The Rhythm of Greek Oral Poetry: Prosody, Accentual Groups and Metrical Anomalies

Kelma! Kelma! Xi tkun inti?
Minn fejn ġejt? Meta tnissilt?
[…]
Ġejt kif ġejt, int l-isbaħ, l-akbar,
fost kemm hawn egħġubijet.

Word, O word! What are you? Where do you come from? How were you born? […] Anyway, you are the prettiest and greatest of all wonders.

Dun Karm (National Poet of Malta), Il-kelma 1–2, 5–6

§0. In October 2018, I applied to Harvard University’s Center for Hellenic Studies (henceforth CHS) with the project “The Rhythm of Greek Oral Poetry: Prosody, Accentual Groups and Metrical Anomalies”. The aim of this project was a study on rhythm through an application of linguistic theory on phonetics and phonology to Ancient Greek oral poetry, exploring the concept of prosodic word, namely on the basis of accentual groups and metrical anomalies, and developing a new approach to investigate the concept of word and how words fit into meters.

1. A matter of “word”

§1.0. In oral cultures a definition of “word” does not suit to the orthographic convention to leave a space. Linguistic anthropologists and ethnolinguists do know this issue very well, in fact they generally use or adapt the International Phonetic Alphabet (henceforth IPA) to graphically represent present-day languages (Duranti 1997:127); this is expressively evident in the case of clitics (marked by =) and tones (grave accents ` mark low tones, acute accents ´ mark high tones), such as in Tommo So (Mali) Àbíjàn=nɛ=wɔ́=gɛ Abidjan=OBL=be.REL=DEF “he who lives in Abidjan”, where a full relative clause in English is a single word in Tommo So (McPherson 2013:227).

§1.1. This fluctuating concept of “word” is, actually, well attested in Proto-Indo-European, where there are different words meaning “word”, which actually have a primary meaning connected with “speech, discourse, tale” or even “voice, sound, utterance”. This shows a concept of word-boundaries in their “acoustic” perception: “prosodic word”, namely a primary accented word that ‘holds’ one or several function words (clitics) with secondary or absent stress(es).

§1.2. Cognitive and neurosciences reveal that extralinguistic factors––context, intonation, syntax––permit the recognition of word boundaries (Kim, Stephens and Pitt 2012 and Ladd 2008:6–7). This is particularly evident in the case of second language learners, who have no clear prosodic and extralinguistic distinctions.

§1.3. In the triad (Turk and Shattuck-Hufnagel 2000) of (a) tune acquire, (b) tuna choir and (c) tune a choir, the samples are phonologically equivalent as /tjunəkwaɪə/. However, on the level of phonetics––I made IPA transcriptions on my native British English speaker––the samples show indeed three different durational patterns, respectively (a) [ˌtj̥unʔəˑˈkwaɪə], (b) [ˌtj̥unəːˈkwaɪə] and (c) [ˌtj̥unʔəˈkwaɪə], that reveal clearly understandable word boundaries for native speakers, not always clear for L2-English speakers. Turk and Shattuck-Hufnagel (2000:426–429) have demonstrated that word boundaries are connected to phonetic––not phonological––issues.

2. Meter, phonology and cliticization

§2.0. Bruce Hayes (1989:219) has shown that, in present-day English, intonational units (henceforth abbreviated as IU and in samples as [ ]_ι) are audible “because each one is aligned with a single ‘tune’ of the intonational system”, such as in the sentence [this is the cát]_ι [that caught the rát]_ι [that stole the chéese]_ι, which witnesses a rhythmic culminative stress (symbolized by an acute accent) on the final syllable of each IU (see Hayes 1995:24). The identification method for IU is based on syntactic and prosodic criteria.

§2.1. Blankenborg’s (2014:§4.4, §Appendix) studies on pause in Homer reveal that syntactic constrains––in simplified terms, the “meaning” of an utterance––create the conditions for underlying pauses, which according to him consist of caesurae (compare below, §4.1). However, his approach remains ‘philogistically’ classificatory, leaving apart a linguistic definition of (phonological) cliticization within a prosodic hierarchy (Anderson 2005:46 and Hayes 1989:208–211).

3. Silent voices: Punctuation in inscriptions

§3.0. As far as ancient languages are concerned, phonetics is inevitably and inescapably lost. Rhythmic features can be reconstructed through a combination of extralinguistic issues (script and epigraphic punctuation), linguistic issues (accentuation and cliticization), and meters in their rules and anomalies.

