Non-representational cognitive science is a promising research field that provides an alternative to the view of the brain as a “computer” filled with symbolic representations of the world and cognition as “calculations” performed on those symbols. Autopoiesis is a biological, bottom-up, non-representational theory of cognition, in which representations and meaning are framed as explanatory concepts that are constituted in an observer’s description of a cognitive system, not operational concepts in the system itself. One of the problems of autopoiesis, and all non-representational theories, is that they struggle with scaling up to high-level cognitive behaviour such as language. The Iterated Learning Model is a theory of language evolution that shows that certain features of language are explained not because of something happening in the linguistic agent’s brain, but as the product of the evolution of the linguistic system itself under the pressures of learnability and expressivity. Our goal in this work is to combine an autopoietic approach with the cultural transmission chains that the ILM uses, in order to provide the first step in an autopoietic explanation of the evolution of language. In order to do that, we introduce a simple, joint action physical task in which agents are rewarded for dancing around each other in either of two directions, left or right. The agents are simulated e-pucks, with continuous-time recurrent neural networks as nervous systems. First, we adapt a biologically plausible reinforcement learning algorithm based on spike-timing dependent plasticity tagging and dopamine reward signals. We show that, using this algorithm, our agents can successfully learn the left/right dancing task and examine how learning time influences the agents’ task success rates. Following that, we link individual learning episodes in cultural transmission chains and show that an expert agent’s initial behaviour is successfully transmitted in long chains. We investigate the conditions under which these transmission chains break down, as well as the emergence of behaviour in the absence of expert agents. By using long transmission chains, we look at the boundary conditions for the re-establishment of transmitted behaviour after chain breakdowns. Bringing all the above experiments together, we discuss their significance for non-representational cognitive science and draw some interesting parallels to existing Iterated Learning research; finally, we close by putting forward a number of ideas for additions and future research directions.

This study explores how Anglophone fiction from the mid-Victorian period to the outbreak of the First World War acted as an imaginative testing-ground for theories of the evolution of language. Debates about the past development and the future of language ranged beyond the scope of empirical data and into speculative narrative. Fiction offered to realize such narratives in detail, building imaginative worlds out of different theories of language evolution. In the process, it also often tested these theories, exposing their contradictions. The lack of clear boundaries between nature and culture in language studies of the period enabled fictions of language evolution to explore questions to which contemporary researchers have returned. To what extent is communication instinctive or conventional? How do social and biological factors interact in the production of meaning? The study traces two opposing tendencies of thought on language evolution, naming them language ‘progressivism’ and ‘vitalism’. Progressivism imagined speakers evolving away from involuntary, instinctive vocalizations to extert rational control over their discourse with mechanical precision. By contrast, language vitalism posited a mysterious, natural power in words which had weakened and fragmented with the rise of writing and industrial society. Certain genres of fiction lent themselves to exploration of these ideas, with utopian tales seeking to envision the end-goals of progressive theory. Representations of primitive language in imperial and prehistoric romances also promoted progressivism by depicting the instinctive, irrational speech from which ‘civilization’ was imagined as advancing away. Conversely, much historical and invasion fiction idealized a linguistic past when speech had expressed natural truth, and the authentic folk origins of its speakers. Both progressivism and vitalism were undermined through the late nineteenth century by developments in biology, which challenged claims of underlying stability in nature or purpose in change. Simultaneously, philologists increasingly argued that meaning was conventional, attacking models of semantic progress and degradation. In this context, a number of authors reconceptualized language in their fiction as a mixture of instinct and convention. These imaginative explorations of the borderlands between the social and biological in communication prefigured many of the concerns of twenty-first-century biosemiotics.

