Reflections on the modern performance of a medieval narrative
Hartmann von Aue's Gregorius
By Benjamin Bagby
Hartmann von Aue's renown as a courtly storyteller spread throughout German-speaking lands in the late 12th century. Gregorius, his lengthy masterpiece of more than 4,000 verses, was much appreciated at aristocratic gatherings, possibly already in the late 1180's. His retelling is closely related to a previous French version of the story, the anonymous Vie du Pape Saint Grégoire (ca. 1180), upon which it expands brilliantly, weaving an exciting new version of an old story which captivated an entire generation of noble listeners and poets. The story was to live on for centuries, in many languages and forms, and in 1952 was even fashioned into a novel by Thomas Mann.
Although Hartmann's version survives in multiple text manuscripts, his tale was not originally circulated as a book to be read silently. It was certainly told orally: sung, spoken and perhaps even accompanied, for gatherings of sophisticated listeners who could never get enough of a good, long story. Sequentia's reconstruction of Hartmann's masterpiece features three singers (performing in the original Middle High German verse), who embody various characters: Gregorius, his wife/mother, the abbot, the fisherman, and of course the storyteller. Two of the vocalists also play Romanesque harps, to retell this story in a style which once captivated binge-listening courtly audiences in the years around 1200.
Since the text of this story survives in numerous manuscript sources from the period following Hartmann's death, we can assume a fairly close link to his original creation. However, these sources contain only the text, no musical notation, and absolutely no indications about a performance of the story. We must search elsewhere for musical sources for our reconstruction, and these will be discussed here.
Not only are no musical sources known to exist for Hartmann's works (or for similar works by other authors), there is no generally agreed-upon solution among scholars and performers today as to how we might reconstruct a vocal performance of Hartmann's narrative texts as they would have been heard in the late 12th century. The performance aspect is certain, but the details of a performance realization are missing.
Transmission and memory
We ask ourselves: how did Hartmann's texts – and through them his fame – come to be so widely known in German-speaking lands, in a time long before printing and general literacy? The answer is simple: by means of the human voice and memory. Although we have difficulty in imagining how such complex texts could be transmitted, learned and enjoyed through oral transmission, and that writing was the exception, we can easily understand that parchment manuscripts were extremely costly, requiring immense resources and effort to produce, and we cannot assume that multiple books containing Hartmann's works were in circulation for the benefit of silent readers. Indeed, most surviving manuscripts containing Medieval vernacular texts were copied long after their authors had died. Only the wealthiest and most powerful noblemen could afford the luxury of having a secular narrative copied, and it is unlikely that a rich monastery (important centers for the technology of manuscript production) would devote its resources to the copying of such a vernacular text. Hartmann's own working copy of the story, which has not survived, was possibly never complete, and there are no known sketches.
The tradition of memoria was powerful and taken for granted among Medieval students, ecclesiastics, intellectuals, poets and performers of all types. We are today incapable of imagining the depth and intensity of this tradition, just as a Medieval intellectual would be unable to comprehend our own age's seemingly endless capacity for electronic data storage, retrieval and instant communication. But memory and manuscripts also inhabited the same space in the 12th century, living in a synergy which allowed a story such as the Gregorius to exist in an informed oral tradition, with the rare manuscript source serving as a point of reference for its human transmitters. And since this tradition was largely oral, we must assume that the written stories were in turn given voice (whether through reciting or singing or some mixture of the two) and heard by active listeners, which implies a use of the human voice as an instrumentum of both content and emotion. If we accept this very simple fact of vocal transmission, we must begin to ask questions about the nature of that transmission, and our answers to these questions will inform the decisions we make today in trying to re-create the mode of a lost oral tradition.
