Crouching TIGER, hidden structure: Exploring the nature of linguistic data using TIGER values

In recent years, techniques such as Bayesian inference of phylogeny have become a standard part of the quantitative linguistic toolkit. While these tools successfully model the tree-like component of a linguistic dataset, real-world datasets generally include a combination of tree-like and nontree-like signals. Alongside developing techniques for modeling nontree-like data, an important requirement for future quantitative work is to build a principled understanding of this structural complexity of linguistic datasets. Some techniques exist for exploring the general structure of a linguistic dataset, such as NeighborNets, δ scores, and Q-residuals; however, these methods are not without limitations or drawbacks. In general, the question of what kinds of historical structure a linguistic dataset can contain and how these might be detected or measured remains critically underexplored from an objective, quantitative perspective. In this article, we propose TIGER values, a metric that estimates the internal consistency of a genetic dataset, as an additional metric for assessing how tree-like a linguistic dataset is. We use TIGER values to explore simulated language data ranging from very tree-like to completely unstructured, and also use them to analyze a cognate-coded basic vocabulary dataset of Uralic languages. As a point of comparison for the TIGER values, we also explore the same data using δ scores, Q-residuals, and NeighborNets. Our results suggest that TIGER values are capable of both ranking tree-like datasets according to their degree of treelikeness, as well as distinguishing datasets with tree-like structure from datasets with a nontree-like structure. Consequently, we argue that TIGER values serve as a useful metric for measuring the historical heterogeneity of datasets. Our results also highlight the complexities in measuring treelikeness from linguistic data, and how the metrics approach this question from different perspectives.

The philosophical interpretation of language game theory

I give an informal presentation of the evolutionary game theoretic approach to the conventions that constitute linguistic meaning. The aim is to give a philosophical interpretation of the project, which accounts for the role of game theoretic mathematics in explaining linguistic phenomena. I articulate the main virtue of this sort of account, which is its psychological economy, and I point to the casual mechanisms that are the ground of the application of evolutionary game theory to linguistic phenomena. Lastly, I consider the objection that the account cannot explain predication, logic, and compositionality.

Bayesian phylogenetic analysis of linguistic data using BEAST

Bayesian phylogenetic methods provide a set of tools to efficiently evaluate large linguistic datasets by reconstructing phylogenies—family trees—that represent the history of language families. These methods provide a powerful way to test hypotheses about prehistory, regarding the subgrouping, origins, expansion, and timing of the languages and their speakers. Through phylogenetics, we gain insights into the process of language evolution in general and into how fast individual features change in particular. This article introduces Bayesian phylogenetics as applied to languages. We describe substitution models for cognate evolution, molecular clock models for the evolutionary rate along the branches of a tree, and tree generating processes suitable for linguistic data. We explain how to find the best-suited model using path sampling or nested sampling. The theoretical background of these models is supplemented by a practical tutorial describing how to set up a Bayesian phylogenetic analysis using the software tool BEAST2.

The emergence of systematic argument distinctions in artificial sign languages

Word order is a key property by which languages indicate the relationship between a predicate and its arguments. However, sign languages use a number of other modality-specific tools in addition to word order such as spatial agreement, which has been likened to verbal agreement in spoken languages, and role shift, where the signer takes on characteristics of propositional agents. In particular, data from emerging sign languages suggest that, though some use of a conventional word order can appear within a few generations, systematic spatial modulation as a grammatical feature takes time to develop. We experimentally examine the emergence of systematic argument marking beyond word order, investigating how artificial gestural systems evolve over generations of participants in the lab. We find that participants converge on different strategies to disambiguate clause arguments, which become more consistent through the use and transmission of gestures; in some cases, this leads to conventionalized iconic spatial contrasts, comparable to those found in natural sign languages. We discuss how our results connect with theoretical issues surrounding the analysis of spatial agreement and role shift in established and newly emerging sign languages, and the possible mechanisms behind its evolution.

Constituent order in silent gesture reflects the perspective of the producer

Understanding the relationship between human cognition and linguistic structure is a central theme in language evolution research. Numerous studies have investigated this question using the silent gesture paradigm in which participants describe events using only gesture and no speech. Research using this paradigm has found that Agent–Patient–Action (APV) is the most commonly produced gesture order, regardless of the producer’s native language. However, studies have uncovered a range of factors that influence ordering preferences. One such factor is salience, which has been suggested as a key determiner of word order. Specifically, humans, who are typically agents, are more salient than inanimate objects, so tend to be mentioned first. In this study, we investigated the role of salience in more detail and asked whether manipulating the salience of a human agent would modulate the tendency to express humans before objects. We found, first, that APV was less common than expected based on previous literature. Secondly, salience influenced the relative ordering of the patient and action, but not the agent and patient. For events involving a non-salient agent, participants typically expressed the patient before the action and vice versa for salient agents. Thirdly, participants typically omitted non-salient agents from their descriptions. We present details of a novel computational solution that infers the orders participants would have produced had they expressed all three constituents on every trial. Our analysis showed that events involving salient agents tended to elicit AVP; those involving a non-salient agent were typically described with APV, modulated by a strong tendency to omit the agent. We argue that these findings provide evidence that the effect of salience is realized through its effect on the perspective from which a producer frames an event.

