Conceptual Metaphors for Jealousy in Latin: A Corpus-Based Dataset

Repository location

The dataset has been deposited in the Zenodo repository and is accessible via https://doi.org/10.5281/zenodo.19653015.

Context

The dataset was developed as part of a doctoral research project on the emotion of jealousy in Latin. More specifically, the project investigates the linguistic and conceptual representation of jealousy in Latin, with particular attention to metaphorical structuring and to the interaction between lexical semantics and conceptual patterns. The dataset underpins the analysis presented in a chapter of the PhD thesis devoted to the identification and classification of conceptual and creative metaphors associated with a set of Latin lexemes semantically related to jealousy, including invidia, livor, aemulatio, obtrectatio, and malevolentia.

The study is grounded in Conceptual Metaphor Theory (Lakoff & Johnson, 1980), according to which abstract concepts (target domains), such as emotions, are structured and understood through more concrete experiential domains (source domains), such as physical objects, substances, or living beings. Within this framework, metaphor is not treated as a purely stylistic device, but as a fundamental cognitive mechanism shaping conceptualisation. This approach has been widely applied to the study of emotional language and cross-cultural semantic patterns (Kövecses, 2000).

The dataset was developed during a research period conducted within the Lexicon of Embodied Experience in Latin (LexeLat: https://lexelat.unige.it/project, last accessed 01/06/2026) project at the University of Genoa (De Felice & Fedriani, 2024). By adopting digital and corpus-based methodologies, this project investigates metaphors for emotions more broadly, including fear, rage, hatred, sadness, happiness, and love.

Finally, the dataset was also partially incorporated into the Lexicon Translaticium Latinum (De Felice et al., 2020), an ongoing digital lexicon of Latin metaphors developed at the University of Exeter, aimed at documenting large-scale metaphorical patterns in Latin language.

Procedure

The dataset was constructed using a corpus-based, lexeme-driven methodology designed to identify and classify metaphorical expressions related to jealousy in Latin. Adopting a target-oriented approach, the analysis focused on a set of lexemes associated with the semantic domain of “jealousy” (Stefanowitsch, 2006; Buccheri et al., 2021). The methodological framework follows the protocol established by the Lexelat project, ensuring consistency across the annotation of different emotions in Latin and enabling systematic comparison of the metaphors used to conceptualise them. The construction of the dataset involved multiple stages, which are outlined below.

First, a set of Latin lexemes semantically associated with jealousy (invidia, livor, aemulatio, obtrectatio, malevolentia) was selected using both lexicographic resources (such as the Oxford Latin Dictionary) and lexico-semantic resources such as Latin WordNet. All occurrences of these lexemes were automatically extracted, in all inflected forms, from the Library of Latin Texts (LLT), specifically the Antiquitas section, a large-scale digital corpus developed by Brepols. This subset includes texts from the second century BCE to the second century CE and allows for systematic full-text querying across a wide range of authors and genres. The extraction yielded a total of 1324 occurrences. Before proceeding to the second stage, a preliminary disambiguation was performed for polysemous lexemes. In particular, livor can denote either an emotional state (“envy” or “jealousy”) or a physical property (“bruise”, “dark colour”); only the 44 instances corresponding to the emotional sense were retained for analysis.

Subsequently, all occurrences were manually examined and classified as either literal or metaphorical. Metaphorical uses were identified on the basis of contextual cues signalling a mapping between an abstract emotional concept and a more concrete domain, in accordance with Conceptual Metaphor Theory (Lakoff & Johnson, 1980). The following passages illustrate this distinction:

1. res inter initia urunt, et livor mora solvitur. (Quint. Decl. 388. 25) Things burn us at the beginning, even jealousy (livor) dissolves with time.

2. Quibus sermonibus satis constat Domitiani quoque aures verberatas, dum optimus quisque libertorum amore et fide, pessimi malignitate et livore pronum deterioribus exstimulabant. (Tac. Agr. 41). All of which gossip, it is certain, beat upon the ears of Domitian no less than of other men, the best of his freedmen seeking from love and loyalty, the worst from malice and jealousy (livor), to stir the emotions of a master who leaned ever to the worst side.

In the first example, livor is conceptualised and linguistically encoded through a metaphorical image, that is in terms of a more concrete entity: specifically, a substance that can be loosened or dissolved with time. In this case, the metaphorical conceptualisation of livor is elicited by the contextual association with the verb solvere. By contrast, the second example does not suggest any metaphorical interpretation of the meaning expressed by the term livor and features instead a literal usage. Cases of this latter type were therefore counted for the purposes of a quantitative examination but were excluded from the annotated dataset.

Finally, in the third stage, the annotation process was carried out. All metaphorical occurrences were systematically recorded in a structured spreadsheet. Each entry includes: (i) the lexical item with morphological information (case and number), (ii) the textual context, (iii) the bibliographic reference, and (iv) the metaphor’s source domain. To ensure analytical consistency, cases in which multiple metaphorical mappings co-occur within the same context were treated as separate annotations. For example, in invidia erat magna ([he] was in a big jealousy cf. Plin. Nat. 18.41.3), the expression simultaneously encodes a container metaphor (“being in jealousy”) and an object metaphor (“jealousy as something with measurable magnitude”); it was therefore annotated twice.

The resulting dataset comprises 831 metaphorical instances. Within this sample, the lexeme invidia accounts for the majority of occurrences (724 instances), whereas livor exhibits the highest proportion of metaphorical usage (84%). On the basis of this annotation, lexemes associated with jealousy were found to be conceptualised through sixteen source domains. These domains were further organised into a hierarchical taxonomy reflecting different levels of abstraction and conceptual relationships (see Figure 1).

Figure 1

Finally, to ensure interoperability and facilitate reuse, a subset of the dataset was encoded in the Lexicon Translaticium Latinum (De Felice et al., 2020). Within this resource, metaphors are formalised as mappings between source and target synsets, following a WordNet-based semantic modelling framework (Miller, 1995; Minozzi, 2017).

Sampling strategy

The dataset adopts a lexeme-driven exhaustive sampling strategy, retrieving all occurrences of selected lexical items rather than sampling texts. This ensures systematic coverage of the target semantic domain and allows for quantitative comparison across lexemes.

Quality control

To ensure reliability, all annotations were independently reviewed by two additional annotators. Disagreements were resolved through discussion and consensus, resulting in a validated and consistent dataset.

Dataset Description

Repository name: Zenodo

Object name: Jealousy Metaphors in Latin_ Dataset

Format names and versions

Microsoft Excel (.xlsx)

Creation dates

Start date: 2019-11-01

End date: 2020-03-01

Dataset creators

Roberta Grazia Leotta, research fellow, Catholic University of Sacred Heart, Milan – conceptualisation, data collection, annotation, analysis

Additional annotators: Chiara Fedriani, Irene De Felice Università di Genova, Università del Piemonte Orientale – validation and quality control

Language

Latin (primary data)

English (metadata and annotations)

License

Creative Commons Attribution 4.0 International (CC BY 4.0)

Publication date

2026-04-2026

Reuse Potential

This dataset offers broad reuse potential across classics, linguistics, digital humanities, and cognitive science. It provides a systematically annotated resource for investigating emotional expression in Latin and supports research on metaphor, lexical semantics, and diachronic language change. Its structured format also enables computational applications, including corpus integration, semantic modelling, and machine learning. However, the dataset is not without limitations. Its scope is inherently constrained by its lexeme-based design and by the interpretative component involved in metaphor identification. Furthermore, it is restricted to the LLT Antiquitas corpus and includes polysemous lexical items, factors which may affect the generalisability of the findings.