Teatro Before the Première : Recording the Performance of Ancient Greek Drama

Ancient Greek theatre, a multimedia spectacle (originally conceived for a unique performance) which involved words, music, gestures and dance, has always been a challenge for scholars investigating its original performance. This paper explores the possibilities of the performative elements of the plays to be recorded during their theatrical staging, that is, before their première. More in detail, it examines the probability that — given the rhythmic and melodic nature of ancient Greek language and the descriptive and/or perlocutionary character of the scenic information within the texts — the authors could inscribe music and gestural expressiveness into the linguistic code. The high level of ‘performativity’ implied in these ancient texts probably delayed the need for a technology that could record their different multimedia components.


Musical Theatre in Ancient Greece
F or several decades, scholars in Classical antiquity have described ancient Greek culture as a performance culture 2 . Indeed, long before the birth of drama (the last decades of the sixth century BC), any kind of poetry had been performed in front of an audience without the intermediation of writing. This situation persisted even when writing was gradually diffused throughout the Hellenic world, supporting -but not replacing -oral performance and aural listening what nowadays we label as 'literature'.
Thanks to the development of the theatrical genre (including tragedies, comedies and satyr plays), ancient poetic communication was soon tranformed into a much more impressive multimedia performance. The actions dramatized on stage by the actors and the chorus -whatever it was, a mythical, historical or utopian plot -2 were scenically realized through words, gestures, dance and music. By means of these tools, the playwrights could effectively represent dispositions, emotions and behaviours of the characters on stage. According to Aristotle, these were the main goals of any kind of poetic communication (see Aristotle Poetics 1447a 26 ff., quoted infra). It is therefore no coincidence that, when in the Renaissance opera first appeared, poets and intellectuals of the Florentine Camerata gave it an intellectual justification just invoking the Hellenic model 3 . Ancient theatrical performance was "a wonderful aural and visual experience," as still Plutarch -between the first and second centuries AD -describes it (On the Renown of the Athenians 5): as if to say, it was more like opera than modern prose theatre.
Tragedy and comedy -the two most important forms of drama -included choral odes called stasima, sung and danced to the accompaniment of the 2 Such an aspect of Greek culture (that has long been neglected) became extremely important for the renewal of modern philological studies, traditionally more closely linked to literary texts (which, despite being the main evidence of the ancient performances, certainly do not fully attest their complexities). Among the most important studies on the topic, see especially HAVELOCK 1963;GENTILI 1988;EDMUNDS-WALLACE 1997;GOLDHILL-OSBORN 1999. 3 Tragedies (and, in a certain way, satyr plays, as far as we know) were mostly based on myths, only occasionally on historical plots (as is the case of Aeschylus' Persians). Comedies, instead, albeit drawing on contemporary Athenian affairs, mainly developed -at least in what has survived -a fictional plot, even if we cannot deny the existence of mythological comedies (that, unfortunately, have survived only in fragments).
4 Ancient Greek tragedy gave to the Renaissance humanists (at the level of suggestion) an influential example of close interweaving of words and music 13 Revista do Laboratório de Dramaturgia -LADI -UnB -Vol. 5, Ano 2 Dossiê Som e Teatro for expression purposes. On this topic, see especially PALISCA 1985.

5
The aulos was a reed-blown pipe, almost always played in pairs, widely diffused in ancient Greece. The orchēstra was, literally, the space in which the chorus dance (from orcheomai, literally "to dance").
6 We know that parts of the original plays (like monologues, solos or excerpta of choral odes) could have well been re-performed in different contexts, for instance the symposium. But this did not imply that all the elements of the original performance (like dance or scenic gestures) were necessarily involved in the new performance.
aulos in the orchēstra, as well as actors' spoken episodes displaying a strong rhythmical beat based on the iambic metre (ROCCONI 2012: 215-218;ROCCONI 2013) 5 . Parts of the actors' roles could also be delivered as a kind of recitative called parakatalogē (that is, katalogē, i.e., recitation, "beside" or "along with" musical accompaniment, see MOORE 2010) or, more often, sung. Since approximately the mid-fifth century BC, in fact, tragic actors (and their parodic counterparts in comedy) started performing solo songs and duets with other actors, or the so-called amoebean songs in concert with the chorus (from ameibō, "to exchange"). The organological innovations in the construction of the aulos, equipped with a mechanism (i.e., metal collars closing and opening the numerous finger-holes) through which pipe-players could produce modulations during performance, led the composers to search for professional actors with growing canorous abilities, as the institution of a prize for the best actor in 449 BC well attests. These and other factors contributed to the development of a more complex and mimetic music style.