Cultural evolution leads to vocal iconicity in an experimental iterated learning task

Experimental and cross-linguistic studies have shown that vocal iconicity is prevalent in words that carry meanings related to size and shape. Although these studies demonstrate the importance of vocal iconicity and reveal the cognitive biases underpinning it, there is less work demonstrating how these biases lead to the evolution of a sound symbolic lexicon in the first place. In this study, we show how words can be shaped by cognitive biases through cultural evolution. Using a simple experimental setup resembling the game telephone, we examined how a single word form changed as it was passed from one participant to the next by a process of immediate iterated learning. About 1,500 naïve participants were recruited online and divided into five condition groups. The participants in the control-group received no information about the meaning of the word they were about to hear, while the participants in the remaining four groups were informed that the word meant either big or small (with the meaning being presented in text), or round or pointy (with the meaning being presented as a picture). The first participant in a transmission chain was presented with a phonetically diverse word and asked to repeat it. Thereafter, the recording of the repeated word was played for the next participant in the same chain. The sounds of the audio recordings were then transcribed and categorized according to six binary sound parameters. By modelling the proportion of vowels or consonants for each sound parameter, the small-condition showed increases of front unrounded vowels and the pointy-condition increases of acute consonants. The results show that linguistic transmission is sufficient for vocal iconicity to emerge, which demonstrates the role non-arbitrary associations play in the evolution of language.

Flexibility in wild infant chimpanzee vocal behavior

How did human language evolve from earlier forms of communication? One way to address this question is to compare prelinguistic human vocal behavior with nonhuman primate calls. An important finding has been that, prior to speech and from early on, human infant vocal behavior exhibits functional flexibility, or the capacity to produce sounds that are not tied to one specific function. This is reflected in human infants’ use of single categories of protophones (precursors of speech sounds) in various affective circumstances, such that a given call type can occur in and express positive, neutral, or negative affective states, depending on the occasion. Nonhuman primate vocal behavior, in contrast, is seen as comparably inflexible, with different call types tied to specific functions and sometimes to specific affective states (e.g. screams mostly occur in negative circumstances). As a first step toward addressing this claim, we examined the vocal behavior of six wild infant chimpanzees during their first year of life. We found that the most common vocal signal, grunts, occurred in a range of contexts that were deemed positive, neutral, and negative. Using automated feature extraction and supervised learning algorithms, we also found acoustic variants of grunts produced in the affective contexts, suggesting gradation within this vocal category. In contrast, the second most common call type of infant chimpanzees, the whimpers, was produced in only one affective context, in line with standard models of nonhuman primate vocal behavior. Insofar as our affective categorization reflects infants’ true affective state, our results suggest that the most common chimpanzee vocalization, the grunt is not affectively bound. Affective decoupling is a prerequisite for chimpanzee grunts (and other vocal categories) to be deemed ‘functionally flexible’. If later confirmed to be a functionally flexible vocal type, this would indicate that the evolution of this foundational vocal capability occurred before the split between the Homo and Pan lineages.

How vocal temporal parameters develop: a comparative study between humans and songbirds, two distantly related vocal learners

Human infants acquire motor patterns for speech during the first several years of their lives. Sequential vocalizations such as human speech are complex behaviors, and the ability to learn new vocalizations is limited to only a few animal species. Vocalizations are generated through the coordination of three types of organs: namely, vocal, respiratory, and articulatory organs. Moreover, sophisticated temporal respiratory control might be necessary for sequential vocalization involving human speech. However, it remains unknown how coordination develops in human infants and if this developmental process is shared with other vocal learners. To answer these questions, we analyzed temporal parameters of sequential vocalizations during the first year in human infants and compared these developmental changes to song development in the Bengalese finch, another vocal learner. In human infants, early cry was also analyzed as an innate sequential vocalization. The following three temporal parameters of sequential vocalizations were measured: note duration (ND), inter-onset interval, and inter-note interval (INI). The results showed that both human infants and Bengalese finches had longer INIs than ND in the early phase. Gradually, the INI and ND converged to a similar range throughout development. While ND increased until 6 months of age in infants, the INI decreased up to 60 days posthatching in finches. Regarding infant cry, ND and INI were within similar ranges, but the INI was more stable in length than ND. In sequential vocalizations, temporal parameters developed early with subsequent articulatory stabilization in both vocal learners. However, this developmental change was accomplished in a species-specific manner. These findings could provide important insights into our understanding of the evolution of vocal learning.

Pantomime as the original human-specific communicative system

We propose reframing one of the key questions in the field of language evolution as what was the original human-specific communicative system? With the help of cognitive semiotics, first we clarify the difference between signals, which characterize animal communication, and signs, which do not replace but complement signals in human communication. We claim that the evolution of bodily mimesis allowed for the use of signs, and the social-cognitive skills needed to support them to emerge in hominin evolution. Neither signs nor signals operate single-handedly, but as part of semiotic systems. Communicative systems can be either monosemiotic or polysemiotic—the former consisting of a single semiotic system and the latter, of several. Our proposal is that pantomime, as the original human-specific communicative system, should be characterized as polysemiotic: dominated by gesture but also including vocalization, facial expression, and possibly the rudiments of depiction. Given that pantomimic gestures must have been maximally similar to bodily actions, we characterize them as typically (1) dominated by iconicity, (2) of the primary kind, (3) involving the whole body, (4) performed from a first-person perspective, (5) concerning peripersonal space, and (6) using the Enacting mode of representation.