[Authors and titles are listed at the end of the review]
The centrality of speech to epic has long been recognised and in this collected volume both the Introduction and de Jong’s opening study on vocatives start by citing the Iliad’s famous injunction that a hero must be “a speaker of words and a doer of deeds” (Hom. Il. 9.443). Each contribution draws on the Digital Initiative for Classics: Epic Speeches (DICES) database, a corpus containing 5000 speeches from Homer to the sixth century CE. As the book is part of the DICES project and explicitly aims to “[offer] new possibilities for research” (p. 4), DICES is worth some initial consideration.
A constructing principle of DICES is to show the evolution of epic, so seminal texts such as Homer and Vergil are accompanied by later and more peripheral work such as the fifth-century CE Eudocia’s poem St. Cyprian. The only exclusive criterion is that the speech must be in hexameters. DICES is not a library of texts in the same way as Beck’s (2013) database of Homeric speech. Instead, there is classifying information, such as the text location, speaker, addressee and speech act, linking to external open sites, primarily Perseus, which provide the full text and analytical tools. As the project team argue (Forstall et al., 2022), theirs represents an update on databases such as Beck’s because linked open data can be manipulated—in particular, parsed and tagged—more readily. I have not explored DICES in full, but navigation was fairly intuitive, and the envisaged ecosystem of linked open data holds considerable potential for a range of users.
The book itself contains an introduction followed by eighteen studies/chapters, grouped thematically into form, tradition, style, emotion, and interaction. While each chapter is separate, methodological collaboration is credited (p. xi) in that contributors attended two workshops exploring DICES and sharing their own approaches. That said, there is, intentionally, wide variety in how the data has been exploited, from classical philology to GenAI, reflecting aims and epistemological stances. To give a representative flavour of the volume, I will discuss a chapter within each theme.
The first chapter dedicated to form is de Jong’s Vocatives in the speeches of Homer and Quintus of Smyrna. DICES allows de Jong to study vocatives synchronically, the distribution of vocatives within/across the Iliad and Odyssey, and diachronically, a comparison between Homer and Quintus. The statistical indices and analyses of individual speeches convincingly support narratological arguments. For instance, multiple vocatives within a speech (Priam’s entreaty to Hector to stay inside the city walls (Il.22.38, 56-57) is de Jong’s example) increase the emotional intensity. This is a broad survey of what could easily be a full monograph so some arguments feel like they need more support from the text, for instance, the suggestion (p. 54) that the influence of the novel genre may account for the relative paucity of vocatives in Quintus. Methodologically, what is valuable is de Jong’s comparison of a manual and digital parsing of the texts. It was found that automatic parsing struggled to distinguish between nominative and vocative case, the two forms often being morphologically identical, as in the frequently attested φίλοι. Because the digital accuracy was as low as 64.5% in one sample, manual rechecking was essential.
Within the tradition section, Chiappiniello’s Desine grande loqui: Characters and speeches in the Psychomachia of Prudentius collates speeches according to their designated speech act in DICES. Chiappiniello illustrates how the verbal duels in the Psychomachia, a Latin fourth-century CE allegorical poem, draw primarily on the Aeneid in order to raise the emotional stakes. For example, Pudicitia in her triumph over Libido (Prud. Psycho. 40-108) is compared to the unassailable Messapus in book 12 of the Aeneid by means of imagery and language patterning that depict the victors as ruthless. There seems some stretch in reconciling the Vergilian landscape with a Christian worldview (“Unlike the new Troy of the Aeneid, the soul’s temple must be rebuilt every day”, p. 199), but close readings of the text, both at sentence and discourse level, make the chapter compelling.
Turning to style, Computational stylometry and speech style in Ovid’s Metamorphoses by Bolt et al. applies quantitative methods to a work that features a higher density of speech than any other epic. Stylometry is not new in the classics—the authors cite an early computer-led foray by Marriott (1979) that claimed single authorship of Historia Augusta using the criteria of sentence length and initial/final lexical items. Interestingly, a more recent analysis (Stover & Kestemont, 2016) of Historia Augusta rejects sentence length as a reliable criterion because punctuation is a modern and often arbitrary imposition on the text. Bolt et al. retain sentence length as a measure, but it is one of twenty-six formal features divided between conjunctions, pronouns and pronominal adjectives, and a miscellaneous category. An example of the latter is relative clauses, and the authors find (1) there is a positive correlation between sentence length and relative clauses, and (2) regardless of sentence length, Metamorphoses deploys more relative clauses than other epics in DICES. To account for the phenomenon, there is a sensitive reading of Macareus’ encounter with Circe (Ov. Met. 14.251-257, 277-287) where the rambling style is interpreted to be stream of consciousness-like and evokes the character’s confused emotions. Such harnessing of quantitative data to inform literary appreciation is a powerful argument for DICES.