Hartmann's predecessors: Germanic epic before the 12th century
Although it is generally accepted – based largely on descriptions of performance situations – that early Medieval epic poetry was the domain of tribal or itinerant entertainers (what we now call the 'singer of tales'), no written musical sources of the epic poems from the early Middle Ages are known to exist. Although Hartmann, working in the wealthy courtly society of the late 12th century, was no longer a part of the older Germanic world of pre-literate tribal singers, the oral nature of performance practice remained in place, and his work is very much in the tradition which had existed since the Romans departed from Germania. After his time, scribes wrote his texts into manuscripts (some since lost), and yet we can be sure that these documents did not include any musical notation nor instructions relating to how his story might be transmitted, whether it was read silently or aloud, sung, chanted, or otherwise made audible in front of listeners. Medieval scribes were not writing with posterity in mind.
Documentation of medieval performance
If we were to hear a Medieval epic performance today, we would probably think of it as a performance style which belongs in the domain of ethnomusicology rather than historical musicology. The surviving musical documents from the period, however, come from the world of the monastic or secular scriptorium, using the available notational symbols (usually the square notation commonly used for notating Gregorian chant) of the 13th and 14th centuries, symbols which were incapable of representing the many variables and subtleties of narrative performance. We have no way of knowing what the motivation might have been for these notated versions, or if the notator was even aware of their relationship with a performance tradition. What are these surviving musical documents, and how might they be related to the performance of a narrative verse such as Hartmann's Gregorius?
From the surviving melodies, we can at least observe that they are syllabic and modally very conservative, melodies associated with the structure of strophic song, of which other examples are known from the 12th and 13th centuries. These melodies are only indirectly associated with epic or narrative verse, but they might provide us with a frame of reference for the working process. Similarly, the lyric lai, with its structural and musical roots in the medieval sequence, was clearly associated with story-telling. There are 12th and 13th-century lais in Latin and Old French which tell Biblical stories: Noah and the Flood, Samson and Dalilah, the Old and New Testaments, and of course Abelard's famous lament of King David on the deaths of Saul and Jonathan. These lyrical lais, which also survive in Old Occitan and Middle High German (the language of Hartmann) with a variety of texts, are striking in their use of simple musical means to express complex textual structures. The melodies are largely syllabic, frequently in the modes based on D or G, and combine clearly defined modal gestures, using repetition and pattern to create an illusion of complexity, variety and narrative energy. They include the masterful large-scale creations of the Minnesänger Heinrich von Meissen (d. 1318), whose stage name was Frauenlob ('Praise of Women'), and other large-scale musical works of the German High Middle Ages, drawing on a long tradition of modal music in Germany. To take an example, Frauenlob's poetic and musical elaboration on the Song of Songs (his 'Frauenleich') is based on a repeating cycle of the eight musical modes, and has inspired Sequentia's work, as has the 'Minneleich' of his contemporary, Alexander, surviving in 13th and 14th-century sources.
Sacred music as a clue to performance
And since we do have many Medieval sources of sung saints' lives, there have also been attempts made to use a similar musical treatment in the setting of narrative texts, involving recitation tones (such as in psalmody) with formulae for beginning, middle and end of each phrase. However, when reading Medieval witnesses to the strong emotional effect of chanson de geste performances on listeners, it is difficult to imagine that the performance being described was as regular and formalized as true psalmody, especially for a text of many thousands of verses. How could a liturgical psalmodic musical setting accommodate the many changes of character, time and place, narrative description and dialogue, and personal expressions of intense emotion? However, it is certainly possible that isolated elements of psalmody (such as the use of a recitation tone, or formulaic beginnings of phrases) were freely employed in the telling of stories, in combination with other techniques.
Finally, there are melodies which are repeated in each strophe. Since many Middle High German narrative epics have such a strophic form (the Nibelungenlied, for example) some modern performers have sought to find a single 'authentic' melody which can be used for performance. But no such melodies have survived with their original texts; all of those which exist were notated later than the texts themselves. The most famous of these is a musical addition to an early 14th-century manuscript of the Jüngere Titurel of Albrecht von Scharfenberg (ca. 1270). We cannot know if this melody was used only for the text to Albrecht's long poem, or for other poems, or if it was in use only at the time of the manuscript's creation in the early 14th century.
