Modality in the Hymns of the Armenian Church

Transcription

1 Modality in the Hymns of the Armenian Church HAIG UTIDJIAN Socharska 5, Praha 7 CZECH REPUBLIC hutidjian@yahoo.com mujweb.cz/www/haig/hu.html Abstract: The paper discusses the meaning and function of modality in Armenian church music, its relation to issues of neumatic notation, oral reconstruction and tonal ambiguity; and the manner in which the latter has influenced nineteenth-century harmonisations, some of which evince a harmonic language anticipating similar effects deployed by Western composers. Matters of performance practice are touched upon, and a critique of efforts to decipher neumatic notation has been provided. Key-Words: Armenian early orthodox chant, Armenian modal systems, performance practice, semiimprovisation, oral transmission and reconstruction, neumatic systems of notation, harmonization, tonal ambiguity, Komitas and Britten. 1 Introduction Subdivision of music into a system of modes goes back to classical antiquity, and in the case of Armenian music too, there are grounds for supposing that pagan tradition already acknowledged four modes or voices, associated with worship of the four elements [1]: Mode 1: earth, Mode 2: water, Mode 3: air, Mode 4: fire.there are various traditions regarding the origins of the modes, but to the modern reader they seem fanciful rather than factual, and do not shed light on musical aspects. For an interesting fourteenth-century exposition, the reader is referred to the Book of Questions of St. Gregory of Tathev ([2], p. 639 ff). The system of modes still in use in Armenian church music entails the subdivision of modes into authentic and plagal, with the result that there are, in effect, eight modes in all (but with a proviso to be stated later, entailing modes possessing variants, which would increase the above count): four voices or authentic modes, and four sides or plagal modes even though in current practice there is no musical significance whatsoever associated with the terms plagal and authentic, and the continued usage of this nomenclature is merely the result of custom. This system appears to have been instituted by SS. Mesrop and Sahak in the early fourth century A.D., and set down in writing shortly after the invention of the Armenian alphabet. 2 The evolution of modal hymns and neumatic notational systems There is compelling evidence to suggest that the current modes are not musically identical to their ancient forms. This is immediately evident [3] upon comparing the oldest items in the hymnal known as the Kanonagloukh (psalm-based liturgical songs of blessing) with more recent melodies of the canonical hymnal to which hymns ceased to be added only at the end of the fourteenth century (with minor exceptions). Prior to engaging in a more detailed discussion of the eight modes, it is appropriate also to consider the history of the hymns themselves [4]. The hymns were composed from the fifth century onwards. It is believed that they were notated neumatically some time afterwards, although as the earliest extant neumatically-notated manuscript fragment dates only to the ninth century [5] and the oldest such hymnbook to the twelfth [6], it is not known exactly when they were first so notated. They must thus have been transferred orally to subsequent generations until such time as they came to be notated. Then, from the seventeenth century or so, the art of reading neumes gradually went into abeyance, and hymns could, once again, be transmitted only orally, with the neumes providing only limited hints here and there (as will be discussed below). This fact, coupled with the political and geographical circumstances of the Armenian nation, rendered the modal melodies particularly vulnerable to distortion due to inexactitude of recollection, non-indigenous influences, or even utter oblivion. In an attempt to prevent a complete catastrophe, attempts were made to give fixity to the melodies by transcribing them into what may roughly be described as an Armenian precursor to the tonic sol-fa notation, devised at the turn of the nineteenth century by Hambartsoum Limonjean (see, for instance, [7]), though the old neumes continued to be respectfully printed in hymn books, for their limited but nonetheless practical value (see, for example, [8]), and continue to be even now. The system of modality is intimately related to contemporaneous systems of neumatic notation. First, there is evidence to suggest that, at least in the case of canonical hymns, the notation itself under-determined the melody to be sung. The melody could be uniquely fixed, at best if at all only when accompanied by two ISSN: ISBN:

