Diachronica - Volume 41, Issue 3, 2024

Historical linguistics has experienced an upsurge in methodological innovation. However, owing to the constraints of contemporary publication, these innovations often appear in print in a format which is technical, dense and less than fully accessible. Moreover, because these innovations are still young and have often arisen in other fields before being translated to historical linguistics, our own discipline has yet to accumulate a stock of trusted explanations of them, presented lucidly and addressed specifically to our field and to its scientific concerns. This special issue is dedicated to promoting the accessibility of new methods. Despite their technical details, most new methods in historical linguistics rest upon a rationale and an essential logic that are perfectly comprehensible to interested historical linguists, so long as they are communicated in a fashion that is sympathetic and respectful to a linguistic audience. This introduction contextualizes the five papers of the special issue.

A perennial conflict in historical linguistics centers around the theoretical and practical virtues of tree-like divergence and wave-like diffusion. This paper presents the Dialect Chain Tree, an extension of the tree model that incorporates both tree-like descent and disintegration of dialect chains in a systematic fashion. As such, it provides a formalization and sharpening of Ross’ (1997: 212–228) linkage concept that allows integration into quantitative approaches.

In contrast to phylogenetic tree inference, wave model approaches are often regarded as difficult to computationally implement for inference of language relatedness. This paper proposes a basic framework for the computational modelling of wave-like diversification in language families and explains the model type of agent-based models for linguistic data. The approach is based on agent-based simulations which allow for the detailed simulation of speaker interactions within speech communities. The proposed framework operates by simulating a large number of possible diversification situations with different parameter settings and selecting those runs that yield a good fit to the linguistic data of the geographical spread of different languages. The model can be fed with the geographical extent of languages and their known innovations in order to computationally reconstruct the most likely diversification scenarios of these languages under the wave model.

We introduce a method for investigating sound change using audio proxies and interpolation between them. Using this method leads us to consider the possible trajectories of attested sound changes in acoustic space and allows us to run speech perception experiments with the interpolated sounds to examine the patterning of perceptual judgements. We discuss the implications for understanding rate and magnitude of sound changes and present the approach as one which might fruitfully be applied to a wide range of studies.

Language contact phenomena have increasingly been researched from different historical linguistic, sociolinguistic and areal-typological perspectives. However, since most of this research is based on case studies, an assessment of contact phenomena from a worldwide comparative perspective has been missing in the literature. In this article, we draw inspiration from historical linguistics and language typology to present a new typological approach for evaluating evidence that given linguistic domains have been affected by language contact. This method has three parts: (1) a new approach to sampling, (2) the analysis of typological data, and (3) making probabilistic inferences about language contact. We argue that this is a parsimonious method for evaluating contact effects that can serve as a starting point for the further development of typological approaches to language contact.

Computational methods have produced meaningful and usable results to study word semantics, including semantic change. These methods, belonging to the field of Natural Language Processing, have recently been applied to ancient languages; in particular, language modelling has been applied to Ancient Greek, the language on which we focus. In this contribution we explain how vector representations can be computed from word co-occurrences in a corpus and can be used to locate words in a semantic space, and what kind of semantic information can be extracted from language models. We compare three different kinds of language models that can be used to study Ancient Greek semantics: a count-based model, a word embedding model and a syntactic embedding model; and we show examples of how the quality of their representations can be assessed. We highlight the advantages and potential of these methods, especially for the study of semantic change, together with their limitations.