Dans cette thèse nous étudions le rôle de la motivation intrinsèque dans l’émergence et le développement des systèmes communicationnels. Notre objectif est d’explorer comment des populations d’agents artificiels peuvent utiliser un système de motivation computationnel particulier, appelé l’autotelic principle, pour réguler leur développement linguistique et les dynamiques qui en résultent au niveau de la population.Nous proposons d’abord une mise en œuvre concrète de l’autotelic principle. Le noyau de ce système repose sur l’équilibre des défis, des tâches à accomplir afin d’atteindre un objectif, et des compétences, les capacités que le système peut utiliser pour accomplir les différentes tâches. La relation entre les deux éléments n’est pas stable mais se déstabilise régulièrement lorsque de nouvelles compétences sont acquises, ce qui permet au système de tenter des défis de plus grande complexité. Ensuite, nous testons l’utilité de ce système de motivation dans une série d’expériences sur l’évolution du langage. Dans le premier ensemble d’expériences, une population d’agents artificiels doit développer une langue pour se référer à des objets ayant des caractéristiques discrètes. Ces expériences se concentrent sur la façon dont les systèmes communicatifs non ambigus peuvent émerger lorsque l’autotelic principle est utilisé pour réguler le développement du langage en étapes de difficulté croissante. Dans le deuxième ensemble d’expériences, les agents doivent créer un langage artificiel pour communiquer sur des couleurs. Dans cette partie, on explore comment le système de motivation peut contrôler la complexité linguistique des interactions pour un domaine continu et on examine aussi la validité de l’autotelic principle en tant que mécanisme permettant de réguler simultanément plusieurs stratégies linguistiques de difficulté similaire. En résumé, nous avons démontré à travers de notre travail que l’autotelic principle peut être utilisé comme un mécanisme général pour réguler la complexité du langage développé de manière autonome en domaines discrets et continus. / This thesis studies the role of intrinsic motivation in the emergence and development of communicative systems in populations of artificial agents. To be more specific, our goal is to explore how populations of agents can use a particular motivation system called autotelic principle to regulate their language development and the resulting dynamics at the population level.To achieve this, we first propose a concrete implementation of the autotelic principle. The core of this system is based on the balance between challenges, tasks to be done to achieve a goal, and skills, the abilities the system can employ to accomplish the different tasks. The relation between the two elements is not steady but regularly becomes destabilised when new skills are learned, which allows the system to attempt challenges of increasing complexity. Then, we test the usefulness of the autotelic principle in a series of language evolution experiments. In the first set of experiments, a population of artificial agents should develop a language to refer to objects with discrete values. These experiments focus on how unambiguous communicative systems can emerge when the autotelic principle is employed to scaffold language development into stages of increasing difficulty. In the second set of experiments, agents should agree on a language to communicate with about colour samples. In this part, we explore how the motivation system can regulate the linguistic complexity of interactions for a continuous domain and examine the value of the autotelic principle as a mechanism to control several language strategies simultaneously. To summarise, we have shown through our work that the autotelic principle can be used as a general mechanism to regulate complexity in language emergence in an autonomous way for discrete and continuous domains.

Written language provides a snapshot of linguistic, cultural, and current events information for a given time period. Aggregating these snapshots by studying many texts over time reveals trends in the evolution of language, culture, and society. The ever-increasing amount of electronic text, both from the digitization of books and other paper documents to the increasing frequency with which electronic text is used as a means of communication, has given us an unprecedented opportunity to study these trends. In this dissertation, we use hundreds of thousands of books spanning two centuries scanned by Google, and over 100 billion messages, or ‘tweets’, posted to the social media platform, Twitter, over the course of a decade to study the English language, as well as study the evolution of culture and society as inferred from the changes in language. We begin by studying the current state of verb regularization and how this compares between the more formal writing of books and the more colloquial writing of tweets on Twitter. We find that the extent of verb regularization is greater on Twitter, taken as a whole, than in English Fiction books, and also for tweets geotagged in the United States relative to American English books, but the opposite is true for tweets geotagged in the United Kingdom relative to British English books. We also find interesting regional variations in regularization across counties in the United States. However, once differences in population are accounted for, we do not identify strong correlations with socio-demographic variables. Next, we study stretchable words, a fundamental aspect of spoken language that, until the advent of social media, was rarely observed within written language. We examine the frequency distributions of stretchable words and introduce two central parameters that capture their main characteristics of balance and stretch. We explore their dynamics by creating visual tools we call ‘balance plots’ and ‘spelling trees’. We also discuss how the tools and methods we develop could be used to study mistypings and misspellings, and may have further applications both within and beyond language. Finally, we take a closer look at the English Fiction n-gram dataset created by Google. We begin by explaining why using token counts as a proxy of word, or more generally, ‘n-gram’, importance is fundamentally flawed. We then devise a method to rebuild the Google Books corpus so that meaningful linguistic and cultural trends may be reliably discerned. We use book counts as the primary ranking for an n-gram and use subsampling to normalize across time to mitigate the extraneous results created by the underlying exponential increase in data volume over time. We also combine the subsampled data over a number of years as a method of smoothing. We then use these improved methods to study linguistic and cultural evolution across the last two centuries. We examine the dynamics of Zipf distributions for n-grams by measuring the churn of language reflected in the flux of n-grams across rank boundaries. Finally, we examine linguistic change using wordshift plots and a rank divergence measure with a tunable parameter to compare the language of two different time periods. Our results address several methodological shortcomings associated with the raw Google Books data, strengthening the potential for cultural inference by word changes.