2 crucial pieces of information: the mode, and the tempo (implying as it did the degree of syllabicism or melismaticism of the appropriate realisation of the given hymn and thus very much more than mere speed of execution). Although the meaning of the neumes is now largely unknown, Bzhshkean [9] and Tntesean [3] were able to establish the time durations associated with syllables bearing various neumes (albeit without information as to internal rhythmic content). Thus, even at present it is not unusual to encounter feats of semiimprovisation whereby hymns are sung convincingly at the requisite tempo, wholly on the strength of one s familiarity with the mode and, to a limited degree (restricted to information as regards the duration of a given syllable), aided by the neumes. For, in Armenian church music, a mode does much more than merely establish the scale on the basis of which a melody belonging to that mode may be formed. The mode immediately suggests certain melodic and rhythmic formulaic motifs, including (but not confined to) standard beginnings and endings. The choice of motif is governed by the number of syllables of text to be covered, the manner in which the verbal text is subdivided (punctuated), its natural accentuation, the presence of any crucial words warranting special emphasis, and so on although, in our experience, most skilled singers do not seem to be explicitly conscious of the processes governing these choices in the course of live performance. Prior to (if one may so designate the process) semiimprovising an appropriate melody, a singer, in addition to knowing what mode a hymn is in, needs also to know as to whether it is meant to be sung in the parent mode or its variant (in a sense to be made precise later) a matter that is often itself evident from the neume groupings over the verbal underlay, as will be discussed later. The situation is sometimes complicated by the fact that in special cases (e.g. the IIIA Magnificat sung on the evening of Good Friday and on Saturday morning) the same hymn might equally be sung in the basic mode or indeed in the associated variant mode. We have already mentioned that the singer also needs to know the tempo at which the hymn is to be sung that is, the degree to which the hymn is to be syllabic or melismatic. Contrary to what one might expect (after all, there are many manuscript sources where the existence of dense neumes over a single syllable is highly suggestive), for most hymns in the canonical hymnal, the same hymn may be sung in differing degrees of melismaticity on different occasions, yet in a manner that is always consistent with the relative time durations associated with different syllables of text as implied by the neumes. Typically, a fairly melismatic hymn is performed syllabically when used as a Midday hymn, and perhaps yet more melismatically whilst sung as a processional hymn. On palm Sunday, the morning Miserere is performed in the evening in a more melismatic manner and so on. But the relative durations of syllables of text remain the same, irrespective of the degree of melismaticity. A particularly intriguing analogy from Western music may be afforded by the splendid arioso Possente spirito in Monteverdi s Orfeo, where the composer provided two versions of the vocal part one consisting mostly of notes of longer duration, and the other of a highly embellished and melismatic version of the same. But the two versions are superimposed over the same basso continuo, which remains constant as we proceed from the plain version to the ornamented. In short, even at present, a modally-aware singer finds that a melody almost suggests itself in real time, the process being helped and partly determined by the neumes over the verbal underlay as well. (Sadly, however, this quasi-improvisational faculty is gradually disappearing, given the current practice of singing hymns from fully worked-out realisations published in Limonjean s or modern Western notation; or that of eschewing the hymns or, even worse, the various divine services in which they are to be sung altogether. Nonetheless, this does not affect the arguments and observations presented in this paper.) There is an interesting consequence to this. We strongly suspect that a high proportion of the less wellknown, and infrequently-sung, items in the hymnal, must have been recorded (using Limonjean s notation) in the nineteenth century through such a process of reconstruction, rather than precise recollection of the exact melodies of those hymns. This is particularly plausible when we consider the huge number of hymns in the canonical hymnal around one thousand. Thus, the power of the modes sufficed to allow a limited recomposition of some of the hymns, and hence their subsequent transmission. Incidentally, in our opinion this fact greatly complicates, and even partly invalidates, attempts to decipher neumatic notation through a mere direct comparison with melodies transcribed in the nineteenth century first, the process does no more than merely confirm the very assumptions (most notably, regarding temporal duration associated with the neumes) made at the time when the melodies were virtually re-composed for transcription, and second, the melodies that were so devised need not have been consistent with the neumes in any other respect at all. Another source of difficulty arises from the sometimes shockingly diverse and highly confused melodic traditions between the various Armenian schools (Constantinople, Jerusalem, New Julfa, etc.): there is no discernible, one-to-one mapping between these and the differently notated manuscripts (the two principal schools of neumatic transcription being those of Cilicia ISSN: ISBN:

