Computational Historical Linguistics

In: Matilde Serangeli & Thomas Olander (eds.). 2020. Dispersals and diversification: Linguistic and archaeological perspectives on the early stages of Indo-European (Brill's Studies in Indo-European Languages & Linguistics). Leiden & Boston: Brill. https://doi.org/10.1163/9789004416192_011

Because of the publisher's embargo I cannot yet upload a digital offprint of the paper.

THE VERSION OF THE PAPER GIVEN HERE IS THE VERSION SUBMITTED TO PEER-REVIEWERS. PLEASE CITE THE ACTUAL PUBLISHED PAPER AVAILABLE AT THE DOI ADDRESS ABOVE.

Abstract: Cladistic hypotheses are ideally based on arguments that use cumulative evidence from a wide range of shared innovations inherited from a more recent ancestor. The majority of historical linguists would agree that the best evidence for subgrouping would be shared phonological and morphological innovations, while evidence from proportions of shared lexical cognacy is less reliable for linguistic subgroup-ing. Recent high-profile studies have appeared, however, that have been based exclusively on comparative lexical material. The results of these methods have been sharply criticised, but in spite of the criticisms to cognacy-based approaches, there remains some potential that the lexical cognacy may provide some useful data to supplement cladistic hypotheses as part of an overall assessment of the complete bundle of available isoglosses. If lexical cognacy judgements can be treated as a potential source of data for cladistic hypotheses, how can they be implemented in a methodologically rigorous way? This chapter focuses on case studies from methodological issues that have arisen in encoding Indo-European lexical cognacy data on the Indo-European Cognate Relationships (IE-CoR) database project based at the Max Planck Institute for the Science of Human History. These issues are illustrated through case studies from problems that have arisen in assigning cognacy coding to lexical data. As such this chapter contributes discussion for improving the reliability of cognacy data for cladistic analyses as a supplement to more traditional analyses based on comparative phonological and morphological criteria.

In the thesis it is discussed in what ways concepts and methodology developed in evolutionary biology can be applied to the explanation and research of language change. The parallel nature of the mechanisms of biological evolution and language change is explored along with the history of the exchange of ideas between these two disciplines. Against this background computational methods developed in evolutionary biology are taken into consideration in terms of their applicability to the study of historical relationships between languages. Different phylogenetic methods are explained in common terminology, avoiding the technical language of statistics. The thesis is on one hand a synthesis of earlier scientific discussion, and on the other an attempt to map out the problems of earlier approaches in addition to finding new guidelines in the study of language change on their basis. Primarily literature about the connections between evolutionary biology and language change, along with research articles describing applications of phylogenetic methods into language change have been used as source material.

The thesis starts out by describing the initial development of the disciplines of evolutionary biology and historical linguistics, a process which right from the beginning can be seen to have involved an exchange of ideas concerning the mechanisms of language change and biological evolution. The historical discussion lays the foundation for the handling of the generalised account of selection developed during the recent few decades. This account is aimed for creating a theoretical framework capable of explaining both biological evolution and cultural change as selection processes acting on self-replicating entities. This thesis focusses on the capacity of the generalised account of selection to describe language change as a process of this kind. In biology, the mechanisms of evolution are seen to form populations of genetically related organisms through time. One of the central questions explored in this thesis is whether selection theory makes it possible to picture languages are forming populations of a similar kind, and what a perspective like this can offer to the understanding of language in general.

In historical linguistics, the comparative method and other, complementing methods have been traditionally used to study the development of languages from a common ancestral language. Computational, quantitative methods have not become widely used as part of the central methodology of historical linguistics. After the fading of a limited popularity enjoyed by the lexicostatistical method since the 1950s, only in the recent years have also the computational methods of phylogenetic inference used in evolutionary biology been applied to the study of early language history. In this thesis the possibilities offered by the traditional methodology of historical linguistics and the new phylogenetic methods are compared. The methods are approached through the ways in which they have been applied to the Indo-European languages, which is the most thoroughly investigated language family using both the traditional and the phylogenetic methods. The problems of these applications along with the optimal form of the linguistic data used in these methods are explored in the thesis.

