References

Ampatzis, C., Tuci, E., Trianni, V., & Dorigo, M. (2010). Evolution of Signaling in
a Multi-Robot System: Categorization and Communication. In S. Nolfi & M. Mirolli (Eds.), Evolution of communication and language in embodied agents (pp. 223-233). Berlin, Germany: Springer.

Castellano, C., Fortunato, S., & Loreto, V. (2009). Statistical physics of social dynamics. Revues of Modern Physics.

Harari, Y., N. (2014). Sapiens: A brief history of humankind. Toronto, CA: McLelland & Stewart.

Kirby, S. (2001). Spontaneous evolution of linguistic structure: an iterated learning model of the emergence of regularity and irregularity. IEEE Transactions on Evolutionary Computation, 5(2), 102–110.

Kirby, S., Griffiths, T., & Smith, K. (2014). Iterated learning and the evolution of language. Current Opinion in Neurobiology, 28, 108-114.

Kirby, S. (2012) Language is an adaptive system: The role of cultural evolution in the origins of structure. In M. Tallerman & K. R. Gibson (Eds.), The Oxford handbook of language evolution (pp. 589-604). New York, NY: Oxford University Press.

Loreto, V., & Steels, L. (2007). Social dynamics: the emergence of language. Nature Physics, 3, 758–760.

Mirolli, M., & Nolfi, S. (2010). Evolving Communication in Embodied Agents:
Theory, Methods, and Evaluation. In S. Nolfi & M. Mirolli (Eds.), Evolution of communication and language in embodied agents (pp. 13-35). Berlin, Germany: Springer.

Neumann, J. V. (1966). Theory of self-reproducing automata. Champaign: University of Illinois Press.

Parisi, D. (2010). Artificial Organisms with Human Language. In S. Nolfi & M.
Mirolli (Eds.), Evolution of communication and language in embodied agents (pp. 13-35). Berlin, Germany: Springer.

Parmiggiani, A., Randazzo, M., Maggiali, M., Metta, G., Elisei, F., & Bailly, G.
(2015). Design and Validation of a Talking Face for the iCub. International Journal of Humanoid Robots, 12(3), 1-20. Doi: 10.1142/S0219843615500267

Serenko, A., Bontis, N., & Detlor, B. (2007). End-user adoption of animated
interface agents in everyday work applications. Behaviour & Information Technology, 26(2), 119-132.

Sierra-Santibánez, J. (2015). An agent-based model of the emergence and transmission of a language system for the expression of logical combinations. Proceedings of the Twenty-Ninth AAAI Conference on Artificial Intelligence, 492-499.

Steels, L. (2010). Modeling the formation of language: embodied experiments. In S. Nolfi & M. Mirolli (Eds.), Evolution of communication and language in embodied agents (pp. 235-262). Berlin, Germany: Springer.

Sukthankar, G. (2008). Creating physically embodied agents using realistic human
motion models. Simulation & Gaming: An Interdisciplinary Journal of Theory, Practice and Research, 39(1), 64-82. https://doi-org.ezlibproxy1.ntu.edu.sg/10.1177/1046878107309686

Ulam, S. (1960). A collection of mathematical problems. New York: Interscience.

Research Limitations

4. Research Limitations

The models mentioned in the previous pages make very weak assumptions about the transmission process itself: no language is easier or harder for learners to acquire than any other. The need for fitness – being able to communicate effectively with others – can be said to be the driving force in the dynamics of, rather than learning.

In addition, computational models might not be able to match the abilities and biases of real human learners to a realistic extent. The use of modern humans is inevitable in laboratory experiments; it is hard for us to determine what humans were like at the point when language emerged. However, it is worth noting replicating the evolutionary history of language is not the objective of such experimental approaches.

Part II: Computational Models

3. Computational models

When studying a social phenomenon such as language, there are certain “predictable” patterns in human behavior that we can try to model with statistical physics or mathematics to represent large-scale population behavior (Loreto & Steels, 2007). As mentioned in Chapter 4, agent-based modelling is a new analytical method for social sciences. It enables one to build models where individual entities and their interactions are directly represented. The collective result (“macro” phenomena) of interactions of interactions among individual agents (“micro” dynamics) can be observed, or inferred from these modeling efforts (Castellano, Fortunato, & Loreto, 2009). Basically, it allows one to represent multiple scales of analysis in a natural and efficient way. In linguistics, we can use agent-based modelling to study the emergence and evolution of language (Sierra-Santibánez, 2015).