§3.1. In Mycenaean the use of word-dividers reflects a concept of “prosodic word” based mainly on accentual criteria, such as a tonic word and clitic(s) (Morpurgo Davies 1987:268). In syllabic script samples, comma (,) identifies word-dividers and the equal sign (=) marks cliticization.

§3.2. In PY Ep 704, Mycenaean e-u-ke-to-qe /e^huk^hetoi=k^we/ “and (s)he claims” ~ Greek εὔχεταί τε (enclisis accentuation); while da-mo-de-mi , pa-si /dāmos=de=min p^hāsi/ “but the community says (that) she” corresponds to (a) Greek δῆμος δέ μίν φησι (while according to Herodianos’ enclisis rules, δῆμός δέ μίν φησι, see Muscianisi 2010:7–8, 79–81) or to (b) δῆμος δέ μιν φησί (Herodianos, δῆμός δέ μιν φησί). Clitic athematic verb forms of φημί and εἰμί might be a secondary Ionic-Attic feature; while Mycenaean /p^hāsi/ (= φησί) seems a tonic word, thus as in (b).

§3.3. A classificatory method, typical of philologism, tries to regularize what is clearly born irregular (see Guarducci 1967:392). Rather, I intend to use a relativistic method to investigate punctuation in archaic Greece (see Lougovaya-Ast 2017:28–30, and previous literature), similar to the observational and descriptive method of cultural anthropology. The individual act of the inscriber must be taken into account, as in an archaic sacral decree from Argos (SEG 11:314, ca. 575–550 BCE, here abbreviated as S), where different intentions of punctuation appear.

§3.4. In alphabetic script samples, colon (:) shows the punctuation in the inscription with normalized text. Specific ‘linguistic’ intentions are: (a) accentual groups (compare Morpurgo Davies 1987), such as S.3 : καὶ τὰ χρήματά τε : “and the treasures too” and S.9 : δᾱμόσιον δέ : “but the state”; (b) deictics, such as S.5–6 : τοῖσι : χρήμασι : τοῖσι : χρηστηρίοισι : τοῖσι : τᾶς θεοῦ : “those treasures that are tools those of the goddess”; (c) “tabular account” (Cuomo 2013:269) in enumerations, such as S.5–10 (column A) Συλεύς : τε, καὶ : Ἐράτυιος, καὶ : Πολύκτωρ, καὶ : Ἐξάκεστος, καὶ : Ἁγίας, καὶ : Ἐρύκοιρος; here, punctuation breaks the accentual criterion of prosodic word (§3.4.a).

§3.5. A similar situation can be found in Cyprian syllabic, in decree ICS² 217.A.2–4 (Idalion, ca. 470 BCE). Here punctuation show an influence from alphabetic epigraphy (see Muscianisi 2010:53–56). Thus, the relativistic method allows to identify (a) the highlighting of deictics (§3.4.b), such as to-se , a-to-ro-po-se , to-se , i-ta-i , ma-ka-i , i-ki-ma-me-no-se /to(n)s ant^hrōpo(n)s to(n)s in=tāi mak^hāi īk^hmāmeno(n)s/ “those men (those) who were wounded in this battle”, and (b) univerbation of some strictly connected words (§3.4.c), such as to-no-na-si-ku-po-ro-ne-to-ni-ja-te-ra-ne /ton=Onāsikuprōn_ton=^(h)īātēran/ “the healer, son of Onāsikupros”.

§3.6. In verse punctuated inscriptions, the syntactic and accentual criteria fall within the metrical and musical ones. An octothorpe (#) before or after a word marks verse beginning or end respectively.

§3.7. In Nestor’s cup inscription (CEG 1.454, end 8^th century BCE, here abbreviated as N), and in an archaic Attic epitaph (CEG 1.26, ca. 540–530 BCE, here abbreviated as A), punctuation shows (a) emphasis of personal names in verse (§3.4.c), against accentual criteria (§3.2, §3.4.a), such as N.1 #Νέστορός : εἰμι and A.2–3 #Εὐκοσμίδης : δέ, #στήλην : δ(έ)’, highlighting the possessor of the cup (Nestor), the craftsman who made the stone (Eukosmides) and the epigram itself; (b) caesurae, and (c) hiatus (^H): N.1 εἰμι ^H εὔποτον (even though here syllabification may show a glide [ei.mi.ieu.pɔ.tɔn]) and N.3 καλλιστεφάνου ^H Ἀφροδίτης, a strictly syntactic and poetic phrase (name + epithet).

