THE HISTORY OF THE BARTÓK-SYSTEM
(Shortened, and compiled version of the introductory study — The Evolution of the Bartók System —
by Sándor Kovács, published in: Béla Bartók, Hungarian Folk Songs. Complete Collection, Vol. I.
(edited by Sándor Kovács, Ferenc Sebő), Akadémiai Kiadó, Budapest 1993, 13–31.)
The adversities of Béla Bartók’s folk-music collections are well-known in professional circles. As regards his important, comprehensive works, a considerable part of the Romanian material was published in 1967 (Rumanian Folk Music), decades after his death; the tunes collected in Turkey were published in 1976 (Turkish Folk Music from Asia Minor), [and the Arabic collection was published in 2006 (Bartók and Arab Folk Music CD-ROM)]; and only two volumes of the three-volume collection of Slovak folk songs (Slovenské l’udové piesne/Slowakische Volkslieder I–II, 1959–1970) are available for research.
The fate of the Complete Collection of Hungarian Folk Songs is, however, the most peculiar. It is, without doubt, Bartók’s main ethnomusicological work. The creation of this collection was of prime importance for Bartók as ethnomusicologist. As early as 1913 Bartók and Kodály had submitted to the Kisfaludy Society their joint draft of this collection; Bartók was busy working out the concept of this collection even when (from 1918) “chances of collection... were practically nil”; finally, because in the early 1920s he had to content himself with publishing a relatively slim volume (of fundamental importance), A magyar népdal (Hungarian Folk Music, 1924), instead of the originally intended comprehensive work, due to lack of publication opportunities; he returned to the project in the fall of 1934, with an official commission to finish this enormous task.
It seems that in 1934, when Bartók was relieved of his professorship, and given, by ministerial decree, a position at the Hungarian Academy of Sciences, Bartók believed it was possible to finish whatever still had to be done with the collection in two years. But this soon proved illusory. In 1935 he estimated that the work would take “about three years”; in a report of September 14, 1937, he wrote about another three years, and in 1939, in a letter addressed to Sándor Veress, he suggested that the work might take “at least one year”.
It is easy to see why certain phases of the work lagged behind the original schedule. Although Bartók himself had stopped field work in 1918, Zoltán Kodály, László Lajtha and others carried on even in the twenties; as a matter of fact, collection began to flourish towards the end of the twenties (when Vilmos Seemayer started to work, and when, in 1929–1930, Pál Péter Domokos and Sándor Veress went to Moldavia), and continued to do so until the mid-thirties. This meant a continuous flow of new material, which hindered classification. Instead of an open catalogue, Bartók had conceived a system of a closed logical structure (which he was able to base on his own previous system used in A magyar népdal, and to develop it further). He could not even think of completing this closed system prior to setting a limit on the number of tunes to be included. Such a step at the time of a rapid accumulation of material would have incurred severe consequences; specifically, a large number of valuable tunes would have been left out of the collection intended to be complete. Quite understandably, Bartók postponed a decision as long as he could; besides certain preparatory classification, he busied himself with revising his earlier phonograph transcriptions (and transcribing a number of new phonograph recordings). In view of the fact that Bartók “closed” the collection in 1938, the decisive phase of the two-years’ work planned in 1934 actually began only after 1938, with nothing added afterwards.
However, at this time other problems arose to hinder the rapid progress of the work. New findings and other reasons (to be discussed later) moved Bartók to reconsider his views over the years. Although he held on to the basic principles outlined in A magyar népdal he no longer regarded his previous solutions to a number of questions as expedient. Finally, he decided to reform the system profoundly.
As a result, on October 10, 1940, two days before he left for the United States, Bartók turned over the Complete Collection to Zoltán Kodály in the presence of his colleagues, Ilona Rácz, György Kerényi, and Sándor Veress, as a work both finished and unfinished. It was finished in the sense that Bartók succeeded in his main goal of presenting a system of classification of tunes. It was unfinished in the sense that he had not prepared an introductory essay. (In all probability, he did intend to write a comprehensive review of the stylistic layers of Hungarian folk music, including an explanation of the principles of classification, as well as a comparative study of the Hungarian material with that of the neighbouring nations together with his conclusions about the comparison.) In addition, he left no instructions concerning a number of problems of editing.
