I am most grateful to Roger Parker for his help and patience with this article, and to Ian Cross for his invaluable comments and advice. Thanks also to John and Polly for introducing me to Herzog's films, and to Dom for watching them all (many times) with me.Google Scholar

1 Herzog directed Lohengrin in 1987. Other productions include Mozart's The Magic Flute (Teatro Bellini, Catania, 1991, conducted by Spiros Argiris; in 1999 he directed another production in the same theatre, this time with conductor Zoltan Pesko) and four productions of Wagner's Tannhäuser (Teatro de la Maestranza, Seville, conducted by Klaus Weise (1997); the Teatro di San Carlo, Naples, conducted by Gustav Kuhn (1998); the Teatro Massimo, Palermo, conducted by John Neschling (1998); the Baltimore Opera Company, conducted by Christian Badea (2000)). During and after these theatre productions, Herzog directed two films explicitly concerned with opera and musical performance. The first, The Transformation of the World into Music (1994), is a 90-minute documentary that focuses on the production techniques behind the scenes at Bayreuth, following rehearsals for three different opera productions. The second, Pilgrimage (2001), is a ‘film-opera’ collaboration with John Tavener.Google Scholar

2 The tale of El Dorado came from Cundinamarca, the ‘Land of the Condor’, now the Andean highlands of the Republic of Colombia, 7,500 feet above the sea. It refers to the ceremony of an Indian tribe, who reconsecrated their king every year by covering his body with resin before he bathed in a lake filled with gold offerings. He emerged from this lake a ‘gilded man’. The legend, first told to the conquistador Sebastián de Benalcázar by local Indians, gradually mutated from El Dorado the gilded man to El Dorado the imaginary kingdom of gold. For a more detailed account of the myth, see Silverberg, Robert, The Golden Dream: Seekers of El Dorado (Ohio, 1967; repr. 1985).Google Scholar

3 The Popol Vuh, believed to be the equivalent of the Egyptian Book of the Dead, was written by the K'iché people after the Spanish Conquest (AD 1524). However, the entire epic was probably formulated during the Late Preclassic period (prior to AD 250). The manuscript, written in codices, was discovered in the nineteenth century: see Popol Vuh: The Definitive Edition of the Mayan Book of the Dawn of Life and the Glories of Gods and Kings, trans. Dennis Tedlock (New York, 1985).Google Scholar

4 Fiachra Gibbons, ‘Herzog Admits he Tried to Kill Kinski’, Guardian (21 May 1999), 24.Google Scholar

5 A threat to which Kinski replies in his recently reissued autobiography: ‘whoever heard of a pistol or rifle with nine bullets?‘ Klaus Kinski, Kinski Uncut (London, 1989), 98.Google Scholar

6 Gibbons, ‘Herzog Admits he Tried to Kill Kinski’.Google Scholar

7 Quoted in Janet Maslin, ‘With Friends Like These …: The Story of Kinski and Herzog’, New York Times (3 November 1999), 12.Google Scholar

8 Stephen Daniels and Denis Cosgrove, ‘Introduction’, The Iconography of Landscape: Essays on the Symbolic Representation, Design and Use of Past Environments, ed. Daniels and Cosgrove (Cambridge, 1988), 1–10 (p. 1).Google Scholar

9 Thomas Parkhill, The Forest Setting in Hindu Epics: Princes, Sages, Demons (Lewiston, Queenston and Lampeter, 1995), 1.Google Scholar

10 Pedro Maligo, Land of Metaphorical Desires: The Representation of Amazonia in Brazilian Literature (New York, 1998), 2.Google Scholar

11 Douglas Davies, ‘The Evocative Symbolism of Trees’, The Iconography of Landscape, ed. Daniels and Cosgrove, 32–42 (pp. 38–9).Google Scholar

12 For discussion of music in Hollywood film, see Gorbman, Claudia, Unheard Melodies: Narrative Film Music (Bloomington, 1987); Caryl Flinn, Strains of Utopia: Gender, Nostalgia and Hollywood Film Music (Princeton, 1992).Google Scholar