[Back to Table of Contents]

3.1 Literature on agent-based modeling

The origins of agent-based modeling started in the 1940s when Von Neumann and Ulam created the concept of cellular automata (Neumann, 1966; Ulam, 1960). In linguistics, earlier work in this area sought to explain the role of interaction and negotiation, or biases of learners in shaping communication systems, focusing mainly on the conditions under which communicatively optimal, socially learnt communication systems would emerge (Kirby, Griffiths, & Smith, 2014). Thereafter, researchers tried to find out how linguistic structure can arise from iterated learning (Kirby, 2001). Emphasis was given on the role of bottleneck learning, which was thought to be the driving force behind the evolution of structure, since language learners must try to learn an infinitely expressive linguistic system on the basis of a small set of linguistic data (Kirby, 2001). A major finding is that compositional languages emerge from unstructured languages due to repeated transmission through the learning bottleneck – language structure appears as an adaptive response by language per se to the problem of being transmitted through a narrow bottleneck, since the presence of compositional rules enables a learner to infer from a small sample rules underpinning the whole language (Kirby, 2001).

There is another model that represents the emergence of systematicity in phonological systems through communicative interaction and iterated learning. One example would be De Boer who looked at the cultural evolution of vowel systems, showing that the universal features of the organisation of all the vowels in the world can arise through repeated interaction between simulated agents under certain reasonable articulatory and perceptual constraints.

Besides the findings on compositionality and vowel systems, there are other areas of communication and language that we can look at using computational models. In the sections below, we present how computational models (in the form of autonomous computer programs in physical robot bodies called “embodied agents”) demonstrate the emergence of certain behaviors that underscore communication (also see: Chapter 4). Computational models can help us understand how a basic level of cooperation can be achieved (via acoustic signalling) among embodied agents with a case study (Ampatzis, Tuci, Trianni, & Dorigo, 2010). This occurs prior to the emergence of language. Additionally, we present an experiment by Luc Steels (2010) to show how embodied agents can form an inventory of spatial categories to communicate by switching perspectives between itself and corresponding agents.

[Back to Table of Contents]

3.2 Embodied agents in communication studies

Embodied agents come in two forms – physical and virtual. A physical embodied agent possesses a physical body, unlike that of a virtual agent. Some of these physical agents are built with human-like or animal-like features which provide them with the physical capabilities to perform certain tasks or actions (Sukthankar, 2008). A humanoid robot such as the iCub robot (Parmiggiani et al., 2015) is an example of a physical embodied agent that resembles a 3-year-old child.

On the other hand, the virtual embodied agent, also known as the interface agent, is represented by a simulated avatar that one sees on a computer screen (Serenko, Bontis, & Detlor, 2007). The computer simulation is achieved with the use of relevant software and artificial intelligence (Serenko, Bontis, & Detlor, 2007). For example, one is able to interact with a conversational agent that provides students who attend online lessons with tutoring services.

These embodied agents form a communication system that resembles human language (Parisi, 2010). In constructing these artificial organisms that behave like real-life organisms, we can further improve our understanding on the behaviours of the latter that aid in linguistic and other scientific research (Parisi, 2010). Before researchers can further progress into the aspects of grammar and lexicon in language evolution using embodied agents, it is important to maximise the potential of these agents’ communication capabilities. Hence, the emergence of communication in embodied agents is a crucial phase in the evolution of language itself. Further, devices or other technological equipment can be created or improved on to assist humans in their daily activities (Parisi, 2010).

[Back to Table of Contents]

3.2.1 General framework of embodied cognition

To enable embodied agents to communicate, Mirolli and Nolfi (2010) emphasise on the importance of the general framework of embodied cognition. The theory is primarily a collection of different ideas in understanding behaviour but are related in challenging the classical cognitive science paradigm where the physicality of these agents and their environments are not taken into consideration (Mirolli & Nolfi, 2010). Hence, they discuss three important aspects with relevance to the physicality of these agents and their environments.