The mechanisms of biological evolution are seen in the thesis as parallel in a limited sense to the mechanisms of language change, however sufficiently so that the development of a generalised account of selection is deemed as possibly fruiful for understanding language change. These similarities are also seen to support the validity of using phylogenetic methods in the study of language history, although the use of linguistic data and the models of language change employed by these models are seen to await further development.

By combining an Ancient Greek semantic domains database with a corpus of annotated Ancient Greek treebanks, one may observe semantic preferences of individual words or word combinations. This information may then be applied to phrases which contain a word of ambiguous sense with the possibility of disambiguating that word. The method was applied to John 1:5, 1 Corinthians 11:15, and the phrase εν τοις επουρανιοις which occurs throughout Ephesians. Due to the small size of the annotated Ancient Greek corpus and the limited capability of the existing Ancient Greek semantic domains database, disambiguation was inconclusive for the first two passages. In Ephesians, however, this method combined with some manual inspection demonstrated that επουρανιοις may very well be referring to heavenly beings rather than a heavenly place. This is a valuable insight which had previously been unanimously rejected by published Ephesians works on the basis of flawed logic and with no consideration of linguistic evidence.

The evidence one can draw from the rhyming behavior of Old Chinese words plays a crucial role for the reconstruction of Old Chinese, particularly for the more recent proposals. Some of these proposals are no longer solely based on the intuition of scholars but also substantiated by statistical arguments which help to assess the probability by which a given set of rhyming instances can be assigned to an established rhyme group. So far, however, quantitative methods were only used to confirm given hypotheses regarding rhyme groups in Old Chinese, and no exploratory analyses that would create hypotheses regarding rhyme groups in a corpus were carried out. This paper presents a new method which models rhyme data as weighted undirected networks. By representing rhyme words as nodes in a network and the frequency of rhymes in a given corpus as links between nodes, rhyme groups can be inferred with help of standard algorithms originally designed for social network analysis. This is illustrated by constructing a rhyme network from the Book of Odes and comparing the automatically inferred rhyme groups with rhyme groups proposed in the literature. Apart from revealing interesting general properties of rhyme networks in Chinese historical phonology, the analysis provides strong evidence for a coda *-r in Old Chinese. The results of the analysis and the rhyme network of the Book of Odes can be inspected in form of an interactive online application or directly downloaded.

The exploration of distant language relationships reaching back to the Neolithic age remains very demanding and is often perceived as controversial. Over the last two decades, significant advances in computational linguistics offer alternatives to the methods in use so far to classify languages, but less quantitative research has been devoted to long-range language relationships. Classical methods reach their limits at a level where the signals they infer are more likely to be due to chance than to relatedness. We created a distance-based method without dependency on human relatedness judgments to set up an automated language classification inference system. We used a backward selection method to optimize the key elements of the system towards existing reference classifications: a word list, sound change rules and other parameters. We arbitrated choices in study design and features, which are beneficial for longer-range inference, if necessary at the cost of the short-range performance. The scope of the project is global: the system processes 1962 languages from all families, the limit for inclusion of more languages being the availability of the material needed by our system. After exploring the methodology, we show that it leads to a reliable language classification, matching the reference listings to a high proportion. We then apply a series of statistical models to target middle- and long-range relationships and explore connections outside the well-established classifications. By shedding light on relationships between top-level language families in the context of chance interference, our results deliver a strong support for two major long-range hypotheses: Eurasiatic (connecting Indo-European, Uralic, Turkic, Mongolic and Tungusic) and Austric (connecting Austronesian, Tai-Kadai, Austroasiatic and Hmong-Mien). Further support is given to parts of the Nostratic macrofamily, to a dozen other long-range hypotheses and to internal classifications of established language families.

In this study, an attempt has been made to use the computer program, PHONO, to develop a computer model which operates on the principle of the regularity of sound change. Surprising though it may seem, since this concept was first coined in modern linguistics, no standard computer program has been developed which could apply cross-linguistically to simulate and study diachronic change. -- We have shown that the program may be used in the study of the Algonquian family of languages. We have developed a working model for the study of phonological change from the reconstructions of Proto-Algonquian to Shawnee. Using this Shawnee Model, we have simulated Shawnee data from reconstructed Proto-AIgonquian words and illustrated a technique to collect or search for evidence for the reconstructed forms. -- Given the nature and status of many languages of the world which are threatened with extinction, this technique may facilitate the preservation of valuable evidence for extinct prehistoric languages.

