasy to write new phrases that, if they do not closely belong to the framework, nevertheless enrich the content of the music. It seems that a long melody in measures 65–68, drawing a large arc, would have been the result of the fantasy, as a striking lyric excursion amidst the music which is continuously jarred by cluster sounds. It is, however, notable that this melody, while almost freely expanding the register first upwards and then downwards, is nevertheless solely made of the same sort of motivic cell that opened the imitative passage in measure 60 (see Example 17). 78 This might be quite an interesting moment in Bartók's compositional process. It can more frequently be observed that he seeks to escape from a strict system that might have facilitated the creative process at a later moment of the composition. 79 In this case, however, while he was doing so, he nevertheless followed another system simultaneously.
Analysis of the long phrase in String Quartet no. 4, mov. I
Citation: Studia Musicologica 62, 3-4; 10.1556/6.2021.00017
We cannot tell to what extent the elaboration of measures 63ff. meant a compositional problem for Bartók; at least, however, it is certain that the composition of these measures was a task markedly different from copying the sketch into the draft (as he did in the previous measures). Thus, it seems likely that he postponed such mechanical work that could have been easily done without inspiration.
At the end of the present section, it is necessary to mention that the insertion of an eighth note might have been related to the notational complexity present in the sketch, too. In the third measure of the third system on page 31, there is a sort of vertical line at the beginning of the measure that might have been intended to mean that the content of the upper three parts should be shifted rightwards, possibly by an eighth note. If this is the case, it is possible that when Bartók picked up the measure and elaborated it, he forgot what he originally intended there. Thus, the insertion of an eighth note might have been caused by mistake. This interpretation, however, does not reduce the extraordinariness of the insertion of an eighth note. The fact that the inserted eighth note triggered the re-organization of the following measures demonstrates that several bar-lines after measure 63 actually do not have the usual function of marking metric accents.
6 Conclusion
The present paper demonstrates the phenomenon of fragmentariness with 10 cases grouped into three categories. As discussed above, there might have been several reasons why Bartók temporarily left some blank space at the time of composition. On the one hand, what he did not write down immediately was generally a section that could easily be written based on already existing material (cases A-1, A-2, B-3, and C-4), and so he chose to start composing the following new section instead. As a busy person occupied with a lot of time-consuming tasks, including folk music research, he might have tried to save time by not fully notating the section that could easily (and possibly, mechanically) be written out later. 80 At the same time, there might have been some specific problems (mainly concerning a “soft part” of composition) with precisely determining the number of repeats that he did not finalize directly during the drafting process (cases B-1 and B-2).
On the other hand, as a composer relying on inspiration rather than a mechanical application of systematic compositional devices, Bartók might have tried to grasp the passing inspiration whenever possible; thus, he might have preferred to work on a new section that required inspiration (especially case C-4). At the same time, however, it might have also happened that the composition of a new passage might have in turn inspired the composition of the preceding measures (especially cases B-2, C-2, C-3). 81
The very concept of discovering fragmentariness in the continuity draft may remain problematic due to the uncertainty of identifying the breaks between the fragments that are not always visible on the surface of the notation. Nevertheless, the concept of fragmentariness might change our understanding of the “continuity draft” with Bartók. If a scholar successfully discovers the fragmentariness in the autograph, it may serve as an additional piece of evidence for the interpretation of the compositional problems Bartók might have faced. 82 Furthermore, the recognition of such compositional problems may enable scholars to conduct an interpretation at a much higher level, that is, the interpretation of “the spirit of the work” that seems to have been a key concept of Bartók's creative thinking. 83
