Skip to main content

Über dieses Buch

Current Work and Open Problems: A Road-Map for Research into the Emergence of Communication and Language Chrystopher L. Nehaniv, Caroline Lyon, and Angelo Cangelosi 1.1. Introduction This book brings together work on the emergence of communication and language from researchers working in a broad array of scientific paradigms in North America, Europe, Japan and Africa. We hope that its multi-disciplinary approach will encourage cross-fertilization and promote further advances in this active research field. The volume draws on diverse disciplines, including linguistics, psychology, neuroscience, ethology, anthropology, robotics, and computer science. Computational simulations of the emergence of phenomena associated with communication and language play a key role in illuminating some of the most significant issues, and the renewed scientific interest in language emergence has benefited greatly from research in Artificial Intelligence and Cognitive Science. The book starts with this road map chapter by the editors, pointing to the ways in which disparate disciplines can inform and stimulate each other. It examines the role of simulations as a novel way to express theories in science, and their contribution to the development of a new approach to the study of the emergence of communication and language. We will also discuss and collect the most promising directions and grand challenge problems for future research. The present volume, is organized into three parts: I. Empirical Investi- tions on Human Language, II. Synthesis and Simulation of Communication and Language in Artificial Systems, and III. Insights from Animal Communication.





1. Introduction

Current Work and Open Problems: A Road-Map for Research into the Emergence of Communication and Language
This book brings together work on the emergence of communication and language from researchers working in a broad array of scientific paradigms in North America, Europe, Japan and Africa.
Chrystopher L. Nehaniv, Caroline Lyon, Angelo Cangelosi

Empirical Investigations on Human Language


2. Evolving Meaning: The Roles of Kin Selection, Allomothering and Paternal Care in Language Evolution

Many contemporary scholars agree that future theories of language evolution need to take a componential approach to language that breaks human language into separate mechanistic components such as vocal imitation, syntactic abilities, and propositional semantics. In this chapter, I discuss the evolution of the last component – the abilities and proclivities underlying honest, complex, propositional meanings. This is both a critical component of language, and one whose evolution is the hardest to explain, precisely because of its apparent uniqueness. Nonetheless, I argue, the comparative approach has important insights to offer in this domain. I briefly discuss the hypothesis that kin selection played an important, but neglected, role in driving the evolution of rich semantic communication. I then review several bodies of comparative data not addressed in previous discussions.
W. Tecumseh Fitch

3. ‘Needs only’ Analysis in Linguistic Ontogeny and Phylogeny

Recently, linguists from several quarters have begun to unpack some of the assumptions and claims made in linguistics over the last 40 years, opening up new possibilities for synergies between linguistic theory and the variety of fields that engage with it. A key point of exploration is the relationship between external manifestations of language and the underlying mental model that produces and understands them. To what extent does it remain reasonable to argue that all humans ‘know’ certain things about language, even if they never demonstrate that knowledge? What is the status of knowledge that is only stimulated into expression by particular cultural input? Many have asked whether the human’s linguistic behaviour can be explained with recourse to less innate knowledge than Chomskian models traditionally assume.
Alison Wray

4. Clues from Information Theory Indicating a Phased Emergence of Grammar

In this chapter we present evidence that there is an underlying local sequential structure in present day language, and suggest that the components of such a structure could have been the basis of a more highly evolved hierarchical grammar. The primary local sequential structure is shown to have its own benefits, which indicate that there could be an intermediate stage in the evolution of grammar, before the advantages of a fully developed syntax were realised.
Caroline Lyon, Chrystopher L. Nehaniv, Bob Dickerson

5. Emergence of a Communication System: International Sign

International Sign (henceforth IS) is a communication system that is used widely in the international Deaf Community. The present study is one of the first to research extensively the origin of both the IS lexicon and grammatical structures. Findings demonstrate that IS is both influenced by naturally evolved sign languages used in grown deaf communities (henceforth SLs) and relies heavily on iconic, universal structures. This paper shows that IS continues to develop from a simplistic iconic system into a conventionalized system with increasingly complex rules.
Rachel Rosenstock