The field of Chinese Historical Phonology is traditionally dealing with a large number of complex and diverse types of data. While the data diversity can be conveniently dealt with in qualitative approaches, computational possibilities that have arisen during the past two decades offer new possibilities and new challenges for the field. In the study, I will summarize the chances and challenges which we face in the discipline and point to some suggestions for future work.

Advances in computer-assisted linguistic research have been greatly influential in reshaping linguistic research. With the increasing availability of interconnected datasets created and curated by researchers, more and more interwoven questions can now be investigated. Such advances, however, are bringing high requirements in terms of rigorousness for preparing and curating datasets. Here we present CLICS, a Database of Cross-Linguistic Colexifications (CLICS). CLICS tackles interconnected interdisciplinary research questions about the colexification of words across semantic categories in the world's languages, and show-cases best practices for preparing data for cross-linguistic research. This is done by addressing shortcomings of an earlier version of the database, CLICS2, and by supplying an updated version with CLICS3, which massively increases the size and scope of the project. We provide tools and guidelines for this purpose and discuss insights resulting from organizing student tasks for database updates.

Advances in computer-assisted linguistic research have been greatly influential in reshaping linguistic research. With the increasing availability of interconnected datasets created and curated by researchers, more and more interwoven questions can now be investigated. Such advances, however, are bringing high requirements in terms of rigorousness for preparing and curating datasets. Here we present CLICS, a Database of Cross-Linguistic Colexifications (CLICS). CLICS tackles interconnected interdisciplinary research questions about the colexification of words across semantic categories in the world’s languages, and show-cases best practices for preparing data for cross-linguistic research. This is done by addressing shortcomings of an earlier version of the database, CLICS2, and by supplying an updated version with CLICS3, which massively increases the size and scope of the project. We provide tools and guidelines for this purpose and discuss insights resulting from organizing stude...

Sound correspondence patterns play a crucial role for linguistic reconstruction. Linguists use them to prove language relationship, to reconstruct proto-forms, and for classical phylogenetic reconstruction based on shared innovations.Cognate words which fail to conform with expected patterns can further point to various kinds of exceptions in sound change, such as analogy or assimilation of frequent words. Here we present an automatic method for the inference of sound correspondence patterns across multiple languages based on a network approach. The core idea is to represent all columns in aligned cognate sets as nodes in a network with edges representing the degree of compatibility between the nodes. The task of inferring all compatible correspondence sets can then be handled as the well-known minimum clique cover problem in graph theory, which essentially seeks to split the graph into the smallest number of cliques in which each node is represented by exactly one clique. The resulting partitions represent all correspondence patterns which can be inferred for a given dataset. By excluding those patterns which occur in only a few cognate sets, the core of regularly recurring sound correspondences can be inferred. Based on this idea, the paper presents a method for automatic correspondence pattern recognition, which is implemented as part of a Python library which supplements the paper. To illustrate the usefulness of the method, we present how the inferred patterns can be used to predict words that have not been observed before.

An etymological proposition is often said to be probable or improbable from the phonetic point of view, and it is not rare for opinions to diverge on which it is. The estimation is typically purely intuitive, based on perceived similarity and no more than a handful of analogous examples. This paper proposes a method for quantifying the phonetic probability of an etymology and comparing it to the alternative hypothesis. It is intended to be used with sizable datasets, to produce a well-supported, objective verdict.

The ASJP (Automated Similarity Judgment Program) described an automated, lexical similarity-based method for dating the world's language groups using 52 archaeological, epigraphic and historical calibration date points. The present paper describes a new automated dating method, based on phonotactic diversity. Unlike ASJP, our method does not require any information on the internal classification of a language group. Also, the method can use all the available word lists for a language and its dialects eschewing the debate on 'language' vs. 'dialect'. We further combine these dates and provide a new baseline which, to our knowledge, is the best one. We make a systematic comparison of our method, ASJP's dating procedure, and combined dates. We predict time depths for world's language families and sub-families using this new baseline. Finally, we explain our results in the model of language change given by Nettle. Citation: Rama T (2013) Phonotactic Diversity Predicts the Time Depth of the World's Language Families. PLoS ONE 8(5): e63238.