13 Quoted in Herzog's 1999 film, My Best Friend: Klaus Kinski.Google Scholar

14 Such an uneasy response to stillness has been exhaustively explored by film critics, who have repeatedly acknowledged that it is the combination of motion with the appearance of forms which provides film with the illusion of reality absent in many other visual art-forms. According to Christian Metz, ‘it is movement that produces the strong impression of reality’. Drawing on Albert Michotte van den Bercks, Metz suggests that motion imparts corporeality to objects and gives them a reality that their still representations could not have. Christian Metz, Film Language: A Semiotics of the Cinema, trans. Michael Taylor (New York and Oxford, 1974), 7. Laura Mulvey discusses the relationship between spectacle (stasis) and narrative (movement) in her article ‘Visual Pleasure and Narrative Cinema’, The Sexual Subject: A Screen Reader in Sexuality (London and New York, 1992), 22–34. See also Rose Theresa, ‘From Méphistophélès to Méliès: Spectacle and Narrative in Opera and Early Film’, Between Opera and Cinema, ed. Jeongwon Joe and Rose Theresa (New York and London, 2002), 1–18; Tom Gunning, ‘The Cinema of Attractions: Early Film, its Spectator and the Avant-Garde’, Viewing Positions: Ways of Seeing Film, ed. Linda Williams (New Brunswick, 1994), 114–33. Brigitte Peuker discusses Herzog's static images in her article ‘Literature and Writing in the Films of Werner Herzog’, The Films of Werner Herzog: Between Mirage and History, ed. Timothy Corrigan (New York and London, 1986), 105–17 (p. 105).Google Scholar

15 Koepnick, Lutz P., ‘Colonial Forestry: Sylvan Politics in Werner Herzog's Aguirre and Fitzcarraldo‘, New German Critique, 60 (1993), 133–60 (p. 140).Google Scholar

16 Dana Benelli, ‘The Cosmos and its Discontents’, The Films of Werner Herzog, ed. Corrigan, 89–103 (p. 94).Google Scholar

17 As Béla Balázs has pointed out, ‘we accept seen space as real only when it contains sounds as well, for these give the dimension of depth’. Béla Balázs, Theory of Film: Character and Growth of a New Art (New York, 1970), 207. However, for some critics, the sounds that accompany images do not have to be realistic. Theodor Adorno and Hanns Eisler (Composing for the Films, London, 1947) argue that music offers more than the speaking voice: in their view, the talkie without music is not very different from a silent movie. For more discussion, see Royal S. Brown, Overtones and Undertones: Reading Film Music (Berkeley, 1994).Google Scholar

18 Walter Benjamin, ‘The Work of Art in the Age of Mechanical Reproduction’, Illuminations (Glasgow, 1979), 219–53.Google Scholar

19 ‘Diegetic’ traditionally refers to film sounds whose source is visible on screen: conversely, non-diegetic sound (most often referred to as the ‘soundtrack') has no footing in the image. The term ‘diegesis’ was coined by the French writer Étienne Souriau; for a more detailed description, see Metz, Film Language, ix.Google Scholar

20 Quoted in Stephen Minta, Aguirre: The Re-Creation of a Sixteenth-Century Journey across South America (London, 1993), 10–11, 196.Google Scholar

21 The account, although technically anonymous, is thought to be by the soldier Custodio Hernández.Google Scholar

22 Although a great Peruvian traitor, Pizarro led an expedition 25 years before Aguirre's journey, during which he successfully seized Peru, where he ruled from 1544 until 1548, when he was finally caught and executed.Google Scholar

23 Peuker, ‘Literature and Writing in the Films of Werner Herzog’, 107.Google Scholar

24 During the explorations of the fifteenth and sixteenth centuries, such methods of conquest were not at all uncommon. Indeed, the historian John Brian Harley has gone so far as to suggest that maps and journals have been the weapons of imperialism as much as guns and warships: ‘insofar as maps were used in colonial promotion, and lands claimed on paper before they were effectively occupied, maps anticipated empire’ (‘Maps, Knowledge, Power’, The Iconography of Landscape, ed. Daniels and Cosgrove, 277–312 (p. 282)).Google Scholar

