Language Evolution and Computation Bibliography

Contemporary disputes about the origins and evolution of language are reviewed. The main issues involved are: how many mutations gave rise to the Language faculty, whether a new cognitive domain was thereby created, how powerful Language was from the beginning, whether the initial function of Language was private thought or public communication, and whether natural selection influenced its rise.

My previous two books were long and academic in tone. This book is shorter (under 60,000 words) and more likely to be read by busy people (I hope). The Origins of Language: A Slim Guide offers a concise and accessible overview of what is known about the evolution of the human capacity for language. Non-human animals communicate in simple ways: they may be able to form simple concepts, to feel some limited empathy for others, to cooperate to some extent, and to engage in mind-reading. Human language, however, is characterized by its ability to efficiently express a wide range of subtle and complex meanings. After the first simple beginnings, human language underwent an explosion of complexity, leading to the very complicated systems of grammar and pronunciation found in modern languages. Professor Hurford looks at the very varied aspects of this evolution, covering human prehistory; the relation between instinct and learning; biology and culture; trust, altruism, and cooperation; animal thought; human and non-human vocal anatomy; the meanings and forms of the first words; and the growth of complex systems of grammar and pronunciation. Written by an internationally recognized expert in the field, it draws on a number of disciplines besides linguistics, including philosophy, neuroscience, genetics, and animal behaviour, and will appeal to a wide range of readers interested in language origins and evolution.

Researches in language evolution have presented some continuity between apes and humans. Researchers have shown that precursors to both sentence meaning (conceptual meaning) and speaker meaning (pragmatic meaning) are present in animals. They have also reported that something other than the ability to comprehend such meanings was necessary to launch language, perhaps shared intentions, and once language was launched it resulted in abilities to create new kinds of meanings. A phenomenon extensively researched in both developmental and comparative psychology is object permanence. There are two forms of the object permanence test, one harder than the other. The easier test, the visible object permanence test, involves simply moving an object, such as a treat, behind a screen, where it is invisible to the animal or child subject, restraining the subject briefly, and then seeing if the subject moves to retrieve the object from behind the screen. The harder test, the invisible object permanence test, involves placing the treat into a small container in view of the subject, then the container is hidden behind the screen, where it is emptied out unseen by the subject and the empty container is then shown to the subject, who passes the test if he/she searches behind the screen. The episodic memory is related to object permanence. This is a kind of memory for specific events that have happened to the individual subject.

This chapter takes the goal of linguistics to be an explanation of how whole languages get to be the way they are. It shows that some insight can be gained into the forces shaping languages by considering them as products of a historical spiral involving both acquisition and production, learning and speaking, and occasionally innovating, over many generations.

That language is shaped to fit the human brain is close to the Chomskyan position. The target article by Christiansen & Chater (C&C) assumes an entity, outside individual heads. What is the nature of this entity? Linguistic niche-construction and co-evolution of language and genes are possible, with some of what evolved being language-specific. Recent generative theory postulates much less than the old Universal Grammar (UG).

This paper argues for an alternative answer to Carstairs-McCarthy's (1999) question ``Why do all languages distinguish between NPs and sentences?'' While agreeing on basic philosophical points with Carstairs-McCarthy, such as the lack of a distinction between truth and reference independent of grammar, I argue that the S/NP distinction has its roots in the basic communicative distinction between Topic and Comment. In the very earliest mental processes, long antedating language, binary structure can be found, with components that one can associate with the functions of identifying or locating an object and representing some information about it. When private thought went public, the earliest messages in any code with rudimentary syntax were of similar bipartite structure, with one part conveying information presumed to be already known to the hearer, and identifying the object that the message is about. The other part of the bipartite message conveyed information presumed to be new to the hearer. This bipartite structure, with its concomitant distinction between types of expression that could fulfil the respective roles, was central enough to the main function of public language, namely communication, that it was never eroded away, and is the basis of the bipartite structure found universally in languages today.

A strong constraint on the arithmetical combinations allowed in compound numerals, called the Packing Strategy, applies very widely to numeral systems across the world. A previous attempt to explain the existence of the strong universal constraint, in terms of a gradual socio-historical process of standardization, will not scale up to higher-valued numerals. It is proposed that the real explanation for the Packing Strategy is that it reflects two natural principles applied in the practical task of counting objects. These two principles, `Go as far as you can with the resources you have', and `Minimize the number of entities you are dealing with', are not specific to the counting task, but are of more general application to practical tasks.