Language is arguably one of the most salient features that distinguish humans from other animal species. However, despite the existence of a large body of relevant theoretical and empirical research, there is currently no consensus as to why language emerged exclusively in the human species or how it evolved its unique communicative structure. In this thesis, I therefore take a multi-pronged approach to analysing and testing several different hypotheses for the biological function and evolution of language. In Chapter I, I review the evidence and theoretical arguments for each of these proposals and provide, in place, a synthetic perspective which integrates or eliminates each of these ostensibly competing hypotheses for the biological function of language. In Chapter II, I employ the first experimental test of the interdependence hypothesis: the unique proposal offered to explain the emergence and potential coevolution of language and cooperation in the human species. In pursuit of this experiment, I employed a cooperative social foraging task using small and large groups to determine what factors enable individuals to make sense of information from others and converge upon a group consensus. In Chapter III, I take an experimental approach to determine whether aspects of human language can be characterised in terms of honest signalling theory. In this respect, I test several different proposals predicted by the sexual selection and deception hypotheses for human language function. In Chapter IV, I divert attention away from biological function to focus more closely on language structure. More specifically, I take an experimental approach to the problem of how and indeed whether recursive syntax evolved to be a consistent feature of human language. In pursuit of this experiment, I utilized the Imposing Memory Task (IMT) and a recursive syntax measure, to determine relative performance on each of these cognitive tasks, thereby testing whether recursive syntax may have evolved in tandem with higher-order intentionality (also known as embedded mindreading). Finally, in Chapter V, I discuss the results and implications of these experiments, and possible suggestions for future studies.

In the years since Joseph Smith completed the translation of the Book of Mormon, the English language has changed; some common phraseologies and word meanings are dissimilar to today's. Often, in reading the Book of Mormon, we impose our current definitions onto terms that in 1830 had a different meaning. Our interpretation of these words, as well as the passages in which they are found, is skewed by our modern definitions. These words, when they occur in the Book of Mormon, demonstrate dialectal and obsolete senses. In the case of some words, the dialectal or obsolete sense is so far removed in meaning from the prevailing modern sense that Book of Mormon readers may be unsure how to interpret the meaning.In this study I investigate words in the Book of Mormon text that have taken on different meanings, thus leading to potential misreadings within some passages of the text.

Depuis des siècles, plusieurs penseurs et scientifiques ont abordé la relation entre la coopération, le langage et la cognition sociale. Parmi eux, Trần Đức Thảo (1917-1993) mérite une mention spéciale. Le but de la recherche qui suit est de présenter au lecteur la réflexion philosophique de Thảo sur le langage humain et son évolution. Nous essaierons de tracer les grandes lignes de la théorie de Thảo sur les origines du langage dans ses Recherches sur l'origine du langage et de la conscience (1973) dans lesquelles il a essayé de trouver une synthèse entre philosophie, linguistique, psychologie et anthropologie physique. La découverte du marxisme-léninisme a conduit Thảo à proposer une approche matérialiste et dialectique au problème de la relation entre corps esprit. De cette façon, Thảo a proposé une sorte de tournant matérialiste et historique de la philosophie de la conscience de Husserl qui était au cœur de ses premiers intérêts philosophiques. La théorie de Thảo met en relief la nature sociale du langage et de la cognition, de sorte que l’évolution du langage est inextricablement liée aux relations sociales. Une telle conclusion reposait sur l’hypothèse que le travail est une caractéristique exclusivement humaine qui distingue les humains des animaux. Pour lui, la genèse du langage est dans le travail humain et donc le langage se développe parmi nos ancêtres pré-humains ainsi que chez les humains en réponse aux problèmes posés par la vie matérielle. En gardant à l’esprit que le langage découle des exigences sociales et des besoins du monde matériel, selon Thảo le langage se transforme lui-même au fur et à mesure que la société humaine change. Et compte tenu des racines sociales de la pensée et du langage, la conscience évolue continuellement avec le temps. Dans ce cadre, Thảo a voulu déterminer la nature du langage et son rôle dans les sociétés préhistoriques et son évolution à travers les relations sociales. / Several thinkers and scientists throughout the philosophical and scientific tradition took up the relationship between cooperation, language, and social cognition. Among them, Trần Đức Thảoʼs (1917–1993) deserves a special mention. The purpose of the following research is to introduce the reader to Thảoʼs philosophical reflection on human language and its evolution. We shall attempt to map out the main lines of Thảoʼs theory of language origins set out in his Recherches sur l’origine du langage et de la conscience (1973) that combines philosophy, linguistics, psychology, and anthropology. The discovery of Marxism-Leninism led Thảo to suggest a materialistic and dialectic approach to the mind-body problem. In this way, Thảo tried to suggest a materialist and historical turn of Husserl’s philosophy of consciousness which was at the very heart of his own first philosophical interests. Thảo’s account threw into sharp relief the social nature of both language and cognition, so that language evolution is linked inextricably to social relations. Such a view depended upon the assumption that labour is an exclusively human characteristic which sets humans apart from animals. And the genesis of language is in human labour. In this way of thinking, language develops among both our pre-human ancestors and present humans in response to problems posed by the material life. Bearing in mind that language arises from the social demands and needs of the material world, language is transformed itself as human society changes. And given the social roots of thought and language, consciousness evolves continuously over time. Within this framework, Thảo wanted to determine the nature of language and its role in pre-historical societies and its making through social relations.