4. “Spin-off” research on the Oracles of Delphi and further activities

§4.0. The inductive classificatory method revealed its limitations; observational method and linguistic (phonological) theory give on one hand the freedom in analyzing within a clear frame. The methodological claims and some case studies presented above  acquire their importance in literary texts, where only linguistic and metrical criteria can be used for scientifically sound prosodic analyses.

§4.1. Caesurae and rhythms (meters) are the only ‘tools’ for prosody; syntax and ‘meaning’ for IU. Following the lead of Hagel (2004:12–13), who considers caesurae as musical and melodic factors, not syntactical; and after observing that clash in melody occurs very often even in present-day song and music, it might be considered that Ancient Greek poetry did witness some clashing diction or meter, on the basis of the composition in performance (Albert Lord).

§4.2. In Ancient Greek rhythm is strongly associated with the mousikē tekhnē and performance (Nagy 2010:373–374). This is particularly evident in orally-composed poetry that has not received the same fine-tuning of Homer or Hesiod, such as the Oracles of Delphi. The responses are quite important for a study of prosody and rhythm, because they witness the specific kairos of the request; there is a statistically high presence of metrical anomalies and clashing diction, due to an oral impromptu composition (as Rossi 1981 has suggested); their cognitive processes of composition per IU are reasonably open to debate.

§4.3. About the war against Xerxes before Thermopylae, the oracle to the Spartans is a good sample for prosody and IU of composition.

To you, O inhabitants of the wide-spaced Sparta: either the great famous city will be destroyed by the descendant men of Perseus or [3–4a] it certainly will not mourn as dead a king from the ancestry of Herakles, protector of Lacedaemon / [3–4b] it will mourn as dead a king not form the ancestry of Herakles, protector of Lacedaemon. Thus, the fury of bulls or lions do not hold him [Xerxes?] in front: in fact, of Zeus he has the might. I claim that he will not be hold earlier than one of these two will been slain totally in pieces.

§4.4. Setting apart the several awkwardnesses in meter and diction for a further research, the linguistically and prosodically based observation of cliticization and IU shows two different interpretations: verses 3–4a say that Delphi recommends fighting against Xerxes otherwise they will be destroyed; verses 3–4b say that Delphi gives no other solution than destruction by or subjugation to Xerxes (the king unrelated to the Heraklidai). Although verses 3–4b witness the absence of trochaic caesura (melodic asymmetry), the interpretation 3–4b fits into the historical support of Delphi to the Persians and its persuasion to following its lead through the oracles. Xerxes would linguistically be the hidden τόν ‘him’ (verse 5), which in interpretation 3–4a has no clear identification. This idea lay the basis for a “spin-off” research on a linguistically sound criterion in identifying the historical authenticity in responses (compare Muscianisi 2012).

§4.5. To conclude my report, I am in particular grateful to Gregory Nagy, Hanna and Joseph Roisman, and Soteroula Constantinidou for their comments and observations during my lecture. My 2019-20 Fellowship in Hellenic Studies has been enriched through the invitation at joining the Delphic Preview: Festival of the Muses, to which I contributed with the following paper:

• 2020.06.11. “The Circle of Fame: Apollo, the Corps de Ballet, and the Song of the Muses at Delphi.” Classical Inquiries. https://classical-inquiries.chs.harvard.edu/the-circle-of-fame-apollo-the-corps-de-ballet-and-the-song-of-the-muses-at-delphi/.

§4.6. Moreover, I had the possibility to continue my theater and ballet journalistic activity, merging it into CHS projects, such as James Baldwin’s celebrations and Black Classicists exhibition; and two articles on ballet performances at The Kennedy Center for Performing Arts:

• 2020.04.06. “When You Are in a Corner, Amen: Black Oedipus is called Sister Margaret” Kleos@CHS. https://kleos.chs.harvard.edu/?p=163703.
• 2020.03.16. “What a Combination! Balanchine + Ashton for The Washington Ballet.” Eastern Sirocco. https://easternsirocco.wordpress.com/2020/03/16/what-a-combination-balanchine-ashton-for-the-washington-ballet/.
• 2020.02.13. “Through Times, Lands, and Lives: Giselle by the American Ballet Theatre with Sarah Lane and Daniil Simkin.” Eastern Sirocco. https://easternsirocco.wordpress.com/2020/02/13/through-times-lands-and-lives-giselle-by-the-american-ballet-theatre-with-sarah-lane-and-daniil-simkin/.

