Speakers
Description
This paper presents a computational and Semantic Web–based model for Explanatory Combinatorial Dictionaries (ECDs), well-established lexicographic resources grounded in Explanatory Combinatorial Lexicology and Meaning–Text Theory (Mel’čuk et al., 1995). The model aims to make ECD structures machine-actionable, FAIR-compliant (Wilkinson et al., 2016), and interoperable with the Linguistic Linked Open Data ecosystem.
The proposal extends ECDs in two directions: by introducing diachronic depth into dictionaries usually conceived synchronically, and by modelling the phenomenon described in the literature as “denominational paths” (Aymerich et al., 2008), i.e. the ways in which lexical units foreground salient conceptual traits within and across languages.
The proposal is illustrated through a computational diachronic ECD, currently under development, devoted to the terminology of marriage in ancient Israel. Marriage is a suitable test case because it is a polythetic concept whose conceptualisation varies across cultures and historical periods, including within the Jewish tradition (Piccini et al. 2024).
While Biblical sources (c. 500-100 BCE) represent marriage as a legal institution based on the acquisition of a woman and the transfer of jurisdiction from her pater familias to the husband—who acquires sexual monopoly and legal paternity—Talmudic sources (c. 150-500 CE) introduce a more sacralised conception, reflected in terminology foregrounding consecration. This conceptual polyhedrality is mirrored by several denominations, each lexicalising a salient trait: acquisition, as in לקח laqaḥ ‘to take for oneself or for another’; possession, as in בעל ba‘al ‘to possess’; consecration, as in קידושין qiddushin ‘consecrations’; and virilocality, as in נשא nasa’ ‘to lead’, הושיב hoshiv ‘to cause to dwell’, and שלח shillaḥ ‘to send’. The modelling of nasa’ provides the illustrative example for the RDF/OWL-based representation presented in Figure 1*.
The model combines a lexical and an ontological layer. The lexical layer relies on three vocabularies: OntoLex-Lemon for lexical units and senses, Lexicog for dictionary-like structure, and LexFom (Fonseca et al., 2016) for Lexical Functions. In Figure 1(b), nasa’ is represented as an ECD entry specifying its canonical form, lexical sense and combinatorial information.
The ontological layer models diachronic conceptual change through a perdurantist, or 4D, approach (Welty et al., 2006), in which change is represented through temporal slices rather than by assigning time-dependent properties to the concept as a whole. Figure 1(a) presents the modelling of Hebrew marriage (HEBREW_MARRIAGE) between 500 and 100 BCE. This temporally bounded phase is represented by the class SLICE_1, defined as a subclass of HEBREW_MARRIAGE and associated with a specific time interval. SLICE_1 is characterised by traits such as legal paternity, male sexual monopoly and virilocal residence, each encoded as an existential restriction on a specific property. At the same time, it inherits more general, time-invariant traits from MARRIAGE, such as the establishment of a contractual bond and a socially regulated relation between sexes. Anchoring ECD entries to such slices the ontoleox:reference property turns the resource into a diachronic ECD.
The second objective, i.e. capturing the selective lexicalisation of salient conceptual traits, is implemented at the lexical–conceptual interface through the property ontolex:isLexicalizedSenseOf, which links a lexical sense (in our case nasa’ lexical sense) to the conceptual trait it lexicalises (virilocality). Since this relation requires an ontolex:LexicalConcept as its object, conceptual traits are represented as lexical concepts within a feature-based schema, consisting of a skos:ConceptScheme and a set of features modelled as SKOS concepts (also typed as ontolex:LexicalConcept).
Such a formalisation supports fine-grained semantic and diachronic queries, from identifying the conceptual facet lexicalised by a term to tracing specific traits across historical phases (Figure 2*).
* Figures see Book of Abstracts