Language Evolution and Computation Bibliography

This methodological article serves as an introduction to this special issue, whose aim is to encourage more and better interaction between empirical studies and computer modeling with regard to the study of language evolution. We argue that research into the field of language evolution is so complex that computer modeling forms an essential tool to generate predictions based on some, possibly descriptive, theory in order to falsify that theory. Falsification should be carried out by comparing the generated predictions with empirical data. In order to improve the quality of the predictions and thus the reliability of the falsification process, we stress the importance of initializing the computer model with empirical data. The papers in this special issue provide some concrete examples and new proposals of applying the suggested methodology.

Using our interdisciplinary research collaboration as a case study, we discuss the question of whether formal modeling and empirical approaches can be successfully integrated into a single line of research. We argue that to avoid an undesirable disconnect between the two, one needs considerable time and patience for a science-humanities collaboration to bear fruit. In our collaboration and, we believe, in science-humanities collaborations in general, certain shared goals are required for success, including: starting with simple models before moving to more complex models; the importance of continually comparing models with empirical data where possible; and a focus on explaining statistical patterns rather than accounting for single data points individually.

This article presents a dense database study of child language acquisition from a usage-based perspective and a new analysis of data from an earlier study on simulating language evolution. The new analysis is carried out to show how computer modeling studies can be designed to generate predictions (results) that can be compared quantitatively with empirical data obtained from the dense database studies. Although the comparison shows that the computer model in question is still far from realistic, the study illustrates how to carry out agent-based simulations of language evolution that allow quantitative verification of predictions with empirical data to validate theories on child language acquisition.

Human speech has been investigated with computer models since the invention of digital computers, and models of the evolution of speech first appeared in the late 1960s and early 1970s. Speech science and computer models have a long shared history because speech is a physical signal and can be modeled accurately. This article gives a brief overview of the use of computer models in the study of the evolution of the vocal tract. We also present a critical case study of one model that has been used to study the vocal abilities of Neanderthals. We argue that this study contains subtle but fatal flaws which invalidate the conclusions drawn from the model, illustrating the dangers of applying computer models outside the area for which they have been developed. Future models need to make use of a broader database of anatomical and physiological data from other animals, especially nonhuman primates, to understand the path leading to modern Homo sapiens.

What are the ``design features'' of human language that need to be explained? Starting from R. Jackendoff's scenario for the evolution of language, we argue that it is the transitions between stages that pose the crucial challenges for accounts of the evolution of language. We review a number of formalisms for conceptualizations, sound, and the mapping between them, and describe and evaluate the differences between each of Jackendoff's stages in terms of these formalisms. We conclude from this discussion that the transitions to combinatorial phonology, compositional semantics and hierarchical phrase structure can be formally characterized. Modeling these transitions is a major challenge for language evolution research.

Observations of alarm calling behavior in putty-nosed monkeys are suggestive of a link with human language evolution. However, as is often the case in studies of animal behavior and cognition, competing theories are underdetermined by the available data. We argue that computational modeling, and in particular the use of individual-based simulations, is an effective way to reduce the size of the pool of candidate explanations. Simulation achieves this both through the classification of evolutionary trajectories as either plausible or implausible, and by putting lower bounds on the cognitive complexity required to perform particular behaviors. A case is made for using both of these strategies to understand the extent to which the alarm calls of putty-nosed monkeys are likely to be a good model for human language evolution.

Abstract A fundamental characteristic of human speech is that it uses a limited set of basic building blocks (phonemes, syllables), that are put to use in many different combinations to mark differences in meaning. This article investigates the evolution of such combinatorial ...

This article proposes an acquisition framework that involves horizontal, vertical, and oblique transmissions. Based on a lexicon-syntax coevolution model, it discusses the relative roles of these forms of cultural transmission on language origin and change. The simulation results not only reveal an integrated role of oblique transmission that combines the roles of horizontal and vertical transmissions in preserving linguistic understandability within and across generations of individuals, but also show that both horizontal and oblique transmissions are more necessary than vertical transmission for language evolution in a multiagent cultural environment.

The iterated classification game (ICG) combines the classification game with the iterated learning model (ILM) to create a more realistic model of the cultural transmission of language through generations. It includes both learning from parents and learning from peers. Further, it eliminates some of the chief criticisms of the ILM: that it does not study grounded languages, that it does not include peer learning, and that it builds in a bias for compositional languages. We show that, over the span of a few generations, a stable linguistic system emerges that can be acquired very quickly by each generation, is compositional, and helps the agents to solve the classification problem with which they are faced. The ICG also leads to a different interpretation of the language acquisition process. It suggests that the role of parents is to initialize the linguistic system of the child in such a way that subsequent interaction with peers results in rapid convergence to the correct language.

Like any other biological trait, communication can be studied from at least four perspectives: mechanistic, ontogenetic, functional, and phylogenetic. In this article, we focus on the following phylogenetic question: how can communication emerge, given that both signal-producing and signal-responding abilities seem to be adaptively neutral until the complementary ability is present in the population? We explore the problem of co-evolution of speakers and hearers with artificial life simulations: a population of artificial neural networks evolving a food call system. The core of the article is devoted to a careful analysis of the complex evolutionary dynamics demonstrated by our simple simulation. Our analyses reveal an important factor, which might solve the phylogenetic problem: the spontaneous production of good (meaningful) signals by speakers because of the need for organisms to categorize their experience in adaptively relevant ways. We discuss our results with respect both to previous simulative work and to the biological literature on the evolution of communication.

We present a novel connectionist model for acquiring the semantics of a simple language through the behavioral experiences of a real robot. We focus on the ``compositionality'' of semantics and examine how it can be generated through experiments. Our experimental results showed that the essential structures for situated semantics can self-organize themselves through dense interactions between linguistic and behavioral processes whereby a certain generalization in learning is achieved. Our analysis of the acquired dynamical structures indicates that an equivalence of compositionality appears in the combinatorial mechanics self-organized in the neuronal nonlinear dynamics. The manner in which this mechanism of compositionality, based on dynamical systems, differs from that considered in conventional linguistics and other synthetic computational models, is discussed in this paper.

This paper shows how phonological structures can be culturally selected so as to become learnable and adapted to the ecological niche formed by the brains and bodies of speakers. A computational model of the cultural formation of syllable systems illustrates how general learning and physical biases can influence the evolution of the structure of vocalization systems. We use the artificial life methodology of building a society of artificial agents, equipped with motor, perceptual and cognitive systems that are generic and have a realistic complexity. We demonstrate that agents, playing the ''imitation game,'' build shared syllable systems and show how these syllable systems relate to existing human syllable systems. Detailed experiments study the learnability of the self-organized syllable systems. In particular, we reproduce the critical period effect and the artificial language learning effect without the need for innate biases which specify explicitly in advance the form of possible phonological structures. The ability of children agents to learn syllable systems is explained by the cultural evolutionary history of these syllable systems, which were selected for learnability.

Much is known about the evolution of speech. Fossil evidence points to modern adaptations for speech appearing between 1.5 million and 500,000 years ago. Studies of vocal behavior in apes show the ability to use combinatorial vocalizations in some species (but not chimpanzees) and some cultural influence on vocalizations, but little ability for vocal imitation. For modern speech, the comparison of many languages shows that speech can become extremely complex, but that it can also be astonishingly simple. Finally, the way in which infants acquire speech is becoming increasingly clear. We therefore know about the starting point of the evolution of speech, its end point and some steps in between.

Much less is known about the exact scenario and the dynamics of the evolution of the acquisition of speech. As it involves co-evolution between culturally transmitted sounds and genetic evolution, the dynamics are complex. Computer models are therefore ideal for studying these dynamics. This paper presents a computer model of the co-evolution between a repertoire of speech sounds and a population of learners that can either represent a lexicon holistically or combinatorially. It shows that cultural influences can change the dynamics of the transition between a population of holistic and combinatorial learners.

Color categories enjoy a special status among human perceptual categories as they exhibit a remarkable cross-cultural similarity. Many scholars have explained this universal character as being the result of an innate representation or an innate developmental program which all humans share. We will critically assess the available evidence, which is at best controversial, and we will suggest an alternative account for the universality of color categories based on linguistic transmission constrained by universal biases. We introduce a computational model to test our hypothesis and present results. These show that indeed the cultural acquisition of color categories together with mild constraints on the perception and categorical representation result in categories that have a distribution similar to human color categories.

Language is a symbolic, culturally transmitted system of communication, which is learnt through the inference of meaning. In this paper, I describe the importance of meaning inference, not only in language acquisition, but also in developing a unified explanation for language change and evolution. Using an agent-based computational model of meaning creation and communication, I show how the meanings of words can be inferred through disambiguation across multiple contexts, using cross-situational statistical learning. I demonstrate that the uncertainty inherent in the process of meaning inference, moreover, leads to stable variation in both conceptual and lexical structure, providing evidence which helps to explain how language changes rapidly without losing communicability. Finally, I describe how an inferential model of communication may provide important theoretical insights into plausible explanations of the bootstrapping of, and the subsequent progressive complexification of, cultural communication systems.

This paper investigates the productive creativity of children in a computational model on the emergence and evolution of compositional structures in language. In previous models it was shown that compositional structures can emerge in language when the language is transmitted from one generation to the next through a transmission bottleneck. Due to the fact that in these models language is transmitted only in a vertical direction where adults only speak to children and children only listen, this bottleneck needs to be imposed by the experimenter. In the current study, this bottleneck is removed and instead of having a vertical transmission of language, the language is -- in most simulations -- transmitted horizontally (i.e. any agent can speak to any other agent). It is shown that such a horizontal transmission scenario does not need an externally imposed bottleneck, because the children face an implicit bottleneck when they start speaking early in life. The model is compared with the recent development of Nicaraguan Sign Language, where it is observed that children are a driving force for inventing grammatical (or compositional) structures, possibly due to a sparseness of input (i.e. an implicit bottleneck). The results show that in the studied model children are indeed the creative driving force for the emergence and stable evolution of compositional languages, thus suggesting that this implicit bottleneck may -- in part -- explain why children are so typically good at acquiring language and, moreover, why they may have been the driving force for the emergence of grammar in language.

The current research describes a functional trajectory from sensorimotor sequence learning to the learning of grammatical constructions in language. A brief review of the functional neurophysiology of the cortex and basal ganglia will be provided as background for a neural network model of this system in sensorimotor sequence learning. Sequential behavior is then defined in terms of serial, temporal and abstract structure. The resulting neuro-computational framework is demonstrated to account for observed sequence learning behavior. More interestingly, this framework naturally extends to grammatical constructions as form-to-meaning mappings. Predictions from the neuro-computational model concerning parallels in language and cognitive sequence processing are tested against behavioral and neurophysiological observations in humans, resulting in a refinement of the allocation of model functions to subdivisions of Broca's area. From a functional perspective this analysis will provide insight into the relation between the coding structure in human languages, and constraints derived from the underlying neurophysiological computational mechanisms.

This article reviews recent progress made by computational studies investigating the emergence, via learning or evolutionary mechanisms, of communication among a collection of agents. This work spans issues related to animal communication and the origins and evolution of language. The studies reviewed show how population size, spatial constraints on agent interactions, and the tasks involved can all influence the nature of the communication systems and the ease with which they are learned and/or evolved. Although progress in this area has been substantial, we are able to identify some important areas for future research in the evolution of language, including the need for further computational investigation of key aspects of language such as open vocabulary and the more complex aspects of syntax.

It has been postulated that aspects of human language are both genetically and culturally transmitted. How might these processes interact to determine the structure of language? An agent-based model designed to study gene-culture interactions in the evolution of communication is introduced. This model shows that cultural selection resulting from learner biases can be crucial in determining the structure of communication systems transmitted through both genetic and cultural processes. Furthermore, the learning bias which leads to the emergence of optimal communication in the model resembles the learning bias brought to the task of communication by human infants. This suggests that the iterated application of such human learning biases may explain much of the structure of human language.

We work with a large spatialized array of individuals in an environment of drifting food sources and predators. The behavior of each individual is generated by its simple neural net; individuals arecapable of making one of two sounds and are capable of responding to sounds from their immediate neighbors by opening their mouths or hiding. An individual whose mouth is open in the presence of food is 'fed' and gains points; an individual who fails to hide when a predator is present is 'hurt' by losing points. Opening mouths, hiding, and making sounds each exact an energy cost. There is no direct evolutionary gain for acts of cooperation or 'successful communication' per se.

In such an environment we start with a spatialized array of neural nets with randomized weights. Using standard learning algorithms, our individuals 'train up' on the behavior of successful neighbors at regular intervals. Given that simple setup, will a community of neural nets evolve a simple language for signaling the presence of food and predators? With important qualifications, the answer is yes.'In a simple spatial environment, pursuing individualistic gains and using partial training on successful neighbors, randomized neural nets can learn to communicate.

Social coordination is studied in a simulated model of autonomous embodied agents that interact acoustically. Theoretical concepts concerning social behavior are presented from a systemic perspective and their usefulness is evaluated in interpreting the results obtained. Two agents moving in an unstructured arena must locate each other, and remain within a short distance of one another for as long as possible using noisy continuous acoustic interaction. Evolved dynamical recurrent neural networks are used as the control architecture. Acoustic coupling poses nontrivial problems like discriminating `self' from `non-self' and structuring production of signals in time so as to minimize interference. Detailed observation of the most frequently evolved behavioral strategy shows that interacting agents perform rhythmic signals leading to the coordination of movement. During coordination, signals become entrained in an anti-phase mode that resembles turn-taking. Perturbation techniques show that signalling behavior not only performs an external function, but it is also integrated into the movement of the producing agent, thus showing the difficulty of separating behavior into social and non- social classes. Structural congruence between agents is shown by exploring internal dynamics as well as the response of single agents in the presence of signalling beacons that reproduce the signal patterns of the interacting agents. Lack of entrainment with the signals produced by the beacons shows the importance of transient periods of mutual dynamic perturbation wherein agents achieve congruence.

Human language is a unique ability. It sits apart from other systems of communication in two striking ways: it is syntactic, and it is learned. While most approaches to the evolution of language have focused on the evolution of syntax, this paper explores the computational issues that arise in shifting from a simple innate communication system to an equally simple one that is learned. Associative network learning within an observational learning paradigm is used to explore the computational difficulties involved in establishing and maintaining a simple learned communication system. Because Hebbian learning is found to be sufficient for this task, it is proposed that the basic computational demands of learning are unlikely to account for the rarity of even simple learned communication systems. Instead, it is the problem of *observing* that is likely to be central -- in particular the problem of determining what meaning a signal is intended to convey.

Abstract Social behaviour and in particular social learning are key mechanisms for the cohesion and evolu tion of primate societies. Similarly, social skills might be desirable for artificial agents who are expected to interact with other natural or artificial agents. We view ...

This article presents a theoretical criticism of current approaches to the study of the evolution of communication. In particular two very common preconceptions about the subject are analysed: the role of natural selection in the definition of the phenomenon and the metaphor of communication as information exchange. An alternative characterization is presented in terms of autopoietic theory which avoids the mentioned preconceptions. In support of this view, the evolution of coordinated activity is studied in a population of artificial agents playing an interactional game. Dynamical modeling of this evolutionary process based on game-theoretic considerations shows the existence of an evolutionarily stable strategy in the total lack of coordinated activity which, however, may be unreachable due to the presence of a periodic attractor. In a computational model of the same game, action coordination evolves, even with individual costs against it, due to the presence of spatial structuring processes. A detailed explanation of this phenomenon, which does not require kin selection, is presented. In an extended game, recursive coordination evolves nontrivially when the participants share all the relevant information, demonstrating that the metaphor of information exchange can be misleading. It is shown that agents engaged in this sort of interaction are able to perform beyond their individual capabilities.

Synthetic ethology is proposed as a means of conducting controlled experiments investigating the mechanisms and evolution of communication. After a discussion of the goals and methods of synthetic ethology, two series of experiments are described based on at least 5000 breeding cycles. The first demonstrates the evolution of cooperative communication in a population of simple machines. The average fitness of the population and the organization of its use of signals are compared under three conditions: communication suppressed, communication permitted, and communication permitted in the presence of learning. Where communication is permitted the fitness increases about 26 times faster than when communication is suppressed; with communication and learning the rate of fitness increase is about 100 fold. The second series of experiments illustrates the evolution of a syntactically simple language, in which a pair of signals is required for effective communication.