1Introduction
Historical Linguistics and Endangered Languages1
Patience Epps, Danny Law and Naâama Pat-El
The discipline of historical linguistics engages with the intersection of the particular and the general â the individual histories of specific languages and the features that define them, and the universal processes and cross-linguistic tendencies that characterize language change. Our reconstructions of particular linguistic histories rely heavily on the predictive power of our generalizations, drawing on such observations as the tendency for stops to give rise to fricatives, and spatial expressions to temporal ones. These generalizations are thus expected to be comprehensive, substantiated, and well established. Yet they have tended to be heavily based on observations drawn from languages that are relatively well studied within the European tradition, and especially on languages with long written histories â centered primarily on the Indo-European language family, which is still the focus of much historical work.
Many widely held views concerning possible and probable pathways and mechanisms of language change have thus taken shape with reference to only a small fraction of the worldâs language families, geographic regions, and sociocultural contexts. Yet just as larger comparative data sets indicate that some features of Indo-European languages, such as relative pronouns and subject-verb inversion in questions, are in fact quite rare cross-linguistically (see e.g. Comrie 2006; Haspelmath 2001), the same is presumably also true of the particular pathways of change through which these and many other features emerge. The converse is also likely, in that some cross-linguistically relevant mechanisms of change may seem vanishingly rare from an Indo-European perspective. Similarly, non-European languages with only modern attestations, such as many languages of Africa and the Americas, have received far less robust diachronic inquiry than languages with a longer written history, such as Chinese. Thus one is more likely to find insights on processes of language change in the literature on Indo-European or Sino-Tibetan languages than in the literature on Chadic languages; and such insights are in turn used to support historical analyses that may be applied far beyond the languages or language families from which the generalizations derive (Pat-El 2020). One need only look as far as glottochronology, with its assumption of constant rates of replacement in basic vocabulary, for an example of an initiative that was based largely on Indo-European languages and quickly foundered when its scope was expanded (Lees 1953; Bergsland and Vogt 1962; Guy 1983). A paucity of well-informed research on lesser-studied languages has consequences.
Such questions of variability across languages and regions relate to the notion of âuniformitarianismâ in language change, and to its predictive power. As Walkden (2019) points out, while methodological uniformitarianism (âactualismâ) remains a key operational null hypothesis for historical linguists â as it is leveraged by the authors in this volume â more substantive assumptions relating to uniformity across states, rates, and outcomes of change are difficult to sustain. As we gain insights into the diversity of geographic regions, population histories, sociocultural contexts, and typological profiles of the worldâs languages, we may raise more nuanced questions regarding characterizations like that of Labov (1972: 275): âthe forces operating to produce linguistic change today are of the same kind and order of magnitude as those which operated five or ten thousand years agoâ; and of Lass (1997: 29): âthe (global, cross-linguistic) likelihood of any linguistic state of affairs (structure, inventory, process, etc.) has always been roughly the same as it is nowâ (see Trudgill 2011). As more and more languages from under-documented regions and language families are brought into the scope of our investigation, we face mounting questions about how our observations relating to uniformity must interact with those of variability across languages, regions, and time periods â and what this diversity implies for our ability to make robust and specific predictions about the rates, pathways, and outcomes associated with language change (Janda and Joseph 2003; Newmeyer 2002; Bergs 2012; Blasi et al. 2019). These questions in turn allow us to refine our toolkit, honing our understanding of what in fact is uniform across languages, and of how our methods should reflect this.
As the chapters in this volume explore, a full response to the questions raised here requires broad and empirically based exploration. It has long been recognized that evidence from understudied languages both complicates and enhances our understanding of linguistic processes and taxonomies (Haspelmath 2007). And just as documentary linguistics has been shifting the focus of linguistic typology from absolute universals to tendencies, work on understudied languages and varieties is providing new insights concerning the dynamics of language change and the structure and fabric of global linguistic diversity (Evans and Levinson 2009; Epps 2010).
