Periodic Reporting for period 1 - PhLex (Phylogenetic Lexification: Patterns of meaning through time)
Okres sprawozdawczy: 2024-06-01 do 2026-05-31
Podsumowanie kontekstu i ogólnych celów projektu
Languages encode knowledge, culture, and identity, yet more than half of the world’s languages are currently endangered, and many are represented only through short, fragmentary wordlists. These sparse records restrict our ability to understand how languages evolve, how meanings change, and how linguistic diversity develops over time. Traditional historical linguistic research relies heavily on large, well-documented datasets and on phonological correspondences, meaning that the world’s most under-documented and often Indigenous languages are systematically excluded from global comparative research. This creates an uneven evidence base and limits the development of methods capable of supporting communities working to document or revitalise their languages. The Phylogenetic Lexification (PhLex) project was established to address these challenges by exploring whether patterns of meaning (rather than only word forms) can provide reliable signals of linguistic history. While meaning has long been considered too variable or unpredictable to be used systematically, new research suggests that semantic structures may persist across many generations of speakers, even when the words themselves change. If semantic patterns can be shown to carry historical information, they offer a promising and inclusive pathway for analysing languages for which extensive documentation is not available. PhLex advances this idea by developing two complementary data resources: EvoLex, a comparative lexical dataset that collates cognate sets across a chosen family of related languages, and EvoSem, a semantic tool that models patterns of meaning alignment, colexification, and semantic divergence. Together, these resources enable large-scale investigation of how meanings shift and how such changes relate to known historical relationships among languages. Although the proposal originally focused on Oceanic languages, the project pivoted to Australian Indigenous languages when it became clear that these offered a feasible and ethically appropriate dataset and stronger opportunities for collaboration. This change did not alter the scientific objectives, but it significantly increased the societal relevance of the work by contributing to a region where documentation gaps are particularly acute. The overarching objective of PhLex is to determine whether meaning-based patterns, identified through lexification and re-lexification, retain a detectable phylogenetic signal. To achieve this, the project pursued three linked goals:
1. Build a comprehensive database of meaning correspondences across a set of related languages (EvoLex/EvoSem).
2. Use phylogenetic techniques to measure the strength of the historical signal in these semantic patterns.
3. Compare a meaning-based model with a traditional cognate-based approach, and assess whether a combined, hybrid approach can enhance historical inference.
These objectives respond to a broader strategic need within European and international research: developing computational and data-driven methods capable of handling incomplete or heterogeneous datasets. Such methods support Open Science goals by increasing the reusability of linguistic data and by extending analytical tools to cases where conventional methods cannot be applied. They also align with the EU’s wider societal priorities concerning cultural heritage, Indigenous knowledge, and equitable access to scientific infrastructure. By focusing on a historically under-studied language family, PhLex tackles both methodological and societal challenges. Scientifically, it tests the evolution of meaning through computational historical linguistics. Societally, it develops resources that may support Indigenous language centres and researchers by improving the accessibility and comparability of lexical information, always following ethical and community-informed protocols for handling sensitive materials. While the project is not expected to produce direct economic outcomes, its cultural and research impacts are potentially significant: improving the representation of under-documented languages in global comparative work and enabling new forms of historical and typological research. The expected impact of PhLex is therefore twofold. First, it contributes directly to the scientific understanding of how languages diversify by introducing semantic data into phylogenetic modelling. Second, it lays the foundations for future comparative work on languages that lack extensive documentation, thereby expanding the inclusiveness and reach of linguistic science. As analyses continue beyond the funded period, the methods and datasets established through the project will support further publications, ongoing collaborations, and long-term contributions to the study of language change worldwide.
1. Build a comprehensive database of meaning correspondences across a set of related languages (EvoLex/EvoSem).
2. Use phylogenetic techniques to measure the strength of the historical signal in these semantic patterns.
3. Compare a meaning-based model with a traditional cognate-based approach, and assess whether a combined, hybrid approach can enhance historical inference.
