Reconstructing prehistoric languages

Abstract

This theme issue builds on the surge of interest in the field of language evolution as part of the broader field of human evolution, gathering some of the field's most prominent experts in order to achieve a deeper, richer understanding of human prehistory and the nature of prehistoric languages. Taken together, the contributions to this issue begin to outline a profile of the structural and functional features of prehistoric languages, including the type of sounds, the nature of the earliest grammars, the characteristics of the earliest vocabularies and some preferred uses, like conversation and insult. By also correlating certain specific features of language with the changes in brain organization during prehistory, the contributions to this issue directly engage the genetic and the neuroscientific aspects of human evolution and cognition.

This article is part of the theme issue ‘Reconstructing prehistoric languages'.

1. Introduction

The last two decades have seen some notable discoveries in a variety of disciplines relevant for language evolution, as well as some attempts at cross-fertilization and synthesis with the goal to move the field forward. This theme issue brings together some of the field's most prominent experts in order to achieve an improved understanding of the nature of prehistoric languages. It thus addresses an important gap in our understanding of human cognitive and behavioural evolution, making use of the most recent scientific developments. In the last decades, palaeoanthropological and especially paleogenetic data have accumulated almost exponentially, enabling a better understanding of the biological changes that prompted the emergence of our species, as well as some key changes in human behaviour. Nonetheless, as far as prehistoric languages are concerned, even though the classic linguistic comparative method has been quite successful in reconstructing certain features of extinct languages, this method cannot confidently reach beyond 10 000 years and thus cannot be used for any detailed reconstructions of the languages spoken by humans in remote times. This issue demonstrates that new methods can indeed be developed, as well as that highly interdisciplinary approaches are needed, engaging a variety of fields of study with widely different methodologies, including: linguistics, language pathologies, anthropology, archeology, evolutionary biology, genetics, computational science and neuroscience.

All contributions offer some projections into human prehistory, shedding light on different aspects of prehistoric languages, and most also offer some predictions and possibilities for testing. This issue includes both theoretical (connecting the dots) proposals and concrete empirical (including methodological) contributions, both of which are essential in advancing the knowledge in this field, especially at this stage when solid hypotheses and frameworks in which they can be interpreted are only beginning to emerge. Overall, the contributions assume gradualism, i.e. incremental, rather than abrupt, emergence of linguistic complexity, whether in sound systems, grammars, the metaphorical domain, the organization of the brain, etc., with some authors also extending this continuity to other species, particularly to our closest relatives, most notably Neanderthals. The question of the relative timing of the emergence of our cognitive hardware for language, including our mode of cognition and our ability for learning and using languages, is also tackled in the issue, as well as different views of this emergence, including the hypotheses that see the origins of language as an enhancer of (other) cognitive abilities, or the possibility that language, cognition, and behaviour were engaged in a sustained feedback loop. In this respect, variation and variability are important topics of this issue, including cross-linguistic typological variation, as well as genetic variability in humans (encompassing also cognitive disorders), both of which are characterized as gradual, continuous clines, rather than abrupt, categorical phenomena. According to one of the contributions to the issue, Dediu et al. [1], this variability in humans, from the molecular, to the anatomical, to the psychological and cognitive, is ‘mostly small and quantitative, and results in a wide range of “normality” that grades into the “pathological”.’ Yet, as these authors point out, this variation is informative enough to merit advancing and testing a variety of hypotheses about human origins that go beyond the simplistic assumption that languages emerged full-fledged in a sudden way at some point in our history.

2. Themes and topics

While the contributions in this issue often address multiple topics, and there will clearly be overlaps among the sections (which can be viewed as a good opportunity for cross-fertilization), the issue can still be roughly organized around the following four general topics:

