User:Jrradney/Temp/Fraser1992.doc

Fraser, Helen. 1992.The subject of speech perception: an analysis of the philosophical foundations of the information-processing model. London: MacMillan Pr.

‘The project reported here has been very much one of analysis and exploration. It is an attempt to understand, clarify and explicate some underlying assumptions of the information-processing model of human speech perception. My original motivation in undertaking this analysis was a general dissatisfaction or unease with information processing models as an account of human speech perception...This book is the result of my attempt to understand the metatheoretical position I disagreed with, and pinpoint the reasons for that disagreement. It can be seen as an extended argument in justification of my reaction: “But people aren’t like that!”’ (xi).

Fraser began with an understanding that people aren’t as the accounts of human behaviour that were based upon the model by which human cognition is seen as a complex version of the circuitry of a computer.

‘The book does not contain reports of any new experiments; it is not about a new way of “doing speech perception”. It is about a new way of thinking about speech perception, from which I hope experimental work will grow’ (xi).

Helen Fraser summarises the process of communication assumed by researchers who follow the information processing model of human cognition in the following way.

‘Speech communication is seen as a process of message transfer, in which the speaker converts a meaning (the message) into sound (articulatory gestures with acoustic consequences) which is transferred to the ear of the hearer. The hearer receives the sound and matches it against meanings, similar to those of the speaker, stored as part of his or her own linguistics knowledge. The goal of speech perception is thus the retrieval of the speaker’s meaning.... Speech perception is the “decoding” operation of matching the sound against meanings, or transforming the sound into meaning, or extracting meaning from the sound’ (3).

This understanding is based upon the conduit metaphor and the code model of communication. Telementation, whereby the thoughts of a communicator are transferred to the mind of another, is a foundational assumption regarding communication.

‘Contrary to expectations, the speech signal is not naturally or obviously divisible into any clearly defined, linguistically relevant units. In an acoustic description, the signal is a quasi-continuous stream; the breaks or segmentations that do occur are not obviously correlated with any of the units or segments that appear to the hearer to be “there”, ready to be matched with their meanings’ (5).

While undoubtedly, there are breaks at larger levels (breath groups, pause groups, rhythm groups) certainly segmentation of the speech stream is a precise and technical process requiring disciplined and lengthy training and well beyond the capability of most communicators.

‘Phoneme monitoring experiments, as well as other evidence, suggests that the phoneme level is not always accessed in order before higher levels – those with larger units... it is not always the case that all stages are traversed in order, with acoustic features being combined to give small linguistic units and larger linguistic units being built up on the basis of information in the intermediate stages’ (21).

An important implication of this is that processing must not be a matter of concatenating sounds into larger and larger units giving sense to the message. If such a procedure is not a given when processing is done, there is also no reason why linearity must be a feature of the generative process either.

Fraser believes that the information-processing model of cognition is inadequate in several ways, which

‘stem precisely from its presupposition of a processing system which is not a possible description of a human being as a perceiver of speech.... There is ... an alternative approach – of starting with an understanding of the nature of human beings and human cognition in general, and then narrowing the focus to account for speech perception in terms of the more general principles’ (36-7).

Fraser uses this approach, beginning with the work of Husserl and Heidegger on human lived experience and building an understanding that cannot be reconciled with the model of human cognition based upon the function of computers.