Thus we have come to the most obscure phase in the history of the Complete Collection, which will be perhaps the most difficult to clarify as well. It is an indubitable fact that, theoretically, nothing ever stood in the way of publication; the considerable, but mostly mechanical work left undone was such that, given normal circumstances, it should not have posed any problems for a research team led by Kodály. However, neither Bartók nor Kodály had any illusions concerning contemporary world events. Work became practically impossible after 1940, and Kodály could do no more than to find a safe place for the material for the duration of the war.
Work on the project was not picked up for several years after the end of the war. Research into Hungarian folk music, under Kodály’s guidance, was soon resumed, but scholarly publication was out of the question during the years of consolidation. Although in the autumn of 1946 the Ministry of Religion and Public Education (which later became the Ministry of Public Education), had set a modest amount aside for financing the engraving of music (initially it paid for the work of only one engraver). “Real work could begin only in the fall of 1949”, wrote Kodály. Until January 1950 the Hungarian Academy of Sciences was unable to provide even temporary quarters for the research team.
Nevertheless, certain arguments arose against the publication of the Bartók-collection around 1949-50 and apparently these were not without foundation from the professional viewpoint. With the passing of time the work could no longer be considered as complete. The fact is that between 1938 and the outbreak of World War II there were still occasional collections, data of which were not, indeed could not be, included in the Complete Collection owing to the nature of the Bartók system. This would not have caused a problem in itself, since, given the scope of the work, small deficiencies were natural. However, it was soon evident that these deficiencies were not at all insignificant. For instance, some of the phonograph cylinders in the Museum of Ethnography had not been transcribed and Kodály also had in his possession such untranscribed cylinders. Furthermore, there were transcriptions that, for some reason, Bartók had never seen, and Kodály’s notebooks also contained unprocessed material. These subsequent discoveries and the data of new collections (which flourished again in the second half of the forties) increased the amount of available material considerably. In 1950 35,000 tunes were on file, which was two and a half times the amount that Bartók had classified, and extensive collections were expected in the years to come. Consequently, the Bartók collection ceased to fulfil its original purpose, namely, totality that Bartók himself had worked to accomplish.
When Kodály (and his research team) decided to put aside the Bartók collection, leaving it to posterity as a historical document, and to work out a new concept of publication, to start a new series, he did not only advocate Bartók’s original ideas, i.e., the creation of a comprehensive, representative encyclopaedia of melodies, but he was also greatly influenced by scientific policy and political considerations. At the end of the forties and in the early fifties Hungarian musicology found itself in a delicate situation. Cultural policy-makers ascribed tremendous importance to the collection and dissemination of folk songs; with growing impatience they pressed for the publication of the collected material. But scholars were inhibited by attacks in all spheres of cultural and scientific life. Since the targets of the “fight” against so-called “formalism” were also Bartók’s original compositions, publication of his scholarly work was deemed undesirable for tactical reasons as well. (It was beyond doubt that leading “ideologists” would condemn Bartók’s system of classification – were it to become widely known – as an exemplar of formalism, with unpredictable consequences for the future of Hungarian ethnomusicology.)
Ethnomusicologists reacted to this pressure partly by assuming a defensive position, talking about the demands of scholarship and the magnitude of tasks and problems. Partly they reacted by considering and changing the publication plan, which meant that instead of the main body of strophic songs, children’s songs were first published (some sort of an independent publication, even given the complete Bartók collection, would have been necessary since Bartók’s system included children’s songs only as exceptions). This was followed by volumes arranged by folk customs, and not by principles of a musical system.
Under the given circumstances this “bridging” solution proved fortunate, because Hungarian ethnomusicology thus avoided the greatest danger, namely, undermining its international status and reputation by publishing useless material. Furthermore, the Academy was able to publish rather bulky material in a relatively short time, without having to state the final word concerning musical classification. The compilers of the Corpus Musicae Popularis Hungaricae [CMPH] (Magyar Népzene Tára) launched in 1951, including Kodály who directed the work till his death, considered it as the successor to the Bartók collection, and to give weight to the fact of legal succession, Bartók was also listed with Kodály as an editor in the first six volumes.