3 and of Eastern Armenia), or at least none has hitherto been discerned or identified. Presumably a hymn was indeed sung differently in Cilicia than in Eastern Armenia, if it was differently notated. But there are reasons to believe that what was sung in the Cilician metropolis of Adana prior to the 1909 catastrophic pogroms there, and what was sung in Etchmiadsin, need not correspond to these respective traditions. The destruction or loss of a high proportion of original sources during the Armenian genocide, as well as the decimation of the oral traditions, is also debilitating. And, perhaps most fundamentally, the existence of several historic systems of neumatic notation is a source of great confusion: neumes initially served merely to help intone biblical chants, being little more than fairly primitive prosody signs; later, to notate modal hymns this being the stage of particular interest to our discussion; and, finally, a more sophisticated system, known as manrousoumn, came into existence in a bid to record very complex and highly melismatic odes, that went beyond the modal system and fell outside the canonic hymnal. It is thus possible that this whole multilayered system came to collapse under its own weight, particularly when it strove to transcend the bounds of the modal system by which it had been underpinned. Yet, as already menioned, a hymn notated in the hymnal with fairly sparse neumes is rendered variously in settings that are syllabic or melismatic.the system of neumatic notation of relevance to canonic hymns is the second (chronologically) of the above three, though even it appears to be multi-layered [6]. Many scholars have attempted to work out the significance of particular neumes in particular modes, with a view to an eventual decipherment and definitive reconstruction of the melodes of hymns; but only very minor successes have been achieved indeed, one might say that there has been little progress since the nineteenth-century work due to Bzhshkean on the temporal durations associated with the various neumes and of Tntesean, who was in addition able to spot characteristic agglomerations and combinations of neumes associated with particular modes, as well as establishing, on the basis of neume combinations used, whether a hymn to be sung in the parent mode or in its variant (in a sense shortly to be made precise). Incidentally, such findings were quite consistent with the melodies sung at the time, thereby improving their credibility. Komitas is reputed to have made progress on neumes, but the destruction of his archives by the Turks, and his loss of mental balance during the Armenian genocide, mean that it is difficult to establish what precisely he was able to achieve in this field. An overview is provided by Atayian [5]. More recently, Tahmizean [6] has done much to disentangle various layers of neumatic notation, of different periods and levels of complexity. But the objective of reading the ancient neumes remained elusive. There are several reasons for this. We have already alluded to some fundamental causes of difficulty perhaps insuperable in themselves. But we have further reservations also. Though Atayian and Tahmizean have undeniably done much useful background work, they have confined themselves to melodies in a version of the hymnal transcribed in the late nineteenth century in Vagharshapat. They have not attempted to bring together the numerous extant melodic versions of each hymn, with a view to seeking a more primitive, fundamental version that might perhaps just come to bear a measure of true correspondence to the neumatically notated sources at least to an extent to which some reasonable deductions about the meanings of given neumes can be made by comparing the primitive melody with the neumatically-notated equivalent. What is even worse is that the Vagharshapat hymnal must undeniably include its fair share of rather unconvincing-sounding melodies, sometimes bespeaking of hasty attempts to impose various successive melodic fragments to the words in a bid to remain consistent with what little was understood of the neumes, particularly in the case of certain hymns of which perhaps no one was able to recall a more natural-sounding melody. Surely this must reduce the usefulness of the Vagharshapat melodies in any attempts to guess the meanings of the neumes through a comparison with these melodies. But similar criticisms can, and should, be made of particular melodies from other traditions. We feel that there is much scope for comparing, classifying, purifying, combining and selecting the various melodic versions of hymns, with a view to improving their usefulness for research on neumes. There is, frankly, little historical evidence to suggest that the Vagharshapat version, or indeed any one extant version, is any more authentic or any closer to the versions that were notated neumatically, than any other. Thus, it is desirable that in some systematic manner, use should be made of all reasonable versions in combination. For instance, it might be reasonable to suppose (if somewhat of an over/simplification) that the original melody of a hymn may have been better preserved if its various known melodic versions associated with the various traditions are close to each other, and less so if the divergence is greater. It is thus regrettable that decipherments of Atayian and Tahmizean are confined to quoting passages from the Vagharshapat version of selected hymns of a given mode (transcribed from Limonjean s to modern Western notation), juxtaposed against the neumatic version from a manuscript hymnal. Moreover, they have at times jumped into all too facile conclusions, e.g. that one neume or another implies an upward leap of some interval in this mode or that. But for their cited ISSN: ISBN:

4 examples, counter-examples can be easily found that are inconsistent with their hypotheses. The procedure is naïve and unsystematic, and they does not prove that the alleged correspondences transcend mere coincidence. The number of such correspondences is not shown to be statistically meaningful, examples cited constitute but a tiny proportion of the total material available, and instances where the evidence is not so helpful to their hypotheses are not explained or even acknowledged. And, finally, their opinions about the meanings or functions associated with the various modes are woefully insufficient to allow one to reconstruct the same (or, for that matter, any other) melodies from the neumes. A more fruitful avenue might mirror the semiimprovisatory practices referred to earlier. Given our experience and knowledge of a mode, and our current knowledge of time durations associated with certain neumes, how would we devise a convincing melody to fit the words? Several candidate melodies could be devised in addition to those already known, and at that stage correlations between the neumes and a number of such melodies be sought. This should be repeated with many hymns. An iterative process would be appropriate, and though arduous, it would at the very least shed light on the melodic structure of hymns. Also, we should endeavour, through all means possible, to come close to the mental and aural frame of those who notated the hymns into Limonjean s notation, or into Western notation, as far as possible. This could help us evaluate different melodic versions of the same hymn, weeding out more dubious examples and retaining more promisng ones and thus provide a kind of telescope with which we could strive, as it were, to peer back a little further into the past. And, conversely, we might put oneself in the position of a medieval singer and scribe: if one already knew a hymn melody by heart say of a particularly celebrated and thus well-known hymn (and, after all, these very hymns would have been less vulnerable to distortion and oblivion over the years), how would one seek to notate it neumatically, to ensure proper performance by one s pupils and colleagues? What information in addition to time duration, would the neumes need to provide the singer? It is even possible that, whilst the modal aspects of hymns, coupled with some of the neumes, were perhaps sufficient to characterise the melodic contours to a considerable degree, certain neumes were crucial, rather, in helping to achieve uniformity of detail in (monophonic) choral performance (for instance, in helping to specify whether or not passing notes should be executed, in instances where some singers might be inclined to render them and others not often a cause for disarray and sloppiness even in the present day wherever singers rely on recollection alone and do not have a score before them). The fact that frequently there is a deliberate misalignement between verbal and musical phrases in Armenian hymns where it would be (from the point of view of the text) unacceptable, yet musically logical and (in the case of a lone singer) physically necessary to breathe in the middle of a word, strongly suggest that many hymns were intended to be sung by several singers in unison. It is reasonable to suppose, therefore, that certain neumes were largely employed to achieve unanimity of detail in performance by choral forces which already knew the basic melodic contours associated with a hymn. If so, such neumes did not supply information about the basic melodic contours. 3 A brief exposition of the principal modal scales For simplicity, we confine the compass of each mode to the minimum required to capture its fundamental tonal characteristics. To save space, we denote authentic by A, plagal by P, and employ the Austro-Germanic nomenclature for sharps and flats, as well as h to mean B natural and b to mean B flat, and a plus sign to denote a quartertone sharp and a minus sign to denote a quarter-tone flat. IA: g ais (a, a+) h c d (d+) e IP: g a (ais) h c d e IIA: a d e fis g e f e d c h a IIP: cis d e f g a IIIA: fis g a b (h) cis d e f IIIP: g a h c d- IVA: a h c (cis) d e f IVP: fis g a h c d Quarter tones (denoted above as a+, d+ and d-) are gradually falling into disuse. Also, note that certain notes denoted above in parentheses, may sometimes be used, in preference to those immediately preceding them in the above table, depending on their position within the phrasal contour, or the position or the phrase within the verse. These effects can be rather subtle, and (for instance) influence the choice between ais and a in the IA mode, or the interplay between fis and f in the IIA mode reminiscent of the Western melodic minor scale, where direction ascent or descent is a determining factor. And some accidentals appear exclusively at the end of a verse, as part of a characteristic formulaic figure, though they are not usually manifested elsewhere in the mode. Examples arise in the IP variant mode (we shall define what we mean by the variant of a mode in the paragraph below), IPV (not shown above), the IIIA (b and h) and the IVA (c and cis) modes. In addition, most of the above modes have associated darstvadzkhs, that is turns or variant modes, with which they sometimes alternate. The variant modes are highly contrasting to the parent mode, incorporating as they do other notes not allowed in the parent mode, and/or a ISSN: ISBN:

5 higher or lower vocal register. Thus, the transition from a mode to its variant can be highly dramatic, especially when occurring within the same hymn. Modes also govern much of the sung material within the Divine Liturgy of the Armenian church (wherein there are only brief and isolated interpolations from the canonical hymnal, constituting variables that are dependent on the church calendar). Thus, much of the Divine Liturgy is in IIAV (the variant of the IIA mode), whereof the scale may be given as: IIAV: cis d e f gis a. 4 Issues of tonal ambiguity in the aural perception of modes Some items from the hymnal, and virtually the whole of the Divine Liturgy, are now most often performed in harmonised versions by choral forces, sometimes even with organ accompaniment. However, should performance be monodic, and without a drone bass (relatively uncommon in current Armenian practice, unlike the case of the byzantine tradition), the listener is free to construe a certain tonal basis (partly analogous to what Tovey [10] called a linear harmony in Bach) emerging from the modality itself. The tonal contours, formulaic endings, and points of rest or stability serve to create a sense of tonal axes (terminology used, interestingly enough, inter alia in analysing late Bartok quartets music that is at the frontiers of tonality and is itself by no means devoid of modal influence!). And it is sometimes not unambiguously clear which tonal axis is principal (or tonic ). For instance, in considering mode IVA, we notice that all the verses bar the last end on what might seem to be a kind of subdominant (d in the above table, with or without a preceding cis to afford an added sense of convincing closure), although much of the time the tonic axis would seem to be a. (Not dissimilar effects can be found in, for example, Andrea Gabrieli!) In the case of the IA mode, it is sometimes possible for the h in the table above to be interpreted as the principal axis, rather than the more obvious g. Comparable subjective effects can be discerned in other modes also, and the manner in which tonal axes as it were compete with each other in the subconscious mental harmonisation that the modern listener may construct, is salutary considerably enhancing the emotional effect of the music, and forfending the danger of monotony. We believe this to be a fundamental, built-in feature of Armenian church music. Such issues were faced by Makar Yekmalean ( ) [11] and archimandrite ( vardapet ) Komitas ( ) [12], when they each set about to make arrangements for chorus of the Armenian Divine Liturgy. Yekmalean s version is in widespread use. We may see examples of tonal ambiguity, and how it was resolved, in the accompaniments he provided to the odes I kouys vimen (where, incidentally, Komitas expressed his strong disagreement with some of Yekmalean s choices in his review of Yekmalean s Liturgy in the contemporary press), Aysor merealks, and to the vesting hymn, Khorhourt khorin (attributed to abbot Khachatur of Taron, from the early thirteenth century). In the case of the latter, Yekmalean s organ accompaniment furnishes much of the hymn (which is in mode IVA) with a pedal g, thereby creating an impression of a major-like tonality, and incidentally accentuating the similarity of this hymn with St. Gregory of Narek s celebrated tenth-century ode, Havun-havun. In contrast, Komitas uses a pedal on the note a, whereby a minor - like tonality is expressed. In these effects, Yekmalean and Komitas are assisted in their respective choice of slightly different versions of the hymn in that the opening anacrusis of the hymn (to the very first syllable Kho-) is itself a g in Yekmalean s case, and an a in Komitas. Incidentally,Atayian s own accompaniment to the tenth-century ode Havun-havun, features an organ part where too, the interplay between these very two axes within a brief timespan, is well exploited. 5 Komitas two versions of Marmin Terunakan We now turn to an example from the Divine Liturgy, evincing the originality of Komitas genius Marmin terunakan. It is derived from that portion of the Liturgy where the mode IIAV predominates. For ease of comparison with Komitas versions, we reproduce our entry in the table above having transposed it down a tone, viz: IIAV: h c d es fis g. In one version (Fig. 1) Komitas treats the piece as if it were in a sort of G minor but with a sharpened third, or, equally, a kind of G major but with a flattened sixth. A broadly similar harmonisation was provided by Yekmalean. However, in another version (Fig. 2), Komitas has provided a realisation where the characteristics of the mode are strikingly exploited, to produce a mysteriously beautiful effect, where one has the feeling that the tonic axis has been displaced from g to h! The resulting harmonies are eerily beatiful, and anticipate chromaticism of a sort to be found in the opening and ending of Schoenberg s Chamber Symphony No. 2, in Frank Martin s Petite Symphonie Concertante, and in Britten s Rejoice in the Lamb (to which we shall return below). A full discussion as to why selected chord progressions from these works are so very similar in effect to those of Komitas is well outside the scope of the present paper. It is remarkable that, even though the Western masters arrived at their results ISSN: ISBN:

6 through very different procedures, were creating original works as opposed to devising harmonisations of given modal melodies, and were working in very different milieux, nonetheless the similarities to the Komitas are striking. Consider the Britten passage, from For I am under the same accusation with my Saviour up to and including For Silly fellow! Is against me and belongeth neither to me nor to my family a passage of great power and poignancy. Remarkably, no transposition is even necessary to enable a comparison with Komitas Marmin terunakan he is be- [sides] and variance with me strongly suggest G minor, whilst [be] -sides himself and For the officers of the peace (Fig. 3) imply a tonal axis of h (B natural), with the chord practically identical to the h-fis-h chord on Ma- of Marmin terunakan. 6 Conclusions We have provided a brief introduction to modality in the music of the Armenian Church, duly stressing the fact that the concept implies much more than a set of scales a fact which allowed neumatic systems of notation to function for centuries, and, later on, enabled more recent versions of the hymns to be devised long after the original melodies might have fallen into oblivion. Herein lies the crucial but little-understood inter-relationship between modality, neumatic notation, quasiimprovisatory performance and semi-aural transmission. We have discussed efforts to decipher neumes in the light of modality, indicated possible reasons why they have met with little success, and suggested potentially fruitful avenues. Finally, we have demonstrated elements of tonal ambiguity accruing from particular modes. We have pointed out the modal similarity between a 10 th - century ode by St. Gregory of Narek and that of a 13 th - century hymn attributed to Khatchatur of Taron, and cited the contrasting harmonisations by Atayian, Yekmalean and Komitas, evincing different tonic axes associated with the same mode. Our final example concerned an excerpt from the Armenian Divine Liturgy, where we have pointed out remarkable similarities between the imaginative, modally-inspired solutions by Komitas and the modal ambiguities accruing from very similar interplays of chords in passages from Britten s Rejoice in the Lamb. Komitas exploited modal ambiguity, creating novel harmonic resources preempting the emergence of comparable chordal progressions in Western music. Indeed, throughout, we have sought pertinent analogies with Western music, adducing, with good cause, Andrea Gabrieli, Monteverdi and Bach, as well as Britten, Martin and Schoenberg consistent with our belief in the unity underlying all music at a fundamental level. To the best of our knowledge, the bulk of our work (except the basic introduction to the modes) appears herewith for the very first time in the literature. It is hoped that it may stimulate further progress in the field. Acknowledgements I wish to express my gratitude to Prof. Eugene Kindler for many illuminating and inspiring discussions and for valuable encouragement; and to the Rev. Dr Vrej Nersessian of the British Museum, who generously gave of his time and kindly made available his translation of Atayian s important and otherwise inaccessible volume. References: 1. Achb. Zareh Aznavorian, Hay yekeghetsakan yergin dzagoumn ou zargatsoum [The origin and development of Armenian church song], in Hay yekeghetsin khsanerort daroun [The Armenian church in the twentieth century], Nicosia St. Gregory of Tathev, Girkh hartsmants [Book of Questions], Constantinople, Y. Tntesean, Nkaragir yergots Hayastaneaytss sourb yekeghetswoy [The nature of the chants of the Holy Armenian Church], Constantinople, Bishop Norayr Bogharean ( Dzovakan ), Foreword to Dzaghkakhagh sharakanats, yergots yev aghothits [Anthology of hymns, songs and prayers], Antelias R. Atayian (transl. V. N. Nersessian), The Armenian Neume System of Notation, Curzon, N. K. Tahmizean, Arthi khazabanoutiwn [Modern neumology], Pasadena, Fr. Aristakes Hisarlean, Patmouthiwn hay dzaynagrouthean yev kensagrouthiwnkh yerazhisht azgaynots, [History of Armenian musical notation and biographies of musician nationals, ], Constantinople, S. Khanjean, Preface to Tonakarg, Constantinople, M. Bzhshkean, Yerazhshtouthiuwn, vor e hamarot teghekouthiwn yerazhshtakan [Music, that is a brief musical manual], Venice 1815 (unpublished study, cited in [5]). 10. Sir Donald F. Tovey, Essays in Musical Analysis, London, Primary sources (selection): 1. Komitas Vardapet [Archimandrite], Dashnaworeal Sourb Patarag [Harmonised Divine Liturgy], Paris, M. Yekmalean, Yergetsoghouthiwnkh Srboy Pataragi [Chants of the Divine Liturgy], Leipzig, ISSN: ISBN:

7 Fig. 1 Fig. 2 Fig. 3 ISSN: ISBN: