phonological processing and reading ability

8/16/2019 Phonological Processing and Reading Ability

http://slidepdf.com/reader/full/phonological-processing-and-reading-ability 1/13

SHORT-TERM MEMORY, PHONOLOGICAL PROCESSING AND

READING

ABILITY*

Susan

Bradyt

Abstract. Verbal

short-term memory

deficits are a

common

chamc

te1,istic

of

children with reading problems an d

ma y

markedly increase

the

difficulty of

learning

to read. Previous

work

suggests that the basis

of

the short-

term memory

deficit

may

involve limitations in phonetic

coding.

In

the present paper, a

series

of

experiments are reviewed

that

examined

the role

of

phonological processes in

sho rt-term m em

ory. First a developmental

study

is descr ibed in which a significant

1 c ationship was found between phonetic processes an d verbal

memory

span

bu t

not between phonetic processes an d nonvC1 bal memory. Sec-

ond additional studies are reviewed

that

collectively found

that

chil-

dren

with

1 eading

problems

a1 e

less

accurate

at phonetic

encoding than

are good readers , a nd t hat pe1,/ormance o n p ho ne tic processing cor1 e-

sp on ds w it h ve1,bal

mem01 y span.

No reading group differences

we1 e

obtained

on

nonverbal

perception 1 mem01 y tasks. The se f indi ngs

suggest that both developmental and individual differences in verbal

memory

span

are related

to

the efficiency of phonological processes.

Practical implications

of

current cognitive research are discussed.

In

the

last 15 years exciting progress has

been made

in understanding the factors in learning

to read. A strong case can now be given for the importance of linguistic skills in reading ability

e.g., Liberman, 1983; Mann, 1984;

Perfetti,

1985; Stanovich, 1985; Vellutino, 1979 .

One

of

the

most robust findings in this area is

that

there is a relationship

between

level of reading abili ty

and performance

on

short-term memory STM

tasks,

with better

reading

skill being associated

with super ior STM. This has been observed in adults Baddeley, Thomson, Buchanan, 97 5;

Daneman Carpenter,

1980 ,

and more

extensively in children see Jorm, 1983, and Stanovich,

1982, for reviews , Deficits in STM for individuals with reading problems have been demonstrated

with digit span measures, letter strings, sentence tasks, and recall for

pictures

of familiar objects.

Interestingly,

th e

memory problem for poor

readers

generally has

been

found to be specific

to

linguistic material or to other mater ial that is easy to represent linguistically. When instead

the memory task consists of nonsense figures, photographs of strangers, or symbols from

an

unfamiliar writing

system, recall

has

not been found to

be

closely

related

to

reading

ability Katz,

Annals

of

Dyslexia.

1986, 36, 138-1.53.

Also, University

of

Rhode

Island

Abwwledgment.

The research

was supported by NIH

Grant

HD-01994 to Haskins Laborat.ories.

The author gratefully acknowledges the comments of Donald Shankweiler and Anne Fowler on

an earlier

draft

= HASKINS LABORATORIES:

Status

Report on Speech Research SR-88

1086

=

145



Short-te7 m Memory Processing and Reading Ability

147

Analysis of

th e

errors

good

an d p oor re ade rs m ak e

on

s hort -te rm me mory tasks

provides

additional evidence that poor re ade rs have less efficient phonetic coding Brady, Mann, Schmidt,

1985;

Brady, Shankweiler, Mann,

1983 . This

interpretation is supported by

th e observation

that both reading groups ma ke e rrors

that

indicate u se o f a phonetic coding strategy. For example,

like g oo d r ead ers, p o or r ea de rs

ar e

apt t o m ak e t ra ns po si ti on err or s, r ecombi ni ng t he p ho neti c

information from a dj ac en t i te ms in th e

l is t, especially

when items h av e p ho ne ti c f ea tu re s

in

co mmo n. W hi le

th e

nature of

errors

suggests that

both

r eadi ng g ro up s ar e emp lo yi ng a phonetic

coding s trat egy, poor

readers low efficiency in

phonetic coding

is revealed

by

a

greater

frequency

of errors.

