Dance Research

Introduction

The twenty-two-minute dance Herodiade created by Martha Graham in 1944 marks a pivotal moment in her choreographic development, bridging the divide between the Americana dances of the 1930s and early 1940s and the mythological works of the next twenty years. Based on Stéphane Mallarmé's poem of the same name¹ (1864/65), Herodiade is a duet for an enigmatic 'Woman' and her 'Attendant' and alludes to the biblical legend of Herodias/Salome. Although the dance seems slight in comparison to the large-scale works, this was Graham's first foray into the field of Judaeo-Christian mythology and led her dance theatre towards the ancient myths from which she drew her later strong, dramatic roles. Embedded within Herodiade are the seeds for the masterworks inspired by Greek mythology from Cave of the Heart (1946), the first in the cycle of Greek dances 1946–1947, to the full length Clytemnestra (1958).

Herodiade was one of the dances which received its first performance onOctober 30 1944 in the Coolidge Auditorium of the Library of Congress in Washington. Appalachian Spring, the last and most widely acclaimed of Graham's Americana works, and the less successful and short-lived Imagined Wing,² also received their premieres that same evening. At its first performance, the duet was called Mirror Before Me, after a key line in Mallarmé's poem, but shortly after the premiere Graham changed the name of the dance to Herodiade at the request of its composer, Paul Hindemith, by which title it is now known.

Contrary to Graham's usual practice it was Hindemith who chose the poem as a basis for his collaboration with Graham. The sonority of its lyric verse appealed to the composer, and Graham's choreographic imagination was caught by the imagery of the poem. Although the negotiations between choreographer and composer were strained, the difficulties within the partnership did not prevent the achievement of a major work. Herodiade has received public and critical acclaim over the years, and it is currently in the Graham company repertory although it has never been performed in Great Britain.

Scholars have overlooked Herodiade, and the writing which has appeared about it (see for example Bentley in Copeland and Cohen 1983; Croce 1978; Lloyd 1947; Stodelle 1984) has not addressed the dance in analytical detail.This article focuses on the genesis of Herodiade, Graham's collaboration with [End Page 1]

Hindemith, and the relationship between the dance and the poem on which it is based. Herodiade repays close study since the separate layers of meaning, drawn from the archetypal sources of myth and history and from Graham's own artistic and personal narratives, overlap, and create a powerful depth and diversity in the dance. This interweaving of different signifying systems can be analysed through intertextuality,a concept which is already familiar in dance studies (Adshead- Lansdale, 1999). [End Page 2] Drawing on Julia Kristeva (1984 and 1986), I shall show later in the article how the dance is an intertext which is 'an act of interpretationin itself' (Adshead Lansdale discussing Frow in Adshead Lansdale 1999, p. 18). I base my analysis of Herodiade on a recorded performance (1991) by Takako Asakawa as the woman and Christine Dakin as her attendant as well as on published photographs and written accounts of the dance.

Graham and Mallarmé: Herodias/Herodiade

In her autobiography and in interviews that have been published concerning Herodiade, Graham refers to the principal role as Herodias (Sears 1982; Graham 1991). According to biblical legend, Herodias was the wife of Herod Antipas, tetrarch of Galilee at the time of Christ's crucifixion. This was an incestuous union, which was denounced by John the Baptist and prompted the famously erotic dance performed by Herodias' daughter Salome. As a reward for her dancing and to avenge her mother's wrath, Salome demands the head of John the Baptist (St Mark 6, 16–17).³

Despite Graham's association of the principal character with Herodias, Mallarmé's heroine is a much more mysterious and abstract creation. In 1865 Mallarmé wrote to Eugène Lefébure about his new venture, a lyric poem intended for dramatic presentation. Mallarmé began this poem at the beginning of his career in 1864 but it remained incomplete at his death in 1898.⁴ The letter suggests that it is Hérodiade's 'nom divin' which has inspired Mallarmé rather than its biblical association. He describes the word Hérodiade' as 'dark and red like an open pomegranate' (Correspondance 1862–71, p. 226), explaining also that his Hérodiade is 'a purely imaginary creature entirely independent of history' uprooting her from her historical and mythological context (ibid.).

In the preparatory note to Les Noces d'Hérodiade (Davies, 1959, p. 51), Mallarmé removed from his central figure all vestiges of Salome and her legendary connections with the fateful dance (ibid.).⁵ But to what extent can the poem's historical and mythological resonance be disclaimed? Scholars have pointed out that Hérodiade is a new form of Herodias, which collapses the mythical Herodias/Salome and their attendant stories into one representative figure (Fowlie 1984, p. 170).⁶ Zagona notes that Heinrich Heine and Théodore de Banville, as well as Mallarmé, 'use the name Hérodiade, though the heroine in each of their works is unmistakably Salome' (1960, p. 20). The 'unmistakable' presence of the legend in Mallarmé's text is supported by Julia Kristeva's notion of intertextuality. Following the Russian theorist Bakhtin, she claims: 'any textis constructed as a mosaic of quotations; any text is the absorption and transformation of another' (in T. Moi 1986, p. 37). On this basis it is impossible to banish all references in Mallarmé's Hérodiade to Salome and the New Testament story of the beheading of John the Baptist.

In fact, Graham's dance Herodiade⁷ resonates with allusions to this grisly scenario. The stage is set with Noguchi's three bone-white sculptures: one represents a mirror, which is, as Graham describes it, a 'structure in bones of the [End Page 3] human body'. The other sculptures are a low stool or 'chair' and an X-shaped form, which bears a length of black cloth (Sears 1982, pp. 28–30).⁸

The duet is choreographed in the form of a danced conversation with alternating solos between the two women, and the tone of the choreography ranges in colour and timbre throughout the dance. From the sombre, almost ritualistic, entrances of each woman, the dancing passes through passages of intensity and passion, as well as moments of reflection and contemplation, anger and desperation. There are chilling moments of tension as the woman crosses the stage, upper body twisting sharply against the feet and hips held in parallel. The attendant⁹ is an intrinsic part of the drama of the dance and is witness toher mistress's emotional agony. She is at times solicitous and at others accusatory. In gestures that are hieratic and ominous she seems to spell out the tragic fate which awaits her wilful mistress. But there are also lyrical passages in the dance, which help to relieve its sense of terror. Halfway through the duet, the woman recalls a time in her life when she was more at ease with her surroundings and with herself. Skirt in hand, and with the piano and oboe playing a gentle tango rhythm, she travels forwards to the mirror and retreats from it in kneeling and lunging movements and with steps that softly rise and fall, as though remembering a ball she once attended.¹⁰ The woman appears to see beyond surface reflections as she bows low to the ground. It is as though she acknowledges a dancing partner whom she conjures from the mirror's depths, or recognises an apparition, a former self, free from the forces of evil and darkness that now envelop her.

Recurring throughout the movement language of Herodiade is a motif which Graham calls 'the dart' (1991, p. 244). This is a walk with each stride linked by a catch step, so that the same foot always leads. Graham etches the woman's body against space in two-dimensional form. The arm is lifted and angled sharply at the elbow as it frames the face. The torso leans slightly forwards, drawing upwards away from the hips as the leg and foot extend to the back of the body. At times in the dance, the woman takes up this body design in a stationary pose, but at others she sweeps round the stage in the dramatic dart walks, upper body directed towards the lifted arm.

The culmination of the duet comes in a solemn ritual of preparation. As her attendant slowly strips away her outer robe, the woman stands in a shimmering white under-garment, which represents her naked and unprotected soul. As the sombre music dies away, clothed now in the black cloak of fate, the woman begins to walk towards the wings of the stage. She raises her arms inside the cloak so that the black cloth forms a pyramid, which extends high over her head. In the final image, she bends slowly backwards in such a manner that her head seems to float disembodied against the black material.¹¹

The Americana Period

Herodiade dovetails between the end of Graham's Americana period and the beginning of the next phase of her career, her turn towards the great myths of [End Page 4] classical Greece. In the 1930s, Graham created stark dances, with plain costumes and little or no sets or props, which often explored her American Puritan heritage. These concert works gradually gave way to a more theatrical use of stage and set. Notable in this transformation is Frontier of 1935. The solomarks Graham's first collaboration with the Japanese American sculptor Isamu Noguchi and her emerging sense of dramatic characterisation. In the minimal setting of Noguchi's gate, placed centre stage with ropes looping upwards to the flies, Graham presents, with economy and a sure sense of choreographic design, her portrait of a nineteenth-century American pioneer woman.

American Document (1938) was another major milestone in the transformation of Graham's dance theatre, not least because it introduced Erick Hawkins asthe first man to dance in the all-female company. The burgeoning partnership with Hawkins provided an altogether new departure for Graham in the form of a lyrical duet or 'telling love passage,' in which she and Hawkins danced the Song of Solomon (de Mille 1991, p. 233).

Graham also broke with her customary methods of using music selected by Louis Horst or composed by him after the dance had been fully choreographed. For American Document the music was commissioned from Ray Green before Graham began work on the choreography (Thomas 1995, p. 133). But Horst's influence on Graham cannot be ignored. From early in her career, he had advised her against seeking to interpret music, stating instead that it was the dance itself that 'should be the centre of interest, the point in tension' (in Madden 1996, pp. 56–7). This sense of artistic independence, which Horst instilled in Graham, is a factor to be borne in mind when considering her collaboration with Hindemith.

Two works, which followed American Document,deepened Graham's narrative explorations and instigated her turn towards an investigation into the internal and external worlds of the personalities she presented on stage. In Letter to the World (1940) with a score by Hunter Johnson and with a sophisticated setting by Arch Lauterer, Graham made a study of the American poet Emily Dickinson. She skilfully presented different facets of Dickinson's nature and personality, weaving together the poet's spiritual and emotional progress (de Mille 1991, pp. 241–2). Deaths and Entrances (1943), a dance which ostensibly concerns Emily Brontë¹² and her siblings, is described by Graham as a 'modern psychological portrait […] of women unable to free themselves to follow their hearts' desires' (in De Mille 1991, p. 252). As a dance for two women, Herodiade chimes withthe 1930s dances for the all-female Group where men were present in relation only to their absence and in which the movement vocabulary was often distinguished by its two-dimensionality and austerity. The archaic motifs and blunt gestures of Herodiade look back to the pre-1938 dances but, with its integration of movement and stage sculpture with specially commissioned music, Herodiade echoes Frontier and American Document. The psychological undertones of the duetreflect the dramatic potency and self-searching intensity of Letter to the World and Deaths and Entrances.