25 A number of possible reasons can be offered for such interpretative stasis. First, the area encouraged less interest during the seventeenth century owing to its geographical position on the periphery. Concerned more with their immediate surroundings, the seventeenth-century revolutions aimed to eliminate magical or superstitious reasoning from their European doorstep before looking to the relatively unexplored ‘other’ world. Second, when attention was finally focused on the Peruvian rainforest, the strong, ingrained belief structures of its inhabitants were more developed and resilient than many of those encountered before. See Godfrey and Monica Wilson, The Analysis of Social Change (Cambridge, 1945), 95.Google Scholar

26 Early in the quest, Hernán Pérez de Quesada discovered El Dorado, a lake called Guatavitá on the Bogotá plateau. In 1540, he became the first of many to drain the lake and recover the golden objects that had been cast into it. Nevertheless, the amount of gold did not conform to his hopes, yielding only 500,000 pesos de ora in all. He thus persuaded himself that Guatavitá was not the true lake of gold, which must lie somewhere east of the plateau: the search continued accordingly, with a steady locational shift towards the Amazon basin. Sir Walter Raleigh and Columbus later joined the pursuit. For a more detailed account of the search, see Silverberg, The Golden Dream.Google Scholar

27 Mary Louise Pratt has shown how Europe has depicted South America as mere ‘nature’, an unclaimed and timeless space. Imperial Eyes: Travel Writing and Transculturalism (London and New York, 1992), 126.Google Scholar

28 Roland Barthes has interpreted performative action as follows: ‘the fact is that writing can no longer designate an operation of recording, notation, representation, depiction; rather it designates exactly what linguists call a performative, a rare verbal form in which the enunciation has no other content than the act by which it is uttered’. Barthes, ‘The Death of the Author’, Image, Music, Text, trans. Stephen Heath (London, 1977), 142–8 (p. 145).Google Scholar

29 The opera scene was directed by Werner Schroeter.Google Scholar

30 Over the short span of cinema's life, the two disciplines have experienced a certain symbiosis, ranging from filmic adaptations of entire operas, with plot, unrealistic setting and operatic gesture kept intact, to snippets of Italian operatic favourites in contemporary Hollywood. The collaboration was particularly prolific during cinema's early years: more than 150 opera-related titles were produced before 1926. Figures quoted in Theresa, ‘From Méphistophélès to Méliès’, 7. See also Jeremy Tambling, ‘Film Aspiring to the Condition of Opera’, Opera, Ideology and Film, ed. Tambling (Manchester, 1987), 41–67.Google Scholar

31 Daniels and Cosgrove, ‘Introduction’, 1.Google Scholar

32 Operatic films often translated theatrical pieces in their entirety, at times even preserving unrealistic settings rather than exploring the greater latitude in location available to screen opera (locations can be indoors, outdoors or animated). See Noël Burch, Life to Those Shadows, trans. Ben Brewster (Berkeley and Los Angeles, 1990); Marcia J. Citron, Opera on Screen (New Haven and London, 2000).Google Scholar

33 Acousmêtre forms a linchpin of Chion's theories of sound in film: it refers to a voice in theatre or cinema that is ‘off-stage’, whose audibility does not depend on seeing. In his earlier writings, descriptions seem clear-cut: if the source of the sound is not visible, then the sound is an acousmêtre, although the source must have ‘one foot’ in the image. An acousmêtre typically moves from an acousmatic to a normal state, or vice versa: thus a sound is acousmaticized or de-acous-maticized. However, in later writing (such as the ‘Epilogue’ to The Voice in Cinema) Chion admits that he had overgeneralized. What happens when a character speaks with back turned, offering a heard voice without moving lips? Chion explores the acousmêtre at length in ‘Part I: The Magic and the Power of the Acousmêtre’, The Voice in Cinema, trans. Claudia Gorbman (New York, 1999), 17–57. Speaking about disembodied sound in her article ‘Debussy's Phantom Sounds’, Cambridge Opera Journal, 10 (1998), 67–96, Carolyn Abbate suggests that acousmatic sound creates a sense of mystery that resists being caught up in interpretation: she also uses the example of God's authority, which is predicated on the presence of his voice in the absence of his body.Google Scholar

