LANGUAGE

Question 1

REVISED DUAL-ROUTE MODEL [Coltheart et al. 2001]

Answer 1

Revised DRM
= printed language - graphemes [sublexical recoding — phonology] - lexicon [non-semantic reading —phonology] - semantic system - phonology - speech

Three-route model = accounts for lexical effects of non-words and regularity effects by means of cross-talk between lexical and non-lexical route

Surface Dyslexia = loss of the ability to make contact with the orthographic lexicon

Question 2

INFLUENCES ON SPEECH PERCEPTION

Answer 2

SEGMENTATION
= Distinguishing separate phonemes and words from a pattern of speech sounds

COARTICULATION
= the ‘b’ phonemes in ‘bill’ ‘ball’ ‘able’ ‘rub’ are all acoustically slightly different

SPEAKER VARIATION
= e.g. dialect, rate of talking etc

RAPID PROCESSING

MCGURK & MCDONALD [1976]
PRIOR CONTEXT [Harley, 2013]
= Interactionist vs. Autonomous
= Interactionist contextual information can influence processing at an early stage and influence the next word perception
= Autonomous states context has its effects late in processing

BROCK & NATION [2014]
=“Alex fastened the button” [butter in the display] -> competitor/constraining
= “Alex chose the button” [lettuce in the display] -> unrelated

WARREN & WARREN [1970]
PHONEMIC RESTORATION EFFECT
= demonstration of context, but as context after target word might be used to support either Interactionist or autonomous accounts

Question 3

TRACE MODEL [McClelland & Elman, 1986]

Answer 3

Individual processing units or nodes at 3 different levels - words, phonemes and features (voicing, manner of production, power etc)

Connections between levels operate in both directions and are always facilitatory

There are connections among units or nodes at the same level - connections are inhibitory

Nodes influence each other in the proportion to their activation levels and strength of their interconnections

Word identified by the listener is determined by the activation level of the possible candidate words

Supports a lot of experimental data = word frequency effect, neighbourhood effects, context effects, phonemic restoration effect and copes extremely well with noisy input [HARLEY, 2008]

LIMITATIONS
Gaskell and Marslen-Wilson [1998] = trace oversteps importance of top-down processes

Perre and Ziegler [2008] = ignores role of orthography of the spoken word plays on its perception

Question 4

REVISED COHORT MODEL [Marslen-Wilson and Tyler, 1990]

Answer 4

Focuses on the processes involved overtime during spoken word recognition

Differs from trace = it focuses more on bottom-up and less on top-down

UNIQUENESS POINT
= point at which only one word is consistent with the acoustic signal

FRAUNENFELDER ET AL [2001]
= found hearing words “focabulaire” activated the word “vocabularie”

Model assumes initial cohort defined by words containing similar initial phonemes

THREE PROCESSING STAGE TO REVISED MODEL
= Access - during which word initial cohort is selected
= Selection - during which word is selected from cohort
= Integrated - the selected words semantic and syntactic properties are integrated within the sentence (in the revised model context effects only influences the later integration stage)

O’ROUKE & HOLCOMBE [2002]
= Found N800 (used to assess speed of word processing)
= Occurred 100ms faster in words with earlier compared to later uniqueness points -> supporting revised cohort model

GROSJEAN [1980]
= The gating task
= found 333ms to identify word in isolation, only 199ms in context

Despite this, context does not appear to have an early effect as words are activated in the cohort that are not compatible with context but are with sensory input [e.g. phonemes and structural features of word input]

LIMITATIONS
= RADEAU ET AL [2000] cast doubt on importance of uniqueness point
= listeners heard nouns with early or late uniqueness points
= found uniqueness point influenced thee identification of nouns which were presented slowly or at a medium speed but not when presented at normal conversation speed

VAN PETTEN ET AL [1999]
= found context sometimes influences word processing earlier than the integration stage - particularly when context is strongly predictive, or speech input is degraded [Wild et al, 2012]

GAGNEPAINE ET AL [2012]
= shown that prediction of upcoming speech segments may b of more importance than the revised cohort model suggests - playing a greater role in the selection of speech perception