Firstly, the aspect of “situatedness” refers to the environment in which the agent is located in (Mirolli & Nolfi, 2010). Parameters must be clearly defined to regulate the interaction between the agents and their external environment and for some cases, the interaction with other agents in the same environment (Mirolli & Nolfi, 2010). In short, “situatedness” provides the agent with the details of the activity and environment (Mirolli & Nolfi, 2010).

Secondly, the aspect of “embodiment” refers to physical properties or characteristics of the body of an agent (Mirolli & Nolfi, 2010). The important characteristics include the agent’s weight, height, shape and size and the type, position, and number of its actuators and sensors (Mirolli & Nolfi, 2010). These properties will affect how the agent behave and solve problems (Mirolli & Nolfi, 2010). In addition, the control system, or its “brain” fundamentally influences the agent’s behaviour. When the agent is required to possess similar characteristics as natural organisms, control systems such as the artificial neural networks are preferred and for agents which are simpler and less bio-mimetic, look-up tables or production rules can be used instead (Mirolli & Nolfi, 2010).

Lastly, the aspect of “adaptivity” refers to the understanding that communication is not the sole action performed by the agents (Mirolli & Nolfi, 2010). The agent’s adaptive value needs to be taken into consideration where communication is investigated as a way of sub-serving other non-communicative behaviours (Mirolli & Nolfi, 2010). It aids in studying the merging development and adaptation between communicative and non-communicative behaviours (Mirolli & Nolfi, 2010).

While these three aspects are crucial in understanding how agents should behave and communicate, it is pivotal to note that they do not more form a clear-cut dichotomy and in setting-up experiments with embodied agents, a continuum exists where all three aspects are either fully present or not present at all (Mirolli & Nolfi, 2010).

[Back to Table of Contents]

3.2.2 Evaluation criteria in the assessment of communication

Based on these agents’ sensory-motor experiences, researchers are able to further study how signals and meanings originate (Mirolli & Nolfi, 2010). These signals play a role in monitoring the progress of communication in embodied agents such as their expressive power and organizational complexity. The number of signals produced will allow us to understand the adaptiveness of the agents. Further, the type of signals emitted provides information on the meaning of the message. For instance, the comparison between deictic and displaced signals presents referential information on the current context experienced by the issuer and the recipient. A set of rules is also necessary in regulating how signals are exchanged among agents whereby the agents are able to function ideally according to the circumstance or environment. While the structure of signals is an important dimension in the evaluation, the development of structured forms of communication yet to be initiated by embodied agents themselves and will represent as an achievement in this field of the research.

Other criteria in monitoring the agents’ communicative performance include understanding the adaptive role of the agents, recording the level of robustness or stability of the communication system, and modelling original or new theories (Mirolli & Nolfi, 2010).

[Back to Table of Contents]

3.2.3 Case study of the evolution of signalling in a multi-robot system

Fig. 1. The set-up of the experiment (a) Environment A and (b) Environment B (Ampatzis, Tuci, Trianni, & Dorigo, 2010, p. 163).

In this experiment (Ampatzis, Tuci, Trianni, & Dorigo, 2010), there are two circular zones created with a diameter of 120 cm – namely environment A and environment B. In each environment, there is a presence of a light source and two s-robots are placed randomly at 75 and 95 cm away from the light source. The light sources in both environments are surrounded by a coloured band each and these colours represent the “danger” zone. However, in environment A, a section of the band is removed to create a “way in” zone which is a path designated for the robots to travel on to reach the light source whereas in environment B, there is an absence of the “way in” zone. The ultimate goal of the experiment is for both robots to discriminate between two different environments where they are able to move towards the light source safely in environment A and away from the danger zone in environment B.

Equipped with light sensors, floor sensors and a sound signalling system, the s-robots will first navigate around the environment with their wheels. When they are faced with “danger”, they will emit a sound to indicate that they are aware of the “danger” and begin to move away accordingly. It was found that these robots did not only communicate with the experimenter via their sound signals but the other robot in the same zone also responded to the sound signal and moved away from the coloured band. The s-robots were not just communicating via sound signalling but were behaving socially. Furthermore, the sound signal shows that it is an encoding sensory information that is integrated over time and contributes to the increase in the reliability of the categorisation process.