Concerning the reconstruction of the order of composition within a group of autographs, the most prominent research findings can be found in László Somfai's monograph, some of the Béla Bartók Complete Critical Edition [henceforth: BBCCE] volumes (especially vols. 24 and 41), and the introduction to the facsimile publications: see László SOMFAI, Béla Bartók Composition, Concepts, and Autograph Sources (Berkeley, CA: University of California Press, 1996); Klára MÓRICZ (ed.), Concerto for Orchestra (München: Henle and Budapest: Editio Musica, 2017) (= BBCCE, vol. 24); Yusuke NAKAHARA (ed.), Mikrokosmos (2) (München: Henle and Budapest: Editio Musica, [2021]) (= BBCCE, vol. 41); Felix MEYER (ed.), Musik für Saiteninstrumente, Schlagzeug und Celesta: Facsimile des Partiturautographs und der Skizzen (Basel: Paul Sacher Stiftung, 2000). A direct predecessor of the present paper dealing with the micro-chronology in a narrower sense is, however, Somfai's chronological reconstruction of the sketches to the First Violin Sonata. He convincingly demonstrated that, for instance, folios 24v and 25r of the “Black Pocket Book” contain several chronologically separate layers even though the notation seems to be continuous; see László SOMFAI, “‘Written between the desk and the piano’: Dating Béla Bartók's sketches,” in A Handbook to Twentieth-Century Musical Sketches, ed. by Patricia HALL and Friedemann SALLIS (Cambridge: Cambridge University Press, 2004), 114–130. See also the reconstruction of a sketch to the third movement of String Quartet no. 2 by Vera LAMPERT: “Bartóks Skizzen zum III. Satz des Streichquartetts Nr. 2,” in Documenta Bartókiana, vol. 6 (Budapest: Akadémiai Kiadó, 1977), 179–189.
The transcriptions are to be included in the Critical Commentary volume of the string quartet volumes: László SOMFAI, Zsombor NÉMETH and Yusuke NAKAHARA (eds.), String Quartets (2) (München: Henle and Budapest: Editio Musica, forthcoming) (= BBCCE, vol. 30).
Here it should be noted that I was not fully aware of this phenomenon when I researched the autographs of the Mikrokosmos. For one of a few possible cases related to no. 133 “Syncopation,” see Yusuke NAKAHARA (ed.), Mikrokosmos (2), 174–176.
Concerning case A-2 dealing with the fifth movement of the Fourth String Quartet, it might be fruitful to analyse the possible formal function of the missing section and compare it with Bartók's analysis, written at the request of Universal Edition as the introductory text for the Philharmonia pocket score editions; for the modern English translation of the analysis, see László SOMFAI and Zsombor NÉMETH (eds.), String Quartets (1) (München: Henle and Budapest: Editio Musica, 2022) (= BBCCE, 29). The apparent difference between the work-in-progress and the later analysis might be due to the lack of the fixed plan of the movement at the time of composition; for a similar case concerning Bartók's Concerto for Orchestra, see Klára MÓRICZ, “Operating on a Fetus: Sketch Studies and Their Relevance to the Interpretation of the Finale of Bartók's ‘Concerto for Orchestra’,” Studia Musicologica 36/3–4 (1995), 461–476. In the case of the fifth movement of the Fourth String Quartet, however, it cannot be excluded that the formal plan could have been of secondary importance, in comparison with the essential concept of the movement that determines not the formal structure but the musical process.
BBA, BAN 494a–b, folio 1v. The source has already been transcribed by Vera Lampert and published in her article accompanied by a facsimile reproduction: see LAMPERT, “Bartóks Skizzen.” (The facsimile reproduction can be found at the end of the publication).
It should be noted that this sketch is written with halved note-values throughout, in comparison with the final version; thus, a half note in the sketch is equivalent to a whole note in the final version. At the same time, however, the time signature is 
in both versions. Consequently, the number of measures differs in the sketch and the final version: seven measures in the sketch might be equivalent to 14 measures in the final version. Here, however, the final version is shorter by one measure for an unknown reason.
The rhythm does not precisely correspond to the final version. In the sketch, the two parts play simultaneously but in the final version, the viola part joins the cello part later. Such difference, however, might be explained by the fact that Bartók merely intended these dyads as a reminder.
