Speech Production Flashcards
How many words do we produce per second in a normal fluent convo?
2-3 words (4 syllables, 10-12 phonemes)
How many words are in the mental lexicon of a normal, literate adult monolingual?
50-100 thousand words
How many times do we err on avg?
No more than once or twice in 1000 words
How many word tokens would we have produced by adulthood (40 mins of talking a day)?
50 million word tokens
What are the types of Meaning-Based Speech Errors?
1) Word Exchanges (Switch 2 words)
2) Blends/Contaminations (Two items fuse together)
3) Deletion (Omission of linguistic material)
4) Lexical Selection (Wrong selection of word)
What are the types of Form-Based Speech Errors
1) Shift (One speech segment disappears from appropriate location & appears somewhere else)
2) Sound/Form Exchange (Two sounds switch places)
3) Anticipation (Later segment takes place of earlier segment)
4) Perseveration (Earlier segment replace a later item)
5) Addition (Add linguistic material)
6) Spoonerism (Switch of initial sounds of 2 separate words)
7) Residual speech errors (distortion of late-developing sounds)
What are meaning-based errors?
A level where syntactic functions are assigned
What are form-based errors?
A level where ordering of forms is organised
What are the 2 modular levels of processing?
1) Dell’s 2-step Interactive Activation Model
2) Levelt et al.’s WEAVER++ Model (Word Encoding by Activation and Verification)
What are the 3 levels of representation in the network organisation of Dell’s model?
1) Semantic (decomposed into features)
2) Words & morphemes
3) Phonemes (Sounds)
What did Dell propose?
Interactivity / Bi-directionality of connections of the 3 levels
What are mixed errors?
Errors that are both semantic and phonological-based
Why does Dell argue against the idea that the network serves in both word pdn & word percepn?
Many aphasic patients show GD auditory word recognition & disturbed phonological encoding
What is the function of the bi-directional connection?
To support fluency in lemma selection
What is the Stroop Task?
Stimuli are differently coloured words, Participants asked to identify colour or say word
What was found in the Stroop Task?
Slowed down colour naming when coloured word is a different colour name
What did Cattell (1885) found?
Naming a list of 100 line drawings of objects took twice as long as naming a list of the corresponding object names
Why was Cattell’s finding so?
Direct access route between word & phonological code
Extra step in activating the object concept
What is the Picture-word interference task?
Participants are to name basic objects as quickly as possible with distractor words embedded in the obj
What was found in the Picture-word Interference Task?
Semantic interference (Meaning-related words slowed down naming the pic while Form-related words speed up processing)
What does the network in WEAVER++ Model consist of?
1) Conceptual Stratum - deciding on the message
2) Lemma Stratum - turning the message into linguistic representations & grammatical encoding
3) Word-form Stratum - Morpho-phonological encoding, prosodification & phonetic encoding and articulation
How does the WEAVER++ Model take place?
1) Lemma Activation
2) Lemma Selection
3) Conversion of Lemma to Phonological Representation
4) Other processes (Stress, phonological segments, prosodification, phonetics & articulation)
What is Prosodification concerned with?
1) Incremental Syllabification (S.. proceeds from left to right with chunking)
2) Implicit Priming Paradigm (Response latencies are significantly shorter in the homogeneous condition than in the heterogeneous condition; tests whether speaker knows which syllable to stress)
What is an Implicit Prime?
Homogeneous words that share beginning part or initial syllable
When does Implicit Priming occurs?
For words that share beginning part of the word but not for words sharing any final part
When does Implicit Primin not occur?
For words that do not share stress position
What is Monitor / Self-perception concerned with?
Conceptual system (produce & perceive speech) is shared with self-monitoring but not phonological system
What is acoustic/phonetic analaysis linked to?
Left temporal lobe
What is phonetic generation of speech controlled by?
Motor & Premotor Areas of the Frontal Lobe & Left Central Gyrus of the Insula
What is the feedback mechanism used by speaker?
Perceptual Loop (Has an ext branch & internal one)
What are some assumptions of the WEAVER++ model?
1) Modularity (later processes cannot affect earlier processes)
2) Only 1 Lemma activates word form
What are the similarities between the 2 models?
1) Similar Representations
2) Frames & Slots (Insert of Representations into the Frames)
What are the differences between the 2 models?
1) Dell (Interactive) vs WEAVER++ (Serial)
2) Dell (Cascaded) vs WEAVER++ (Modular)
3) WEAVER++ has the Perceptual Loop