Distributed learning is the iterative process of decision-making in the presence of other decision-makers. In recent years, researchers across fields as disparate as engineering, biology, and economics have identified mathematically congruous problem formulations at the intersection of their disciplines. In particular, stochastic processes, game theory, and control theory have been brought to bare on certain very basic and universal questions. What sort of environments are conducive to distributed learning? Are there any generic algorithms offering non-trivial performance guarantees for a large class of models? The first half of this thesis makes contributions to two particular problems in distributed learning, self-assembly and language. Self-assembly refers to the emergence of high-level structures via the aggregate behavior of simpler building blocks. A number of algorithms have been suggested that are capable of generic self-assembly of graphs. That is, given a description of the objective they produce a policy with a corresponding performance guarantee. These guarantees have been in the form of deterministic convergence results. We introduce the notion of stochastic stability to the self-assembly problem. The stochastically stable states are the configurations the system spends almost all of its time in as a noise parameter is taken to zero. We show that in this framework simple procedures exist that are capable of self-assembly of any tree under stringent locality constraints. Our procedure gives an asymptotically maximum yield of target assemblies while obeying communication and reversibility constraints. We also present a slightly more sophisticated algorithm that guarantees maximum yields for any problem size. The latter algorithm utilizes a somewhat more presumptive notion of agents' internal states. While it is unknown whether an algorithm providing maximum yields subject to our constraints can depend only on the more parsimonious form of internal state, we are able to show that such an algorithm would not be able to possess a unique completing rule--- a useful feature for analysis. We then turn our attention to the problem of distributed learning of communication protocols, or, language. Recent results for signaling game models establish the non-negligible possibility of convergence, under distributed learning, to states of unbounded efficiency loss. We provide a tight lower bound on efficiency and discuss its implications. Moreover, motivated by the empirical phenomenon of linguistic drift, we study the signaling game under stochastic evolutionary dynamics. We again make use of stochastic stability analysis and show that the long-run distribution of states has support limited to the efficient communication systems. We find that this behavior is insensitive to the particular choice of evolutionary dynamic, a fact that is intuitively captured by the game's potential function corresponding to average fitness. Consequently, the model supports conclusions similar to those found in the literature on language competition. That is, we expect monomorphic language states to eventually predominate. Homophily has been identified as a feature that potentially stabilizes diverse linguistic communities. We find that incorporating homophily in our stochastic model gives mixed results. While the monomorphic prediction holds in the small noise limit, diversity can persist at higher noise levels or as a metastable phenomenon. The contributions of the second half of this thesis relate to more basic issues in distributed learning. In particular, we provide new results on the problem of distributed convergence to Nash equilibrium in finite games. A recently proposed class of games known as stable games have the attractive property of admitting global convergence to equilibria under many learning dynamics. We show that stable games can be formulated as passive input-output systems. This observation enables us to identify passivity of a learning dynamic as a sufficient condition for global convergence in stable games. Notably, dynamics satisfying our condition need not exhibit positive correlation between the payoffs and their directions of motion. We show that our condition is satisfied by the dynamics known to exhibit global convergence in stable games. We give a decision-theoretic interpretation for passive learning dynamics that mirrors the interpretation of stable games as strategic environments exhibiting self-defeating externalities. Moreover, we exploit the flexibility of the passivity condition to study the impact of applying various forecasting heuristics to the payoffs used in the learning process. Finally, we show how passivity can be used to identify strategic tendencies of the players that allow for convergence in the presence of information lags of arbitrary duration in some games.

The language contact of the inflectional Romani and Spanish, Catalan and other languages of the Iberian Peninsula began in the first half of the 15th century. A long- term and immediate contact between Romani and a language of the majority in several places of Europe resulted in emergence of so called Para-Romani varieties, mixed languages which predominantly make use of the grammar of the surrounding language, whereas the Romani-derived vocabulary is at least partly retained. In the Iberian Peninsula several Para-Romani varieties emerged. The process of their phonological, morphological and lexical evolution from the inflectional Iberian Romani is described in the first part of this thesis. The original form of the Iberian Romani of the 15th century was - according to our opinion - similar to so called early Romani of the Byzantine period that is supposed to end by the 14th century when the Romani language started to disperse throughout Europe. The influence of Spanish (and other languages of the Iberian Peninsula) is evident not only in the morphology and syntax, but also in phonology and semantics of the Iberian Para-Romani varieties. On the other hand, the Romani elements conserved in these varieties show archaic features as well as specific innovations and many of them can hardly be assigned...

No description available.