The new expressive codes, together with the actorial gestures and the scenographic settings, gave rise to a highly multimedia performance, originally conceived not for being replicated (not, at least, in the first instance), but for winning a competition within the religious festivals in honour of the god Dionysus: the Dionysia (comprising the Rural Dionysia and the City Dionysia) and the Lenaia, which took place in different parts of the year (PICKARD-CAMBRIDGE 19882 3 ). During these festivals, the poets selected by the archon participated in the competition, always with new dramas.
Within this context, it is clear that the need for recording 'afterwards' the performance of shows originally conceived as one-off events was not a priority, at least regarding the possibility to replicate them in all their different components 6 . The memory of these performances after the theatrical staging was fixed by official and public epigraphic inscriptions that took into account only practicalities, such as the playwrights, the winners and the sponsors (cf. MONTANA 2016), i.e., the choregoi, wealthy Athenian citizens who, appointed by the archon, assumed the public duty of financing some aspects of the dramatic productions (on choregia, see especially WILSON 2000).
This, however, does not exclude that, before going on stage (that is, before the 'première'), some of the performative components of the theatrical event might well have been fixed or recorded, at least for compositional or staging needs. Firstly, we should take into account the poetic text, which we could also call libretto: once the funding had been obtained, the poet had to refine in some ways the preliminary text and, then, teach it to the actors and the chorus (this latter being instructed by the author himself or by a chorodidaskalos). Secondly, we 7 We can identify scenic scripts of different sorts, firstly in some Hellenistic and early Roman Imperial papyruses containing the text of mimes (anonymous performances sung and recited by actors, cf. GAMMACURTA 2006) with occasional indications suggesting the accompaniment of musical instruments (like drums or pipes); secondly, in papyruses of tragic texts with musical scores, as the famous Vienna Papyrus G 2315 containing seven lines of the first choral ode of Euripides' Orestes (PÖHLMANN-WEST 2001: 12-17). This latter document, however (to be dated around 200 BC, at least two centuries later than the first mise en scène of the Euripidean drama), should not be read as the original script of the play, but as a later document attesting the performance of a virtuoso. Indeed, Hellenistic tragōidoi often performed as soloists extracts of older plays, sometimes also providing new music (see PRAUSCELLO 2006).
8 We have the complit set of melodic symbols (consisting of letters of the Greek alphabet, variously modified) only thanks to late works like Alypius' Eisagōgē mousikē (approximately fourth century AD). The earliest fragments, however, are from the Hellenistic period.
9 The possibility that some rhythmical values deviated from the metrical scheme of the text has been (and is still) a debated topic among scholars. Particularly difficult to interpret are those verses which do not show a perfectly corresponding responsion between strophe and antistrophe (as, for instance, cretic ~ iambus: -U-~ U-U-), where the missing "time" (chronos) could have easily been provided by the music, cf. Fragmenta Neapolitana 21 (remarks based on Aristoxenian material): "Keeping the same rhythm does not mean keeping should consider the possibility that the music has been partially written down since earlier times (this is the hypothesis of Pöhlmann 1988, who argues, however, that these preliminary scores were only drafts). Finally, we should consider other essential elements of the show, like the choreography of the chorus' members who danced in lines (4x3 o 5x3 in tragic choruses, consisting of twelve or fifteen men, 6x4 in comic choruses, consisting of twenty-four men) and the scenic movements of the actors (such as entrances and exits from the skēnē door or from the lateral entrances called eisodoi; body and hand gestures to compensate for the lack of facial expression, due to the presence of the mask; and so on).
But how could ancient technology support the need for 'recording' these multimedia components?