[Back to Table of Contents]

3.3 Embodied agents in language studies

Besides communication studies, certain aspects of language such as name formation, spatial inventories, and grammatical case can also be investigated using embodied agents. Due to the constraints in this wiki chapter, only the spatial language and perspective reversal game (Steels, 2010) will be briefly introduced.

Consider this example: When two people are standing opposite each other, they see the world differently because of the different vantage points that each person is adopting. However, the body of the self remains a constant landmark that can be used as a reference point. Thus, one could use various spatial terms such as left, right, up, and down (with respect to her own body) to communicate with the other about directions. How can embodied agents with no predefined inventory of spatial terms convey meaningful utterances to another?

[Back to Table of Contents]

3.3.1 Methodology

In this experiment, the researchers test the assumption of egocentric perspective transformation (EPT) (Steels, 2010). EPT is the agent’s ability to transform certain features of an object with respect to the position of another object. Over the course of approximately 5000 games, the experiments revealed that the agents are able to self-organize a communication system that includes the formation of an inventory of spatial categories (Steels, 2010). The embodied agents used in the study were five AIBO robot dogs that could move freely within an indoor laboratory.

Fig. 2. Graph of experimental results comparing the communicative successes of the AIBO robots (a) without any need for EPT, (b) required EPT, but could not perform it (c) could perform EPT but did not have spatial language lexicon and (d) could perform EPT and had spatial categories marked in language (Steels, 2010, p. 251).

[Back to Table of Contents]

3.3.2 Results

It was found that EPT in these robots were highly required for communicative success because EPT reduced the amount of cognitive effort needed for language users (Steels, 2010). A seen in the graph above, robots without EPT (graph b) only achieved 15% communicative success. In graph c, we see that robots with EPT unmarked in language achieved greater success, at the expense of greater cognitive effort because the hearer needs to adopt the speaker’s perspective and then perform EPT in the hearer’s world model (Steels, 2010). Robots with EPT marked in language (graph d) have an added “perspective indicator” in their conceptualization. Cognitive effort drops significantly, and the hearer knows which perspective to use instantly (Steels, 2010).

[Back to Table of Contents]

3.3.3 Discussion

EPT was found to be essential for communicative success (Steels, 2010). Spatial categories emerged as predicates in the robots’ conceptualization which helped reduce the cognitive effort of the hearer. This finding is significant for language evolution studies. From an evolutionary perspective, humans come from the superfamily of Hominoidea, of which the members are characterized by having no tails, being biped, and most significantly, having large brains. This comparatively large brain consumes a quarter of the body’s energy even when at rest; thinking was thus, a “costly” activity (Harari, 2014). It is possible that marking spatial categories in language was an evolutionarily more efficient way of using the brain.

[Back to Table of Contents]

Part I: Iterated Learning Model (ILM)

 

2. Iterated Learning Model (ILM)

2.1 Theory of Iterated Learning

iterated-learning-21-300x216

As the key mechanism of the cultural evolution of language, the aforementioned models are largely based on iterated learning.

Iterated learning refers to the process by which an individual acquires a behavior by observing a similar behavior in another individual who acquired it in the same way.

Early studies of iterated learning that observed human behaviour in a laboratory setting were designed to learn about cultural transmission. One of the pioneer studies in iterated learning is Bartlett’s (1937) ‘serial reproduction’ experiment, in which participants were exposed to some stimulus (such as drawings) and were then asked to reproduce the same material from memory. Their reproduced work served as the stimulus for a second participant, and so on. Bartlett observed that the material that was transmitted in this manner had changed as participants impressed their expectations about what they deemed was the right and appropriate content onto the material, thus causing it to be restructured.  For example, if one was shown a picture of an apple and told to draw it, another would observe the resultant drawing and produce a new drawing of an apple (as pictured below). An interesting observation that came out of the study was that drawings could change toward conventional, prototypical forms of the object drawn. (Bartlett, 1932)

apples-300x103

Spoken and signed languages, birdsong, and music are transmitted via iterated learning as opposed to explicit teaching. One’s linguistic behavior is thus a product of one’s observation of others’ similar behavior, which was induced by that of those who came before. The chain of diffusion occurs in not only cultural transmission, but also horizontal negotiation of conventions between peers of the same generation. Iterated learning is thus manifested both along a cross-generational chain of different individuals and back-and-forth within a dyad.