In this, the first of two ground-breaking volumes on the nature of language in the light of the way it evolved, James Hurford looks at how the world first came to have a meaning in the minds of animals and how in humans this meaning eventually came to be expressed as language. He reviews a mass of evidence to show how close some animals, especially primates and more especially apes, are to the brink of human language. Apes may not talk to us but they construct rich cognitive representations of the world around them, and here, he shows, are the evolutionary seeds of abstract thought - the means of referring to objects, the memory of events, even elements of the propositional thinking philosophers have hitherto reserved for humans. What then, he asks, is the evolutionary path between the non-speaking minds of apes and our own speaking minds? Why don't apes communicate the richness of their thoughts to each other? Why do humans alone have a unique disposition to reveal their thoughts in complex detail? Professor Hurford searches a wide range of evidence for the answers to these central questions, including degrees of trust, the role of hormones, the ability to read minds, and the willingness to cooperate.

Expressing himself congenially in consistently colloquial language the author builds up a vivid picture of how mind, language, and meaning evolved over millions of years. His book is a landmark contribution to the understanding of linguistic and thinking processes, and the fullest account yet published of the evolution of language and communication.

Table of Contents

Part I Meaning Beford Communication
1. Let's Agree on Terms
2. Animals Approach Human Cognition
3. A New Kind of Memory Evolves
4. Animals Form proto-propositions
5. Towards Human Semantics
Part II Communication: What and Why?
6. Communication by Dyadic Acts
7. Going Triadic: Precursors of Reference
8. Why Communicate? Squaring With Evolutionary Theory
9. Cooperation, Fair Play and Trust in Primates
10. Epilogue

Before the evolution of languages as public conventional communication systems, pre-humans had somewhat complex private mental schemes for representing the external world. What is known about human and some animal vision suggests that proposition-like cognitive structures existed for the mental representation of perceived scenes before the advent of complex language. The structures traditionally adopted by formal Logic can be modified to conform to known constraints on the visual representation of scenes. While this modification slightly reduces the expressive power of representations (in that the meanings of some complex sentences cannot naturally be represented), it provides a unified, ontologically parsimonious, primitive notation for cognitive representations, suitable for later recruitment by complex syntactic language. The most basic semantic elements later mapped onto sentences are all present in the prelinguistic mental representation, which reflects the workings of the visual attention system

The last quarter of the 20th century saw a surge in research in the evolution of language, and this activity continues to grow and extend its influence in the present century. This article is a personal review of some conclusions that can be deemed to have been established in that period. Many of these modern conclusions had ancient precursors as speculative hypotheses with little empirical backing. Modern empirical research in a range of fields has driven foundations deeper, and careful theoretical work has begun to weave a more consistent network of ideas across disciplines. Many mysteries remain, but some clear outlines of the evolutionary bases of humans? most distinctive capacity have begun to emerge. Often the clearer outlines have revealed more complex problems than was vaguely suspected earlier. Three propositions have been selected here, and each will be briefly discussed in a separate section. The three propositions are: 'Language' is not a single monolithic behaviour; Animals have rich conceptual systems; Primates are not necessarily the closest to human-like capacities.

Human language is qualitatively different from animal communication systems in at least two separate ways. Human languages contain tens of thousands of arbitrary learned symbols (mainly words). No other animal communication system involves learning the component symbolic elements afresh in each individual's lifetime, and certainly not in such vast numbers. Human language also has complex compositional syntax. The meanings of our sentences are composed from the meanings of the constituent parts (e.g. the words). This is obvious to us, but no other animal communication system (with honeybees as an odd but distracting exception) puts messages together in this way. A recent theoretical claim that the sole distinguishing feature of human language is recursion is discussed, and related to these features of learned symbols and compositional syntax. It is argued that recursive thought could have existed in prelinguistic hominids, and that the key step to language was the innovative disposition to learn massive numbers of arbitrary symbols

Pure synonymy is rare. By contrast, homonymy is common in languages. Human avoidance of synonymy is plausibly innate, as theorists of differing persuasions have claimed. Innate dispositions to synonymy and homonymy are modelled here, in relation to alternative roles of speaking and hearing in determining fitness.

