What phonological deficit? Franck Ramus and Gayanesh Szenkovits.

What phonological deficit?What phonological deficit?Franck Ramus and Gayanesh SzenkovitsFranck Ramus and Gayanesh Szenkovits

Different theories regarding Different theories regarding developmental dyslexiadevelopmental dyslexia

Different theories co-exist in an attempt to capture the Different theories co-exist in an attempt to capture the mechanisms which cause developmental dyslexia- the mechanisms which cause developmental dyslexia- the magnosystem hypothesis, the temporal processing magnosystem hypothesis, the temporal processing impairment and the phonological deficit theory. (Habib impairment and the phonological deficit theory. (Habib 2000)2000)

Here we will focus on the phonological deficit theory and Here we will focus on the phonological deficit theory and try to present it in a different light. try to present it in a different light.

In Ramus own words:‘’ The present paper is, at last, In Ramus own words:‘’ The present paper is, at last, about a first significant attempt to get to grips with the about a first significant attempt to get to grips with the phonological deficit’’phonological deficit’’

Phonological cognitive architecture Phonological cognitive architecture assumedassumed

BaBa

Banana

The 3 main dimensions to the The 3 main dimensions to the phonological deficit rolled into 1phonological deficit rolled into 1

1. Poor phonological awareness 1. Poor phonological awareness (phoneme deletion task - (phoneme deletion task - say ‘say ‘rrock’ without ‘ock’ without ‘rr’ – ‘ock’).’ – ‘ock’).

2. Poor verbal short term memory2. Poor verbal short term memory( poor digit span - ( poor digit span - 1425367 – 1425???).1425367 – 1425???).

3. Slow lexical retrieval 3. Slow lexical retrieval ( rapid automatic naming task - ( rapid automatic naming task - ….. Owl). ….. Owl).

The poor performance of dyslexics in verbal The poor performance of dyslexics in verbal tasks is related to one or more of these tasks is related to one or more of these dimensionsdimensions..

The three dimensions have something in The three dimensions have something in common: They all implicate phonological common: They all implicate phonological representations, each in its own way.representations, each in its own way.

The phonological representations in The phonological representations in developmental dyslexia are developmental dyslexia are

degradeddegraded The most commonly accepted hypothesis regarding the The most commonly accepted hypothesis regarding the

phonological deficit in dyslexia is that the phonological phonological deficit in dyslexia is that the phonological representations are degraded representations are degraded (Fuzzier, or noisier, or are not (Fuzzier, or noisier, or are not sufficiently categorical and preserve to many acoustic or allophonic details).sufficiently categorical and preserve to many acoustic or allophonic details).

Nevertheless, there is much more to phonology than Nevertheless, there is much more to phonology than these three dimensions.these three dimensions.

First, the locus and the nature of the phonological deficit First, the locus and the nature of the phonological deficit must be explicitly defined. Moreover, the vast knowledge must be explicitly defined. Moreover, the vast knowledge on phonology is seldom brought to bear on the on phonology is seldom brought to bear on the phonological deficit hypothesis. Secondly, the phonological deficit hypothesis. Secondly, the phonological deficits predict deficits in heard language phonological deficits predict deficits in heard language ( Ramus 2001).( Ramus 2001).

In search of a locus-experimentsIn search of a locus-experiments

In all the presented experiments, subjects are dyslexic In all the presented experiments, subjects are dyslexic and non dyslexic adults.and non dyslexic adults.

In the first set of experiments, there was an attempt to In the first set of experiments, there was an attempt to find the deficient levels of phonological representations find the deficient levels of phonological representations depicted in the suggested cognitive model.depicted in the suggested cognitive model.

Lexical and sublexical levels were contrasted using words versus non-Lexical and sublexical levels were contrasted using words versus non-words.( ton vs. zan) words.( ton vs. zan)

Input and output pathways were contrasted by comparing repetition (both) Input and output pathways were contrasted by comparing repetition (both) tasks with auditory discrimination task (only input representations).tasks with auditory discrimination task (only input representations).

Results show deficiency in both lexical and sublexical Results show deficiency in both lexical and sublexical levels and both pathways. Furthermore, a substantial levels and both pathways. Furthermore, a substantial impairment in the input pathway. Short term memory impairment in the input pathway. Short term memory load was an important factor. load was an important factor.

Representations vs. working Representations vs. working memory processes-experimentsmemory processes-experiments

Why do dyslexics fail to discriminate and repeat verbal Why do dyslexics fail to discriminate and repeat verbal material, when the short term memory load is material, when the short term memory load is significant?significant?

Could be because their phonological representations are degraded. Could be because their phonological representations are degraded. Alternatively, it could the phonological representations are intact but the Alternatively, it could the phonological representations are intact but the short term memory is limited.short term memory is limited.