Ferenc BÓNIS, “Quotations in Bartók's Music: A Contribution to Bartók's Psychology of Composition,” Studia Musicologica 5/1 (1963), 371. The relationship between Bartók's string quartet and Debussy's opera is, however, recently questioned: see Péter TORNYAI, “Idézi-e Bartók Mélisande halálát a 2. vonósnégyesben?: Kísérlet egy bő fél évszázados pontatlanság korrekciójára” [Does Bartók quote Mélisande's death in his Second String Quartet?: An attempt to correct a semicentennial inaccuracy], Magyar Zene 58/2 (2020), 230–237.
SOMFAI and NÉMETH, String Quartets (1).
There are several problems concerning the status of the passage as a quotation. First, only the melody of the first five measures (mm. 88–92, according to the published score) largely corresponds to Pelléas. Consequently, it is possible that Bartók did not merely quote Pelléas but intended to write his own music based on it, by incorporating the elements derived from Pelléas. Second, it is also possible that not only this passage but also at least some part of the movement is related to Pelléas: for instance, Tornyai argues the possible relationship between interlocking major thirds in the string quartet (e.g., in mm. 68, 72, etc.) and Pelléas; see TORNYAI, “Idézi-e Bartók,” 233–236. To highlight the passage as the only quotation may obscure the possible relationship of other elements to Pelléas.
The original autograph of the Fourth String Quartet is currently located in the Paul Sacher Stiftung, Basel (CH-Bps), Béla Bartók Collection, deposited by Peter Bartók (hereinafter: PSS), shelfmark 62FSS1 (photocopy in the Budapest Bartók Archives). Bartók's autographs in PSS occasionally bear more than one pagination. The most important two in this manuscript are the following: (1) an incomplete series of handwritten pagination given by the composer himself and (2) an almost complete one stamped by the staff at the former New York Bartók Archive. For the sake of better orientation, the present paper only refers to the latter one. It should be noted that these paginations do not always follow the actual order of the pages; consequently, four pages of a bifolio may have non-consecutive page numbers.
In fact Bartók left five blank staves on page 27 without actually grouping these staves into a system, unlike two blank systems on page 28, where a brace is added at the beginning of each system. Nevertheless, it is obvious that the five staves on page 27 are also intended as a full system, with an additional blank staff that is usually left between adjacent systems.
The existence of this blank space has already been mentioned by Somfai: “Movement V is written on two bifolios …; it is a continuity draft with lots of shorthand notation and just some empty spaces on the last page to indicate the structure.” See László SOMFAI, “Bartók and the Paper-Studies: The Case of String Quartet no. 4,” Hungarian Music Quarterly 1/1 (1989), 10.
The fair copy is currently located in PSS, shelfmark 62FSFC1 (photocopy in the Budapest Bartók Archives).
It remains an open question concerning the chronological relationship between the extensive revisions and the space left on pages 27–28. It is naturally possible that the revisions written not in the system but in the blank spaces (i.e., on single blank staves between systems or in the margin) might have been done considerably later.
The draft is currently located in PSS, shelfmark 71FSS1 (photocopy in the Budapest Bartók Archives).
For a full description of the paper types, see SOMFAI, NÉMETH and NAKAHARA, String Quartets (2).
The draft (71FSS1) contains four folios of the 24-lined music paper (eight pages in total: pp. 34, 50–51, 53–54, 62, 64, and a blank page). Except for page 34 that eventually contains the final version of the draft of the fifth movement, all other pages have preliminary or side sketches that became the basis for the final version of the continuity draft. See also the examination of case C-2.
For instance, Bartók used paste-ups on several Mikrokosmos manuscripts, in order to make an easily readable final version ready for use at concert performances as well as at the publisher. See, for instance, Yusuke NAKAHARA, “Genesis and the ‘Spirit’ of Bartók's Mikrokosmos” (PhD diss., Budapest: Liszt Academy of Music, 2020), 99–100.
