Abstract
AbstractIn this work, I address the connection of phonetic structure with phonological representations. This classical issue is discussed in the light of recent neurophysiological data which – thanks to direct measurements of temporal and spatial brain activation – provide new avenues to investigate the biological substrate of human language. After describing principal techniques and methods, I critically discuss magnetoencephalographic and electroencephalographic findings of speech processing based on event-related potentials and event-related oscillatory rhythms. The available data do not permit us to clearly disambiguate between neural evidence suggesting pure acoustic patterns and those indicating abstract phonological features. Starting from this evidence, which only at the surface represents a limit, I develop a preliminary proposal where discretization and phonological abstraction are the result of a continuous process that converts spectro-temporal (acoustic) states into neurophysiological states such that some properties of the former undergo changes interacting with the latter until a new equilibrium is reached. I assume that – at the end of the process – phonological segments (and the related categorical processes) take the form of continuous neural states represented by nested cortical oscillatory rhythms spatially distributed in the auditory cortex. Within this perspective, distinctive features (i.e., the relevant representational linguistic primitives) are represented by both spatially local and distributed neural selectivity. I suggest that this hypothesis is suitable to explain hierarchical layout of auditory cortex highly specialized in analyzing different aspects of the speech signal and to explain learning and memory processes during the acquisition of phonological systems.
Publisher
Cold Spring Harbor Laboratory