Although we can learn much by studying and adapting these tunes, the search for an 'authentic' melody for the singing of such narratives – barring an astonishing discovery – will never yield a definitive result. The performer today who decides to take one of the surviving late-medieval German melodies as a fixed model will soon come up against the hard reality of its rigidity in performance. Such a fixed 'melody' (in the sense of a crafted chanson tune) cannot fully accommodate the flexible structure and emotional energy of a narrative, the many changes of scene, the different qualities of time, the rhetorical idiosyncrasies of the various speaking roles and characters, or the ever-changing role of the narrator himself, who gives the listeners information and comments on the action at the same time. Most importantly, a fixed melody cannot accommodate the one aspect of the human voice which makes it so supple and brilliantly adapted to telling stories: the use of different vocal registers, colors, and the passage from speech into song (including all the gradations in-between). A fixed melody will make use of only a small part of this large spectrum of vocal possibility, whereas the telling of a story traditionally demands the use of the voice as a complete instrumentum.
A view from the modern performer's workshop
Barring the discovery of time-travel, we shall never hear Hartmann or a storyteller of his tradition. Deprived of this essential face-to-face musical experience, we are encouraged to confront our own past musical cultures indirectly, with no guarantee of 'authenticity'.
How can we find a vocal and musical means of re-telling the Gregorius which allows the complex story to unfold naturally over a long period of time? Many years of working on a possible solution to this problem, in a variety of similar repertoires, have led me to believe that a performance solution must indeed be possible, a solution which respects Hartmann's text, the needs of the story itself, the audience, the vocal engagement of the performer and also a plausible 12th century idea of mode, melodic gesture and structure.
The search for performance solutions takes place in a fluid workshop environment which is difficult to reproduce on paper. It is easy to understand the idea behind a given performance after having heard it – a brief demonstration of a spoken or sung text, a modal gesture played on an instrument, or a visual examination of the instrument itself, can only make sense in the context of performance, and can hardly be replaced by written words. But in this format the crucial element of sound itself – the audible and visual presence of performer and instrument in live performance – will be lacking. And so we must enter this virtual workshop which consists of printed words, where listeners (and potential listeners) can hopefully understand the practical issues confronting a modern-day 'singer of tales' and examine some of the factors leading to one possible reconstruction of Medieval narrative poetry in performance.
Modal gesture as a vehicle for text
My first thought was to discover the essential elements in those surviving Medieval melodies, and to understand them as the result of a modal storytelling tradition maintained by rhetorically-gifted singers which, in oral culture, would have been much more flexible and adaptable than anything we find written in the later manuscripts. What are the most basic elements – the building blocks or modal gestures – from which melodies, tunes, 'Weisen', can be built up? Might it not be possible, by working backwards from existing Medieval melodies in manuscripts, to extract a kind of 'modal DNA' which could then be used as a tool to give the singer a more fluid access to the text of the story being told, to make use of variation and repetition within a small repertoire of modal gestures, and to combine these gestures with the text itself, and with all the possibilities of vocal utterance? I am convinced that this is a potential working process which today's singer can use in the reconstruction of lost oral storytelling traditions. The result which it would yield is something supple and flexible, adaptable to the text and yet modally coherent – it would give the singer a freedom of shaping the text from one performance to the next, while at the same time maintaining a highly structured modal integrity. I could even imagine that – if carefully developed – some of the verses might end up sounding uncannily like one of those surviving Medieval melodies which we know from manuscripts.