The Fifth String Quartet was commissioned by Elizabeth Sprague Coolidge in June 1934, and the deadline for its completion was set as December of the same year. As usual, however, Bartók had time for the composition only during his summer vacation; consequently, he had only one month. For the deadline, see a letter by Carl Engel to Bartók, June 4, 1934 (PSS, PB, BB–LIBCONGR; photocopy in the Budapest Bartók Archives); see also SOMFAI and NÉMETH, String Quartets (1) concerning the genesis of the work.
In the published score, the cello part slightly deviates from the strict canon: in measure 60, the first three notes are D♯‒E‒F, instead of F‒E‒D♯ (i.e., the melody does not descend but ascends; see also mm. 62 and 64). This is, however, the result of an intentional revision by Bartók. It is possible to observe the traces of revision on page 25.
It can be considered a similar issue that Bartók made some preliminary sketches to some sections of the first movement, where an inverted form of an early part of the movement appears (see pp. 54–53, and 56; here the order of pp. 53–54 is reversed, because the notation continues from p. 54 to p. 53).
Worthy of attention is the fact that in a part of the preliminary sketch on page 53, Bartók meticulously marked the vertical alignment of the four parts by vertical lines. The difficulty in notation can be underlined by the fact that there is probably a slip of the pen in the second violin part, where he might have written the part erroneously, by inadvertently omitting an eighth note; as a result, a considerable part of the third and fourth measures was shifted rightwards.
The partial use of the hand-written staff in the margin should not be regarded as strong proof, as Bartók occasionally did so when he did not want to break the content of a measure at the end of a system. However, considering that it is a rather time-consuming task to write staves in the margin, he might not have done so if there had been available space in the next system, especially when he had rather limited time for composition.
The issue of exchanging voices is usually related to the “re-orchestration” of a short section. See, for instance, page 41 of the draft of the Fifth String Quartet (mm. 411ff. of mov. V), where the upper two parts are exchanged for some reason, probably in order to assign the higher register to the first violin. Nevertheless, there may be several exchanges of the parts that may serve as a proof of the irregular order of composition.
Concerning the difficulty in interpreting the notation in the margin, see a draft page from the Second Violin Concerto (PB 76VPS1, page 1); the page can be found in the form of a facsimile reproduction in an article by Somfai; see László SOMFAI, “Diplomatic Transcription versus Facsimile with Commentaries: Methodology of the Bartók Edition,” in De editione musices: Festschrift Gerhard Croll, ed. by Wolfgang GRATYER and Andrea LINDMAYR (Laaber: Laaber Verlag, 1992), 79–97. Somfai calls the readers' attention to the fact that the passage of the solo violin written in the margin is neither a second thought nor a later insertion, as there is a preliminary sketch that already contains the violin theme.
Concerning the genesis of the Fifth String Quartet, see footnote 20.
It is likely that page 23 was clipped to page 24, based on the trace of iron rust on page 25 (the other side of p. 24) and pages 22 and 26 (the adjacent pages of the folio containing pp. 24–25). Note that page 24 contains two kinds of continuity draft. One is an abandoned draft belonging to the original version of the Scherzo da capo, written from the beginning of the page and occupying about one system and two-thirds of a second system. The other is the part of the final version of the Scherzo da capo, written directly after the abandoned draft until the end of the page.
It should be noted that Bartók did not cross out the replaced section on page 24. Considering the fact that he did not add pagination to page 23, it is likely that this replacement took place at the very last moment of the draft process. It is even possible that Bartók made this replacement during the preparation of the fair copy when he no longer needed to cross out the discarded section and to paginate the draft page for the sake of his own orientation. Notably, there is a contrary case: another page (p. 29) contains a replacement, but this page bears a remark related to the orientation (“27. laphoz” [to p. 27]; note that this refers to Bartók's own pagination and the current page 28 bears his page number “27”). On page 28, the replaced section was eventually crossed out, and then a reference was added concerning where the valid version can be found. For the pagination issue of the source, see SOMFAI, NÉMETH and NAKAHARA, String Quartet (2).
There are some additional examples within the draft of the Fifth String Quartet: e.g., pages 20, 21, and 37. All these pages contain a replacement of the material within the continuity draft.