The section on emotion feels superordinate as the editors (p. 27) note that this is a prominent theme throughout the volume. Schirner’s Speech and emotion in Valerius Flaccus’ Argonautica acknowledges that emotion is difficult to categorise, but perhaps it would have been fruitful first to define exactly what is meant by emotion, as opposed to other affective conditions such as feeling and mood, and I would suggest Scherer’s (2005) behaviouristic model as robust. Schirner pulls from DICES 193 speeches from Valerius Flaccus’ Argonautica and then produces heat maps of the emotions of the addressees/audiences and speakers before and after the speech. It is reported that for addressees/audiences the most frequent emotion before a speech is “fear/worry”, while “undefined” is most common after a speech, with “fear/worry” dropping to third place. Case studies of particular speeches then account for the emotional dynamics, justifying the holistic observation that there is a tendency for a pre-speech negative state to improve to a positive mindset by the end of the speech. There is an issue with the concept of audience, no distinction seeming to be made between contemporary and modern-day recipients of the poem, but the heat maps are a visual and systematic way of tracking emotional states and thus unlocking the force of the speeches.
The final section concerns interaction and for the title alone I have to pick out the book editors’ Listen to mummy! Epic mother speech and persuasion from Homer to Nonnus. Using DICES to filter all speeches by mothers (94 in all, with Aphrodite/Venus most prominent), Forstall and Verhelst combine close reading, manual analysis of intertextuality, and distant reading, ranking the frequency of lemmata through natural language processing software, to answer the question, “What does it mean to speak like a mother?” (p. 431). A “motherly mode” is explored in terms of discourse function—in particular, persuasive strategies—and sentence-level features, for instance, a predilection for personal pronouns. The results are not conclusive as there is considerable variation in mothers’ speeches, as well as overlap with other speech genres, but the methodology suggests compatibility between close reading and corpus-scale analysis.
What unites these diverse contributions is the recognition that speech is central to epic and that our understanding of its role is enhanced by the availability of a rich and developing resource such as DICES. In keeping with the exploratory spirit, the full text of the book is open access, and it is a valuable resource even for those, like this reviewer, who are digital visitors rather than residents. Technology cannot provide all the answers, and as de Jong’s contribution highlights, it is not infallible, but perhaps digital tools can open up more paths to the answers. This volume does a commendable job of highlighting the methodological options available as well as quelling any discussion that the last word has been said on speech in epic.
Authors and titles
- Christopher W. Forstall and Berenice Verhelst – Introduction
- Irene J.F. de Jong – Vocatives in the Speeches of Homer and Quintus of Smyrna
- Ombretta Cesca and Matteo Romanello – A Computational Approach to Characters’ Intentional Repetitions in Homeric Epic
- Christoph Schwameis and Jan R. Telg genannt Kortmann – Exhortations in Silius Italicus’ Punica
- Valéry Berlincourt – Between Epic and Epideictic: Direct Speech in Claudian’s Political Poems through a Digital Lens
- Martina Delucchi – Epic Speech and Intertextuality: The epyllium Telephi in Its Literary Context
- Charles W. Oughton – Hannibal, Hercules, and the Livian Narrators of Silius Italicus’ Punica
- Roberto Chiappiniello – Desine grande loqui: Characters and Speeches in the Psychomachia of Prudentius
- Patrick J. Burns – Measuring Dialogism in Latin Epic
- Francesco Mambrini – The (Annotated) Language of the Homeric Heroes: Towards a Treebank-Based Approach
- Thomas J. Bolt, Pramit Chaudhuri, and Joseph Dexter – Computational Stylometry and Speech Style in Ovid’s Metamorphoses
- Rebekka Schirner – Speech and Emotion in Valerius Flaccus’ Argonautica
- Bernhard Söllradl – Quantifying Genre: A Study in Tragic Voices in Roman Epic
- William J. Dominik – Characterizational “Swerves” in the Speeches of Achilles in Greek and Latin Literature
- Deborah Beck – κερτομέω in Context(s) in Greek Epic Poetry
- Melissande Tomcik – Speaking Fake News in Flavian Epic
- Elizabeth Minchin – Evaluating Workplace Relationships in the Homeric Iliad: Bringing Together Digital Approaches and Social and Cognitive Theory
- Christopher W. Forstall and Berenice Verhelst – Listen to Mummy! Epic Mother Speech and Persuasion from Homer to Nonnus
- Jeff Rydberg-Cox – Using Networks to Explore the Relationships between Characters and the Words They Speak in Homer’s Iliad
References
Beck, D. (2013). Speech presentation in Homeric epic. The University of Texas at Austin. https://homeric-speech-beck.la.utexas.edu/home
Forstall, C. W., Finkmann, S., & Verhelst, B. (2022). Towards a linked open data resource for direct speech acts in Greek and Latin epic. Digital Scholarship in the Humanities, 37(4), 972–981. https://doi.org/10.1093/llc/fqac006
Marriott, I. (1979). The authorship of the Historia Augusta: Two computer studies. The Journal of Roman Studies, 69: 65–77. https://doi.org/10.2307/299060
Scherer, K. R. (2005). What are emotions? And how can they be measured? Social Science Information, 44(4): 695–729. https://doi.org/10.1177/0539018405058216
Stover, J. A., & Kestemont, M. (2016). The authorship of the “Historia Augusta”: Two new computational studies. Bulletin of the Institute of Classical Studies, 59(2): 140–157. https://doi.org/10.1111/j.2041-5370.2016.12043.x