(a). Part One: prehistoric sounds and gestures

The contributions in this section address the medium in which prehistoric linguistic communication took place, including sounds and gestures. Dediu et al. [1] report on the experiment that extracted detailed anatomical information about the hard structures of the vocal tract from the remains of three historical populations from The Netherlands, which was then used to conduct simulations of their vowel production. According to the authors, this method can be applied to probe even deeper into prehistory, including analysing remains that are quite old, and not necessarily best preserved, with a goal to draw informed inferences about the speech abilities of long-gone humans, as well as the range of variability. Dellert et al. [2] explore the role of vocal iconicity in contributing to the stability of sound inventories. Building on automated homologue clustering and sound sequence alignment of basic words from more than one hundred Eurasian languages, they show that the most stable sounds are found in the most iconic words. As the authors discuss, one possibility accounting for this finding is that it reflects how first words were coined when language evolved, with high-stability sound groups surviving in iconic mappings. Everett [3] provides some more specific inferences about the nature of sound inventories and prosodic features of prehistoric languages. Based on the sound inventories of extant languages, he reconstructs phonological and phonetic constraints, mostly biomechanical in nature, that are suggestive of Upper Paleolithic languages having medium size inventories, preferring nasal and voiceless stop consonants over posterior voiced consonants, and relying on peripheral vowels [i], [a] and [u] and their variants. Overall, he sees prehistoric sounds as adapting to subtle pressures, cognitive, behavioural and environmental, that have varied with time across populations. Moran et al.'s [4] contribution explores the putative effect of one of the common mechanisms of language change, namely language contact, on the speech sounds of prehistoric languages. More specifically, these authors examine the impact of language contact on sound inventories of languages during the past two millennia, finding evidence for differences between archaic languages and present-day languages, with language contact homogenizing initial diversity in recent millennia. Finally, Etxepare & Irurtzun [5] explore the possibility that hand stencils (i.e. representations of the hand that are formed by placing them against a wall and covering the surrounding area in pigment) encode a sort of basic sign language. Their hypothesis is based on the finding that many prehistoric hand stencils appear with missing fingers, suggesting that each different shape might represent a different hand configuration as observed in sign languages. More specifically, they hypothesize that these hand stencils with missing fingers might be part of an ‘alternate sign language’, that is, a simplified sign language used by hunter–gatherer bimodal communities. This contribution is also of interest for broader hypotheses about language evolution, particularly for those claiming that language might have evolved from gestural systems, and only later switched to orality. Gestures are further discussed in Żywiczyński et al. [6], as mentioned in Part Three.

(b). Part Two: prehistoric grammar and lexicon

While two other papers touch on the properties of earliest grammars and vocabularies (Hartman & Pleyer [7] and Benítez-Burraco & Progovac [8]), the contributions by Calude [9] and by Gil [10] focus most directly on these questions. Calude provides an overview of the research on number words, concluding that numerals 1–5 are the most stable across cultures, often being preserved for thousands of years, and that they are thus the ones that should be reconstructed for prehistoric languages, while higher numerals have emerged as an innovative tool in more recent times. As for Gil's [10] contribution, drawing on a quantitative analysis involving the sample of 868 languages, he postulates a covariance relationship between grammatical complexity and certain aspects of cultural/socio-political complexity, inferring that earliest languages in prehistoric times had measurably simpler syntax, including simpler expressions of predication and tense-aspect marking. This consideration is consistent with the postulated proxies of proto-syntax in Benítez-Burraco & Progovac [8] (more below), who offer some concrete examples of compounds serving as proxies or ‘living fossils' of the earliest grammars. Likewise, according to the framework of Construction Grammar, adopted and discussed by Hartman & Pleyer [7] (more below), there is some continuity as well between the grammar and the lexicon from prehistoric times to present days.

(c). Part Three: prehistoric behaviour, cognition and the brain

The contributions in this section focus on reconstructing prehistoric behaviour, cognition and/or the brain organization with the aim of inferring structural and functional aspects of prehistoric languages. Two contributions (Gil & Shen [11] and Benítez-Burraco & Progovac [8]) address the question of the concrete versus abstract nature of the first vocabularies and the role of metaphoricity/cross-modality in the shaping of the first grammars. Gil & Shen [11] consider a hypothesis that unidirectionality/asymmetry in the metaphorical domain evolves gradually, in several steps, with the first step correlating with the emergence of asymmetrical grammars universal to all humans (but not found in other species). In this respect, the authors give grammars an active role in shaping cognitive abilities. Furthermore, in an experimental setting, Gil & Shen [11] found that there is variability between speakers of two different languages in this respect, concluding that the development of unidirectionality in the metaphorical domain is a gradual, ongoing process. As for Benítez-Burraco & Progovac [8], they focus on the nature of selected proxies (fossils) of the earliest grammars, specifically those relevant to metaphoricity (cross-modality) and verbal aggression (insult), as found in living languages, including in disordered language. The authors purport to solve the puzzle of the link between aggression management, cross-modality and language complexity, and particularly, a persistent co-occurrence found in language disorders between heightened aggressive impulses and diminished metaphorical/figurative language. Invoking the evolutionary framework of self-domestication (SD), their solution outlines a specific scenario for the evolution of the cortico-striatal brain networks, as it was engaged in a feedback loop with the taming of aggression and the emergence of the simplest grammars and metaphorical language. Disorders are also the focus of Code's [12] contribution. Through his discussion of aphasia resulting from brain damage, Code's paper reinforces the crucial role of cortico-striatal brain networks for language evolution, including specifically the basal ganglia, and proposes that the specific ways in which patients recover from aphasia and use certain vocalizations and speech automatisms mimics the early emergence of language in prehistory. This includes the use of reduplicated simple syllables, the structures also characteristic of much of ideophonic language, ideophones arguably being of direct relevance for the evolution of language, as discussed as well by Benítez-Burraco & Progovac [8]. Overall, both Code [12] and Benítez-Burraco & Progovac [8] consider the role of (dis)inhibition in cortico-striatal circuits, whether in brain damage or in cognitive disorders, where evolutionary older subcortical systems get (partly) released from the inhibitory control of the later-evolved higher cortical levels. Finally, the contributions by Żywiczyński et al. [6] and Ferretti & Adornetti [13] address broader aspects of language communication in the past with the aim as well of inferring the nature of prehistoric languages and how they might have evolved. Żywiczyński et al. [6] examine two prerequisites for human communication, namely, the ability to use our body for conveying meanings (mimesis), and a form of sociality that enables a default level of honest communication (platform of trust). According to these authors, these two innovations afforded the emergence of pantomime, which, in turn, enables the transmission of objective information via gestures. With time, gestures became less iconic and more compositional, including the emergence of combinatorial rules for conveying more complex meanings. In this respect, similarly to some of the features of the disordered speech observed in patients with developmental or acquired language disorders (as depicted by Benítez-Burraco & Progovac [8], and Code [12], respectively), present-day pantomimic manifestations can also be regarded as a language fossil, according to these authors. Pantomime is also the object of interest of the contribution by Ferretti & Andornetti [13]. These authors explore in more detail the further steps in the transition from pantomimic communication to full-fledged oral languages. According to their hypothesis, this transition was fuelled by the adaptive value of having more efficient tools for persuading and convincing others, for which storytelling (first achieved through pantomime and later via oral forms of narrative) can be regarded as a prominent resource.