34 There has been much discussion concerning recorded sound in film. Balázs contends that ‘what we hear from the screen is not an image of the sound but the sound itself, which the sound camera has recorded and reproduced again… . There is no difference in dimension and reality between the original sound and the recorded and reproduced sound, as there is between real objects and their photographic images.’ Opposing this argument, Rick Altman, Alan Williams and Thomas Levin maintain that recorded sound's fidelity to the original is an illusion: if recorded sounds were reproduced in a different acoustic space – not in a street but inside a theatre – it would constitute a different sound. Williams contends that every sound, whether original or reproduced, is unique since it is ‘spatio-temporally specific': every sound is historical in that every sonic event is inseparable from the time and space in which that event is made. Quoted in Jeongwon Joe, ‘The Cinematic Body in the Operatic Theatre: Philip Glass's La belle et la bête’, Between Opera and Cinema, ed. Joe and Theresa, 59–73.Google Scholar

35 Freud has discussed the effect of a figure meeting itself: the person is apt to take an instant dislike to the image as it seems like a ghost and conjures up the spectre of his/her own death. See Freud, Sigmund, ‘The Uncanny’, Art and Literature: Jensen's Gradiva, Leonardo da Vinci and Other Works, trans. and ed. James Strachey and Albert Dickson, The Penguin Freud Library, 14 (Harmondsworth, 1985), 334–76.Google Scholar

36 For discussions regarding synchronization as loss of intimacy between sound and image, see Adorno and Eisler, Composing for the Films. Slavoj Žižek also renounces the possibility of united sight and sound in film in ‘I Hear You with my Eyes; or The Invisible Master’, Gaze and Voice as Love Objects, ed. Renata Salecl and Žižek (Durham, 1996), 92. For other accounts of the loss inflicted on film with the achievement of synchronized speech, see Arnheim, Rudolf, Film as Art (Berkeley, 1957; repr. 1966); Sergei Eisenstein, Film Form and The Film Sense, trans. and ed. Jay Leyda (New York, 1957); Rick Altman, ‘Moving Lips: Cinema as Ventriloquism’, Yale French Studies, 60 (1980), 67–79. In Music and Silent Film (New York and Oxford, 1997), 42, Martin Marks discusses the seeming contradictions of film based on opera during the era before standardized synchronization of sound.Google Scholar

37 As Abbate points out, it is often only in moments of narration (often the aria) that a particular character can become aware that they are singing: whilst Fitzcarraldo changes the semantic function of music, creating pure music from song, he maintains the self-awareness of its original dramatic setting. Carolyn Abbate, Unsung Voices: Opera and Musical Narrative in the Nineteenth Century (Princeton, 1991).Google Scholar

38 See, for example, Warhol's Sleep and Kiss, in which images have no subjectivity or psychological depth.Google Scholar

39 Fitzcarraldo was filmed in English and dubbed into German. Although the dialogue appears more realistic in the English version, Herzog prefers the German-dubbed version.Google Scholar

40 See Abbate, Unsung Voices. Peter Conrad, speaking about the operatic voice, contends that ‘one of the bequests of film to opera is its demonstration that song is soliloquy, not overt statement: that the voice is consciousness – or the yearning subconscious – overheard’. Conrad elaborates on the soliloquy through a discussion of the love duet in Jean-Pierre Ponnelle's film Madame Butterfly (1974), in which the characters occupy two different locations. A Song of Love and Death: The Meaning of Opera (St Paul, 1996), 273.Google Scholar

41 Michal Grover-Friedlander, ‘“There Ain't No Sanity Claus!”: The Marx Brothers at the Opera’, Between Opera and Cinema, ed. Joe and Theresa, 19–37 (p. 26).Google Scholar

42 Benjamin, ‘The Work of Art in the Age of Mechanical Reproduction’.Google Scholar

43 Metz, Film Language, 7.Google Scholar