We can trace the continuing progress of Graham's train of thought through [End Page 5] the correspondence concerning the commissioning of the three works that followed: Appalachian Spring, Herodiade and Imagined Wing. These three dances formed the Elizabeth Sprague Coolidge project 1942–44.

The Genesis of Herodiade and Thecoolidge Project

It was Erick Hawkins who had the idea in 1942 to approach the connoisseur and patron Elizabeth Sprague Coolidge and propose that her music foundation for the promotion of chamber music should commission scores for Graham and subsidise her choreography (Shirley 1999, pp. 64–5). Hawkins suggested that Coolidge should commission composers of 'the caliber of [Aaron] Copland and [Paul] Hindemith'¹³ to write music for 'Martha Graham and her Company, and sponsor the first dance performance of these works' (Hawkins quoted in Shirley 1999, p. 64 and p. 93, note 1).¹⁴

In his analysis and discussion of the Coolidge correspondence, Wayne D. Shirley (1999) points to a number of developments in Graham's career. Firstly, in 1942 Graham was on the cusp of a major shift. The correspondence reveals that her thoughts were already turning towards the Greek myths and archetypes of antiquity, subjects which were to engage her imagination throughout her career. This can be glimpsed first in the exchange between Graham and the composer Carlos Chavez, who was originally commissioned for the Coolidge project but who was replaced in 1944 by Paul Hindemith.¹⁵ In the Chavez correspondence, Graham talks about the idea of a dance based on the legend of Medea, anticipating that Chavez would provide music with a 'secret violence' (Graham in Shirley 1999, p. 81). The Coolidge correspondence also reveals that Graham had established a method for her collaborations, which involved the preparation of what Graham called a 'script'. These scripts comprised:

An introduction describing the purpose of the dance. This is followed by a list of characters, with each character described in detail. Then there follows a one-page 'outline of action' giving the sections of the dance with an indication of the length of each section. Finally there is a detailed description of the action of the dance' .

In the correspondence, Graham also confirms her authorship as a dancer:

I have always chosen my subject when I was asking a composer to write for me. I submitted an idea and a detailed script, not of the dance steps, etc., but of the idea and the action. The reason that I have worked this way is that I find I only do things well when I can feel my way into them as a dancer. I have done those things only that I could feel and understand, not in a verbal sense, perhaps, but in my medium, my instrument, my body .

When Hindemith challenged Graham's pre-emptive scripts by choosing the poem and the concept of Hérodiade,it was predictable that this collaboration between the established European composer and the independently minded American choreographer would encounter some turbulence along the way. [End Page 6]

Sources, including the Coolidge correspondence, reflect in some detail the testing relationship between Graham and the powerful personalities of Paul Hindemith and his wife Gertrud. The composer had been forced to leave Germany by the National Socialist Party in 1934. His personal, moral, and ethical views conflicted with those of the Nazi government, and his compositions were banned because of their dissonance and modernity. He left Germany in 1938 for Switzerland and England and finally settled in the United States (Skelton 1975). Graham met the composer for the first (and possibly only time) in March 1944 when he was teaching at Yale University. She was so struck by 'the energy of the man' and by the intensity of Hindemith and his wife Gertrud, '[she] almost fainted, [she] felt so weak'. According to Graham she was riveted by Hindemith's presence as he stared at her throughout 'the whole time he sat at the piano and played' (in Shirley 1999, p. 89 and de Mille 1991, p. 260). Gertrud Hindemith was no less formidable and telephoned Graham soon after their meeting to enquire if they had been 'too strong' for her. Graham was herself a force to be reckoned with, but assured Gertrud Hindemith that they were 'almost'[my italics] too overpowering (Graham in de Mille 1991, p. 260), as if to signal that she could survive even the most intense artistic pressure.

At the very outset of the Coolidge commission in 1942, Graham had favoured the Brazilian Heitor Villa-Lobos as the composer for one of her works. When this idea failed to take root (Shirley 1999, pp. 66–9), Coolidge's proposal that Hindemith receive the commission worried Graham. She wrote to Coolidge expressing her concern that the composer had not seen her work and thatshe understood him to be 'quite independent' (ibid., p. 69). Coolidge then approached the Mexican composer Carlos Chavez,¹⁶ because she knew that his Sinfonia de Antigona (1933) based on the story of Antigone had impressed Graham (ibid., p. 70). Accordingly Graham prepared a script which reflected both Chavez's experience of writing music to accompany a myth and her own developing interest in the subject. She chose the myth of Medea and called her proposed work Daughter of Colchis.

In Graham's script for Daughter of Colchis she explains that 'Colchis was the barbaric land from which Medea came' and gives a personal interpretation of the myth:

This is a legend of emotional conflict.
The legend concerns a man and a woman
The man is led away from the woman
   by a dream in the guise of a muse.
The woman is dominated by a fury.
The passer-by is the spectator, the chorus.