6. Distributed Language: Biomechanics, Functions, and the Origins of Talk

Emphasizing that word-forms are culturally selected, the paper takes a distributed view of language. This is used to frame evidence that, in ontogenesis, language emerges under dual control by adult and child. Since parties gear to each other’s biomechanics, norm-based behaviour prompts affective processes that drive prepared learning. This, it is argued, explains early stages in learning to talk. Next, this approach to external symbol grounding (ESG) is contrasted with ones where a similar problem is treated as internal to the agent. Then, turning to synthetic models, I indicate how the ESG can be used to model either populations of agents or dyads who, using complex signals, transform each other’s agency.
Stephen J. Cowley

Synthesis of Communication and Language in Artificial Systems


7. The Recruitment Theory of Language Origins

Tremendous progress has been made recently on the fascinating question of the origins and evolution of language (see e.g. (55), (7), (9), (31)). There is no widely accepted complete theory yet, but several proposals are on the table and observations and experiments are proceeding. This chapter focuses on the recruitment theory of language origins which we have been exploring for almost ten years now. This theory argues that language users recruit and try out different strategies for solving the task of communication and retain those that maximise communicative success and cognitive economy. Each strategy requires specific cognitive neural mechanisms, which in themselves serve a wide range of purposes and therefore may have evolved or could be learned independently of language.
Luc Steels

8. In silico Evolutionary Developmental Neurobiology and the Origin of Natural Language

It is justified to assume that part of our genetic endowment contributes to our language skills, yet it is impossible to tell at this moment exactly how genes affect the language faculty. We complement experimental biological studies by an in silico approach in that we simulate the evolution of neuronal networks under selection for language-related skills. At the heart of this project is the Evolutionary Neurogenetic Algorithm (ENGA) that is deliberately biomimetic. The design of the system was inspired by important biological phenomena such as brain ontogenesis, neuron morphologies, and indirect genetic encoding. Neuronal networks were selected and were allowed to reproduce as a function of their performance in the given task. The selected neuronal networks in all scenarios were able to solve the communication problem they had to face. The most striking feature of the model is that it works with highly indirect genetic encoding–-just as brains do.
Eörs Szathmáry, Zoltán Szathmáry, Péter Ittzés, GeroŐ Orbaán, István Zachár, Ferenc Huszár, Anna Fedor, Máté Varga, Szabolcs Számadó

9. Communication in Natural and Artificial Organisms: Experiments in Evolutionary Robotics

In the field of ethological studies many efforts of researchers are devoted to understand how animals communicate and what is the role of communication from an evolutionary and functional point of view. Progress in this area might also have an impact on our understanding of human communication since animal and human communication systems share several features (Hauser, 1996). The social function played by human language, for instance, is one of the first traits that allows us to place language in the same evolutionary field as other animal communication systems. Moreover, recently also the idea of the uniqueness of human language regarding the representational fashion of the knowledge and the compositionality of signals is challenged by new findings in primate research that indicate that, in baboons, knowledge is representational, based on properties that have discrete values and, from a certain point of view, propositional (Seyfarth et al., 2005).
Davide Marocco, Stefano Nolfi

10. From Vocal Replication to Shared Combinatorial Speech Codes: A Small Step for Evolution, A Big Step for Language

Humans use spoken vocalizations, or their signed equivalent, as a physical support to carry language. This support is highly organized: vocalizations are built with the re-use of a small number of articulatory units, which are themselves discrete elements carved up by each linguistic community in the articulatory continuum. Moreover, the repertoires of these elementary units (the gestures, the phonemes, the morphemes) have a number of structural regularities: for example, while our vocal tract allows physically the production of hundreds of vowels, each language uses most often 5, and never more than 20 of them. Also, certain vowels are very frequent, like /a,e,i,o,u/, and some others are very rare, like /en/. All the speakers of a given linguistic community categorize the speech sounds in the same manner, and share the same repertoire of vocalizations. Speakers of different communities may have very different ways of categorizing sounds (for example, Chinese use tones to distinguish sounds), and repertoires of vocalizations. Such an organized physical support of language is crucial for the existence of language, and thus asking how it may have appeared in the biological and/or cultural history of humans is a fundamental questions. In particular, one can wonder how much the evolution of human speech codes relied on specific evolutionary innovations, and thus how difficult (or not) it was for speech to appear.
Pierre-Yves Oudeyer