Before the Première: Recording Multimedia
One could start from some general comments on the poetic text. The length of ancient dramas (approximately from 1000 to 1700 lines) certainly required the script writing. We should, then, expect the existence of an autograph and of some scenic scripts used for the mise en scène. Unfortunately these documents have been completely lost. The textual tradition thanks to which some theatrical works have eventually survived to this day does not directly derive from such a type of documents: they most probably stemmed from consumable copies contemporary to the author (on which, however, no music has ever been recorded) 7 .Furthermore, also the melodies accompanying the lyric parts of these texts are substantially lost. In fact, we do not know for certain when a system of musical notational signs was fully elaborated by the Greeks. It is plausible to assume (following HAGEL 2010) that it was initially conceived in the fifth century BC, as if to say, in the age of the greatest Hellenic tragic poets: Aeschylus, Sophocles and Euripides. 8 But the musical element that, more than any others, left a trace on the poetic texts and that can be potentially investigated is rhythm. In ancient Greek language, the syllabic quantities (long vs. short) were clearly detectable and had a distinguishing value. Musical rhythm and prosodic quantities were strictly linked, the alternation of long and short syllables of the text being the rhythmical basis for musical scores, albeit melodic rhythm could occasionaly vary the prosodic lengths (ROCCONI 2008). Therefore, just by virtue of the deep roots of musical rhythm in the logos, we are still able to identify which parts of the dramas were once recited (i.e., the iambic trimeters, more closely reproducing -according to Aristotle's Poetics 1449a 24 f. -the ordinary pitch of conversation) and which parts were sung (i.e., the lyric verses), also establishing their main rhythmical patterns. 9 There is another aspect of ancient Greek language, however, which allowed the (at least partial) recording of some performative aspects of Classical dramas. Thanks to the pitch accent, syllables carrying the acute accent ( / ) were in fact pronounced at a higher pitch than the others, syllables carrying the grave accent ( \ ) returned to the original intonation, while syllables carrying a circumflex accent ( ^ ) showed a rise and a fall in pitch (DEVINE-STEPHENS: 172 f.). 10 The main evidence for reconstructing the functioning of the melodic movements of Greek speech comes from the musical theorist Aristoxenus of Tarentum (late fourth century BC): "While every vocal sound can move in the manner mentioned, there are two forms of this movement, the continuous and the intervallic. In the continuous form the voice seems to perception to traverse a space [topos] in such a way as never to stand still even at the extremities themselves [...] whereas in the other, which we call intervallic, it seems to move in the opposite way. During its course it brings itself to rest at one pitch and then at another: it does this continuously (I mean continuously in respect of time), passing over the spaces bounded by the pitches, but coming to rest on the pitches themselves and sounding them alone, and is described as singing, and as moving in intervallic motion [...] We say that continuous movement is the movement of speech, for when we are conversing the voice moves with respect to place in such a way that it seems never to stand still." (Aristoxenus Harmonic Elements 13,8 ff., transl. in BARKER 1989: 132 f.) But what happened when words were set to music for the purpose of creating song? From the analyses of some surviving scores of Greek vocal music (none of which earlier than the late third century BC) it emerges that, more times than not, musical melodies tended to follow the pitch accent. Some ethnomusicological parallels confirm that, in any language with similar characteristics, the tones of the words do usually have an influence on shaping the melody (cf. NETTL 1973: 138 f.). A problem arises when the words that were set to music are in strophic responsion, as if to say, when two series of verses are rhythmically symmetrical. In fact, in ancient lyric poetry, strophe (lit. "turn") and antistrophe (lit. "counter-turn") were groups of verses with the same rhythmic structure and, as it is generally assumed, the same melody. This arrangement of verses had probably the scope to facilitate choral dance, since during the antistrophe the the same count of vowels and syllables. It depends on the chronoi, bidding us lengthen some of them and shorten others, and equalize some that were unequal. It can do this while the syllables and the letters remain unchanged" (transl. in PEARSON 1990: 31).
10 In the first century BC, Dionysus of Halicarnassus (probably paraphrasing Aristoxenus) points out that the maximum interval that the melody of spoken language could cover was a fifth (On Literary Composition 11,73 ff., transl. in USHER 1985: 77): "Now the melody of spoken language is measured by a single interval, which is very close to that which is called a fifth. When the voice rises towards the acute, it does not rise more than three tones and e semitone; and when it falls towards the grave, it does not fall from this position by more than this interval." technique of singing see http://www. oeaw.ac.at/kal/sh/index.htm, where Dr. Stefan Hagel performs Homeric verses guided by the accentual structure and sentence-intonation of ancient Greek language.