Iterated learning can have profound effects on linguistic structure. A study on such effects involved the learning of an artificial language by participants who were organized into diffusion chains. Such studies show that iterated is an adaptive process, in which the linguistic behavior being transmitted gains input from each generation to overcome key constraints. Some constraints to a language’s transmission over time include error rate and ambiguity.

[Back to Table of Contents]

2.2 Simulating cultural transmission of language

langevo-1024x681

The question of how language emerged concerns itself with both the biological evolution of the various cognitive capacities deemed necessary for language and the cultural evolution of languages, beginning from theorized proto-languages. It should be noted that cultural evolution should not be considered in isolation from its biological counterpart, as the cognitive adaptations an individual is equipped with bear implications for social interaction and learning.

Upon establishment, a language requires to be learned by subsequent generations, each from a prior generation, in a cross-generational manner, also known as vertical cultural transmission.

People learn a language from other people who once learned that language themselves.

Studies on cultural transmission seek to explain the changes an emergent language system undergoes.

The cultural transmission of a language can be said to give rise to design without intention and designer.

Exposure to linguistic behavior exhibited by members of one’s speech community induces one’s production of particular language properties. The resulting language used by one in turn translates to observable linguistic behavior which shapes the language of further members. Cultural evolution of the language is thus enabled by this cycle of repeated induction and elicitation of linguistic behavior.

Simulations of cultural transmission are based on the belief that when language is culturally transmitted, it develops:

  • Structure
  • Key design features unique to language over other communication systems
  • Enhanced learnability through minimization of errors

The models of reserach include computational agent-based simulations, mathematical models, and most recently, laboratory experiments.

Pioneer work on agent-based simulations sought to explain how negotiation due to learners’ biases and interaction between them influence communication systems greatly. Subsequent studies focused largely on the development of linguistic structure as a byproduct of cultural learning (specifically iterated learning), despite poverty of the stimulus. Work on mathematical models followed, supplementing the findings of such agent-based simulations through mathematical characterizations of changes effected by cultural transmission.

To support prior computational and mathematical models empirically, laboratory experiments aim to demonstrate how cumulative, adaptive, and non-intentional the cultural evolution of language is, by using human participants.

[Back to Table of Contents]

2.2.1 Challenges

One significant challenge faced by studies on cultural evolution remains to be the arguably reductionist approach taken. Any given language is constituted by thousands of language systems (capturing pragmatic, semantic, morphological, and phonological distinctions) and language strategies, of which all are intertwined. No model can hope to closely replicate all aspects of language evolution through simulation.

Despite availability of real-life observations of cultural transmission with the likes of Nicaraguan Sign Language, study of genuine emergence remains limited by the lack of direct, natural, data. Hence, only indirect evidence can be drawn.

[Back to Table of Contents]

2.3 Laboratory experiments with human subjects

A couple of studies have combined iterated learning techniques with artificial language learning or communication game paradigms in a bid to explore how languages and other communication systems evolve through learning and use. In language evolution, iterated learning has become a paradigm which involves experimentation with artificial languages. Human participants learn a set of items in the language, and then produce linguistic behavior which subsequent individuals learn from and so on. This was introduced by researchers Kirby S, Cornish H, and Smith K. A learning bottleneck is also imposed on transmission: participants are asked to learn a target language based on exposure to a smaller set of language items from the original set of stimuli, with the language produced by the nth participant. In this chain, the nth participant provides the input to participant n +1 and so forth. In short, participants have access to only a limited set of data.

[Back to Table of Contents]

2.3.1 Literature on laboratory experiments with humans

Many types of experiments relating to exploring the phenomenon of language evolution have been conducted in the laboratory. They are divided into experiments focusing on signal creation, the emergence of communication systems, and cultural transmission itself.