In addition, Kodály had a further reason for not pressing the rapid publication of the Bartók collection, which we must discuss briefly, delicate as the subject is.
Kodály disagreed with Bartók’s principles of classification, which he had once or twice hinted at. It is noteworthy that in praising Bartók the ethnomusicologist, he wrote in great detail about the collecting trips and about Bartók’s transcriptions (which “represent the limit of what the human ear is capable of without technical instruments”), but made no mention at all of Bartók’s work of classification. In his conversations with Lutz Besch he spoke with unusual tact about the antecedents of CMPH saying that “previously they worked with rhythm types”, without referring to Bartók in any way.
More of this attitude came to light in the early eighties. Benjamin Suchoff published the greater part of the letter that Kodály had written in 1941 and given to Péter Bartók in Budapest to deliver to his father in the United States. To understand the letter we must know that in doing the analysis of the collected material of Hungarian folk songs, Bartók and Kodály had worked together assisting one another, but in a sense they had also worked in parallel, making several copies of each record card (the so-called támlap, each of which contains the notation of one folk song). Thus, as regards their system, two very different versions of the Complete Collection were created: the above described Bartók-system and the Kodály-system, which is a catalogue of melodies based on the principle of cadence. In the thirties Kodály, as co-editor of the Complete Collection, had assumed the task of reviewing and copying the material of old, historically valuable publications and manuscripts of folk song collections, and he let Bartók have the decisive word on the question of classification, thus acknowledging in effect Bartók’s priority. At the same time, as co-editor he had every right to expect Bartók to at least consult him before making the final decision, but he never did. Kodály faced a fait accompli when he received the collection in October 1940. Consequently, he had to spend long weeks in order to become thoroughly acquainted with Bartók’s brand new system. In doing so he wrote a number of critical comments on the explanatory tables of the system. Kodály let Bartók know his more important reservations in the letter mentioned above, which he composed in English because of the war, so that American authorities could easily make sure its content was harmless:
After you le[f]t us [...] I asked my[self], what restrained you to [=from] tell[ing] me a sincere word in due time, about your projects. Sorry enough you becam[e] more and more mi[=y]sterious, which caused many mistakes. So we did not fulfill we engaged us: instead of one, but good Collection we made two, both incomplete and unready. I was not very busy in [=on] my part, true, but I had not the impression, you wish[ed] as early as possible [to] return to teach. [...]
But also the material to inspect proved to be tremendously greater as [=than] we ever meant. No copyist is able to decide, what to copy, I must see all. And you “closed” and “numbered” your collection, not paying attention to this part which must be printed, containing important and unic [=unique] dates; or at least, must be related of. If ‘you had the intention to make the whole thing alone, why not tell it to me [in] 1934? I had plenty of other work. [...]
You did a great work in ritmics [= rhythmics]. But your system A B C in the new form, although more logical, is hardly more able as a frame of edition. The division is made from heterogeneous points of views: style, metrical, formal. It is too complicated and difficult. I found a song, 3 variants of which you put respectively under 3 different classes A, B, C. Your Roumanien order is fare [=far] more logical and simple.
The work advanced slowly. I worked through all avalaible [=available] (sic.) manuscripts (detected some new). The printed material is on the half [=by looking at half of the publications of historical value]. Not ready my phonos. The gramaphon [=gramophone] reached to 200 bearing meny [=much] important material. Veres[s] is not yet ready with the Polish cataloguy [=catalogue], you committed him. He did nothing other.
I wonder how you think to use this foreign material? I found any [=some] references in your copies to slovakian [=Slovakian], roumanien [=Romanian] or serb [=Serbian] variants. The minimum should be to remark if there are any. What became [of] your Slovak[ian] collection? – I found no synoptic table of your new diacritical signs etc. You see your work is not finished here, and you must use [=find] occasion to let me knew [=know] your ideas thereabout. I could accept a metrical division, although the many secret melodical affinities come to sight only in a melodical system. Sorry, you did not spend an hour or two, to verify this. Metrics, as a common feature to words and music, can very well stay in first line, especially today, as connection of both factors is often neglected (Vaisenän [= Väisänen], Kannisto).