In

sum, past

re se arch has shown

that p oo r r ea de rs

have a specific difficult.y

with

verbal

memory tasks; they show reduced sensitivity to rh yme ; verbal ST M deficient regardless of

whether

verbal or

auditory

presentation is used;

and poor

readers

produce

a

greater

frequency

of

phonetic errors

of

transposition.

This

evidence converges

to

suggest that poor

readers have a

language problem in th e phonological domain that is no t restricted to r ead in g t asks.

Processing in Short Term emory

In

recent

years my

colleagues

an d

I

have be en i nte res te d

in

t ry in g t o

find

ou t

more

about

th e nature of

th e

phonetic coding problem of poor readers. Ou r aim, as stated

earlier,

has

been

to

u nd er st an d b et te r t he basis

of

th e short-term memory

deficits

common

for

these

individuals.

If we consider

th e

requirements of an STM task, we can, at least

conceptually,

consider different

processes that

ar e

involved, anyone of whic h might be th e s our ce o f a defici t. F or examp le,

in

order

for

something to b e r em em be re d,

it must f irst have

been

perceived.

The stimulus,

w he th er i t

is seen or heard,

ha s

to be perceived

an d represented

phonetically. So a deficit might

arise from

problems in in it ia l

phonetic

encoding. In m akin g this assumption,

it

perhaps

n ec es sa ry t o u nd er sc or e

th e

evidence

that humans

have a ubiquitous

tendency

to use

phonetic

codes

whenever

possible. Even

when alternative strategies m ig ht h av e s ee me d

likely, such as in

reading

by

congenitally

deaf

i nd iv id uals o r for

readers

of a l og og raph ic scri pt such as Chinese,

evideu.:e

of

phonetic

coding

h as b ee n o bt ai ne d

[Hanson,

Liberman,

Shankweiler,

1984;

Tzeng,

Hung, Wang, 1977] . Or , once a phonetic representation is

c re at ed , i t h as

t o be r eta in ed ,

an d

a deficit

might stem

from a

limitation

in r ehearsin g o r

retrieving th e information

in memory.

We

have

viewed

short-term memory,

or

working memory,

as a

l imited capacity

system

Bad

deley Hitch, 1974

that

briefly stores

information

in a phonetic code. Since

th e am ou nt of

information

it

c an h an dl e

is

limited,

th e

functional capacity

of

STM can b e e xp ec te d

to affect

performance

on a

number

of cognitive

tasks

such

as

l isteni ng co mp reh en si on , d ecod in g w ri tt en

language,

an d

reading comprehension.

Th e

role of

memory

in re ading has be en discussed in

detail

elsewhere

[e.g., see L ib e rm a n e t 801 1977; or Perfetti Lesgold, 1977] . Furt.her, within

ST M

th e

resources or

attention

required by

aspects

of processing such as encoding, storage,

an d

retrieval)

ar e

a ss um ed t o

become

more

automatic

with

experience

and to

require

less resource

allocation LaBerge Samuels, 1974; Perfetti, 1985 . These characteristics,

widely

supported

in

th e cognitive l it er at ur e, a re t ak en to reflect normal constraints on

human

information processing.

Within this framework, we have explored th e question of w het her t he

poor

reader s phonetic

coding deficit in

m em or y m ig ht

arise from less efficient encoding processes. Part. of t he i mp et us

for

this a pp r oa ch c am e from experimental

evidence w ith a du lt s

that perceptual

difficulty

ca n

d et er m in e m em o ry p erform anc e. In a

study

by Rabbitt 1968 , m em or y s pa n was measured



148

r dy

for lists of digits that were presented clearly

o r d eg ra de d

by th e addition of noise. I n a dd it io n,

th e a bil ity o f subjects

to

identify th e individual d igi ts w as determined in a

condition

in which

t he s ub je ct s r ep ea te d t he

i ndi vidual i t ems . Noise levels

that c au se d n o

effect

on

perception of

th e

single

items

significantly

impaired

recall of strings of

th e same

stimuli.

Therefore,

adding

noise,

an d

increasing

th e

perceptual difficulty, reduced memory span.

Rabbitt

argued that

in

a

l imit ed c a pa c it y

system,

if perception or encoding) is difficult, fewer resources will be available

for

memory.

R es ea rc h o n i nd iv id ua l differences in adults ha s f ur th e r s u pp o rt e d a c onne ct ion bet we en

phonetic processing

an d

memory

span.

is known that

adults

memory span

can

b e p re di ct ed

on

th e

basis of

the num be r

o f wor ds that a subject

can read

in approximately t wo s ec on ds

Baddel e y e t al., 197.5 .

In th e same

vein, Hoosian 1982)

reported

a

negative correlation between

naming

speed for digits

an d

digit span.

In both

s tudie s , phone ti c processes involved in

encoding

and ar ticulating a response were statistically

related

to m em ory performance.

These results

ar e

c ompa ti ble w it h

th e

view that perception

an d memory

processes in ST M

share

limited resources.

Developmental

Factors

S h o r t- Te r m

M e m o r y

Similar arguments have

been

m ade to explain

th e

c ommonl y obse rve d deve lopme nt al dif-

ferences in m em or y s pa n. is well known tha t measures of memory span sho w co nsid erabl e

improvement

f ro m e ar ly c hi ld ho od

to

adulthood.

is estimated

that

span

approximately triples

from an a ve ra ge of slightly more than

tw o items

at ag e

tw o

to

an aver ag e o f

nearly seven

items

at

aduIt.hood Dempster, 1981). Related

to

t he a p pr oa c h that

Rabbitt s

work suggests,

on e

hypoth

esis

about

th e c au se of th is

increase

is that, with

experience,

th e

component

processes in STM

come t

operate

more

efficiently,

reducing th e resource demands an d resulting

in a developmental

i nc re as e i n func ti ona l storage space.

P er ti ne nt t o o ur i nt ere st s, encoding

processes

ar e p re su me d t o increase

in

efficiency, con-

t ri bu ti ng t o a concomitant

improvement in

memory

capacity. These

developmental

c ha ng es i n

ST M

a re a ss um ed

to occur a t a n au tom atic ,

nonstrategic

level. We were interested in pursu

ing this

explanation

as will be des cribed s hort ly) be ca use looking at

th e

basis

of developmental

improvements offers a wa y to examine

th e

role

of

phonetic processes in

STM.

Tw o recent experiments

have

produced

compelling developmental evidence that

perceptual

processes

may be

importantly related t o m em ory

capacity.

T es ti ng c hi ld re n f ro m three to six

2 A second c at eg or y o f

explanation

of developmental i ncreases i n

memory

focuses on s t ra t egi c

o r m ne mo ni c variables that ar e under conscious direction. T h es e i nc lu de r eh ea rs al , c hu nk in g,

organization,

an d retrieval

strategies. According

to

this view, a s c hil dre n ge t

older

t he y b ec om e

increasingly aware that

particular strategies

will ai d recall,

and

increasingly able

to

employ

th e

appropriate

procedure.

is clearly

th e

case

that

ch il dr en d o

improve

in

th e

use

of

t.hese

metacog

nitive skills Chi, 1976, reviews

this

evidence). However, it doesn t look as though

this

is

th e

sole basis for

th e

ag e

differences

in

span. For example, Samuel 1978 equated children s

us e

of t.he organization control

process

by t em po ra ll y g ro u pi ng i te m s

in

recall lists . This procedure

improved recal l for all

th e

ag e g ro up s t es te d r an gi ng in ag e from 6

to

29 , and

thereby

preserved

th e ag e differences. Similar results have

been obtained

by H

uttenlocher an d

Burke 1976

and

by Engle

an d

Marshall 1983 . So strategic variables, while showing

increments

with age, do not

seem

to

account

completely

for developmental differences in span.



Short term

Memory

Processing

and

Reading Ability

149

years of age, all of whom are unlikely to use mnemonic strategies, Case, Kurland, and Goldberg

1982) observed a

strong

correlation between how fast

th e children

could

repeat words

and

the

size

of

their

memory span: th e

older

chi ldren, who could repeat faster, remembered more

words.

The

authors argued that th e

basic encoding

and

retrieval

operations

in STM show a development.al

increase in operational efficiency, resulting in more functional storage space.

In

a convincing

test

of this, Case

and

his colleagues equated six-year-olds and

adults

on speed of counting by

manipulating word familiarity (adults were forced to count in an unfamiliar language). Corre-

spondingly, memory span with these stimuli was also

demonstrated

to be

no different for t.he two

age groups. In a second

study, it has

likewise

been reported that

alt.hough younger

children

recall

less, t.he same l inear function relates

speaking

rat.e to

short-term

memory for subjects ranging in

age from four years old to adulthood

(Hulme,

Thomson,

Muir, Lawrence, 1984). The results of

t.hese studies indicate that the buffer component of working memory is phonologically organized,

and

that developmental

increases in memory

span

ar e related to

th e

efficiency of phonological

processes.

Current

Research

Reiterating the

main

points of

t he paper thus

far,

there are three interes ting

conclusions

to

be drawn.

First,

poor readers have a

short-term

memory deficit

that

is specific to tasks requiring

phonetic coding. Second , adults performance in

verbal

short-term

memory is

r el at ed to

their

perceptual abilities. Third, there is some evidence that

developmental

increases in STM may be

the by-produc t of increases in th e efficiency of phonetic encoding. Influenced by

these

findings,

ou r

experimental

work

has proceeded in

two

directions: 1) we focused on the developmental link

between

phonetic

processes and verbal

STM

to determine

more

precisely how closely they are

connected with

one another, and 2) we examined the role of perceptual processes in the short-term

memory

deficits

of

poor

readers.

In

th e

remainder

of the paper, th e

results

of

these

experiments

will be

summarized.

Development

of

Phonetic

k lls

and

Memory

Span

The

developmental relationship between phonetic

processes

and memory

span

was examined

II I

a

study with 4 ~ y e a r o l d s 6 ~ y e a r o l d s and 8 ~ y e a r o l d s (Merlo

Brady, in

preparation).

A number

of ta sks were used to

evaluate

the

correspondence of different

aspects

of phonetic

and memory processes. The phonetic tasks included word repet it ion with monosyllabic and

multisyllabic

words,

and repeated

production

of

disyllabic tongue twisters e.g.,

sishi/).

The

responses were scored for reaction time

and

for

the

number of correct productions. In this way

both

speed and accuracy of phonetic processing were measured, allowing

separate

analyses for

these

two

elements

of phonetic

efficiency.

In th e word

repet it ion tasks , response

time

would

reflect both encoding and

articulation.

We were interested in isolating these

components

of word

repetition speed. To that end, a control task was included in which the children would

hear

a

nonspeech

tone to which they

were

to respond

wit.h a prespecified

word

(either

monosyllabic

catl

or multisyllabic Ibanana/). In

this

task, no speech signal has to be encoded; hence it provides

an

indication

of th e

output

or articulatory process.

The goal in this experiment was to examine the

developmental

link between the phonetic

processes

of

encoding and

articulation, and

verbal

short-term memory

capacity. To assess verbal

STM, word strings of familiar concrete

nouns

were given. A

nonverbal

STM task , the Corsi block

t.est (Corsi, 1972), was included for comparison

to

help

ascertain

whether increases in phonetic



5

Brady

processes were related to increases in

verbal

memory. In this test nine identical black blocks

are

mounted in a scattered arrangement on a

wooden

surface. The experimenter points, one by one,

to a series of blocks.

The

subject

mus t then repeat

this sequential

pattern.

All

of the measurements

on th e

phonetic

and

memory tasks

showed improvement

with

age.

The linear

correlations between verbal STM

VSTM and each of

th e

measures of

phonetic

pro

cess ing skill were all

found

to be s ignificant : as

VSTM

increased, verbal

reaction and response

time decreased,

as

did th e

number

of errors.

was

necessary to

check that

th e

var iation in

age

or

some cognitive factor that improves with age was

not

a spurious basis for the

reported

significant relationships among

VSTM

and the phonetic

processing

measures. To assess this,

a series

of

correlational analyses were conducted with the effects of age controlled statistically.

The resulting partial correlation coefficients showed that when th e effects of age were

removed,

a statistically significant l inear relationship continued to exist between VSTM and phonetic pro

cessing.

The highest correlations

were

obtained on th e accuracy measures

for

th e

more

demanding

t asks: mul ti sy llab ic words and tongue twisters.

On

the speed

of

processing measures there was

an interesting split.

The

react ion t ime scores were significantly

corre la ted with verbal memory

span

after age was partialled

out,

with two exceptions.

These

were th e

control

tasks in which

th e

subjects

said

a

word

as

soon as they

heard

a

tone.

This

lack

of relationship

demonstrates

that

th e process

tapped

by these measures speaking rate was not closely related to memory. In any

case

from

th e overall results, it se ems reasonable to conclude that th e ability to

process

phonetic

information efficiently plays

at least

a

partial

role in verbal short-term memory functioning.

The importance of th is connection between phonet ic

processing

a nd VSTM

is shown

by

contrasting these results

with

th e

correlations

with

nonverbal memory

the

Corsi

task . When

age was partialled

out,

verbal and nonverbal memory

spans

were no t significantly

correlated,

nor

was

nonverbal

memory

performance

closely related to the phonetic processing

measures.

Only

one

variable, monosyllabic react ion t ime,

produced

a significant

correlation

with nonverbal

memory

w ith this

number

of comparisons that could be a chance resul t . Certa inly

th e general picture

for

th e relationship

of phonetic skills to nonverbal

memory

was very different .

Demonstration of th e close ties between performance on th e phonetic tasks and verbal STM

span strengthens the evidence that phonetic encoding processes in memory contribute to th e basis

of individual

differences

in memory capacity. These results both corroborate and extend previous

research

findings Case

et

a1. 1982;

Hulme e t

a1 1984 that showed evidence for developmental

improvements in

word repetition

speed and in verba l STM.

In

th e Merlo

and Brady s tudy

in

preparation , these f indings were

confirmed

with a

variety

of phonetic tasks , and

there

was also

evidence for age-related increases in accuracy. Thus older children had longer memory spans

and were able to

repeat

verbal stimuli more accurately and rapidly than younger children. The

importance of

th e

relationship

between

the phonetic variables

and

memory was underscored

by

demonstrating that relationship even when age was removed as a factor, and showing no such l ink

with nonverbal

memory.

Thus

when

we

examine

short-term

memory

developmentally,

there

is

strong

evidence that

th e

efficiency of

phonetic processes

is

centrally

related to

verbal short-term

memory span.

Phonetic Processes, Memory

Span,

and

Reading

Ability

As

noted in th e

section on background

evidence, th e resul ts of

several

studies indicate

that

the problems poor readers demonstrate

in

verbal STM

are

related t o the ir

use

of phonetic

coding.



Short term Mem01 y

Processing

and

Reading Ability

In light of th e evidence of th e developmental relationship between phonetic

processes

an d

verhalmemory,

we c an ask more

analytic

questions

about

th e phonetic basis of th e

poor re aders

memory deficit.