12 Particularly interesting is the comparison (for which I thank Giovanni Giuriati) with the Cambodian theatre. The scripts of the Royal Ballet of Cambodia, the dominant genre of dance theatre in the country, contain only the texts and some general indication of performance, certainly not the scores (GIURIATI 2003). This fact confirms that the interpretation of musical notation as a 'prescriptive' tool is an idea that is completely extraneous to oral cultures. (10, 453c-f, transl. in OLSON 2009: 171-173): "Callias of Athens was a bit earlier than Strattis and wrote the so-called Literal Review, which he organized as follows. [...] His chorus consisted of women who represented pairs of letters and sang in meter, in a lyric style, in the following way: bōta alpha ba, bōta ei be, bōta ōta bō, bōa iōta bi, bōta ou bo, bōta u bu, bōta ō bō, and likewise in the antistrophic portion of the song and the meter: gamma alpha ga, gamma ei ge, gamma ēta gē, gamma iōta gi, gamma ou go, gamma u gu, gamma ō gō, and so on similarly through each of the other syllables, all of which have the same metrical and lyrical structure organized in antistrophic form. Euripides is accordingly not only suspected of having composed his entire Medea using this as his source, but has also patently borrowed the song itself (to melos auto).

D' Angour hypothesis is based on the interpretation of a passage in Athenaeus' The Learned Banqueters
[...] As a consequence, apparently, all the other (poets) adopted antistrophic chorus reversed the movement performed during the strophe. The strophic pair was, then, often followed by an epode of different rhythmic structure, realizing the triadic structure -strophe, antistrophe, epode -commonly used in much choral lyric poetry. Since there was no strict responsion of words accents between the text of the strophe and that of the antistrophe, it is unlikely that the same melody could match words accent in responding metrical structures. Some scholars have hence assumed that, in choral odes of ancient Greek drama -always arranged in strophic pairs -, the melody did not take the words accent into account (COMOTTI 1989).
An alternative solution has been recently advanced by D'ANGOUR 2006. This scholar suggests to extend to all the lyric poetry of the Archaic and Classic periods the practice suggested long ago by Martin West, as far as the aedic improvisation is concerned (cf. WEST 1981: according to him, the syllables of the epic verses were disposed over the limited set of fixed notes available on the phorminx, the stringed instrument of ancient bards, the rise and fall of the voice being governated by the melodic accents of the words). 11 Ancient Greek poets would have then created their melodies or accomodated traditional regional formulas (the so-called nomoi and harmoniai) just following the natural accents of the words. As a consequence, musical composition -and transmission -of lyric poetry would have been smoother and more natural, with no need for the poets to record melodies through writing. 12 A great novelty would have been later introduced by Euripides, who firstly in the Medea (performed in 431 BC) would have realized the melodic identity of strophe and antistrophe paying no attention to words accents and, thereby, innovating a centuries-old musical practice. 13 For further evidence, see the remarks of Dionysius of Halicarnassus on Euripides' disregard of the melodic flow naturally suggested by words accents: "Music requires that the words should be subordinate to the melody, and not the melody to the words. Many lines of verse illustrate this, but none better than the lyric which Euripides makes Electra address to the Chorus in the Orestes (408 BC): σίγα σῖγα, λευκὸν ἴχνος ἀρβύλης τίθετε, μὴ κτυπεῖτ'· ἀποπρὸ βᾶτ' ἐκεῖσ', ἀποπρό μοι κοίτας. In these lines the words σῖγα σῖγα λευκόν are sung on one note; and yet each of the three words has both low and high pitch." (On Literary Composition 11,93, transl. in USHER 1985: 79-81) If it were possible to prove beyond any doubt this hypothesis, we could easily explain not only why ancient sources insisted so much on the revolutionary songs into their tragedies from this source." The date of Callias' play is controversial and has often been connected (since PÖHLMANN 1971, see now GAGNÉ 2013 to the adoption of the Ionic alphabet in Athens in 403 BC. The reliability of the information given by Athenaeus in this passage remains the subject of much debate among scholars.
15 On this expression (which seems to imply that orchestic schēmata were conceived as crystallized postures which helped to understand the underlying storytelling in virtue of their mimetic nature), see now Rocconi 2017. impact of the Euripidean musical language (shared by other late fifth-century authors, like the tragedian Agathon, the dithyrambographers Melanippes and Cinesias, the kitharodes Phrynis, Philoxenos and Timotheus, all labelled by modern scholars as practitioners of the so-called 'New Music'). 14 We could also say why, in the late fifth century, the need for a technique that would enable the recording of these authorial melodies became more urgent. Neverthelesss, the creation and diffusion of musical notation was a slow and gradual process. Until replicas of earlier theatrical plays were not fully established (thanks to the introduction, from 386 BC, of palaia dramata into the competition at the City Dionysia), a repertoire of dramas was not set up. Consequently, writing music was not considered absolutely necessary beyond the specific environments of professional performers. I would like, now, to conclude my inquiry on the possibilities of recording the performative components of Greek theatrical events in the phase preceding the first official staging with some remarks on dance and scenic movements.