These objectives respond to a broader strategic need within European and international research: developing computational and data-driven methods capable of handling incomplete or heterogeneous datasets. Such methods support Open Science goals by increasing the reusability of linguistic data and by extending analytical tools to cases where conventional methods cannot be applied. They also align with the EU’s wider societal priorities concerning cultural heritage, Indigenous knowledge, and equitable access to scientific infrastructure. By focusing on a historically under-studied language family, PhLex tackles both methodological and societal challenges. Scientifically, it tests the evolution of meaning through computational historical linguistics. Societally, it develops resources that may support Indigenous language centres and researchers by improving the accessibility and comparability of lexical information, always following ethical and community-informed protocols for handling sensitive materials. While the project is not expected to produce direct economic outcomes, its cultural and research impacts are potentially significant: improving the representation of under-documented languages in global comparative work and enabling new forms of historical and typological research. The expected impact of PhLex is therefore twofold. First, it contributes directly to the scientific understanding of how languages diversify by introducing semantic data into phylogenetic modelling. Second, it lays the foundations for future comparative work on languages that lack extensive documentation, thereby expanding the inclusiveness and reach of linguistic science. As analyses continue beyond the funded period, the methods and datasets established through the project will support further publications, ongoing collaborations, and long-term contributions to the study of language change worldwide.
Prace wykonane od początku projektu do końca okresu sprawozdawczego oraz najważniejsze dotychczasowe rezultaty
The project centred on building a comparative lexical and semantic dataset and preparing it for use in phylogenetic modelling. The technical work began with assembling the full lexical dataset from a wide range of primary and secondary sources. This required standardising formats, cleaning metadata, and aligning entries across languages to ensure they could be compared reliably. The curated dataset now forms the basis of EvoLex, the cognate-focused component of the project.
In parallel, the project developed EvoSem, a semantic comparison framework designed to model patterns of meaning correspondence, colexification, and semantic divergence across the language sample. The coding of meaning relationships has been completed, and initial tests have confirmed that the semantic patterns can be expressed in a form suitable for downstream phylogenetic analysis. These preliminary outputs represent one of the project’s central achievements, demonstrating that semantic information can be systematised at a scale not previously attempted for this language family.
The project also advanced the analytical pipeline. Early computational analyses were run to explore the structure of the semantic data and to identify areas requiring additional refinement. These preliminary tests provide a foundation for the full phylogenetic modelling, which will proceed once the final version of the dataset is incorporated. Although the full modelling phase extends beyond the funded period, the project has completed the preparatory work needed to carry it out.
These technical outcomes establish the core infrastructure for comparing meaning-based and cognate-based models and for assessing the extent to which semantic patterns carry historical signal. They represent the main scientific achievements of the project and provide a clear basis for the publications planned in the next phase of work.
In parallel, the project developed EvoSem, a semantic comparison framework designed to model patterns of meaning correspondence, colexification, and semantic divergence across the language sample. The coding of meaning relationships has been completed, and initial tests have confirmed that the semantic patterns can be expressed in a form suitable for downstream phylogenetic analysis. These preliminary outputs represent one of the project’s central achievements, demonstrating that semantic information can be systematised at a scale not previously attempted for this language family.
The project also advanced the analytical pipeline. Early computational analyses were run to explore the structure of the semantic data and to identify areas requiring additional refinement. These preliminary tests provide a foundation for the full phylogenetic modelling, which will proceed once the final version of the dataset is incorporated. Although the full modelling phase extends beyond the funded period, the project has completed the preparatory work needed to carry it out.
These technical outcomes establish the core infrastructure for comparing meaning-based and cognate-based models and for assessing the extent to which semantic patterns carry historical signal. They represent the main scientific achievements of the project and provide a clear basis for the publications planned in the next phase of work.
Innowacyjność oraz oczekiwany potencjalny wpływ (w tym dotychczasowe znaczenie społeczno-gospodarcze i szersze implikacje społeczne projektu)
The project set out to test whether patterns of meaning can offer reliable historical information in cases where traditional lexical comparison faces practical limits. Although full analytical modelling is still underway, the project has already advanced the state of the art in several ways.
First, the development of the EvoLex and EvoSem resources provides a structured foundation for investigating meaning change at a scale not previously attainable for Australian Indigenous languages. EvoLex brings together cognate information within a unified comparative format, while EvoSem models patterns of meaning correspondence, colexification, and semantic divergence. The combination of these resources offers a new way to examine how meanings reorganise across related languages. This represents a clear methodological step forward, especially for language families with uneven documentation. Second, the project demonstrates that semantic information can be integrated into phylogenetic workflows rather than treated as supplementary or unpredictable. Preliminary analyses show that semantic patterns can be systematised in ways compatible with existing computational tools. This opens new directions for historical linguistics, allowing semantic evidence to contribute to classification and reconstruction in families where phonological or morphological comparison is limited. It also offers a way to cross-validate cognate-based models with meaning-based ones, which strengthens confidence in historical interpretations. Third, the project contributes to global typology and comparative linguistics by preparing a dataset that aligns an understudied language family with broader cross-linguistic samples. This makes it possible to situate semantic patterns in Australian Indigenous languages within a worldwide comparative context and to examine whether certain types of meaning change are regionally distinctive or follow broader tendencies.
