Producing sentences

This (1976) model distinguishes between 2 major stages of syntactic planning: Functional - word order is not yet explicitly represented. Semantic content of words is specified and assigned to syntactic roles such as subject and object (mixing him and her). Positional - words explicitly ordered (stranding errors). There is a dissociation between syntactic planning and lexial retrieval
1. Garrett argued that content (noun verbs semantic) and function (syntactic) words play v different roles in language production. Content words are selected at the functional level wheras functional words selected at the postional level
  1. Evidence for Garret- diff levels cannot exchange errors, observed that content words almost always only exchange with other content words and same for function words. Robust finding Harley (2008) found this too. Supports the idea that content and function words are from computationally distinct vocanularies
2. Errors can occur within any of the levels ie. word substitution errors occur at the functional level, sound sub at sound level
A stranding error: once slotted in wrong place must sound them, s matches word but not intended word
Agreement errors: plural subject requires plural word, if goes wrong you get error. Typical one is attraction- influence on verb subject agreement cos of proximity of puppies very attracted for verb to agree "the boy with the puppies play in the garden"
1. Occurs more often than chance can tell us about scope of syntactic planning
Serial model of planning: is word by word and errors are influenced by transitional probability ie. puppies play has high transitional properties
Hierarchical models: planning occurs in linguistic units larger than word, errors with more likely to occur within such units than between
1. Garret (1980) proposed that the relevant unit is a clause
  1. Clause basically a sentence, one sentence may contain multiple sentences inside. It consists of a verb plus its arguments. Info in one clause unlikely to lead to error in another clause (HM- 2 function assignment processes). sentence sandwich bread wont affect filling.
2. Frequency of attraction errors depends on what kind of pharses clauses are in the sentence
  1. Bock & Cutting (1992) phrase wouldn't be seperately planned like a clause so if syntactic planning occurs in clause domains there should be fewer attraction errors for the clauses than the phrases
    1. Contrast: serial would predict attraction error as likely in both cases
    2. A clause will always have a verb and a subject, in exp number of nouns were manipulated. More attraction errors for phrases than clauses. supports HM, errors not likely to occur across clauses. Not serial account: agreement error simply depends on the previous words, which is the same whether a clause or phrase
Garretts model says alot about syntactic planning but little about the processes themselves (Bock et al 1982) asked if people can be biased to produce particular constructions - word order in speech is determined by a number of factors that interact (ie. noun, accessible items)
Vigliococco & Nicol (1998) participants had to complete an incomplete sentence, in another exp they had to form a question. both liable to agreement error response. Same number of errors iin the two exp. According to linguistic theory, same underlying structure for both, evidence that syntactic planning involves abstract structures and suggests that abstract syntactic structure is planned seperately from linear position
How much do we plan?
1. Incrementality hp: planning and articulation happen in parallel: can articulate eariler words before fully planning later words or whilst planning later words
  1. Predictions: time to start speaking is unaffected by aspects of later part of sentence but may be affected by aspects of the earlier parts of the sentence
2. Meyer (1996) picture/word interference task. has to name pairs of objects in pic, auditory distractor presented (related i nsound or meaning) . time to strart speaking was slower when distractor was semantically related to either noun. sound only had effect if distractor was related to first noun, only 1st word word form planned in advance. consistent with radical incrementality. start speaking before you know sound of secoond word, lemmas are planned. Also consistent with 2 stage lexicalization only lemma of later words retrieved, not sound. brocas aphasia not able to put subject and verb together
3. Smith & Wheeldon (1999) less time to speak when first phrase is simple compared to complex (even though sentences same complexity overall). lemmas planned in advance if you had to plan subject in advance take longer with more subjects (2) If you planned whole sentence you SHOULD have been equally slow on both sentences
4. Schnur et al (2006) Picture/word interference task. had to describe a picture , pic also included a written word. word could be phonologically related to verb of sentence or unrelated. Phono related made people faster to start speaking. People must have started phono encoding of the verb before starting to say the first phrase (whole sentence first) doesnt fit with incrementality (first then second)
5. Discrepancy between meyer and schnur could be explained if the main verb of a sentence is planned very early . Meyer didn't discuss noun, maybe that first noun of first phrase is planned at the same time as the verb

Media attachments

garrett (image/jpg)

Next up

Producing sentences

Description

Resource summary

Media attachments

Similar

	Created by issy_hinds about 10 years ago

	Copied by issy_hinds about 10 years ago