Signal Creation

Understanding the origins of language will involve the uncovering of the necessary cognitive capacities used for linguistic communication and detecting communication intentions. Studies on signal creation investigate how individuals recognise the communicative nature of certain behaviour, before even questioning how meaning is created from the signals. Scott-Phillips et al. (2009) ‘s embodied communication game (ECG) is a two-player game designed to serve this investigation. It requires participants to travel around a 2X2 grid with movement as their only communicative resource. This forces participants to find ways in revealing which movements are communicative in nature rather than acts of travel. The difficulty of this task revealed that common ground serves is especially key to the emergence of communication channels. Related work (as cited in Scott-Phillips & Kirby, 2010) also lead to similar conclusions. The challenge of these games is highlighted in the difficulty of communicating one’s communicative intent.

Emergence of communication systems

Once communicative intent is established, individuals now face the task of negotiating the forms and meanings of symbols to create a communication system. In a pioneer study done by Galantucci (2005), pairs of participants were tasked to invent and agree on a set of signs to use to solve a coordination problem. The study aimed to illustrate how human communication can be understood as a form of joint action. An example of a line of research that has been spawned from the pioneering studies on the role of interaction in the emergence of communication system is the use of graphical communication tasks. Such tasks are advantageous in its provision of a medium which allows the invention of new signs to be used in an interactive context.  Examples of studies that illustrates this include those done by Garrod et al (2007, 2010).

Cultural transmission

Following the establishment of some sort of language, cultural transmission, which is an instance of iterated learning, takes place. In Kirby et al’s (2008) experiment, participants were asked to learn labels for coloured moving shapes, where the initial artificial language provided a randomly generated, unique label for the shapes. Their findings revealed a structured language that developed from the initial unstructured set of meaning-signal associations as a result of the iterated learning. By the 10th participant, each label had consisted of a prefix which specified colour (e.g. ne– for black, la– for blue), a stem for shape (e.g. –ho– for circle, ­-ki­- for triangle) and an affix specifying motion (e.g. –plo for bouncing, –pilu for looping). As predicted by mathematical modelling results, languages developed over time to ones that facilitated generalisations and rules. Compositional languages also developed where components or sub-parts of each complex label or word specified components of the picture that the label had referred to.

[Back to Table of Contents]

Chapter 8 – Language Evolution in the Laboratory

2019: Nur Amirah Bte Rosman, Joanne Tan Hui San, Cheng Wei Cong Jonathan
2014: Goh Xiao-Qing, Ng Si, Ning You Jing

evomon_welcome_by_zaratus

This is a site dedicated to Language Evolution, mainly focusing on the topic of Language Evolution in the Laboratory. We’ve divided this topic into pages in a chronological order. Hope you’ll have a good read!

Start reading about experiments on language evolution and the ideas behind them now!

1. Introduction

1.1 Observing Language Transmission

The origins of natural language cannot be observed directly. In recent years, evolutionary linguists have designed experiments in the laboratory to study the human cognitive capacities necessary for language and the emergence of new languages. There are three main methodologies to understand how the acquisition of language occurs among populations of individuals. Firstly, computational/robotic models where embodied agents (see also: Chapter 4) interact with one another in simulated environments (Kirby, 2012). Less commonly, there are also mathematical models which focus on mathematical techniques. Lastly, the iterated learning model (ILM) which uses human subjects, is based on the principle that individuals learn by observing instances of others’ behavior and population level behavior is a collective result of the interaction of individual subjects (Kirby, 2012). These models represent and explore the plausible hypotheses about the historical origins of languages.

1.2 Research Significance

Language evolution experiments thus focus on the emergence of new languages and how they are used by human participants. A central theme in language evolution laboratory research is the transformation of individual-level behaviours observed in each participant into linguistic phenomena that can occur at the level of an entire population (Kirby, 2012). The emergence of languages or linguistic phenomena cannot be explained only by reference to the evolution of our biological capacities such as our cognitive mechanisms. As such, we need to consider factors of interaction that occur from generation to generation such as cultural transmission and feedback in providing a considered account for the evolution of language. Research in the laboratory is thus able to provide with empirical data that can further the investigation on the role of cultural transmission and feedback.

In Part I of this chapter, we will explain iterated learning as the underlying concept of laboratory experiments on language evolution. It explores the main hypotheses of and previous work on the cultural transmission of language. In Part II, we present some studies of computational models which help us corroborate the validity of certain theories in language. Lastly, the limitations faced by this area of research are briefly discussed.