By the end of the funded period, the following results have been achieved:
1. A curated comparative lexical dataset (EvoLex) that supports downstream semantic and phylogenetic analyses.
2. A semantic comparison framework (EvoSem) capable of modelling meaning correspondences across a large sample of languages.
3. Preliminary analyses demonstrating the feasibility of detecting structured semantic patterns suitable for phylogenetic testing.
4. Conference presentations and seminar outputs that have disseminated emerging findings to both European and Australian research communities.
5. A clear pathway for forthcoming peer-reviewed publications, including one article on the historical dimension of the project and further papers on typology and semantic modelling.
To ensure that the results reach their full potential, several further steps are needed:
1. Completion of full phylogenetic modelling once the finalised dataset is fully integrated.
2. Continued methodological development to refine how semantic patterns are coded and compared, particularly for languages with heterogeneous documentation.
3. Sustained collaboration with Indigenous language centres to ensure that data handling remains culturally appropriate and that outputs are useful for community-led research.
4. Dissemination through high-impact journals in historical linguistics, typology, and cognitive linguistics to encourage wider scientific uptake.
5. Long-term maintenance of the EvoLex/EvoSem infrastructure, ideally through future funded projects, to support open and reusable research across languages.
These developments position the project to make a substantive contribution to historical linguistics and to expand the methodological tools available for studying languages with limited documentation. As the remaining analyses are completed, the project is expected to generate findings that extend far beyond its original scope and offer durable benefits to both the academic field and community-based language research.
First, the development of the EvoLex and EvoSem resources provides a structured foundation for investigating meaning change at a scale not previously attainable for Australian Indigenous languages. EvoLex brings together cognate information within a unified comparative format, while EvoSem models patterns of meaning correspondence, colexification, and semantic divergence. The combination of these resources offers a new way to examine how meanings reorganise across related languages. This represents a clear methodological step forward, especially for language families with uneven documentation. Second, the project demonstrates that semantic information can be integrated into phylogenetic workflows rather than treated as supplementary or unpredictable. Preliminary analyses show that semantic patterns can be systematised in ways compatible with existing computational tools. This opens new directions for historical linguistics, allowing semantic evidence to contribute to classification and reconstruction in families where phonological or morphological comparison is limited. It also offers a way to cross-validate cognate-based models with meaning-based ones, which strengthens confidence in historical interpretations. Third, the project contributes to global typology and comparative linguistics by preparing a dataset that aligns an understudied language family with broader cross-linguistic samples. This makes it possible to situate semantic patterns in Australian Indigenous languages within a worldwide comparative context and to examine whether certain types of meaning change are regionally distinctive or follow broader tendencies.
By the end of the funded period, the following results have been achieved:
1. A curated comparative lexical dataset (EvoLex) that supports downstream semantic and phylogenetic analyses.
2. A semantic comparison framework (EvoSem) capable of modelling meaning correspondences across a large sample of languages.
3. Preliminary analyses demonstrating the feasibility of detecting structured semantic patterns suitable for phylogenetic testing.
4. Conference presentations and seminar outputs that have disseminated emerging findings to both European and Australian research communities.
5. A clear pathway for forthcoming peer-reviewed publications, including one article on the historical dimension of the project and further papers on typology and semantic modelling.
To ensure that the results reach their full potential, several further steps are needed:
1. Completion of full phylogenetic modelling once the finalised dataset is fully integrated.
2. Continued methodological development to refine how semantic patterns are coded and compared, particularly for languages with heterogeneous documentation.
3. Sustained collaboration with Indigenous language centres to ensure that data handling remains culturally appropriate and that outputs are useful for community-led research.
4. Dissemination through high-impact journals in historical linguistics, typology, and cognitive linguistics to encourage wider scientific uptake.
5. Long-term maintenance of the EvoLex/EvoSem infrastructure, ideally through future funded projects, to support open and reusable research across languages.
These developments position the project to make a substantive contribution to historical linguistics and to expand the methodological tools available for studying languages with limited documentation. As the remaining analyses are completed, the project is expected to generate findings that extend far beyond its original scope and offer durable benefits to both the academic field and community-based language research.