But most important is, that you feel well and content, (if ever possible) and we hope, one of the chief results of your trip will be the writing of more “Divertimento” and other works to be performed “more then [=than] once”.
Bartók had never replied nor was it possible since postal services between the United States and Hungary were suspended for several years at the turn of 1941-42. Thus the argument between the two of them remained unresolved, and this had a profound influence on Kodály’s later behaviour (as well as on the course of Hungarian ethnomusicology).
As regards musical classification, Kodály never assumed the responsibility of making a decision; he expected younger scholars to provide the final system. In the end this was performed by Pál Járdányi, and his system was adopted in volume six of the CMPH, forming to this day the basis of the scholarly compilation and publication of Hungarian folk songs.
It is no longer necessary to prove the gains that the launching of the CMPH and the publication of successive volumes meant for Hungarian ethnomusicology and for Hungarian musical culture, in general. Without doubt, Bartók’s fundamental idea can be realized only with this type of truly complete series. It is a peculiar (and perhaps it is no exaggeration to say, tragic) paradox of Hungarian folk-music research that for the sake of progress, of making the great step forward Bartók had dreamt of, Béla Bartók’s major scholarly work had to be discarded. In this way one of the most interesting, most original systems of melody classification of 20th-century ethnomusicological scholarship became inaccessible, whereas it is a system that reveals the most, more than any document, about the scientific – and, indirectly, artistic – views and thinking of one of the greatest composers of the 20th century.
It is more revealing than anything else, because (as mentioned above) it is the product of several decades of work. There is no other work (including his compositions) that took Bartók as much time to create and finish; there is no other work in which the creative process (even if it is not clear in every detail) is so easy to follow, often step-by-step, and through so many phases.
The first important documents that contain the already concrete concepts of classification are dated from before the outbreak of World War I. These are, besides the brief introduction to Magyar Népdalok (Hungarian Folk Songs) published in 1906, Bartók’s “Az összehasonlító zenefolklore” (Comparative Music Folklore), an article written in 1912, and the already mentioned draft prepared for the Kisfaludy Society. True enough, the latter remained just a plan, since no reply was received from the Society; henceforth both Kodály and Bartók considered collection as their main task. But the situation changed fundamentally at the end of the war. The collapse of the Austro-Hungarian Monarchy, the partitioning of the territory of the Hungarian Kingdom (sanctioned later by the Treaty of Trianon) had made collection impossible precisely where the most valuable, most ancient Hungarian material was expected to be found. Under these conditions Bartók had to consider scholarly elaboration, i.e., systematization as the primary task of research, which the amount of material collected up to then justified, in fact, demanded.
The first study which must be regarded as expressing the basic features of the Bartók system, was Die Melodien der madjarischen Soldatenlieder published in 1918 (Die Melodien der madjarischen Soldatenlieder). The article is noteworthy for two reasons. First, because Bartók makes a firm distinction between tunes of the “old style” and the “new style”; and second, because in each basic section he first analyses rhythm. If we study the antecedents, the setting up of the basic sections and the stress laid on the examination of rhythmic factors both seem to have been significant steps. The preface written in 1906, and the article written in 1912, both still speak about the Krohn system, suggesting the idea of a melody collection based on the principle of cadence. This very same concept was elaborated by Kodály in the draft of the project he co-signed with Bartók. It seems, however, that the mechanical, vocabulary-like order no longer satisfied Bartók once he became deeply involved with questions of classification. He wanted to develop a method of classification which would, to an extent, also reflect the “evolutionary history” of styles. In the case of the general division, the proven “newness” of the “new style”, which still flourished around the turn of the century, and the “ancientness” of pentatonic-descending melodies, with much evidence supporting it, provided the basis for this method. In this respect recent views are more refined and supplemented but not modified decisively by Hungarian ethnomusicology. It is no accident that in doing further research on “historical” environment Bartók’s attention turned towards rhythm. It is probable that Kodály’s well-known study of the strophic structure of Hungarian folk songs was the strongest motivating force in this work. In his study, Kodály discusses, among other things, the changes in “evolutionary” tendencies of certain verse types and rhythms that may be “derived from” simple formulas; some of his reasoning was taken over by Bartók almost verbatim in A magyar népdal (Hungarian Folk Music).