This Medea, though, was fated to go no further. Chavez continually missed his deadlines and in September 1943 a postponement of the entire project was inevitable (1999, pp. 72–9). The original date scheduled for the performances in the Coolidge auditorium came and went. But Graham was keen not to lose sight of the opportunity and negotiations went ahead to replace Chavez¹⁷ (Shirley [End Page 7] 1999, pp. 81–2). Hindemith, Ernst Toch and Igor Stravinsky¹⁸ were proposed (ibid., p. 83), and Graham, overcoming her concern about his 'independence' chose Hindemith (ibid., 84).

Graham and Hindemith

Graham was familiar with Hindemith's score for Massine's ballet Nobilissima Visione¹⁹ (1938) and was optimistic that he would follow a similar 'theatrical line' when writing the music for her dance (letter 19 March 1944 in 1999, p. 85). There was a clash of personalities and artistic processes from the start of the collaboration, which is evident from a letter that Graham wrote on 19 March 1944 to Harold Spivacke, Chief of the Music Division of the Library of Congress. She told Spivacke that she had received a letter from Hindemith in which he rejected her script on the grounds that 'he prefers not to do this kind of music at this time' (Graham in Shirley 1999, p. 85). She continues:

I do not know whether the idea of a script worried him or not. He did say that if I wanted to do something with more abstraction for abstraction's sake perhaps we could collaborate sometime in the future […].

Hindemith was in a neo-classical phase at this time. In the 1940s he had produced 'Theme with Four Variations […]' (Shirley 1999, p. 85),a score which Balanchine purchased from the composer in 1940 (Taper 1984, pp. 208–9²⁰ ) but did not use until 1946 for his ballet The Four Temperaments (1999, p. 85). As an expressionist choreographer with abstract leanings, Graham was also interested in the formal and compositional elements of dance, but, as she told the critic Anna Kisselgoff (1984), 'I have to have a 'dramatic line even in the most abstract things I've done' (in Kisselgoff 1984). Graham's notion of the abstract in dance requires more attention and I shall return later to this topic.

Seeking to accommodate Graham's requirement for music that would offer an emotional core and provide a dramatic edge, Hindemith proposed a score based on Stéphane Mallarmé's poem Hérodiade. In so doing, he deflected Graham from her burgeoning interest in Greek mythology and Medea, though the idea had settled firmly in Graham's mind and Medea would become in 1946 Cave of the Heart with music by Samuel Barber. Now, though, she responded positively to Hindemith's interest in another archetypal heroine, and, withher experience of working from poetry in Letter to the World,turned towards Mallarmé's Hérodiade. Hindemith had a process in mind whereby he would 'explore word by word every nuance' of the poem, and he planned to 'set the poem as though to be sung, then score the setting for (voiceless) chamber orchestra'²¹ (Shirley 1999, p. 86). In a letter of 1948 to the conductor, Willy Strecker, Hindemith himself confirms his intention: 'The melodic lines in the orchestra are the recitation, which follows the poem word by word' (in Skelton 1975, p. 212).

Siglind Bruhn²² has since explained that the process followed by Hindemith is musical ekphrasis,a term deriving from ekphrastic poetry: 'poems inspired by [End Page 8] paintings or other works of visual art, including etchings and drawings, sculptures and architecture, photographs, films, etc.' (Bruhn 2000, p. xviii). Drawing on a quotation from an article by Claus Clüver (1997) Bruhn defines ekphrasis as 'a representation in one medium of a real or fictitious text composed in another medium' (2000, p. 5). Bruhn demonstrates how Hindemith approximated the sonorous qualities of Mallarmé's vers libre, and employed the musical techniques of 'pitch contour and rhythm' so literally that 'they are found to speakquite eloquently.' She goes on to say that the music is 'by no means in some symbolic, non-verbal language, but in Mallarmé's French' (ibid.).

The notion of musical ekphrasis is appealing and even useful since the process of what Bruhn calls 'transmedialization' is the basis on which she claims that several composers have translated literary and visual texts into music. This theory allows Bruhn to provide scholarly and insightful analyses (2000) into these works, including Hindemith's Herodiade and the close relationship he achieved between his music and the words and symbols of Mallarmé's poem. But herview of Graham's dance is both unfounded and contentious. Relying on critical and eye-witness accounts of the duet but without giving any evidence of having herself seen it, Bruhn states that it is 'not an interpretation of Mallarmé's poem or [sic] Hindemith's music […]' (2000, p. 546). The concept of transmedialisation raises several issues amongst which there is the question of different levels of interpretation and the extent to which there is more than one method of treating literary or visual stimuli. Not all artists set out to attempt to translate the host medium into their own.

According to Bruhn, Hindemith used the sonic resonance of the poetic words to transform them into the musical tones and textures of piano, string and woodwind instruments. Hindemith himself claims that his musical idea was to make the sense of the words implicit within what he referred to as the 'musical Gestalt.' This drawing together or fusion of instrumental cadences sought 'to mirror the language/text, which strives towards the most secret springs of poetic expression' (in Briner 1971, p. 163). These words suggest that Hindemith had a sensitive approach to Mallarmé's poetry. As a consequence he created a score in which he uncharacteristically gave free reign to 'sensuous feelings' (Skelton 1975, p. 211) and which provided the core of emotional depth and expression to which Graham could respond. This development in Hindemith, away from formal abstract structure and patterning and towards more emotionally charged musical composition, should have drawn him closer to Graham's artistic sensibility.

