This week we’re reading
Agnes and Michael will be presenting work and plans at this event next week; abstracts below. Good luck, have fun!
An ecologically-inspired real-time visual feedback training system to support novel vowel production
During infancy, the explosion of possibilities afforded by increasing control over language resources are paradoxically accompanied by a significant narrowing of flexibility when it comes to the building blocks of speech itself. Powerful perceptual biases are at play before we can produce words, directing attention towards a native phonology which ultimately constrains the sounds that come to auditorily define us in the world. For these reasons, second language learners struggle to develop novel accents without influence from their native tongue and demonstrate poor perceptual discriminability and categorisation of non-native sounds (Best, 1995; Flege, 1995).
Augmented feedback systems employed for speech-training frequently include rich anatomical, or abstract acoustic feedback which fail to appreciate the tight coupling of perception and action, yielding displays that are difficult to interpret and use. This talk considers the motivation and design for a feedback system providing real-time information on novel speech-sound production along task- relevant perceptual dimensions.
An ecological approach characterises the linguistic entities to be learned as vocal tract gestures (e.g. Fowler, 1986) and proposes that the learning process involves both action execution and exposure to relevant perceptual control information. This feedback is a simple 2D visual display in which speakers’ productions are represented by a continuously moving dot which they must drive towards a target, taking advantage of the mapping between tongue position and vowel formants.
With small periods of training, augmented access to speech gesture information is fast and effective at driving and stabilising production. Crucially, these gains hold in the presence and absence of feedback. This work acts as a proof-of-concept of the power of the ecological approach to increase our theoretical and practical understanding of speech-sound acquisition.
The co-evolution of language and object throwing in hominins
Language has a multiplicity of uses for humans today and many causes have been posited for its origin, e.g. tool use, information transmission, social interaction, abstract thought. In 1982, the neurophysiologist WH Calvin proposed a novel hypothesis for the origin of language capacity in hominins: Morphological adaptations for arboreal living in ancient apes were exapted for object throwing and led to the emergence of hominins as an action-at-a-distance predator. Throwing accurately however requires precise timing in the coordination of various effectors, placing a strong selective pressure for multiple morphological changes including more neurons and increased encephalisation. These timing requirements are similar to those required for speech and Calvin hypothesised that the neural mechanisms which supported this timing in throwing were then exapted for language.
Calvin’s hypothesis has never been experimentally tested in humans. Multiple recent strands of paleoanthropological, archaeological, genetic, linguistic and psychological research indicates both language and throwing are more ancient than previously thought. Moreover, evidence for throwing behaviours often predates evidence for language. Many researchers are understandably sceptical of the potential pitfalls of such investigations, especially the creation of just-so stories. However, if language is viewed as an embodied activity rather than a mental process, then both speaking and throwing can be viewed as the activities of organisms and become open to investigation in a similar manner to any other behaviour.
This talk will explain Calvin’s hypothesis, highlight recent language and throwing evolution research, make a case for its investigation and outline a specific experimental method combining linguistics and movement science.
We’ll be reading this in two weeks