It is interesting to note that Kodály himself did not draw any conclusion from the theses conceived in 1905 concerning classification. He aimed primarily at putting a “catalogue” together in which the various folk-music data could be listed according to a “mechanical” system of arrangement and which could easily be traced back by anyone with some experience in folklore; he did not regard the presentation of historical-evolutionary aspects as the main task of the classification.
Essentially, ever since Kodály became acquainted with the Krohn system he remained faithful to its basic principles or rather to its modified version, which had been applied to the Hungarian material. Bartók started to move gradually away from Kodály’s ideas of classification largely because he followed the direction outlined in Kodály’s study and continued thinking along that line.
It would be a mistake, however, to interpret the separation of the “old” and “new” style as a sign of a break between them. The difference between these two styles can be well illustrated on the basis of the cadences as well (the new and the old melodies would almost automatically separate in a melody collection based on the principle of cadences).
But the cadence-system is in no way suitable to reliably separate those melodies from the old and the new melody groups, which – for various reasons – do not belong to either one. So when Bartók created an independent section for the latter (according to our present knowledge, first in an essay intended for Musikblätter des Anbruch), he was already on the way to breaking with the “amended Krohn system”.
Actually, the integration of this class into the “historical” system was not at all easy, and Bartók had great difficulties even finding a name for it. At first he used lengthy and complicated descriptions to characterize this group of “most heterogeneous” melodies, mostly of foreign – Western European – origin, apart from those on which “the imprint of Eastern Europe is decidedly perceptible”.
Finally, while doing preparatory work on A magyar népdal he found a phrase, “miscellaneous class” which has been adopted ever since by Hungarian folk-music research. It must have caused him a great deal of stress to decide the sequence of the sections. First he put the “heterogeneous” melodies in third place. This, however, turned out to be in contradiction with the “historical” concept. As far as the “chronologial sequence” was concerned, the “new style” had to be the last in line and in this order Bartók recognized, and rightly, a 19th century phenomenon (in the light of contemporary research it seems that he had been too cautious when putting the birth of dome-structured melodies around the 1840s and 1850s). At the same time he created a peculiar theory about how the melodies of foreign origins got into the mainstream of Hungarian folk music. Although with a characteristic cautiousness he was eager to point out that these melodies dated back “to widely different ages”, at the same time he regarded the majority of them as stemming from the Slovak language-region as well as coming through “Czechoslovakian transmission” from the West to Hungary during the 18th century, when, according to Bartók, Hungarian culture came under strong foreign influences.
At the same time the question whether it was worthwhile to pursue the “historical” aspects at all levels of classification, was well justified. The question also arose whether it would be more useful – as a first step – to classify the melodies according to their Hungarian or foreign origin. How much Bartók had struggled with this dilemma is clearly visible from his subsequent essays. In the German preliminary version of the article (at the beginning of the 1920s) intended for the Musical Courier [The Peasant Music of Hungary, 1931] for instance, the basic classification originally followed the one used in the Anbruch article. Bartók, however, crossed out the word “Neue Melodien” which stood in second place and wrote instead: “Melodien entstanden infolge westeuro- peischen [sic!] Einfluß [sic!]…” and he kept this sequence for a short while as can be seen in the article written for the Revue Musicale [La Musique populaire Hongorise, 1921] at around the same time as the one above, and published in 1933 in Musical Quarterly [Hungarian Peasant Music]. In the end Bartók decided to revert to the previous version.