In the present paper, I occasionally refer to the asymmetric meter of this movement as “Bulgarian rhythm,” an authentic term derived from Bartók's usage; see his essay, “The So-called Bulgarian Rhythm” in Béla Bartók Essays, ed. by Benjamin SUCHOFF (London: Faber & Faber, 1976), 47. Note that modern ethnomusicology prefers the term “aksak” instead of “Bulgarian rhythm,” and Timothy Rice offers the term Bulgarian “meter” instead of “rhythm” as a more precise expression; see Timothy RICE, “Béla Bartók and Bulgarian Rhythm,” in Bartók Perspectives, ed. Elliott ANTOKOLETZ et al. (Oxford: Oxford University Press, 2000), 197. However, it is justifiable to use the authentic term “Bulgarian rhythm” in relation to Bartók's own compositions as a particular case within “aksak.”
Concerning the term “soft part,” see SOMFAI, Béla Bartók: Composition, 283. According to Somfai's explanation, a “soft part” means a passage of “repeated figures, between two themes, or in the coda,” and there Bartók sometimes varied the number of repetitions.
The problem is probably best represented by Mikrokosmos no. 153 “Six Dances in Bulgarian Rhythm (6).” This piece contains six measures solely consisting of repeated notes (mm. 69–74). In fact, Bartók gradually increased the number of repetitions from one to six. Nevertheless, he seems not to have followed the published score at concerts, as he played five measures on the recording instead of six. For details, see NAKAHARA, Mikrokosmos (2).
See, for instance, an unusual instruction “rep. ad libitum” for measures 37–38 in Mikrokosmos no. 103 “Minor and Major,” where the number of repetitions of these two measures is left to the performers' liberty. As a related case, see also measure 80 in the “Musettes” from the suite Out of Doors (BB 89, 1926), where Bartók marked the second beat as “due o tre volte ad lib.” [two- or three-times ad libitum].
In fact, page 19 contains a revised form of the Trio; consequently, the function of page 19 itself should be regarded as a “replacement” of the original version of the Trio (on pp. 18, 67, etc.), as is discussed in the case B-0. Nevertheless, this does not affect the relationship between systems 1–2 on page 19. These systems are intended to be continuous, regardless of the existence of a blank space at the end of system 1.
As the original version of the Trio is written in a different tonality, the ostinato passage is a whole tone higher than in the final version. However, it should be regarded as a remarkable difference that the original Trio completely lacked the sustained dyad that produces a Bartókian major-minor chord together with the ostinato passage. Here, it deserves a brief remark regarding a statement by Marcia Beach. According to her, the C2/A1 dyad, implying the tonality of F, had already existed at the end of the Scherzo section and preceded the original Trio; then, later, the dyad eventually triggered the revision of the tonality of the Trio from G to F; see Marcia BEACH, “Bartók's Fifth String Quartet: Studies in Genesis and Structure” (PhD diss., Rochester, NY: Eastman School of Music, 1988), 68–69 and 72. However, it should be mentioned that in the last three measures of the Scherzo the viola part originally had a whole rest instead of a dyad, and this version is followed by the original Trio. Thus, the dyad did not affect the choice of the tonality of the revised Trio but, conversely, is directly chosen due to the tonality of the revised Trio and added in order to make a smooth connection between the Scherzo and Trio sections.
In the score, the MM numbers are written in smaller values so that an ordinary metronome should indicate the tempo: 
= 46 and 
= 120. The calculated values of the MM number are presented here for easier comparison.
It is, however, possible that at the beginning of the Trio the transition might have originally been intended to consist of not five but four measures. The width of the second and third measures is too narrow, even if they only have a simile symbol. It seems that these measures originally constituted a single measure (their total width is largely equivalent to the width of the fifth measure that also has a single simile symbol). If this is the case, Bartók introduced the simile symbols at a later moment of the draft process.
