MODELS OF SPEECH PERCEPTION - LLE 2 (Language comprehension involves…
MODELS OF SPEECH PERCEPTION - LLE 2
Language comprehension involves certain sub-operations: segmenting continuous speech signals
G. updating model to reflect the meaning of the new sentence
F. making a mental model of discourse
choosing among different possible structures when there are syntactic ambiguities
D. parsing into tree structures
C. finding appropriate meanings of ambiguous words
looking up up words and morphemes in mental lexicon,
A. into phonemes, morphemes, words and phrases, :
BOTTOM UP PROCESSING
Listener waits for 'the' and 'boy' before constructing a NP, tehn waits for next word etc...
Each step of building meaning is based on sensory data and accompanying lexical info.
moves step by step from incoming acoustic (or visual) signal -- to phonemes, morphemes -- to words and phrases --- to semantic interpretation
TOP DOWN PROCESSING
has a role for context (top down processing) in segmentation
eg sequence of phonemenes compatible with 2 segmentations - but phonetic details of pronunciation can signal where word boundary is
ret] as in
:check: night rate - t is part of the coda of the first syllable and aspirated or
:check:nitrate t begins the onset of the second syllable, which is stressed and t is aspirated
:check: phonetic sequence [g,r,e,d,e] in context of meat or eggs will be heard as 'Grade A'
in a discussion of weather, it is heard as 'grey day'
eg: when hearing 'the', listener begins constructing a a Noun Phrase (NP) - expects that next word could be a noun , eg 'boy' --- USES KNOWLEDGE OF PHRASE STRUCTURE as source of info
Higher level info to predict what is to follow in the signal
from semantic and syntactic info
to lexical info gained from sensory input
Language perception and comprehension involves both top-down processing and bottom-up processing