An experiment of phonological similarity was run to An experiment of phonological similarity was run to differentiate between the explanations.differentiate between the explanations.

The more phonologically similar a sequence of words are the harder it is to The more phonologically similar a sequence of words are the harder it is to recall the sequence. If the phonological representations of dyslexics are recall the sequence. If the phonological representations of dyslexics are degraded ,there will be a different in the similarity effect compared to degraded ,there will be a different in the similarity effect compared to controls.controls.

Dyslexics show the same similarity effect as controls.Dyslexics show the same similarity effect as controls. Thus, short term limitations are argued for.Thus, short term limitations are argued for.

Universal or hypernative Universal or hypernative phonology-experimentsphonology-experiments

Language specific phonological representations are Language specific phonological representations are acquired very early in life by a child.( rice – lice)acquired very early in life by a child.( rice – lice)

As the developmental dyslexia deficit is congenital, one known hypothesis is As the developmental dyslexia deficit is congenital, one known hypothesis is that dyslexics incompletely acquire the specific phonological category of that dyslexics incompletely acquire the specific phonological category of their language. Thus, their phonology is less specific to any particular their language. Thus, their phonology is less specific to any particular language.( Mody et al 1997)language.( Mody et al 1997)

Thus, one prediction is that dyslexics will be able to Thus, one prediction is that dyslexics will be able to perceive and perhaps produce foreign speech sounds.perceive and perhaps produce foreign speech sounds.

An experiment was performed using dyslexic French An experiment was performed using dyslexic French speakers, who were tested in discrimination and speakers, who were tested in discrimination and repetition of foreign Korean voicing categories. repetition of foreign Korean voicing categories.

Again there was an assessment of short term memory load, by Again there was an assessment of short term memory load, by contrasting conditions of one foreign nonwords compared to contrasting conditions of one foreign nonwords compared to conditions with two or three nonwords.conditions with two or three nonwords.

Universal or hypernative Universal or hypernative phonology-continuedphonology-continued

The results showed that dyslexic were similar in The results showed that dyslexic were similar in performance to controls in single nonwords, but not in performance to controls in single nonwords, but not in the two or three non word sequence.the two or three non word sequence.

The results suggest that dyslexics are equally unable to The results suggest that dyslexics are equally unable to represent foreign speech sounds as controls. Group represent foreign speech sounds as controls. Group differences appear only when short term memory load differences appear only when short term memory load increases.increases.

Thus this is compatible to the hypothesis that Thus this is compatible to the hypothesis that phonological representations of dyslexics are intact and phonological representations of dyslexics are intact and short term memory processes to operating on them short term memory processes to operating on them seem to be impaired. seem to be impaired.

Phonological grammar-experimentsPhonological grammar-experiments

This is the whole host of rule like processes that apply in This is the whole host of rule like processes that apply in speech production when phonological lexical items and speech production when phonological lexical items and assembled to make phrases. These phenomena are assembled to make phrases. These phenomena are related also to speech perception, as an adaptation to related also to speech perception, as an adaptation to native phonological structure.native phonological structure.

Most of these differences are language specific and Most of these differences are language specific and learnt during acquisition of language. Do dyslexic learnt during acquisition of language. Do dyslexic children learn them as well as controls?children learn them as well as controls?

In one experiment, assimilations by phonotactic In one experiment, assimilations by phonotactic constraints were investigated.constraints were investigated.

Phonological grammar- continuedPhonological grammar- continued

In English, clusters like [dl] or [tl] can never occur in the begging of a word. In English, clusters like [dl] or [tl] can never occur in the begging of a word. The consequence is when subjects hear a non word such as [dla] or [tla], The consequence is when subjects hear a non word such as [dla] or [tla], they most often assimilate it to the closest legal cluster( [gla] or [kla])they most often assimilate it to the closest legal cluster( [gla] or [kla])

In such a discrimination task it was found that dyslexics In such a discrimination task it was found that dyslexics listeners fell victims to this perceptual illusion as much as listeners fell victims to this perceptual illusion as much as controls, hearing [gla] instead of [dla]. controls, hearing [gla] instead of [dla].

Thus, their speech perception is constrained by the Thus, their speech perception is constrained by the same phonotactic of their native language as controls. same phonotactic of their native language as controls.

In conclusion, the aspects of phonological grammar that In conclusion, the aspects of phonological grammar that was investigated seem perfectly intact.was investigated seem perfectly intact.

Unconscious speech processing-Unconscious speech processing-experiments experiments

A problem inA problem in psycholinguistics is that tasks require psycholinguistics is that tasks require attention to the stimuli, and introspection, which may blur attention to the stimuli, and introspection, which may blur the effects observed, particularly when the population the effects observed, particularly when the population testedtested have problems with phonological awareness. have problems with phonological awareness.