The Genesis of Herodiade

There do not seem to have been many opportunities for Graham and Hindemith to exchange ideas or to discuss their approaches to Mallarmé's poem. Details of the interaction between them are available only in fragments of information, including an article by Graham, which was translated into German for the opera and ballet magazine Operenvelt (1964).²³ The article and extracts [End Page 9] from it have since been quoted in other German and English sources (see for example Briner 1971, pp. 163–4; Skelton 1975, p. 211). Graham states that Hindemith proposed that she should choreograph Hérodiade at a meeting between herself and Hindemith on 19 June 1944 but she wrote these words almost twenty years after the events she describes. A letter from Graham to Hindemith of 19 June 1944 refers to an afternoon that she spent with the composer and his wife Gertrud and expresses how overwhelmed she felt by their strength (Graham, in Shirley 1999, p. 89). The Coolidge correspondence (Shirley 1999, p. 85) demonstrates that Graham and Hindemith in fact met on 29 March 1944 and it was during this personal encounter between composer and choreographer that Hindemith suggested Mallarmé's poem as a basis for their project. Graham's letter of 19 June demonstrates that she was already working on Mallarmé's poem and engaging with Hindemith's music:

I have thought so much about your music and I do know how deeply moving and gratifying it is. I am beginning to see my way with Herodiade [that is, the poem] more clearly. I find the poem fuller and richer every time I come to it. Hearing it in French, as Mrs. Hindemith read it to me, helped me a great deal in understanding the sonority that is in the poem in the original which the English lacks.

In this letter Graham shows that she was excited by the sounds of the poem in its original French. However, Erick Hawkins, Graham's professional partner and lover at this period, remembers reading the poem in French to Graham, and confesses that he found its esoteric nature baffling and could never work out what it meant. He describes how Graham went 'through real agony' when working on the dance and arriving at an interpretation of its meaning (in Helpern 1999,p. 36).²⁴ A choreographer intent upon exploring the heart and mind of the vengeful Medea, now had to grapple with the mysterious and elusive Hérodiade – a self-absorbed young woman, who was as much a product of Mallarmé's linguistic explorations as she was a phantom from antiquity.

The sources that I have found on Hindemith's response to the dance itself, suggest that he was not concerned about the staging of his musical composition. A letter to friends of 13 September 1944, states that the score for Herodiade was complete, although Hindemith writes that neither he nor Graham had been in correspondence with one another. This lack of collaboration during the artistic process did not seem to trouble him, since all the financial and contractual arrangements for the Commission were in order, and the project appeared for Hindemith to have been little more than an affair 'now done with' (in Skelton 1995, p. 185). Earlier in the letter he refers to Hérodiade as one of his 'pet' compositions, but the letter shows no interest in its theatrical realisation and there is nothing to suggest that Hindemith ever saw the dance performed. The following remarks written in a letter of 1948, four years after the premiere, hint at his scepticism about Graham's work, but seem based on an assumption of what the dance might look like rather than on an instance of actually viewing it:

As a ballet (only two dancers) it will probably suffer the sad fate of being done only in an idiotic way, for what nowadays hops around on the stage is for the most part too limited [End Page 10] to cope with a work demanding the highest degree of concentration and dedication […] .

In an interview that she gave over forty years later (1985) Graham states that the Hindemiths did not like the dance.²⁵ But she refers only to the words of Gertrud Hindemith who asked:

Where are the bottles of perfume that you're supposed to be passing in front of? Where is the leopard on the stage? […] You didn't consult the composer at all about this.

Stodelle,²⁶ writing about Herodiade in 1984, does not disclose the sources of her evidence, neither does she state that she was herself present at the performance, yet she claims that Paul Hindemith did indeed see Graham's Herodiade. In a passage that bears a resemblance to Gertrud Hindemith's reaction, she describes his perplexity at the lack of literal staging ('where is the mirror, where are the legends?' 1988, p. 132). If Hindemith did see the dance, Graham never knew this, as she indicates in 1964, though she explains how she followed his instructions about controlling the volume of sound of the chamber orchestra instruments and maintaining the accuracy of tempo:

I always wished that I could have said that to him myself but it never happened. I don't know whether he ever saw my choreography, in any case he never mentioned it. Solely his wife came to me after the performance and made me understand that she didn't exactly like my 'Herodiade'. Seemingly Hindemith had a [W]agnerian idea of the choreography, in that case my version was too abstract for him .

The major factor that emerges from Graham's statement is that again there is a clash between composer and choreographer about notions of abstraction. Graham's dance theatre tends towards a form of abstraction for which she has a specific theory. Abstraction in dance for Graham is not mere formalism or, as she states, 'a matter of lines' (Graham 1963, p. 26). She uses the word abstract more in the sense of a concentrated form of expressionsince she explains that it means 'taken from.' 'Orange juice,' she says, 'is the abstract of an orange' (ibid.). Graham's narratives are the expression of underlying currents rather than of surface events. It is the mental processes and internal dynamics of her protagonists' experiences to which her dances give representation and expression. Graham's metaphoric movements with their associative meanings, and Noguchi's symbolic tower of bone-like shapes, which represents the all-important mirror of the poem (Fowlie 1984²⁷ ), must have seemed abstruse and contrary to the Hindemiths. No doubt the composer or his wife expected a dance which would offer a more literal interpretation of the dramatic action of the poem and its imagery.