Looking at the manuscript of the longer essay mentioned above written in German with an English title we find exactly the opposite to what we have seen in the article intended for the Musical Courier: the original essay, completed towards the end of 1920, was supposed to finish with the introduction of the new style, but (sometime during the spring and summer of 1921) in the course of alterations and additions, Bartók decided to use the sequence “old style – new style – melodies of foreign origin” (the expression “miscellaneous” was not yet used here). Undoubtedly, this solution had the advantage that at least as far as the Hungarian material was concerned, the “historical” aspects could be more clearly enforced. The succession of old and new melodies without the confusing insertion of “heterogeneous melodies” informed the reader about two especially significant phases in the evolution of the Hungarian material. As far as the “old” class or section was concerned Bartók tried to create further “historical”, or more precisely, “evolutionistic” sequences. It is an indisputable fact that Bartók, like so many of his contemporaries, tried to explain history on the basis of the theory of evolution; he attempted to fill the gap between historically proven (or seemingly proven) facts with an evolutionistic logical construction. Proof of the evolutionistic character of his thinking may be noted in the way he dealt with the classification of old melodies constructed of six-syllable lines. While establishing the subsections of the old style Bartók’s reasoning followed from the assumption that the “ancient layer” of strophic Hungarian folk songs is mostly represented by parlando-rubato melodies with lines of eight and twelve syllables. He believed that all the other types originated from the above – especially the ones with eight syllables – with one exception, namely, those with six syllables. (Later though, in his essay published in Musical Quarterly, he noted that those melodies with lines of six syllables were “not so old”, but he could not produce any proof for this statement.) This “subclass”, therefore, did not at all fit in the “evolution chain”, and so Bartók was faced with the dilemma of sequences once again. In an uncompleted Hungarian outline for an essay full of insertions and corrections entitled Hungarian Folk-Song Types with Respect to their Process of Evolution [Bartók Archives Budapest (BH 5128)], immediately after the melodies with lines of eight and twelve syllables, he listed those of eleven, ten and seven syllables. When writing the German-language article with the English title, however, he changed their order, i.e., after the subclasses of melodies built on twelve- and eight-syllable lines he put those with six syllables thus hinting at the fact that he assumed the latter – as opposed to all other syllable types – to belong to the ancient stratum of Hungarian folk music, but that it is not quite equal with the basic twelve- and eight- syllable types. Furthermore, he made another alteration in the sequence: he brought the seven-syllable melodies forward as the immediate descendants of the eight-syllable ones and put the eleven-syllable ones later, as formation deriving from the eight-syllable melodies via the seven-syllable ones. The sequence closed by the ten- and nine- syllable melodies. In A magyar népdal all these sections and subclasses were marked with letters of the alphabet and with Roman numerals. The structure of the complete system is as follows:
There is no doubt that considering the information available at the time, a great many arguments could be put forward in favour of this system. Nevertheless, there remained problems, which Bartók did not manage to solve at the time. The most important among these was the problem of the so-called adjustable tempo giusto melodies. Bartók, who – following in Kodály’s footsteps – quite clearly saw a difference even in evolution between parlando-rubato and tempo giusto rhythms, thought to have discovered a kind of style in the adjustable tempo giusto melodies, which anticipated the new style and which acted as a bridge between the new and the old. This sequence of evolution, however, could be proven basically only in the subdivisions of Division A (and even there only as a version of further classification based on cadence and rhythm-formulas) which was not at all a good solution for two reasons: partly, because contrary to the principles of classification of Division A, Bartók created two independent subdivisions for those adjustable tempo giusto melodies which had been put under C because of certain stylistic characteristics (this has been perhaps the most obvious inconsistency in his system); and, partly, because this way already in the very first A-subdivision the “most ancient” parlando-rubato and the most definitely “late” adjustable tempo giusto-material was juxtaposed.
In order to ease the latter contradiction somewhat, Bartók described his method in a summary about the old style, as a kind of postscript. The entire process of evolution, meaning the assumed historical sequence which his classification system had been able to suggest was accomplished only by using two different threads which cross one another only casually (see sequence of the subclasses and the description of the characteristics of melody types belonging to this or the other subdivision). He wrote, “the oldest of these [melodies] are the twelve-, eight-, and six-syllable parlando-rubato tunes. More recent are the eight-, six-, and seven-syllable tunes in invariable tempo giusto rhythm. Comparatively recent are the eight-, seven-, and eleven-syllable tunes in variable tempo giusto rhythm, and all the ten- or nine-syllable tunes”.
