Language is Not About Language: Towards Formalizing the Role of Extra-Linguistic Factors in Human and Machine Language Acquisition and Communication

Despite the large research efforts in understanding early language acquisition (LA), it is still unclear how young children learn to transform their noisy and ambiguous auditory experience into a symbolic compositional representational system known as language. This paper argues that a major obstacle towards a more comprehensive picture of LA is the lack of a unified conceptual framework that would capture the full extent of factors critical to language learning in real world contexts, and that we should pursue such a framework in order to be able to place individual behavioral studies and computational models into a mutually compatible context. As an example of the issue, the widely used standard model of the speech chain—a description of the information flow from talker’s idea to listener’s interpretation of the meaning of the spoken message—is shown to be insufficient for characterizing learning and communication in natural contexts. Instead, a realistic model should account for the inherent multimodality and contextual dependency of communication and learning by formally acknowledging the role of a shared communicative context, interlocutors’ subjective representations of the shared situation, and how these factors drive message generation and speech perception in order to acquire information on the external world. By understanding how language is connected to the more generic sensorimotor and predictive processing principles of human cognition, we can also start to understand the core forces driving language learning in natural environments and through varying individual developmental trajectories.