A possible alternative interpretation is that Bartók decided to use a more regular phrase structure made of binary divisions (2+2, 4+4, etc.) that most performers can easily perceive rather than irregular ones. Concerning a possible “concession” to the performers, see also below the discussion of case B-3. It is, however, notable that Bartók occasionally used an irregular phrase structure in a piece in Bulgarian rhythm; for instance, see NAKAHARA, “Genesis and the ‘Spirit’,” 339–344.
Concerning the extension of the number of repeated measures, it is possible that Bartók imagined different tempos for the original and final version of the Trio, and the different tempos might have required different numbers of repetition. On page 67 in the draft, there is a calculation of the tempo based on his measurement that gives slower tempo (530 per eighth) than the final version (600 per eighth).
The draft of the Sixth String Quartet is currently located in PSS, shelfmark 79FSS1 (photocopy in the Budapest Bartók Archives).
The fair copy of the Fifth String Quartet is currently located in the Library of Congress, Music Division, Washington, DC (US-Wc), shelfmark ML 29c B29 (photocopy in the Budapest Bartók Archives).
Even though the draft lacks the indication of the time signature at the beginning of the Scherzo da capo (and 
occurs after a temporary change to 
in m. 11), it is very likely that Bartók intended the asymmetric 
meter as in the first Scherzo section.
Concerning this phenomenon, see Bartók's letter to Rudolf Kolisch, quoted below (case C-4).
It is, however, enigmatic that Bartók used a different time signature in mm. 43–46: 
instead of Even if this exceptional time signature is not an error but expresses Bartók's specific intention, the argument of the present paper may not be affected. Regardless of the indicated time signature, the music of mm. 41 and 48–51 (in ) is quite similar to that of mm. 43–46.
This hypothesis could be supported by the additional piece of evidence that the current measure 47 and part of measure 48 might have originally constituted a measure with 10 eighth notes if these measures were written continuously. In a section written in the so-called Bulgarian rhythm, Bartók never applied a time signature that contains a greater number of beats than the basic meter of the section.
When Bartók finalized page 23, it seems that a connection issue was caused by the shift of bar-line by three eighth notes. It appears that Bartók solved the problem with minimal effort, by lengthening three quarter notes in measures 52–53 to dotted quarter notes.
It can be observed in the facsimile reproduction that there might have been some problem concerning the cello part in the previous measures (mm. 46–47); however, this problem might not have affected the irregular order of composition. It is likely that Bartók first wrote the leading part (here, the first violin) and then filled the accompanimental parts somewhat later; indeed, there are no similar problems in the leading parts in the previous measures. In this case, Bartók might have simply made a mistake when he wrote the cello part.
The revision in the second violin part is irrelevant here, as it merely suggests that Bartók originally intended to write the first comes in the second violin part instead of the viola part.
The parts are paired by the instruction “col I” or “col vla” with a wavy line following it.
See, for instance, the autograph fair copy of Mikrokosmos no. 147 “March” on ordinary music paper (PSS, PB 59PS1, pp. 4–5). This autograph is written in the inner side of a bifolio, and while Bartók consistently used the left margin on page 4, he did not do so on page 5. It seems that he might have tried to write as much as possible on page 4 so that the inside pages of a bifolio should accommodate the entire piece. When he started writing on page 5, he certainly realized that there was enough space; consequently, he did not use the margin and notated the rest of piece in a slightly more spacious manner.
In fact, there is a preliminary sketch containing the elaboration of the motif in a four-part canon on page 59, staves 5–8. This sketch, however, differs in several aspects. The voices regularly enter from top to bottom, and the intervals between the voices are the following: a perfect fifth, a perfect fifth, and a major sixth. When Bartók drafted measures 43ff., he might have been able to rely on this sketch to some extent.
Concerning the paper type, see footnote 17.
For instance, see the case of the Music for Strings, Percussion, and Celesta in the introduction by Felix Meyer to the facsimile edition: MEYER (ed.), Musik für Saiteninstrumente, 44–45. The use of three different types of paper is graphically represented in a two-page-wide diagram.
