How does language emerge from a prior state in which no language exists? This is the central question for the field of language evolution. Although early attempts to address this question focussed on biological evolution, the current consensus is that cultural evolution plays a fundamental explanatory role (Tamariz & Kirby 2015). To understand language evolution, we need to understand how individual humans improvise solutions to communicative challenges, how groups of individuals create conventions through interaction, and how these conventions are transmitted over time through learning.
The manual modality provides the best hope we have of understanding how these processes work and answering the central question of language evolution. It offers a broad range of phenomena, from fully conventionalised sign languages to cases where a conventionalised system has not yet been established. In particular, research into homesign systems and emerging sign languages such as Nicaraguan Sign Language, but also the silent gesture paradigm in the laboratory, allow observation of human communication systems from their point of origin, and directly allow us to investigate how linguistic structures evolve.
We recognise that it is essential to have clear terminology, and to be aware of the differences between sign and gesture. However, the rigid dichotomy between gesture as pictorial and sign as categorical is problematic when it comes to determining the characteristics of the aforementioned cases: silent gesture, homesign, and possibly also emerging sign languages. Because on which side of the dividing line do these fall? As Goldin-Meadow & Brentari (G-M&B) note (sect. 7.1 and 7.2), homesign and silent gesture are language-like in some respects but not in others.
An evolutionary perspective shifts emphasis away from problematic questions about the status of silent gesture, homesign, and the early stages of emerging sign languages as being either pictorial and gesture-like or categorical and sign-like. Instead, we argue that the emphasis should be on the continuity of cultural-evolutionary processes involved in shaping these various systems.
These phenomena are ultimately rooted in situations of communicative stress; they emerge because no existing conventional language system is available. Where they differ is in which cultural forces have the upper hand in the situations in which they emerge. For example, silent gestures elicited in laboratory experiments are not subject to the routinisation that occurs from using a system repeatedly, whereas homesign systems are. This may be the explanation behind differences found between the two phenomena, such as that for motion events mentioned by the authors (sect. 7.2): Silent gesturers do not break their gestures for motion events into path and manner components, whereas homesigners do.
This shift of emphasis that is at the heart of an evolutionary approach to language – a shift from considerations of state to considerations of process – can be extended to the silent gesture laboratory paradigm. We propose augmenting this paradigm by implementing different cultural processes, such as communicative interaction and cultural transmission. To do this, we can borrow from the experimental techniques developed in the field of language evolution more broadly.
The iterated learning paradigm (Kirby et al. Reference Kirby, Griffiths and Smith2014), in which a participant learns a language from the output of a previous participant, has been used to probe the role that learning plays in shaping linguistic structure, specifically through modelling the transmission of language to new learners. More recently, this experimental framework has been expanded to investigate the effects of interaction in conjunction and in comparison with transmission to new learners. Kirby et al. (Reference Kirby, Tamariz, Cornish and Smith2015) studied pairs of participants organised into transmission chains (a condition with both interaction and transmission) compared with isolated pairs of participants (an interaction-only condition). Their results showed that when both transmission and interaction processes were at play, the compositional structures found in natural languages emerged. Isolating these processes, however, had different effects: The participants in the interaction-only condition produced “holistic” systems, useful for expressive communication, but not compositionally structured. Similarly, studies looking only at transmission to new learners, but without interaction between pairs (Cornish et al. Reference Cornish, Smith, Kirby, Knauf, Pauen, Sebanz and Wachsmuth2013; Kirby et al. Reference Kirby, Cornish and Smith2008), found that easily learnable but nonexpressive, unstructured languages were the result.
We have now begun to apply this framework to the manual modality, assessing the effects of cultural processes in the laboratory alongside data from homesign and emerging sign languages. Following research into motion events in Nicaraguan Sign Language, Smith et al. (Reference Smith, Abramova, Cartmill and Kirbyunder revision) examined the effect of cultural transmission on motion events in silent gesturers. Supporting previous results, the gestures produced by participants became more regular and structured as they were transmitted to new learners, showing increasingly language-like properties.
Expanding this paradigm, Motamedi et al. (Reference Motamedi, Schouwstra, Smith and Kirbysubmitted) studied the emergence of systematic category structures in silent gesturers, looking at the effects of iteration alone, interaction alone, and the effects of both processes working together. Corroborating the work by Kirby et al. (Reference Kirby, Tamariz, Cornish and Smith2015), Motamedi et al. (Reference Motamedi, Schouwstra, Smith and Kirbysubmitted) concluded that participants in the iteration and interaction condition produced fully systematic systems, which did not emerge in the conditions where these processes were isolated.
These findings make it clear that silent gesture elicited from single participants is a very temporary phenomenon: They are the structures that participants produce in the lab when they are asked to do this for the first time. The patterns that are observed can be seen as representative of an individual's cognitive preferences for structuring information (Schouwstra & de Swart Reference Schouwstra and de Swart2014). When these utterances are subject to cultural processes such as communicative interaction and cultural transmission, they will be transformed to become more systematic and increasingly regular. Being able to witness these processes at play in the lab is extremely exciting and informative. At the same time, we are convinced that we should not restrict our view to laboratory data alone (Schouwstra Reference Schouwstra2012). Combining the precision of laboratory experiments with the naturalness of field data is a promising next step in uncovering the cultural processes that shape emerging language.