Auditory masked priming have shown phonological lexical representations Auditory masked priming have shown phonological lexical representations processed unconsciously (hospital- nurse).processed unconsciously (hospital- nurse).

A new question can be how efficient are unconscious A new question can be how efficient are unconscious lexical access to phonological representations in lexical access to phonological representations in dyslexics? dyslexics?

A masked priming tasks in dyslexics replicated the results of controls.A masked priming tasks in dyslexics replicated the results of controls.

The results support the notion that phonological The results support the notion that phonological representation in dyslexics are not degraded and representation in dyslexics are not degraded and dyslexics rely on them and not on the acoustical level.dyslexics rely on them and not on the acoustical level.

A new hypothesisA new hypothesis Findings converge to a conclusion: the phonological Findings converge to a conclusion: the phonological

representation of dyslexics are normal.representation of dyslexics are normal. Although, subjects do have a phonological deficit Although, subjects do have a phonological deficit

measured by traditional tests: nonword repetition and measured by traditional tests: nonword repetition and rapid naming tasks.rapid naming tasks.

It seems that short term memory load is paramount to It seems that short term memory load is paramount to the deficiency of dyslexics in different tasks studied: the deficiency of dyslexics in different tasks studied: such as digit span or phonological awareness tasks, such as digit span or phonological awareness tasks, which require holding segmented phonological units in which require holding segmented phonological units in short term memory and conscious access to those short term memory and conscious access to those representations.representations.

An alternative explanation for the phonological task An alternative explanation for the phonological task deficits is deficits is phonological access. phonological access.

A new hypothesis-continuedA new hypothesis-continued Phonological accessPhonological access:: Verbal short term memory requires access to Verbal short term memory requires access to

phonological representations to copy them into buffers, phonological representations to copy them into buffers, then access for them retrieval, as well as to access input then access for them retrieval, as well as to access input representation to copy them to output representation. representation to copy them to output representation.

Phonological awareness tasks involve conscious access Phonological awareness tasks involve conscious access to phonological representations, which may place special to phonological representations, which may place special demands on the system.demands on the system.

Rapid naming tasks require fast access to lexical Rapid naming tasks require fast access to lexical representations.representations.

Thus, dyslexics seem to fail in tasks that are particularly Thus, dyslexics seem to fail in tasks that are particularly demanding in phonological access.demanding in phonological access.

DiscussionDiscussion Are the subjects dyslexic enough? Yes, the subjects had Are the subjects dyslexic enough? Yes, the subjects had

a history of reading disability. Furthermore, their a history of reading disability. Furthermore, their performance in phonological tasks (digit span, rapid performance in phonological tasks (digit span, rapid naming) did not match controls.naming) did not match controls.

Is working with adult subjects enough? It can be argued Is working with adult subjects enough? It can be argued that children may have deficient phonologicalthat children may have deficient phonological representations, but these are recovered in adulthood. A representations, but these are recovered in adulthood. A replication study with children is needed. This criticism replication study with children is needed. This criticism steel don’t explain why representations have recovered, steel don’t explain why representations have recovered, and additional load on phonological access have not.and additional load on phonological access have not.

Other past articles have been coarse in defining the Other past articles have been coarse in defining the phonological deficit as shown here.phonological deficit as shown here.

Discussion- continuedDiscussion- continued Its seems that deficient phonological access can be Its seems that deficient phonological access can be

viewed as an executive function deficiency. This is a viewed as an executive function deficiency. This is a specific type of executive function, specific in processes specific type of executive function, specific in processes and in modality.and in modality.

In relation to sensory theories, others have found that In relation to sensory theories, others have found that dyslexic never seem to have a problems in perceiving a dyslexic never seem to have a problems in perceiving a certain kind of stimulus, be it auditory or visual: rather certain kind of stimulus, be it auditory or visual: rather the difficulty is related to task requirements, and the difficulty is related to task requirements, and specifically to tasks that demand short term memory, or specifically to tasks that demand short term memory, or when the stimulus is noisy.when the stimulus is noisy.

Dyslexic can be viewed as people with dysfunctional Dyslexic can be viewed as people with dysfunctional sensory presentation access: most just to phonology and sensory presentation access: most just to phonology and others also to other domains: visual or auditory. others also to other domains: visual or auditory.

But…But…

Dyslexia: a deficit in visuo-spatial attention, not in Dyslexia: a deficit in visuo-spatial attention, not in phonological processing (vidyagaskar et al. 2010)phonological processing (vidyagaskar et al. 2010)

Speech Perception Abilities of Adults With Dyslexia: Is Speech Perception Abilities of Adults With Dyslexia: Is There Any Evidence for a True Deficit?( HazanThere Any Evidence for a True Deficit?( Hazanet al. 2009)et al. 2009)

חזרה לראשי