Recent decades have seen an explosion of work focused on the documentation and description of endangered and lesser-studied languages, with many initiatives led by and/or carried out in close collaboration with their speakers and community members. At the same time, however, there has been a sharp increase in the rates of language endangerment and loss: Of some 7000 languages spoken worldwide today, 45% of these can be considered endangered, and nearly a quarter of the worldâs approximately 420 known language families are now without speakers (L. Campbell 2016: 256; see also Krauss 1992, inter alia).
Just as more linguistic evidence is bound to contribute to a more nuanced understanding of the mechanisms of language change, the loss of linguistic diversity robs us of the diversity of that linguistic evidence. This point holds not only on the level of languages and language families, but from a more fine-grained perspective as well â a dwindling speaker population is likely to drastically reduce or alter patterns of variation within a given speech community, which may relate directly to processes of language change (Weinreich, Labov, and Herzog 1968; Labov 1994, inter alia). In addition, as an integral part of a larger cultural and expressive ecology, languages communicate and reproduce aspects of the cultural context in which they are spoken; and the loss of languages, varieties, styles, and registers entails the loss of key aspects of the cultural context in which to understand and interpret language change (Woodbury 1993). Recent decades have seen a return to a focus on language in its cultural context, and an interest in how grammar and lexicon intersect with discourse and verbal art (see e.g. Epps, Webster, and Woodbury 2017). A deeper awareness of language history in its cultural context may also be meaningful for the communities in which the languages are spoken, as we observe below.
As we add to our knowledge of understudied and endangered languages, we come across more and more novel and important insights concerning pathways of grammaticalization, forms and dynamics of contact-driven change, and further diachronic relationships among lexical and grammatical categories. In many cases, these new data provide additional support to our understanding of cross-linguistically established processes and methods, and help to confirm previous predictions. In a particularly celebrated example, data from Swampy Cree corroborated Bloomfieldâs (1925, 1928) reconstruction of an additional consonant cluster in Central Algonquian, attested in its sister languages only via a distinct correspondence pattern. Another example can be seen in Bowern et al.âs (2011) finding that lexical borrowing rates among hunter-gatherer languages are broadly consistent with borrowing rates cross-linguistically, despite prior proposals to the contrary. Data from such understudied languages provide a plethora of new opportunities to apply the Comparative Method, and to test its limits (see Pick, this volume).
Studies of diachrony in endangered languages are also expanding our understanding of the range of trajectories of change, and of the diversity of possible sources for emergent morphological forms and categories. For example, Gerdts and Hinkson (2004) show that the dative applicative in Halkomelem (Salishan, North America) derives from a noun meaning âface,â a previously undocumented source for an applicative. Epps (2008) demonstrates that the future tense marker in Hup (Naduhupan family, Amazonia) derives ultimately from a noun meaning âstick, woodâ â a source that appears highly improbable at first glance, but is in fact consistent with our wider understanding of diachronic pathways in each of the steps that connected these two endpoints in a grammaticalization chain (from noun > classifier > nominalizer > purpose adverbial > future tense marker). Phonological systems can also reveal diachronic surprises, as seen in Dufourâs (this volume) exploration of a typologically unusual form of interaction between stress and vowel quality.
Some of these insights do more than simply corroborate or expand our understanding of the dynamics of change; they may also challenge our assumptions about the universality of particular rates and pathways. Work on endangered and understudied languages is feeding a growing list of ways in which processes of language change are sensitive to social, cultural, and/or typological variables. For example, Sweetserâs (1990) generalization that verbs of visual perception (i.e. âseeâ) are the principal source of cognition verbs (âthink,â âknow,â etc.) was taken to be a universal fact about human language, though based on a sample of mostly Indo-European languages. However, Evans and Wilkins (2000) demonstrate that in Australian languages the primary source is aural perception (âhearâ), and observe that this pathway is correlated with a...