In this study, an attempt has been made to use the computer program, PHONO, to develop a computer model which operates on the principle of the regularity of sound change. Surprising though it may seem, since this concept was first coined in modern linguistics, no standard computer program has been developed which could apply cross-linguistically to simulate and study diachronic change. -- We have shown that the program may be used in the study of the Algonquian family of languages. We have developed a working model for the study of phonological change from the reconstructions of Proto-Algonquian to Shawnee. Using this Shawnee Model, we have simulated Shawnee data from reconstructed Proto-AIgonquian words and illustrated a technique to collect or search for evidence for the reconstructed forms. -- Given the nature and status of many languages of the world which are threatened with extinction, this technique may facilitate the preservation of valuable evidence for extinct prehistoric languages.

Advances in computer-assisted linguistic research have been greatly influential in reshaping linguistic research. With the increasing availability of interconnected datasets created and curated by researchers, more and more interwoven questions can now be investigated. Such advances, however, are bringing high requirements in terms of rigorousness for preparing and curating datasets. Here we present CLICS, a Database of Cross-Linguistic Colexifications (CLICS). CLICS tackles interconnected interdisciplinary research questions about the colexification of words across semantic categories in the world’s languages, and show-cases best practices for preparing data for cross-linguistic research. This is done by addressing shortcomings of an earlier version of the database, CLICS2, and by supplying an updated version with CLICS3, which massively increases the size and scope of the project. We provide tools and guidelines for this purpose and discuss insights resulting from organizing stude...

Advances in computer-assisted linguistic research have been greatly influential in reshaping linguistic research. With the increasing availability of interconnected datasets created and curated by researchers, more and more interwoven questions can now be investigated. Such advances, however, are bringing high requirements in terms of rigorousness for preparing and curating datasets. Here we present CLICS, a Database of Cross-Linguistic Colexifications (CLICS). CLICS tackles interconnected interdisciplinary research questions about the colexification of words across semantic categories in the world’s languages, and show-cases best practices for preparing data for cross-linguistic research. This is done by addressing shortcomings of an earlier version of the database, CLICS2, and by supplying an updated version with CLICS3, which massively increases the size and scope of the project. We provide tools and guidelines for this purpose and discuss insights resulting from organizing stude...

The proposal of new quantitative methods supposed to handle problems in historical linguistics has created a gap between what one could call "classical" approaches to historical language comparison and the "new and innovative" automatic approaches. Classical linguists are often skeptical of the new approaches, partly because the results differ from those achieved by classical methods (Anthony and Ringe 2015, Holm 2007), but also because the majority of the new approaches work in a black box fashion and do not allow inspecting the concrete findings in detail. Computational linguists, on the other hand, complain about classical historical linguists' lack of consistency when applying the classical methods. How to cite List, J.-M. (2021): Computer-assisted approaches to historical language comparison. Habilitation. Friedrich Schiller University: Jena.

Like many other research fields, linguistics is entering the age of big data. We are now at a point where it is possible to see how new research questions can be formulated-and old research questions addressed from a new angle or established results verified-on the basis of exhaustive collections of data, rather than small, carefully selected samples. For example, South Asia is often mentioned in the literature as a classic example of a linguistic area, but there is no systematic, empirical study substantiating this claim. Examination of genealogical and areal relationships among South Asian languages requires a large-scale quantitative and qualitative comparative study, encompassing more than one language family. Further, such a study cannot be conducted manually, but needs to draw on extensive digitized language resources and state-of-the-art computational tools. We present some preliminary results of our large-scale investigation of the genealogical and areal relationships among the languages of this region, based on the linguistic descriptions available in the 19 tomes of Grierson's monumental Linguistic Survey of India (1903-1927), which is currently being digitized with the aim of turning the linguistic information in the LSI into a digital language resource suitable for a broad array of linguistic investigations.

In this paper, we investigate how the prediction paradigm from machine learning and Natural Language Processing (NLP) can be put to use in computational historical linguistics. We propose word prediction as an intermediate task, where the forms of unseen words in some target language are predicted from the forms of the corresponding words in a source language. Word prediction allows us to develop algorithms for phylogenetic tree reconstruction, sound correspondence identification and cognate detection, in ways close to attested methods for linguistic reconstruction. We will discuss different factors, such as data representation and the choice of machine learning model, that have to be taken into account when applying prediction methods in historical linguistics. We present our own implementations and evaluate them on different tasks in historical linguistics.