In the computer model, linguistic signs are acquired via different genetically determined strategies, variously (in)tolerant to synonymy or homonymy. The model defines communicative success as the probability of a speaker getting a message across to a hearer; interpretive success is the probability of a hearer correctly interpreting a speaker's signal. Communicative and interpretive success are compared as bases for reproductive fitness. When communicative success is the basis for fitness, a genotype evolves which is averse to synonymy, while tolerating homonymy. Conversely, when interpretive success is the basis for fitness, a genotype evolves which is averse to homonymy, while tolerating synonymy.

The human capacity for language and the structures of individual languages can best be understood from an evolutionary perspective. Both the biological capacity and languages owe their shape to events far back in the past. Biological steps toward language-readiness involved preadaptations for modern phonetics, syntax, semantics and pragmatics. Once humans were language-ready, ever more complex language systems could grow, relatively fast, by cultural transmission, generation after generation. This latter process is profitably studied by grammaticalization theory and computer modelling.

Models of the cultural evolution of language typically assume a very simplified population dynamic. In the most common modelling framework (the Iterated Learning Model) populations are modelled as consisting of a series of non-overlapping generations, with each generation consisting of a single agent. However, the literature on language birth and language change suggests that population dynamics play an important role in real-world linguistic evolution. We aim to develop computational models to investigate this interaction between population factors and language evolution. Here we present results of extending a well-known Iterated Learning Model to a population model which involves multiple individuals. This extension reveals problems with the model of grammar induction, but also shows that the fundamental results of Iterated Learning experiments still hold when we consider an extended population model.

(In Linguistic Evolution through Language Acquisition: Formal and Computational Models, edited by Ted Briscoe, Cambridge University Press. pp.301-344. Note: This HTML version may differ slightly from the printed version; the printed version is the `authorized' version. See a ...

As language users humans possess a culturally transmitted system of unparalleled complexity in the natural world. Linguistics has revealed over the past 40 years the degree to which the syntactic structure of language in particular is strikingly complex. Furthermore, as Pinker and Bloom point out in their agenda-setting paper Natural Language and Natural Selection ``grammar is a complex mechanism tailored to the transmission of propositional structures through a serial interface'' (Pinker and Bloom, 1990, 707). These sorts of observations, along with influential arguments from linguistics and psychology about the innateness of language (see, e.g. Chomsky, 1986; Pinker, 1994), have led many authors to the conclusion that an explanation for the origin of syntax must invoke neo-Darwinian natural selection.

``Evolutionary theory offers clear criteria for when a trait should be attributed to natural selection: complex design for some function, and the absence of alternative processes capable of explaining such complexity. Human language meets these criteria.'' (Pinker and Bloom, 1990, 707)

Since Pinker and Bloom made these arguments there have been many attempts to put forward a coherent evolutionary story that would allow us to derive known features of syntax from communicative selection pressures (e.g. Nowak, Plotkin, and Jansen, 2000; Newmeyer, 1991 and discussion in Kirby, 1999a). One problem with this approach to evolutionary lin- guistics is that it often fails to take into account that biological natural selection is only one of the complex adaptive systems at work.

Language emerges at the intersection of three complex adaptive systems:

• Learning: During ontogeny children adapt their knowledge of language in response to the environment in such a way that they optimise their ability to comprehend others and to produce comprehensible utterances.
• Cultural evolution: On a historical (or glossogenetic) timescale, languages change. Words enter and leave the language, meanings shift, and phonological and syntactic rules ad- just.
• Biological evolution: The learning (and processing) mechanisms with which our species has been equipped for language, adapt in response to selection pressures from the environ- ment, for survival and reproduction.

There are two problems with this multiplicity of dynamical systems involved in linguistic evolution. Firstly, we understand very little about how learning, culture, and evolution inter- act (though, see Belew, 1990; Kirby and Hurford, 1997; Boyd and Richerson, 1985), partly because language is arguably the only sophisticated example of such a phenomenon. There clearly are interactions: for example, biological evolution provides the platform on which learning takes place, what can be learnt influences the languages that can persist through cultural evolution, and the structure of the language of a community will influence the selec- tion pressures on the evolving language users (see figure 1).

Secondly, it is not clear what methodology we should use to study this problem. Mathe- matical techniques for looking at the interaction of dynamical systems and linguistic behaviour are in their infancy (though, Nowak, Komarova, and Niyogi, 2001, take a potentially valuable approach). We feel that computational modelling is currently the most appropriate method- ology, but although simulations of language learning have a long history, and there are many methods from the A-life field that can be used for modelling evolution, models of the cultural transmission of learned behaviour are relatively sparse (see Steels, 1997 for a review). This is unfortunate, since we will argue in this chapter that it is through this particular mechanism that the most basic features of human language syntactic structure can be explained.