For instance, the sketches related to the first movement can be found on pages 53–54 and 56. See also footnote 18.
It is, however, also possible that other bifolios of 18-lined paper had already formed gatherings, and a single folio of the 24-lined paper was the only available folio at Bartók's hand. For the verification of this hypothesis, it is necessary to reconstruct the exact paper structure by a close examination of the original manuscript at PSS.
If this is the case, the dynamics on page 34 were elaborated at a later stage of composition. Another possible (yet somewhat problematic) interpretation is that at least a part of page 34 was already prepared as a preliminary sketch with dynamics, and later incorporated into the main part of the continuity draft. In fact, other preliminary sketches also contain a canonic phrase that Bartók sometimes worked out in advance (see pp. 50, 59, 60, and 62). The fact that the top two parts on page 34 originally began with a whole rest supports this possibility (i.e., at least the beginning of p. 34 was written without having fixed the previous measures on p. 33). It is nevertheless problematic that the theme on page 34 is already written with flats, whereas Bartók seems to have notated this theme with sharps on page 59 (assumedly the first notation of the theme), and, moreover, on page 33 where he changed the notation enharmonically, with flats instead of sharps, by a verbal instruction “át b-be” [rewrite with flats].
A similar explanation has already been offered in relation to the previous case C-1 (see footnote 48), where the first violin might have been written throughout, and then the lower three parts were added later.
Concerning the possibility that Bartók might have had a different formal plan, see also the interpretation by Beach; BEACH, “Bartók's Fifth String Quartet,” I, 109–110 and II, 70. It is, however, problematic that she reconstructed a less convincing original version in which the upper two parts are fragmentary: for instance, the descending motif of the first violin (doubled by the second violin) in measure 62 is suddenly followed by four blank measures, without the concluding note of the motif.
In the list, the terms of the formal analysis are borrowed from Bartók's own analysis written at the request of Gaston Verhuyck-Coulon (impresario of the Pro Arte Quartet) for the December 13, 1935 performance in Marseilles. For the annotated modern English translation of the analysis, see SOMFAI and NÉMETH, String Quartets (1).
The special status of the fourth occurrence may require some explanation. In this case, the motif is preceded not by repeated notes but by an F/E dyad in half note, and the half note opens the new system. Here, however, different from other occurrences, the half note is an integral part of the motif, in order to better prepare a varied form of the motif that is in inversion and has an E upper pedal point as an additional note.
For the term “soft part,” see footnote 32.
Concerning the fifth occurrence of the motif on page 10 (where the location of the motif and the beginning of a system do not coincide), Bartók later inserted several repeated notes in the margin.
For instance, see pages 10–12 in the draft of the first movement of the Fifth String Quartet.
For some possible exceptions, see cases A-3 and C-2.
Concerning the formal plan, see Bartók's own analysis in SOMFAI and NÉMETH, String Quartets (1).
Bartók to Rudolf Kolisch, October 23, 1934, private collection, photocopy in Budapest Bartók Archives. It should be noted, however, that this remark was made in relation to the notational problem concerning the first movement of the Fifth String Quartet, where Bartók introduced dotted bar-lines within regular ones to unambiguously mark the metric accents that do not coincide with the actual bar-lines. In the first movement of the Fourth String Quartet, he usually used the marking for up- and down-bowing to mark figures in up- or downbeats.
In the reproduction, however, only the trace of scratching-out can be observed through the existence of something similar to white lines that break pre-printed staff lines.
This is the direct proof that Bartók stopped writing the draft at the first measure of the third system, even before drawing the bar-line at the end of the measure. It is not likely that he continued writing the draft somewhat further and only later made the measure coherent.
Cf. case C-1, where the leading first violin part might have been written throughout.
Essentially all the previous cases have a new beginning (that may also have an additional function as a point of arrival) after some blank space; the only exception is case A-1 that “continues” the Pelléas phrase.
For a detailed examination of the sketch, see SOMFAI, Béla Bartók: Composition, 155–158.