In ancient Greece, dance was conceived as a mimetic art included in the broader notion of mousikē technē. In the Poetics (1447a 26 ff.), Aristotle describes dance as a mimēsis of specific states of the characters, experiences and actions realized through pure rhythm: "rhythm alone, without melody, is used by the art of dancers, since they too, by means of figured rhythms (dia tōn schēmatizomenōn rhythmōn), 15 imitate characters (ēthē), emotions (pathē) and actions (praxeis)." In Greek drama, chorus was always a collective character involved in action (Aristotle Poetics 1456a 25 f.) who sang and danced during the stasima, while it was the chorus-leader (called koryphaios) who dialogued with actors during the episodes. But which were the specific qualities of tragic and comic dances (respectively called emmeleia and kordax) and how, or if, have they ever been 'recorded'?
Ancient sources do not provide information on such an important component of theatrical performance. In The Learned Banqueters, Athenaeus of Naucratis (second-third century AD) provides some clues about choreographers, who were the same playwrights, sometimes acting even as trainers: "Aeschylus [...] also created many dance-steps (schēmata orchēstika) himself and passed them on to the members of his choruses. Chamaeleon, at any rate, says that he was the first to arrange the dances, and that he did not use special trainers (orchēstodidaskaloi), but worked out the dance-steps for his choruses himself and generally took on the entire management of the tragedy. Most likely, therefore, he acted in his own plays. Aristophanes, at any rate -there is credible information about ERCOLANI 2000: 24). According to this terminology -applied to ancient theatre -, the adjective 'performative' refers to an utterance that imposes to realize the action it describes: in other words, it has a 'perlocutionary' function.
17 Also the beating of the breast was a sign of mourning: hence the name given to the kommos (from koptō = "I strike" the breast), a mourning song that the chorus and the actor sang together. Greek dance was a highly codified way to express and underline the meaning of the words that the chorus sang during the odes, whose content was symbolically represented (so it seems) by hand gestures. But, unfortunately, we have no information on the possibility to 'record' these ancient postures (schēmata). We only know that the training of the chorus -whose members, unlike actors, never became professionists -was long and expensive, involving the main part of the expense in theatrical productions, and that it was financed by the choregos (cf. supra).
As far as the scenic movements and actorial gestures are concerned, their rudimentary recording was, in a certain way, inserted within the poetic text. Besides providing a set of spatial and temporal information on theatrical setting and character identity, textual indications sometimes describe the actions on stage (ERCOLANI 2000). We may interpret them as scenic suggestions inserted by the poet for actors, to provide them with staging indications that could serve as 'performative utterances' (adopting a terminology borrowed from Austin's 'speech acts theory'). 16 But they could have also been inserted simply for helping the audience to better decode what happened on stage, that is, with a pure informative function. If a tragic character was crying on stage, for example, spectators could not see tears flowing down his face because of the mask, that covered any facial expression (MEDDA 2013). Hence interjections (e.g.: aiai) or verbal description of weeping, preferably made by another character and sometimes enhanced by body gestures like covering the head by a veil or keeping the eyes cast down, 17 had the scope to compensate for the lack of realism on stage, drawing attention to important scenic events. The announcement of exits and entrances of characters from the scene had the same information purposes (TAPLIN 1977;HALLERAN 1985), even if sometimes it was realized for more specific needs. 18

After the 'Première': Towards the Supremacy of the Text
The description of the many components of ancient theatrical performance has allowed me to point out the high level of 'performativity' inscribed within the dramatic texts. This is probably the reason why, in ancient Greek culture, there has never been need of a specific technology supporting the recording of these components.
But what happened to the plays after their first participation in the dramatic competitions? How has the memory of these multimedia events been preserved over time?
The traces left on contemporary or subsequent evidence (both textual and iconographic) are numerous and their routes quite complex to detect (for a specific treatment of these issues, see MONTANA 2016). What, nowadays, we know for certain is that the surviving texts of ancient Greek dramas cannot fully reveal the complexity of their original performance, since they are unable to convey their multimedia components. Language (logos) is no longer "embellished" (hēdysmenos) -as Aristotle defines it in the Poetics (1449b 24 ff.) when it is associated with rhythm and melody -when it is 'recorded' through writing.