Despite their different approaches, the first performance evidently displayed a coherence which convinced the critic Robert Sabin. Reviewing the premiere of the duet as Mirror Before Me,hepraises the way that the 'musicand dance blend flawlessly', remarking that the 'counterpoint' of Graham's movement was 'as strong and logical as the development of the musical texture' (1944, p. 121). [End Page 11]

Considering the dance now, Graham's choreography closely follows the range of the composer's rhythmic gestures and tonal devices, which fluctuate between 'the lowest notes of the contrabass to the highest flute' (Hindemithin Briner, p. 162). As each musical speech reaches a crescendo, Graham's choreography grows vehement. When the musical 'words' hover in the air the women hold each other's gazes. At other times a lingering gesture coincides with the music as it dies away. Graham also follows Hindemith in setting the dances for the woman to woodwind, string, or piano solos, and for the attendant to the 'gentle themes of the string instruments' which represent her character. Graham does not just reflect the music, since the attendant punctuates the softly intermingling sounds of strings with violent split falls, high kicks to the sideor dramatic hand gestures, representing the spewing forth of some ominous message. In the dances for the woman, Hindemith's tonal 'words' cascade through her body in a torrent of arm and leg movements, or rolling on the floor in an agony of passion. At other times she echoes the almost whispered tremors and vibrations of musical sound with small, delicate movements of the head and hands. When the music grows insistent and urgent, mounting in volume, Graham matches the musical climaxes with forceful movement – the 'sparkle' jumps with arms reaching desperately upwards or sudden dives in arabesque to the floor. Graham often infuses the musical score with an extra layer of dramatic intensity but also responds to it with subtle details of movement which follow Hindemith's variations in dynamic range and richly textured orchestration.

Synergy: The Choreographic Interpretationof the Poem

Creating a dance based on Mallarmé's Hérodiade was a task which Graham originally found daunting, working hard to respond to Mallarmé's vers libre. At the heart of the poet's practice is the process he underwent in striving to create lyric verse that was entirely self-referential, a 'pure' poetry which would achieve Mallarmé's ideal of perfect beauty (Millan 1994). It was the activity and the creation of a 'coherent, self-sufficient construct' with 'tight structures, and deliberately imposed rhymes,' which preoccupied him (Killick in Temple, p. 37; Tufail in Temple, 1998, p. 148). Mallarmé's research into language revealed that words are little more than a collection of meaningless and conventional signs constructed through human agency rather than dictated by an omnipotent or divine order (Millan, pp. 134–9). The arbitrary relationship between the word and its referent (signifier and signified), as postulated by the Swiss linguist Ferdinand de Saussure in 1906–11,²⁸ was already apparent to Mallarmé in the late nineteenth century. He commented wryly on the idiosyncrasies of the French language when he considered that the word for the insubstantial 'shadow' is the rather solid sound 'ombre,' whereas the word for opaque darkness in French is the light and wispy 'ténèbres'. The lingering quality of ténèbres seems better suited to the intangibility of 'shadow' than to the density [End Page 12] of darkness (ibid.). For Mallarmé, it was the sonic quality of French words which was of paramount importance, and, as we saw earlier, he chose the name Hérodiade not so much for its legendary associations but for the beauty of its rich tones. Mallarmé's use of the interplay between the sounds of words and their sense is evident, however, in a key line from the poem, when the princess Hérodiade says to her nursemaid: 'Assez! Tiens devant moi ce miroir'. The assonance of 'moi' and 'mirroir' reflect one another phonetically and semantically, investing even this one image with tonal and symbolic resonance.

It was Graham's task to embody the formal structures and linguistic sophistication of Mallarmé's poetry in her own dance language and she consulted the scholar of French literature, Wallace Fowlie, in her search for an interpretation of the poem. To some extent, she followed the myth enmeshed within it, but for Graham, the poem also evokes the notion of fear. She described it as '[…] the fear we all have, the fear of the artist, of a blank white page when writing a composition, the fear of the empty studio when starting a dance' (1982, p. 28). Clive Barnes responded to this reflexive aspect of the work when he wrote of Herodiade as a 'wonderful piece of art within art, of drama within drama, the mirror image of the mature woman as an artist […] (1990, p. 17).

Barnes (1990, p. 17) and Croce (1978) recognise that the duet follows the structure of French classical theatre with 'long movement speeches in turn for each woman, with little or no physical contact between them' (1978, p. 191). This aspect of the dance is in line with Mallarmé's original intentions since he had conceived Hérodiade as a poetic tragedy to be theatrically staged (in Millan 1994, p. 120). The poem is structured as a dialogue between the young princess and her old nursemaid, and presents Hérodiade as an almost inhuman figure 'combining inviolate beauty and uncompromising aspiration to attain it' (Killick, in Temple 1998, p. 37).

