Covert Editing Process Introduction There is intriguing evidence indicating that editing processes intervene between the planning of an utterance and its articulation. The editing operations might provide a last check to determine whether the planned utterance is linguistically and socially acceptable Evidence: Laboratory-induced speech errors o Phonological bias effect ◆ Ball doze ◆ Bash door ◆ Bean deck ◆ Bell dark ◆ Darn bore ◆ RESPOND co The target is darn bore, but the preceding four pairs increase the likelihood of the spoonerism barn door: It was found that the spoonerism occurs about 30%o of the time
Covert Editing Process ◼ Introduction ◆ There is intriguing evidence indicating that editing processes intervene between the planning of an utterance and its articulation. The editing operations might provide a last check to determine whether the planned utterance is linguistically and socially acceptable. ◼ Evidence: Laboratory-induced speech errors ◆ Phonological bias effect ⧫ Ball doze ⧫ Bash door ⧫ Bean deck ⧫ Bell dark ⧫ Darn bore ⧫ RESPOND The target is darn bore, but the preceding four pairs increase the likelihood of the spoonerism barn door. It was found that the spoonerism occurs about 30% of the time
◆ The lexical effect: ◆ Big dutch ◆ Bang doll ◆ Bill deal ◆ Bark do ◆ Dart board ◆ RESPOND co The spoonerism bart doard occurs only 10%o of the time induced speech errors that result in words are more frequent than errors that result in nonwords co Explanation: During speech, we sometimes develop two or more competing speech plans for production, but we use the editing process to suppress the less desirable plan, which leads to the relatively low frequency of certain types of errors
◆ The lexical effect: ⧫ Big dutch ⧫ Bang doll ⧫ Bill deal ⧫ Bark dog ⧫ Dart board ⧫ RESPOND The spoonerism bart doard occurs only 10% of the time -- induced speech errors that result in words are more frequent than errors that result in nonwords. Explanation: During speech, we sometimes develop two or more competing speech plans for production, but we use the editing process to suppress the less desirable plan, which leads to the relatively low frequency of certain types of errors
Parallel models Introduction o Serial models assume that the linguistic message is organized at semantic, syntactic, morphological, and phonological levels. Activation of a node at one level may trigger activation of nodes at other levels, and feedback may occur from morphological and phonological levels back to higher levels of processing account for several important research findings(Dello Models organized along these lines have been shown 1985,1986,1988; MacKay1982,1987 Stemberger 1985) ■ Illustration
Parallel models ◼ Introduction ◆ Serial models assume that the linguistic message is organized at semantic, syntactic, morphological, and phonological levels. Activation of a node at one level may trigger activation of nodes at other levels, and feedback may occur from morphological and phonological levels back to higher levels of processing. Models organized along these lines have been shown to account for several important research findings. (Dell 1985, 1986, 1988; MacKay 1982, 1987; Stemberger 1985) ◼ Illustration
e Suppose a person activated the word reset at the syntactic level this simply means that the person intended to place this noun in the syntactic frame being developed. This activation at the syntactic level then triggers activation of the component morphemes, re-and set, at the morphological level. These morphological nodes further spread the activation to the phonological level as well, activating the node for the phoneme /r/ Evidence-It can account for(pp 203-204) ◆ Lexical bias effect Speech errors favor true words because true words have morphological nodes (nonwords not), which can receive feed back from phonological nodes
◆ Suppose a person activated the word reset at the syntactic level; this simply means that the person intended to place this noun in the syntactic frame being developed. This activation at the syntactic level then triggers activation of the component morphemes, re- and set, at the morphological level. These morphological nodes further spread the activation to the phonological level as well, activating the node for the phoneme /r/. ◼ Evidence—It can account for (pp.203-204): ◆ Lexical bias effect ⧫ Speech errors favor true words because true words have morphological nodes (nonwords not), which can receive feedback from phonological nodes
o Phonemic similarity effect The tendency for intruding phonemes to be phonemically similar in their distinctive feature composition to the target phonemes- because each phoneme that is activated spreads its activation to the corresponding set of distinctive features, in turn, the features then activate a number of phonemes that share one or more of these features This increases the probability that an intrusion will be phonologically similar to ne target Slow speaking rate effect o Slow speaking rate is associated with fewer speech errors because there is more time for activation to spread from the current morphemes to the correct sounds and for the activation of the previously activated sounds to decay
◆ Phonemic similarity effect ⧫ The tendency for intruding phonemes to be phonemically similar in their distinctive feature composition to the target phonemes – because each phoneme that is activated spreads its activation to the corresponding set of distinctive features, in turn, the features then activate a number of phonemes that share one or more of these features. This increases the probability that an intrusion will be phonologically similar to the target. ◆ Slow speaking rate effect ⧫ Slow speaking rate is associated with fewer speech errors --- because there is more time for activation to spread from the current morphemes to the correct sounds and for the activation of the previously activated sounds to decay