Admittedly, this is a group of gestures based on my own intuition of how this song is constructed, but also based on my long experience in Medieval song. A different singer or scholar might make slightly different choices, but in the end there would not be much variance in their essential qualities. This attests to the composer's conservative use of very traditional modes. Making use of this material in an ideal workshop situation, the performer should interiorize these vocal gestures and be able to recall them by memory in any order, set to any text, with the possibility for repetition or unusual combinations, thereby creating an endless series of longer or shorter melodies. All of these will maintain the essential modal quality of the written source while at the same time remaining flexible, ready to serve any text which might be sung. Just as we see how various melodies (Weisen or Töne) in the German song tradition can serve for the singing of a variety of unrelated texts, we have here a further 'deconstruction' which breaks one Ton into smaller gestural components, so that they can be re-assembled at will in a supple and variable manner, in order to tell a story. The Ton keeps its integrity as a micro-structure and as a modal unit. In a sense, this entire process has brought us close to an actual narrative melodies found in German sources ca. 1300. We now possess an organic and flexible modal tool which can be used to shape texts and to perform a complex and lengthy text.
The role of the voice
Combining this technique with the many qualities of the human voice brings yet another level of richness and variety to the performance. When performing a narrative text, the voice enters into a world which is informed as much by the rhetorician and actor's art as by the singer's, and in that world the performer can rarely make use of the techniques suited to the needs of what we might call lyric song of the same period (for example, Minnesang, which was also known by Hartmann, or the songs of French trouvères and Occitan troubadours which he certainly knew). These lyric techniques – which call for vocal consistency, the nuanced 'delivery' of a large strophic form with its intricate structure of rhyme and versification, its almost dreamlike disregard for time – are perfectly suited to the re-creation of a formalized emotional state which, for a few moments, conjures up in the listeners' minds their own similar experiences, their own memories and desires. But for the narrative storyteller's art, in which time passes at various speeds, and in which real-time events are recalled, relived, commented upon and sometimes quite literally inhabited by the 'singer of tales', the use of lyric techniques might be applied rather selectively, especially for those moments which seem to call for reflection, introspection or lamentation (again, Sigune's lament from the Jüngere Titurel comes to mind).
In the world of the storytelling singer, the voice is used over the entire range of its technical possibilities: from something we would identify as lyric singing (at times including melismatic passages, formalized and complex melodies, long notes, ornaments) through a more syllabic song, to more approximate, informal song, to sung speech ('Sprechgesang'), heightened speech (exaggerated and hyper-realistic), formal or public speech (before a large group), private speech (a small group), intimate speech (confidentially, between two persons), and even whispering; but also at the other end of the spectrum of human sound, there is shouting, groaning, roaring, screaming and finally, animal noises. Although these widely-varied vocal sounds have been represented by descriptive notation in vocal works since the mid-20th century and even before, we would not expect to find accurate representations of any these sounds in a Medieval manuscript. They were the domain of performers in oral tradition.
Rhythm and the metrical structure of narrative language
Few musical sources of medieval monophonic song from the period before 1300 contain notation which gives specific rhythmic information to the performer. In our musical manuscript sources from the 13th century there are no rhythmic indications in the notation. Generally, it is the rhythmic nature of the poetic text itself which will provide information to the singer when the notation does not specify a rhythm. On a larger rhythmic level, metrical structures will inform the performance and will play an important role in the management of time, both real performance-time and the more fictional passage of time in the story itself. In performance of a work such as the Gregorius, the issue which interests the singer is this: how would such metrical structures have expressed themselves orally? This explains my changing of the verse structure of the text, also as an aid to memorization. Today, our relationship to Medieval texts and music is based on a 'literate' course of study: readings, analyses and exercises, using textbooks, editions, translations, transcriptions into modern notation and sometimes manuscript facsimiles. There is no one alive today who has experienced this poetic art as a uniquely oral phenomenon. As literate beings, we are fascinated by poetic structures which can be indicated in writing, and music students are programmed at an early age with a great reverence for notation, metrics and the unassailable authority of 'the score'. But any singer who has been involved with Medieval European modal song has hopefully procured more differentiated tools for the shaping of texts and melodies, for the telling of stories.