This paper presents the results of an exercise in lexical comparison between language families: In order to test the statistical significance of the quantity and quality of the author's lexical comparisons between reconstructed Proto-Uralic and Proto-Eskimo roots, as presented in a previous paper, in this paper the author undertakes a thorough investigation of the quantity and quality of lexical comparisons that can be found between Pre-Basque and Proto-Eskimo roots, which clearly belong to unrelated language families. For the purpose of classifying the quality of semantic connections in lexical comparisons, the author proposes a provisional system of 5 Tiers, in which Tier 1 represents the highest level of quality of semantic connection and Tier 5 represents the lowest level. On this basis, the paper presents the results of a thorough comparison of Proto-Eskimo roots with initial *q-and Pre-Basque roots with initial *g-. It is shown that there are in fact very few such lexical comparisons of any significant phonological and semantic similarity that occur by random chance coincidence: Only 3 of 44 Pre-Basque roots with initial *g-can be found to exhibit a valid phonological resemblance with a Proto-Eskimo root with initial *q-and a Tier 1 or Tier 2 level of quality in their semantic connection. By contrast, the author has identified 14 of 76 Proto-Uralic roots with initial *k-that exhibit a comparable phonological correspondence with a Proto-Eskimo root with initial *q-and a Tier 1 or Tier 2 level of quality in their semantic connection. The paper concludes that this investigation points to a significantly strong possibility that the Uralic-Eskimo lexical comparisons are the result of a genetic language relationship rather than the result of mere random chance coincidence.

A new approach is proposed for studying Arabic morphosemantics, set in the belief that language evolves, has tendency for economy and simplification, and avoids synonymy. It is a corpus-based approach that utilizes, for theorization, modern written and spoken usage trends and native speaker questionnaires to support its claims. In demonstration, it is used here to study the meanings of verb-stem templates; it has established that each morphological pattern has a unique meaning in Modern Standard Arabic: I: eventuality; II: repetitiveness; III: commencement; IV: causation; V: reflexivity of repetitiveness; VI: reciprocity; VII: reflexivity; VIII: exertion; IX: acquisition of a color or trait; X: requisition.
Key Words: Arabic; morphosemantics; morphological pattern; corpus-based approach; pattern frequency; linguistic survey.

In order to better support the text mining of historical texts, we propose a combination of complementary techniques from Geographical Information Systems, computational and corpus linguistics. In previous work, we have described this as 'visual gisting' to extract important themes from text and locate those themes on a map representing geographical information contained in the text. Here, we describe the steps that were found necessary to apply standard analysis and resolution tools to identify place names in a specific corpus of historical texts. This task is seen as an initial and prerequisite step for further analysis and comparison by combining the information we extract from a corpus with information from other sources, including other text corpora. The process is intended to support close reading of historical texts on a much larger scale by highlighting using exploratory and data-driven approaches which parts of the corpus warrant further close analysis. Our case study presented here is from a corpus of Lake District travel literature. We discuss the customisations that we have to make to existing tools to extract placename information and visualise it on a map.

This repository contains the nexus files and MrBayes command files needed for running the experiments to determine the optimal word list size required for inferring the best phylogenies. The paper is forthcoming at <strong>The 27th International Conference on Computational Linguistics (COLING 2018), Santa Fe, New-Mexico, USA</strong>.

Like many other research fields, linguistics is entering the age of big data. We are now at a point where it is possible to see how new research questions can be formulated - and old research questions addressed from a new angle or established results verified - on the basis of exhaustive collections of data, rather than small, carefully selected samples. For example, South Asia is often mentioned in the literature as a classic example of a linguistic area, but there is no systematic, empirical study substantiating this claim. Examination of genealogical and areal relationships among South Asian languages requires a large-scale quantitative and qualitative comparative study, encompassing more than one language family. Further, such a study cannot be conducted manually, but needs to draw on extensive digitized language resources and state-of-the-art computational tools. We present some preliminary results of our large-scale investigation of the genealogical and areal relationships am...