Here I refer to the analytical terms applied in Bartók's own analysis (see footnote 4).
For instance, it is remarkable that what seems to be the “proto-motif” of the movement is altogether missing until the very end of the sketch. See SOMFAI, Béla Bartók: Composition, 156–158. I borrowed the term “proto-motif” from János KÁRPÁTI, Bartók's Chamber Music (Stuyvesant, NY: Pendragon Press, 1994), 343.
In the final version, 10 semitones appear in four parts; nevertheless, they constitute a sound mass rather than a chord in the traditional sense.
While my interpretation might be more valid for the sketch version, the musical process of the final version is described better by Kárpáti, who recognizes the reconstruction process made of the struggle between “the shivering trills and the proto-motif” stepping into the foreground. See KÁRPÁTI, Bartók's Chamber Music, 347.
Based on the final version of the movement, the recapitulation indeed begins from this very pitch F (but an octave higher).
For an analytical interpretation of this passage, see also Elliott ANTOKOLETZ, The Music of Béla Bartók: A Study of Tonality and Progression in Twentieth-Century Music (Berkeley, CA: University of California Press, 1984), 231. It is striking that he recognized the octatonic collection in this passage (C–D–D♯–E♯–F♯–G♯–[A]–B). It is, however, a pity that he failed to include the last note, B♭, that actually falls outside of the supposed octatonic collection.
For instance, see my conference paper “
感と理性の狭間で ── バルトークの《ミクロコスモス》の作曲過程に対する一考察” [Between the Intuitional and Rational Compositional Approaches: On the Compositional Process of Béla Bartók's Mikrokosmos], a paper read at the annual conference of The Musicological Society of Japan on November 13, 2021. The paper deals with the case of Mikrokosmos no. 125 “Boating,” where Bartók had to revise a phrase repeatedly until he was finally able to free himself from the compositional system he had established.
In fact, he frequently used various kinds of musical shorthand or verbal remarks to indicate repeats and doubling of parts. Besides the usual simile symbols indicating repetition, Bartók frequently used a horizontal wavy line following “col [the name of an instrument]” meaning that the given part plays the same material as the named instrument. It can, however, be occasionally observed that he did not give any reference concerning how the missing notes should be complemented. One of the best examples is the sketch to the second movement of the Second Piano Concerto, where he wrote only the outer voices of the six-part string harmony consisting of perfect fifths. For the facsimile of the sketch, see SOMFAI, Béla Bartók: Composition, 67.
It has been discussed by Somfai that a theme in the previous composition inspired a new theme for a new composition; for instance, see SOMFAI, “‘Written Between the Desk and the Piano’”. Such relationship can also be found elsewhere, even among the short pieces written one after another; see NAKAHARA, “Genesis and the ‘Spirit’ of Bartók's Mikrokosmos,” 204–251.
It seems necessary to call attention to a specific problem concerning the methodology of the transcription of the manuscript. If a scholar tries to reconstruct the content of the manuscript in a transcription, it is necessary to consider whether the given section originally constituted a continuous passage. This is especially important when transcribing a continuity draft by Bartók. Because the content of the continuity draft is almost identical to the final version, the scholars' attention falls not on the transcription of the actual layer of the draft but on the reconstruction of the compositional process and how the composer developed his idea through (possibly) a lot of revisions. However, if scholars ignore the possible break between the sections, they may connect the phrases that had not been written in that order and might produce a version that never existed.
Bartók used the term “the spirit of the work” only once in his life, in one of the lectures held at Harvard University; see his “Harvard Lectures” in Béla Bartók Essays, 376. He neither gave a precise definition of it nor described it clearly; nevertheless, it is frequently possible to observe the existence of such a “spirit” affecting the structure of the work, even in the very first works such as the symphonic poem Kossuth (BB 31, 1903), Scherzo op. 2 (BB 35, 1904), and so on. See, for instance, László SOMFAI, “Invention, Form, Narrative in Béla Bartók's Music,” Studia Musicologica 44 (2003), 291–303.