Close analysis reveals that Graham's dance follows Mallarmé's structural devices by including the three indiscretions or 'triple impieties' committed by the nurse towards Hérodiade (Fowlie 1953, p. 131; 1984, p. 167). These occur in the first half of the poem, near the beginning, when the nurse tries to kiss the princess's heavily ringed hand, and Hérodiade commands her to withdraw – 'Reculez' (line 4). Later the nurse offers perfumes, only to be told to 'Laisse-là ces parfums !' – leave the perfumes ! (line 57) and thirdly when she attempts to adjust the tresses of her mistress's magnificent hair, only to be told 'Arrête dans ton crime' – restrain your crime […] (line 80) (Fowlie 1953, p. 131). Hérodiade is irritated by, and rejects, her nurse's attentions, because the 'crimes' are a foreboding of the disaster that awaits her (ibid.).

Graham registers the three places in the poem where the nurse's transgressions or crimes take place. The first occurs in the early moments of the dance. The attendant reaches towards the woman but she recoils with a series of childlike springs. The second crime or transgression comes when the attendant tries to touch her mistress as she kneels in the dart pose. Again the woman reels away from her maid this time in a dance of fury and venom. The third crime comes when the attendant reaches out and touches the woman gently as she [End Page 13] stands, hip thrust forwards and upper body twisted away from the supporting leg. For a third time the maid is repulsed as the woman bursts into an explosive and almost violent dance. The Hindemiths may have been disappointed that the choreography has no narrative representation of images such as the heroine's ringed hands, bottles of perfume or stray strands of hair. But, by registering the three dramatic climaxes, Graham respected the structure of Mallarmé's poem thereby creating a sense of mounting tension which pointed towards the mysterious fate awaiting her heroine.

Croce's view that Graham 'takes the poem's characters and its iconographic setting and transforms them into elements of her personal mythology' (1978, p. 192) is tantamount to the principle of intertextuality.The literary theorist Terence Hawkes draws on the work of Julia Kristeva (1984) to point out that 'no text can ever be completely free of other texts'. Even those literary texts which '[emit] messages that refer to themselves, also make constant reference to other works of literature' (p. 144). Mallarmé's Hérodiade is the site of intersecting texts drawn from the centuries' old myth and from Flaubert's novel Salammbô (1862). According to Fowlie, Salammbô and Hérodiade are heroines who share similar characteristics of aloofness. 'Their beauty is mysterious and hermetic. It is shattered or would be shattered by marriage' (1953, p. 126).

This inviolate beauty so integral to Mallarmé's poem is an idea difficult to access since it is closely bound up with death and with poetry itself. If beauty is to survive and to be preserved in all its purity, then it must remain aloof from and untainted by life and hence immune from death. But such a denial of the forces of life is tantamount to death, as Hérodiade states early in the poem: 'A kiss would kill me, woman,/If beauty were not death'. The narrative texts, drawn from the Salome myth, play alongside the poet's quest for the beautiful andhis search for the ideal. These intertexts create a poem rich with analogy and allusion verging often on the esoteric.

We can treat Graham's Herodiade as a text, because it intersperses and incorporates strands drawn from the ancient myth, from Mallarmé's poem and from previous dances based on the Salome legend. It is intertextual in the way that Graham as the 'speaking subject' (creative artist) takes over preceding or older signifying material and re-presents it in a new form (Kristeva 1984, pp. 59–60).

Graham took what after all is a literary construct and gave it a fully-fledged embodied presence. She did so through the two-dimensional and regal motif described earlier in the article. The movement permeates the dancer's bodywith lines of energy that flow into the space that surrounds her. Within the dart pose, Graham's woman/Herodias maintains the imperious and commanding presence redolent of the nobility and remoteness of the poem's heroine with her protective 'torrent of […] immaculate hair' (line 5). When Mallarmé's Hérodiade gazes into the depths of her mirror transfixed in an endless searchfor her memories she finds only 'une ombre lointaine' – a distant ghost (Fowlie 1953, p. 131). [End Page 14]

        O miroir!
Eau froide par l'ennui dans ton cadre gelée
Que de fois et pendant les heures, désolée
Des songes et cherchant mes souvenirs qui sont
Comme des feuilles sous ta glace au trou profond,
Je m'apparus en toi comme une ombre lointaine,
Mais, horreur ! des soirs, dans ta sévère fontaine,
J'ai de mon rêve épars connu la nudité!

These lines represent more than a young woman's anxious hours spent gazing fruitlessly at her own image. It is the mirror as symbol of art into which she peers. For Mallarmé the mirror reflects not the material world of external reality but the immaterial realm of the inaccessible ideal. It is the mirror of art which can focus and hold the 'otherwise scattered Dream and allow for a glimpse of the Ideal' (Hassold 2003).