The amount of data from languages spoken all over the world is rapidly increasing. Traditional manual methods in historical linguistics need to face the challenges brought by this influx of data. Automatic approaches to word comparison could provide invaluable help to pre-analyze data which can be later enhanced by experts. In this way, computational approaches can take care of the repetitive and schematic tasks leaving experts to concentrate on answering interesting questions. Here we test the potential of automatic methods to detect etymologically related words (cognates) in cross-linguistic data. Using a newly compiled database of expert cognate judgments across five different language families, we compare how well different automatic approaches distinguish related from unrelated words. Our results show that automatic methods can identify cognates with a very high degree of accuracy , reaching 89% for the best-performing method Infomap. We identify the specific strengths and weak...

Advances in computer-assisted linguistic research have been greatly influential in reshaping linguistic research. With the increasing availability of interconnected datasets created and curated by researchers, more and more interwoven questions can now be investigated. Such advances, however, are bringing high requirements in terms of rigorousness for preparing and curating datasets. Here we present CLICS, a Database of Cross-Linguistic Colexifications (CLICS). CLICS tackles interconnected interdisciplinary research questions about the colexification of words across semantic categories in the world’s languages, and show-cases best practices for preparing data for cross-linguistic research. This is done by addressing shortcomings of an earlier version of the database, CLICS2, and by supplying an updated version with CLICS3, which massively increases the size and scope of the project. We provide tools and guidelines for this purpose and discuss insights resulting from organizing stude...

Advances in computer-assisted linguistic research are greatly influencing and reshaping linguistic investigation. With the increasing availability of interconnected datasets created and curated by researchers, more and more interwoven questions can now be investigated. Such advances, however, are bringing high requirements in terms of rigorousness for preparing and curating datasets. In this work we present CLICS, a Database of Cross-Linguistic Colexifications which aims to both tackle interdisciplinary and interconnected research questions as well as showcasing best practices in preparing data for cross-linguistic research. This is done by addressing shortcomings of an earlier version of the database, CLICS2, and supplying an updated version with CLICS3 which massively increases the size and scope of the project. We provide tools and guidelines for this purpose and discuss insights resulting from organizing student tasks for database updates. PREPRINT: This draft has not been peer-...

In this paper, we describe the problem of cognate identification and its relation to phylogenetic inference. We introduce subsequence based features for discriminating cognates from noncognates. We show that subsequence based features perform better than the state-of-the-art string similarity measures for the purpose of cognate identification. We use the cognate judgments for the purpose of phylogenetic inference and observe that these classifiers infer a tree which is close to the gold standard tree. The contribution of this paper is the use of subsequence features for cognate identification and to employ the cognate judgments for phylogenetic inference.

Resumen: El estudio del cambio gramatical que se aprecia en la lengua constituye un área de justificado interés en la lingüística diacrónica. En este ámbito, los córpora lingüísticos pueden constituir una herramienta idónea para el análisis cuantitativo de las tendencias de uso registradas. Una de las características primordiales de este tipo de enfoques viene dada por la posibilidad de realizar comparaciones estrictas entre los datos de un conjunto representativo de corpora. De este modo, es posible efectuar un análisis estadístico apropiado. En el caso de las construcciones progresivas del inglés, se hace necesario extender la dimensión temporal sobre la que analizar el cambio gramatical que puede atestiguarse. Así, el análisis de las ocurrencias de la construcción desde el período del inglés antiguo permite establecer el alcance y la naturaleza del cambio lingüístico que ha tenido lugar. Por tanto, en esta tesis se describe el cambio gramatical que se observa en las construcciones progresivas del inglés desde que existen registros de esta lengua germánica. El estudio de los patrones de desarrollo del aspecto progresivo en la lengua inglesa se lleva a cabo mediante la comparación de los datos obtenidos del grupo de corpora York-Toronto-Helsinki-Penn. Esto permite sentar las bases para la caracterización completa de la construcción desde su origen. Así, se analiza la evolución de la construcción desde el período del inglés antiguo para concluir con la descripción de la expansión y consolidación que experimenta durante el inglés moderno. Esta caracterización pormenorizada se conduce bajo el marco de los modelos gramaticales y del cambio gramatical de la Lingüística Cognitiva, más concretamente, las Gramáticas de Construcciones, la Gramática Cognitiva, y los estudios sobre la gramaticalización y la construccionalización.