It is not clear whether Bartók carried on work on the Hungarian material immediately after the completion and publication of A magyar népdal. It is a fact that he gave a reference number to all melodies (record cards) collected until 1919; in other words, he put all the available material in numerical order. It is fairly certain that these reference numbers could not have been assigned before the autumn of 1920 (that is before the completion of the German-language article), and there are indications that not before October 1921, either. It is probable, therefore, that the provision with reference numbers may be regarded as the final phase in the classification work during the twenties, which included also – even if only partly – the revision of some aspects of A magyar népdal.
Whichever hypothesis we are willing to accept as regards the origin of the reference numbers, it seems most probable that Bartók did no serious work on the classification of the Hungarian folk song material during the second half of the twenties and the beginning of the thirties.
The case of the publication of the Complete Collection didn’t get along until 1933 when the newly formed Subcommittee for Folklore of the Ethnographic Committee of the Hungarian Academy of Sciences submitted a proposal for the publication of the Collection and the Board of Directors passed a long overdue resolution. The so-called “horseshoe” room was placed at Bartók’s disposal by the Academy and after some bureaucratic procrastination he could start his officially sanctioned work, which he originally planned for two years and which took him six years in all.
He was relieved of his teaching duties on August 28, 1934, only a few days before the start of the new academic year at the Academy of Music on September 14; he transported all the materials – notes, cards, books, recordings – from his home to the Academy and on the 18th he started working regularly, usually three times a week.
In one sense we know a great deal, and in another, hardly anything at all about how he spent his working days and how the work progressed. For instance there is a copy of a diary (kept by his colleague György Kerényi), which shows the days when Bartók spent the afternoon (usually from 2 to 7 pm, sometimes to 8 pm) in the horseshoe room. There are a number of memoirs describing the room, its furnishings, some of Bartók’s habits and events of greater and lesser importance. A number of interviews published around that time give the reader information about the environment and the circumstances (as well as details on the project itself, etc.). With the help of letters and reports written by Bartók we can make a more or less adequate picture of the work in progress.’
Such articles and essays, however, which Bartók had written before completing A magyar népdal and which touched also on problems of classification, no longer accompanied his work. The story of the development of the “system” can, therefore, be reconstructed mainly with the help of the notes written on the record cards.
It seems likely that Bartók thought of making certain changes quite soon, perhaps as early as the autumn of 1934. The newly collected melodies were classified, nevertheless, according to the old system for some time (this is obvious from the notes attached to the old reference system); but he must have regarded this classification as a temporary one: the notes to be found on many of the record cards reveal pondering (for instance, “is it CI or pentatonic?”) and small corrections (for instance, signs referring to the breaking up of certain groups of variants), signs like “apart” next to the reference number and letter of the melody in question, etc.).
There is also a continuous line of references with the help of which one can obtain a perfectly correct picture about the state of perhaps not the whole system, but at least a significant part of it, immediately prior to the final revision. It is made up of the so-called “P numbers”. According to Ilona Rácz, Bartók wanted the P numbers (together with the “Kol” numbers) marked on the record cards of the pentatonic melodies (and those with Kolomeika-rhythms) immediately before the rearrangement, so that in case the new system turned out to be useless the cards could be filed back again relatively quickly.” But this possible reversal would have no longer meant the re-establishment of the original system of A magyar népdal. The P numbers progress according to the increasing number of syllables (namely, P 1 – 79: 6 syllables. P 80 – I 16: 7 syllables, etc.) and not according to the subclass sequence of class A in A magyar népdal. Bartók, therefore, abandoned this complicated system long before he made a decision on a final and radical reorganization, which would have also involved a revision at the highest level of classification.
It is almost impossible to determine when this final, radical revision actually took place. Ilona Rácz talks about 1938 but that seems too early. Bartók seems to have made the decision most probably during 1940 (perhaps at the end of 1939, the beginning of 1940).
The essence of his revision can be summed up as follows:
These rhythmic formulas constitute the framework of the Bartók system. The catalogue material complete with reference numbers conforms to this framework. There may be small inconsistencies, gaps and some confusion in the numbering, but the fact that the record cards follow an order basically corresponding to the tabulation of the rhythmic formulas, is a decisive argument for regarding the Complete Collection as a scientific opus, completed and in no need of major revision, and, therefore, ready for publication.