How does language emerge from a prior state in which no language exists? This is the central question for the field of language evolution. Although early attempts to address this question focussed on biological evolution, the current consensus is that cultural evolution plays a fundamental explanatory role (Tamariz & Kirby 2015). To understand language evolution, we need to understand how individual humans improvise solutions to communicative challenges, how groups of individuals create conventions through interaction, and how these conventions are transmitted over time through learning.
The manual modality provides the best hope we have of understanding how these processes work and answering the central question of language evolution. It offers a broad range of phenomena, from fully conventionalised sign languages to cases where a conventionalised system has not yet been established. In particular, research into homesign systems and emerging sign languages such as Nicaraguan Sign Language, but also the silent gesture paradigm in the laboratory, allow observation of human communication systems from their point of origin, and directly allow us to investigate how linguistic structures evolve.
We recognise that it is essential to have clear terminology, and to be aware of the differences between sign and gesture. However, the rigid dichotomy between gesture as pictorial and sign as categorical is problematic when it comes to determining the characteristics of the aforementioned cases: silent gesture, homesign, and possibly also emerging sign languages. Because on which side of the dividing line do these fall? As Goldin-Meadow & Brentari (G-M&B) note (sect. 7.1 and 7.2), homesign and silent gesture are language-like in some respects but not in others.
An evolutionary perspective shifts emphasis away from problematic questions about the status of silent gesture, homesign, and the early stages of emerging sign languages as being either pictorial and gesture-like or categorical and sign-like. Instead, we argue that the emphasis should be on the continuity of cultural-evolutionary processes involved in shaping these various systems.
These phenomena are ultimately rooted in situations of communicative stress; they emerge because no existing conventional language system is available. Where they differ is in which cultural forces have the upper hand in the situations in which they emerge. For example, silent gestures elicited in laboratory experiments are not subject to the routinisation that occurs from using a system repeatedly, whereas homesign systems are. This may be the explanation behind differences found between the two phenomena, such as that for motion events mentioned by the authors (sect. 7.2): Silent gesturers do not break their gestures for motion events into path and manner components, whereas homesigners do.
This shift of emphasis that is at the heart of an evolutionary approach to language – a shift from considerations of state to considerations of process – can be extended to the silent gesture laboratory paradigm. We propose augmenting this paradigm by implementing different cultural processes, such as communicative interaction and cultural transmission. To do this, we can borrow from the experimental techniques developed in the field of language evolution more broadly.
The iterated learning paradigm (Kirby et al. Reference Kirby, Griffiths and Smith2014), in which a participant learns a language from the output of a previous participant, has been used to probe the role that learning plays in shaping linguistic structure, specifically through modelling the transmission of language to new learners. More recently, this experimental framework has been expanded to investigate the effects of interaction in conjunction and in comparison with transmission to new learners. Kirby et al. (Reference Kirby, Tamariz, Cornish and Smith2015) studied pairs of participants organised into transmission chains (a condition with both interaction and transmission) compared with isolated pairs of participants (an interaction-only condition). Their results showed that when both transmission and interaction processes were at play, the compositional structures found in natural languages emerged. Isolating these processes, however, had different effects: The participants in the interaction-only condition produced “holistic” systems, useful for expressive communication, but not compositionally structured. Similarly, studies looking only at transmission to new learners, but without interaction between pairs (Cornish et al. Reference Cornish, Smith, Kirby, Knauf, Pauen, Sebanz and Wachsmuth2013; Kirby et al. Reference Kirby, Cornish and Smith2008), found that easily learnable but nonexpressive, unstructured languages were the result.
We have now begun to apply this framework to the manual modality, assessing the effects of cultural processes in the laboratory alongside data from homesign and emerging sign languages. Following research into motion events in Nicaraguan Sign Language, Smith et al. (Reference Smith, Abramova, Cartmill and Kirbyunder revision) examined the effect of cultural transmission on motion events in silent gesturers. Supporting previous results, the gestures produced by participants became more regular and structured as they were transmitted to new learners, showing increasingly language-like properties.
Expanding this paradigm, Motamedi et al. (Reference Motamedi, Schouwstra, Smith and Kirbysubmitted) studied the emergence of systematic category structures in silent gesturers, looking at the effects of iteration alone, interaction alone, and the effects of both processes working together. Corroborating the work by Kirby et al. (Reference Kirby, Tamariz, Cornish and Smith2015), Motamedi et al. (Reference Motamedi, Schouwstra, Smith and Kirbysubmitted) concluded that participants in the iteration and interaction condition produced fully systematic systems, which did not emerge in the conditions where these processes were isolated.
These findings make it clear that silent gesture elicited from single participants is a very temporary phenomenon: They are the structures that participants produce in the lab when they are asked to do this for the first time. The patterns that are observed can be seen as representative of an individual's cognitive preferences for structuring information (Schouwstra & de Swart Reference Schouwstra and de Swart2014). When these utterances are subject to cultural processes such as communicative interaction and cultural transmission, they will be transformed to become more systematic and increasingly regular. Being able to witness these processes at play in the lab is extremely exciting and informative. At the same time, we are convinced that we should not restrict our view to laboratory data alone (Schouwstra Reference Schouwstra2012). Combining the precision of laboratory experiments with the naturalness of field data is a promising next step in uncovering the cultural processes that shape emerging language.