11. Learning and Transition of Symbols: Towards a Dynamical Model of a Symbolic Individual

The remarkable feature of linguistic communications is the use of symbols for transmitting information and mutual understanding. Deacon (1997) pointed out that humans are symbolic species, namely, we show symbolic cognitive activities such as learning, formation, and manipulation of symbols. In research into the origin and the evolution of language, we should elucidate the emerging process of such symbolic cognitive activities.
Takashi Hashimoto, Akira Masumi

12. Language Change among ‘Memoryless Learners’ Simulated in Language Dynamics Equations

In general, all human beings can learn any human language in their first language acquisition. One of the functions of language use is to communicate with others. In the work described here we investigate situations in which learners are exposed to more than one language. Our study leads us to suggest how creole languages could emerge. We make the assumption that the language learners come to acquire one of the languages that is optimal for communication, which would vary according to the environment. It is postulated that the most preferable language in the community would eventually survive and become dominant in competition with other languages, depending on how large a proportion of the people speak it. Accordingly, language change can be represented by population dynamics, examples of which include an agent-based model of language acquisition proposed by Briscoe et al. (2002) and a mathematical framework by Nowak et al. (2001), who elegantly presented an evolutionary dynamics of grammar acquisition in a differential equation, called the language dynamics equation.
Makoto Nakamura, Takashi Hashimoto, Satoshi Tojo

13. The Evolution of Meaning-Space Structure through Iterated Learning

One of the most striking aspects of human linguistic communication is its extensive use of compositionality to convey meaning. When expressing a complex meaning, we tend to use signals whose structure reflects the structure of the meaning to some degree. This property is the foundation upon which the syntax of language is built. It is natural, therefore, that an evolutionary account of human language should contrast compositional communication with a non-compositional, holistic alternative where whole signals map onto whole meanings in an arbitrary, unstructured way. Indeed, Wray (1998) has argued that holistic communication (which is still in evidence in particular contexts today) can be seen as a living fossil of an earlier completely non-compositional protolanguage.
Simon Kirby

14. The Emergence of Language: How to Simulate It

The emergence of language in populations of primates that initially lacked language can be simulated with artificial organisms controlled by neural networks and living, evolving, and learning in artificial environments. Some simulations have already been done but most of the necessary work is a task for the future. We discuss language evolution under two topics: language is learned from others on the basis of genetically inherited predispositions, and language has important influences on human cognition. We propose an evolutionary sequence according to which bipedalism and the emergence of the hands represent a selective pressure for developing an ability to predict the consequences of one’s actions, this ability is the basis for learning by imitating other individuals, and learning by imitating other individuals is applied to learning to imitate their communicative behaviour.
Domenico Parisi, Marco Mirolli

15. Lexical Acquisition with and without Metacommunication

A central concern of work on the evolution of language has been to offer an account for the emergence of syntactically complex structure, which underwrites a compositional semantics. In this chapter we consider the emergence of one class of utterances which illustrate that semantic expressiveness is not correlated with syntactic complexity, namely metacommunicative interaction (MCI) utterances. These are utterance acts in which conversationalists acknowledge understanding or request clarification. We offer a simple characterisation of the incremental change required for MCI to emerge from an MCI-less linguistic interaction system. This theoretical setting underpins and motivates the development of an ALife environment in which the lexicon dynamics of populations that possess and lack MCI capabilities are compared.
Jonathan Ginzburg, Zoran Macura*

16. Agent Based Modelling of Communication Costs: Why Information Can Be Free

What purposes, other than facilitating the sharing of information, can language have served? First, it may not have evolved to serve any purpose at all. It is possible that language is just a side effect of the large human brain — a spandrel or exaptation — that only became useful later. If language is adaptive, this does not necessarily mean that it is adaptive for the purpose of communication. For example Dennett (1996) and Chomsky (1980) have stressed the utility of language in thinking. Also, there are different ways to view communication. The purpose of language according to Dunbar (1993), is to replace grooming as a social bonding process and in this way to ensure the stability of large social groups.
Ivana Čače, Joanna J. Bryson

17. Language Change and the Inference of Meaning

The natural state of living human languages is one of continuous gradual change, underpinned by variation in both form and meaning (Trask, 1996). Small differences in the contexts in which particular utterances are used, or changes in the way in which words are pronounced, accumulate over generations of use to such an extent that the language itself can become unrecognisable in only a few generations (Deutscher, 2005).
Andrew D. M. Smith

18. Language, Perceptual Categories and their Interaction: Insights from Computational Modelling

How do humans acquire perceptual categories? This question is far from being resolved. Specifically the balance between the influence of nature and nurture on perceptual categories remains the topic of heated debate. We present a computational model and take as case study colour categories to study two issues in perceptual category acquisition. The first issue is the effect of linguistic communication on categories during their acquisition: we demonstrate how categories can become coordinated under the influence of language. The second issue concerns the amount of coordination needed between the categories of individuals in order to achieve unambiguous communication. We show that, depending on how strictly linguistic utterances are interpreted, coordination of the individuals’ categories is not always a prerequisite for successful communication.
Tony Belpaeme, Joris Bleys

Insights from Animal Communication


19. Emergence of Linguistic Communication: Studies on Grey Parrots

Most studies on the evolution of communication systems concentrate on the primate lineage, ignoring the concept of parallel lines of evolution. Although phylogenetically remote from humans, some birds—particularly Grey parrots—share many cognitive and communicative abilities with humans. On certain tasks, they demonstrate processing abilities comparable to 5-6 year-old humans; they learn very simple vocal syntactic patterns and referential elements of human communication, but only through social interaction and in a manner that proceeds in ways similar to those of humans. Given this knowledge of vocal learning in birds, of the effects of social interaction on such learning, and of birds’ complex cognitive abilities, we should not ignore the avian line if we wish to determine the evolutionary pressures that purportedly affected the evolution of complex communication systems—particularly vocal systems—and develop theories and models that can be tested.
Irene M. Pepperberg

20. A Possible Role for Selective Masking in the Evolution of Complex, Learned Communication Systems

The human capacity for language is one of our most distinctive characteristics. While communication systems abound in the natural world, human language distinguishes itself in terms of its communicative power, flexibility and complexity. One of the most unusual features of human language, when compared to the communication systems of other species, is the degree to which it involves learning. Just how much of language is innate and how much is learned is an ongoing controversy, but it is undeniable that the specific details of any particular language must be learned anew every generation. We do, of course, bring a great deal of innate resources to bear on our language learning process, and the results these innate biases have on the development of languages may explain a great deal about the structure of the languages we see today. But still every child in every new generation must go through a lengthy process of language acquisition if they are to become normal language users.
Graham R.S. Ritchie, Simon Kirby

21. The Natural History of Human Language: Bridging the Gaps without Magic

Human languages are quintessentially historical phenomena. Every known aspect of linguistic form and content is subject to change in historical time (Lehmann, 1995; Bybee, 2004). Many facts of language, syntactic no less than semantic, find their explanation in the historical processes that generated them. If adpositions were once verbs, then the fact that they tend to occur on the same side of their arguments as do verbs (“cross-category harmony”: Hawkins, 1983) is a matter of historical contingency rather than a reflection of inherent structural constraints on human language (Delancey, 1993).
Bjorn Merker, Kazuo Okanoya

22. Neural Substrates for String-Context Mutual Segmentation: A Path to Human Language

Linguistic structures are products of biological prerequisites and historical processes. Here we consider a number of neural, behavioral, and learning mechanisms that serve necessary or facilitating roles in the initiation of historical processes. We hypothesize that if mutual segmentation of strings and contexts is promoted by particular biological adaptations and ecological pressures, this could initiate a subsequent historical process of linguistic elaboration. To enable this mutual segmentation, three biological sub-faculties are indispensable: vocal learning, string segmentation, and contextual segmentation. Vocal learning enabled intentional control of vocal output via the direct connection between face motor cortex and medullary vocal nuclei.
Kazuo Okanoya, Bjorn Merker


Weitere Informationen

Premium Partner