The performer does experience the rhythm of the words and the poetical structures as she sings and speaks the story; they are influencing and shaping the use of voice, decisions about a possible instrumental accompaniment, timing, speed and rhetorical gesture – in short, all of the variables of performance. But these details of performance could never be notated precisely without creating a 'score' which would be overly complex and intrusive. The poetic understanding of the singer must be the place of exchange where rhythmic values are assigned in the context of the story being told.
Assuming a small medieval audience of cognoscenti who had heard a given story already dozens of times, there would be among the listeners a subtle appreciation of the text's rhythms and structures, even a certain delight in the performer's masking of the obvious, and in the interplay of sounds, patterns, and meaning. A 'poetically correct' performance of Medieval narrative text, based on a written source, can still aspire to re-create the spirit of an oral poem performed in a notationless culture. The poetic structure maintains an important role in the text, but should function subtly, almost subconsciously. All elements of measured time must be free to help shape the story: from the smallest unit of the individual syllable to the single, long pulse of an entire performance. And this can be accomplished without notation.
Instrumental participation
Should narrative performance be accompanied by an instrument? There is evidence for the participation of instruments in the performance of a narrative such as Hartmann's Gregorius, but ultimately this decision is not mandated and will have to be left to the individual modern performers. An instrumental component brings not only modal stability to the performance, but is a potentially strong storytelling tool as well, where the instrument can assist in setting the scene, changing the scene, implying the passage of time and creating mood, density, accent, confusion, clarity, or a fixed rhythm. Clearly, when working in a manner such as described here, we would expect the singer to prefer the freedom of self-accompaniment in the shaping of the story, but it doesn't rule out the possibility of a separate accompanist. It would also be logical to suppose that an accompanying instrument, if used, would be a type of instrument widely-known at the time and associated with vocal performance. The two main candidates for this would be the fiddle (vielle) and the harp. The tuning system of any accompanying instrument will be closely related to the mode which the tradition of the song demands, so that the instrument must sometimes be re-tuned to accompany in a new mode (it should be noted that there were no medieval standards of absolute pitch in the tuning of instruments). Regarding playing technique for the harp, it hardly needs stating that an instrument of 15 strings or less is not suited to playing the elaborate melodies with accompanying chords which we tend to associate with later Medieval and Renaissance harps. Instead, we have here a harp type which has as its means of expression the use of pattern, inversion and variation, and on the 'playing out' of modal vocabularies over a fixed set of lower tones basic to the given mode. In the spirit of our 'workshop', we must admit that it would be almost impossible to notate such an accompaniment, since it would be constantly evolving, reactive to the vocal line and the movement of the story, and never exactly the same from performance to performance. However, we can safely say that just as the singer will rely on a small repertoire of potent modal gestures for the vocalization of the text, the harp will make a virtue of its seeming limitations and, like an interlaced medieval design, bring a richness of articulation to the expression of the mode, and hence to the telling of the story. In our reconstruction, the courtly performative framework is enhanced by the accompaniment of Romanesque harps, copies of instruments based on sculpture created during Hartmann's lifetime.
Conclusion
We can never know if our performances precisely duplicate the art of a particular medieval performer of narrative, in Hartmann's circle or elsewhere; nor can we ever claim to reconstruct a verifiable 'original melody' to which any narratives were sung in the Middle Ages, since it is highly likely that such an original melody never existed for any one narrative or story. This certainly varied according to each performance style with its specific oral tradition, and according to the language and dialect of the performer. A retelling of Gregorius in Hartmann's homeland of Swabia might have sounded radically different from the performance heard in Thüringen or the Rhineland. However, I am convinced that by making careful use of the specific techniques described here – coupled with an intuitive spirit based on a working knowledge of both medieval song and the essence of sung oral poetry – it is possible to reconstruct highly plausible models which allow these magnificent stories to live again in performance.