To remedy this situation, we introduce here the Iterated Learning Model (ILM), a gen- eral approach to exploring the transmission over a glossogenetic timescale of observationally learned behaviour. We will illustrate the ILM with a few examples of simulations that lead to two conclusions:

• There is a non-trivial mapping between the set of learnable languages (i.e. the lan- guages allowed by our innate language faculty), and the set of stable languages (i.e., the languages we can actually expect to see in the world).
• Under certain circumstances, cultural evolution leads inevitably to recursively compo- sitional (i.e., syntactic) languages.

Describes an attempt to cast several abstract properties of natural languages in the framework of Kauffman's (1993, 1995) random Boolean nets (RBN). The properties are complexity, interconnectedness, stability, diversity, and underdeterminedness. A language is modeled as a Boolean net attractor. (Groups of) net nodes are linguistic principles or parameters as posited by Chomskyan theory, according to which the language learner sets parameters to appropriate values on the basis of very limited experience of the language. The setting of one parameter can have a complex effect on the settings of others. A RBN is generated to find an attractor. A state from this attractor is degraded, which represents the degenerate input of language to the learner, and this state is then input to a net with the same connectivity and activation functions as the original net to see whether it converges on the same attractor. Many nets degenerate into attractors representing complete uncertainty. Others settle at intermediate levels of uncertainty, and some manage to overcome the incompleteness of input and converge on attractors identical to that from which the original inputs were (de)generated. Finally, an attempt was made to select a population of such successful nets, using a genetic algorithm where fitness was correlated with an ability to acquire several different languages faithfully. This has so far proved impossible, supporting the Chomskyan suggestion that the human language acquisition capacity is not the outcome of natural selection.

There is evidence for increase, followed by decline, in synaptic numbers during development. Dendrites do not function in isolation. A constructive neuronal process may underpin a selectionist cognitive process. The environment shapes both ontogeny and phylogeny. Phylogenetic natural selection and neural selection are compatible. Natural selection can yield both constructivist and selectionist solution to adaptuive problems.

Evidence suggests that there is a critical, or at least a sensitive, period for language acquisition, which ends around puberty. The existence of this period is explained by an evolutionary model which assumes that (a) linguistic ability is in principle (if not in practice) measurable, and (b) the amount of language controlled by an individual conferred selective advantage on it. In this model, the language faculty is seen as adaptive, favoured by natural selection, while the critical period for language acquisition itself is not an adaptation, but arises from the interplay of genetic factors influencing life-history characters in relation to language acquisition. The evolutionary model is implemented on a computer and simulations of populations evolving under various plausible, if idealized, conditions result in clear critical period effects, which end around puberty.

Most linguistic theories assume lexicon entries which incorporate the idea of the Saussurean sign, a bidirectional mapping between a phonological form and some representation of a concept. This Sign, like grammars generally, is unbiased with respect to perception or production, and provides part of the cognitive map which speakers use both in speaking and in interpreting the utterances of others. This bidirectionality of the Sign, or code, is a design feature of human language although workable communication does not necessarily incorporate such a feature.

Part of the Language Acquisition Device is a mechanism which mentally constructs such a bidirectional mapping, on the basis of observed samples of communicative behaviour (transmission and reception). This basic aspect of the LAD presumably evolved because of its superiority over other conceivable mechanisms for acquiring a basis for communicative behaviour from a sampling of observed transmission and reception data.

Three conceivable strategies for acquiring the basis of communicative behaviour, here labelled the Imitator, Calculator, and Saussurean strategies, are defined as functions from samplings of observed behaviour to acquired behaviour patterns. Essentially, the Imitator imitates the transmission and reception patterns in the observed sample; the Calculator constructs optimal reception responses to the transmissions in the observed sample and optimal transmission responses to receptions in the observed sample; the Saussurean imitates the transmission behaviour in the sample, but shapes his reception behaviour to mirror the acquired transmission behaviour, thus internalizing the equivalent of a set of bidirectional Saussurean signs. Extensive simulations of populations endowed with these innate strategies show the Saussurean strategy to be a winner in evolutionary terms; it produces individuals who communicate more successfully than individuals endowed with the other two strategies.