Graham approximates the lines from Mallarmé's poem although she adapts them towards a more narrative function. She borrows the desolate imagery and tone from the poem, echoing their sombre prescience, but as an artist from a different age, she is less concerned with the elusive ideal and more with the search for the self. She emphasises the subjective dimension of Mallarmé's text, but at the same time she redirects the narrative towards a more pronounced sense of place and moment of self-realisation as the programme note demonstrates:

The scene is an ante-chamber where a woman waits with her Attendant. She does not know what she may be required to do or endure, and the time of waiting becomes a time of preparation. A mirror provokes an anguish of scrutiny; images of the past, fragments of dreams float to its cold surface, adding to the woman's agony of consciousness. With self-knowledge comes acceptance of her mysterious destiny; this is the moment when waiting ends … Solemnly the attendant prepares her. As she advances to meet the unknown, the curtain falls (Programme note, 1944 in Sears 1982, pp. 27–8).²⁹

Noguchi's stage sculptures invoke a woman's retiring room or even a theatrical dressing room (Barnes 1990, p. 17). The tower of bones which represents the mirror is the 'sévère fontaine' (austere pool) of Mallarmé's poem, but for Graham it is also a metaphor for the artist's soul. Concealed within the depths of this bleak structure is a spark of affirmation, a small mobile object or [End Page 15] 'bird'.³⁰ Graham describes how the woman's dart walks are directed towards the mirror with its kernel of hope in the form of the bird, symbolising Herodias' heart, 'vibrating and exposed to life' (1991, p. 224).

The correlation between the dart, the mirror and the bird, provides a new field of signification, which infuses the deathly myth and Mallarmé's pessimism with a more optimistic tone. Graham states that the centre of the mirror retains a constant focus as the place where there is 'a living potential'. She explains how the potency of the bird symbol relates to the notion of the artist's quest for self-fulfilment:

Whenever I dance Herodiade,it was always to that animating force that I moved across the stage. It was for me the core of Isamu's heart as well, that part of each artist which, as he becomes as one with the sacrificial animal, he exposes to the world

Graham creates her own symbolic allusion to the artist and to life although she follows Mallarmé and redefines or 'redistributes' (Kristeva 1984) the signifying material from his self-reflexive but sensuous poetry. [End Page 16]

Conclusion

In this article, I have endeavoured to show that the dance, the music and the poem spring from pre-existing artistic and narrative codes. Each is already a space in which sign systems and artistic processes enter into a free play with one another. Graham's dance, for example, merges the archaic two dimensionality of her early movement vocabulary with the later development of characterisation and dramatic staging. The fusion of codes is generative in the production of the enigmatic heroine's dialogue with her sibylline attendant. But dance, music and poem are also constituted as intertexts because, as Worton explains, they are:

bound up with a host of other texts, some known and intended by the author, others known only by the reader and evoked as reference points by him or her as he or she engages in the process that is reading

In their exploration of the poem, Hindemith and Graham responded to different aspects of Mallarmé's poetic sensibility. Hindemith chose to extend what was perhaps a natural affinity between music and poetry as both art forms address the ear. An aspect of Mallarmé's poetic endeavour was to approximate music's shaping of sound and exploration of rhythm (Millan 1994). In turn Hindemith explored tonal variation and the reverberations of human dialogue. Graham's choreography engages the eye and, I suggest, echoes Mallarmé's description of the dancer he saw in the 1880s. He remarks on the ballerina's 'miraculous lunges and abbreviations' as a way of 'writing with her body' suggesting 'things which the written work could express only in several paragraphs of dialogue or descriptive prose'. The dancer's bodily configurations produced a poem 'written without the writer's tools' (1984, p. 112). Mallarmé realised that dance can tap into the pre-linguistic space of human thought and imagination and, that unlike poetry, it transcends the conventions of representation and rules of syntax. Dance produces the fluency of form and pliancy of meaning which Mallarmé constantly sought for his own artistic practice.³¹

Graham's motifs are a form of bodily inscription in the way that they etch enigmatic hieroglyphs against the visual field of the stage space in which the dancer moves. The motifs in Herodiade condense and distil a narrative of the mythological heroine's proud and unassailable nature, the emotional and spiritual journey she undertakes and the drama of her situation which generates the courage she needs to confront the fate that awaits her. The motifs, movement devices and stage sculpture in Graham's Herodiade exceed the merely representational. Their bodily forms are infused with a rich store of signification reflexive of dance itself and of the mythical heroine who permeates the poem, the dance and the music. The princess Hérodiade is an elusive and ineffable creation because she is a symbol for 'the poem itself, so difficult to seize and possess that the poet ultimately despairs of knowing it' (Fowlie 1953, p. 126). In this respect, it seems to me that Graham succeeded in her interpretation of the enigmatic heroine and in following Mallarmé by meditating on her own art form.

Hindemith's choice of Mallarmé's Hérodiade spurred Graham on to develop [End Page 17] a role of archetypal significance. The woman/Herodias was to become the crucible from which she forged the panoply of fearless heroines from Greek mythology, notably Medea in Cave of the Heart,Jocasta of Night Journey (1947) and the range of mythical characters portrayed in Clytemnestra (1958).

Herodiade was the first of these mythological masterworks, and those who have seen it remark on its power and dramatic intensity. I conclude with the words of Edwin Denby who reviewed the first performance of Herodiade in New York in 1945. He wrote of 'its sustained wonder, its amazing human power and sense of human knowledge' (1949, p. 321), qualities, which surely have contributed to the longevity of Herodiade and assured its place within Graham's repertory.