Select bibliography

Anderson, S.R. 2005. Aspects of the Theory of Clitics. Oxford.

Blankenborg, R. 2014. Rhythm without Beat. Prosodically Motivated Grammarisation of Homer. Washington, DC. http://nrs.harvard.edu/urn-3:hul.ebook:CHS_BlankenborgR.Rhythm_without_Beat.2014.

CEG 1 = Hansen, P.A., ed. 1983. Carmina epigraphica Graeca saeculorum VIII–V a.Chr.n. Berlin.

Cuomo, S. 2013. “Accounts, Numeracy, and Democracy in Classical Athens.” In Writing Science: Medical and Mathematical Authorship in Ancient Greece, ed. M. Asper and A.-M. Kanthak, 255–278. Berlin.

Duranti, A. 1997. Linguistic Anthropology. Cambridge.

Guarducci, M. 1967 [1992]. Epigrafia greca. 1. Roma.

ICS² = Masson, O. 1983. Les Inscriptions chypriotes syllabiques. 2nd ed. Paris.

Hagel, S. 2004. “Caesura and Melody.” In Autour de la césure. Actes du colloque Damon des 3 et 4 novembre 2000, ed. F. Spaltenstein and O. Bianchi, 11–22. Bern.

Hayes, B. 1989. “The Prosodic Hierarchy in Meter.” In Phonetics and Phonology. 1. Rhythm and Meter, ed. P. Kiparsky and G. Youmans, 201–260. San Diego, CA.

———. 1995. Metrical Stress Theory: Principles and Case Studies. Chicago.

Kim, D., J.D.W. Stephens, and M.A. Pitt. 2012. “How Does Context Play a Part in Splitting Words Apart? Production and Perception of Word Boundaries in Casual Speech.” Journal of Memory and Language 66:509–529.

Ladd, D.R. 2008. Intonational Phonology. 2nd ed. Cambridge.

Lougovaya-Ast, J. 2017. “Some Observations on the Usage of Punctuation in Early Greek Inscriptions.” In Signes dans les textes, textes sur les signes: érudition, lecture et écriture dans le monde gréco-romain. Actes du colloque international (Liège, 6–7 septembre 2013), ed. G. Nocchi Macedo and M. C. Scappaticcio, 27–42. Liège.

McPherson, L. 2013. A Grammar of Tommo So. Berlin.

Morpurgo Davies, A. 1987. “Folk-linguistics and the Greek Word.” In Festschrift for Henry Hoenigswald on the Occasion of His Seventieth Birthday, ed. G. Cardona and N. H. Zide, 263–280. Tübingen.

Muscianisi, D.G. 2010. La parola metrica. Studio linguistico dei nessi ortotonica–appositiva attraverso il miceneo e i grammatici antichi. BA thesis. Università Cattolica del Sacro Cuore di Milano.

———. 2012. Esame linguistico e testuale degli oracoli delfici di età arcaica. MA thesis. Università Cattolica del Sacro Cuore di Milano.

Nagy, G. 2010. “Language and Meter.” In A Companion to the Ancient Greek Language, ed. E. J. Bakker, 370–387. Malden, MA. http://nrs.harvard.edu/urn-3:hlnc.essay:Nagy.Language_and_Meter.2010. 2nd ed.

PW 2 = Parke, H. W., and D.E.W. Wormell. 1956. The Delphic Oracle. 2. The Oracular Responses. Oxford.

Rossi, L.E. 1981. “Gli oracoli come documento di improvvisazione.” In I poemi epici rapsodici non omerici e la tradizione orale. Atti del convegno (Venezia, 28–30 settembre 1977), ed. C. Brillante, M. Cantilena and C. O. Pavese, 203–230. Padova.

Turk, A.E., and S. Shattuck-Hufnagel. 2000. “Word-boundary-related Duration Patterns in English.” Journal of Phonetics 28:397–440.

---