(d). Part Four: modelling prehistoric languages

In truth, most if not all the papers in the issue suggest new experimental approaches to test their predictions and to improve future research. The contributions in this section do so more prominently and systematically, advancing a variety of new methods to model prehistoric languages. These include Hartman & Pleyer's [7] reliance on the framework of Construction Grammar; Ceolin et al.'s [14] reliance on a formal syntactic framework; Schreyer & Adger's [15] insights from languages constructed for popular culture media; and Wichmann & Rama's [16] statistical modelling of geographical homelands of language families. Hartman & Pleyer [7] advocate using the framework of Construction Grammar to understand how new constructions can arise from prelinguistic systems, as this framework considers a continuum of constructions, ranging from single words to sentences, and from concrete to highly abstract, also positing continuity between the lexicon and the grammar. On the other hand, Ceolin et al. [14] used an algorithm based on a formal syntactic theory and proposed that this method can be used to provide a proof of historical relations between different families, even in the absence of a common lexicon from which to draw sound correspondences. According to the authors, this method, based on formal syntactic theory, can significantly expand the time limits imposed by the classical comparative method (see Introduction). Schreyer & Adger's [15] contribution relies on the insights obtained from their own construction of prehistoric languages for popular culture media, including for the film Alpha (2018). Finally, Wichmann & Rama [16] compare different Bayesian and non-Bayesian methods commonly used for inferring the homeland of language families, with the aim of finding the most useful one in terms of performance.

3. Concluding remarks

Overall, all the contributions in the theme issue offer some projections into prehistory, shedding light on different aspects of prehistoric languages. While certainly these postulates will benefit from further testing and confirmation, and while many other aspects of early human languages are yet to be reconstructed, these contributions not only begin to outline a profile of earliest human languages, as well as the organization of the brain, the cognition and the behaviour needed to support such languages, but they also offer a variety of methods and experimental paradigms enabling further probing of these questions. This profile includes the nature of the sounds and gestural systems; the nature of early vocabularies (e.g. concrete versus abstract; expressive vocabulary); the nature of earliest grammars, and the uses to which they were subjected (e.g. conversation; insult); the nature and extent of metaphorical extension; the nature and extent of physical versus verbal aggression in prehistoric populations; and the nature of the biological prerequisites, including speech organs, brain organization and behavioural features.

Even though it is often claimed that language evolution cannot be studied scientifically because language does not fossilize, the contributions to this issue provide evidence that it is indeed possible to reconstruct several aspects of early stages of language evolution with some confidence, using various precise methods to draw informed inferences about the past from the present (variation), often invoking the notion of linguistic ‘fossils' or proxies for prehistoric languages. Even small improvements and adjustments in our understanding of language origins will directly impact upon the assumptions and postulates of many related disciplines, opening avenues for fresh hypotheses and findings. As this issue purports to demonstrate, the key to success in this field is to engage in aggressively interdisciplinary endeavours, cross-fertilizing research coming from vastly different fields and methodologies, and then advancing and testing novel hypotheses, including those that at first sight may seem implausible.

Data accessibility

This article has no additional data.

Authors' contributions

The authors wrote this introduction equally.

Competing interests

We declare we have no competing interests.

Funding

We received no funding for this study.