O ne s uc h q ue st io n

is whether

r ea di ng g ro up

differences in

memory

span arise

from

th e

pe rc e pt ual re qui re ment s of STM: do good and p oo r r ea de rs

differ in

th e

efficiency

of

phonetic encoding? 3

Effects of noise on phonetic

encoding A few

y ea rs a go

we

b eg an t o investigate th e phonetic

skills

of

children

varying

in

r ea di ng p er fo rm an ce . I n B ra dy e t

al. 1983 ,

third-grade

children

v ar yi ng i n

reading

ability

were tested

on a word repetition t ask con si st in g o f

monosyllabic

nouns

chosen to

control for

phonetic and

syllabic

construction and

for word frequency. Since

th e

percep

tual

abilities of poor

readers

ar e

within

normal limits, we

anticipated that detecting

a phonological

deficit

i n p oo r r ea de rs

with a

perceptual t as k m ig ht

require an especially demanding t as k. T he re

fore, th e words were recorded in

tw o

noise conditions to vary

th e

perceptual difficulty. In one

condition,

th e w or ds w er e presented

clearly; in

th e

second,

amplitude-matched

noise was

added

to

each

stimulus. On the

no-noise task,

b oth the good and

t he p oo r readers

ha d

nearly perfect.

performance. The

addition

of noise made it harder

for all

th e

s ub je ct s t o identify th e

words, but

reduced

th e

poor readers accuracy significantly more than

that of th e

g oo d r ead er s.

T he s tu dy showed that the perception

skills

o f p oo r readers ar e

less effective,

b ut t ha t

this

difference is

only

observable

when the task

is difficult. To

d et er mi ne w he th er p oo r

readers have

a general

problem

with

auditory perception,

these same

subjects

were tested o n p ercept io n

of

environmental

sounds

e.g.,

human

activities [knocking on a door]; animal noises [frogs croaking];

mechanical

sounds

[a

ca r

starting

up

an d

driving

off] .

O nc e a ga in

th e

stimuli

were

presented

against

a

background of

silence

and

in white ) noise. As was f ou nd w ith

th e speech

stimuli,

pe rforma nc e on e nvi ronme nt a l s ounds was hig h i n

th e

noise-free condition

and deteriorated

with

noise. However,

th e reading

groups

di d

no t differ significantly f ro m each

other

on

either

condition.

anything, th e p oo r r ea de rs identified th e sounds-in-noise s ome w ha t more

accurately

than their

better-reading

peers. This

pattern of

r esul ts sug gests

that th e

difficulty poor

readers

manifest

i n p er ceiv in g speech- in -n oi se is

not the

consequence

of

generally deficient a udi t ory pe rc e pt ua l

ability,

but

is

related

specifically

to

th e

processing requirements

for

speech.

Thus,

paralleling th e

m e mo ry r es ul ts , p o or r ea de rs

have a

problem

specific

t o l an gu ag e that entails

phonetic coding.

Effects of

phonological

difficulty on phonetic

encoding We

hypothesized

that

poor readers

may have a

problem

forming a

phonetic representation

in

short-term

memory,

w het her t he y

ar e

establishing

a speech l abel for a visual

stimulus a letter

or a word or a

nameable

picture),

or

w hether they are listening to

a

spoken

word.

Given that

however

language

is

perceived

it

is

necessary to

encode th e

material, we

conjectured that the

deficiency

of

poor

readers

in process

ing phonological

structures

is a

constant

feature. Further, this p ro bl em m ay limit

th e

resources

in memory and impair

recall on

memory tasks.

We

speculated that th e p oo r re ad ers encoding

3

As

in

th e

developmental literature,

researchers have

also

asked

whether

differences in

reading

skill

might be related to

differences

in th e

use

of

cognitive

strategies

i .e ., conscious

control

processes

such

as rehearsal.)

As Stanovich 198.5

h as n ot ed

in a review see also

Jorm,

1983 ,

there is evidence

that

poor

r ea de rs a re

less likely to employ cognitive

strategies

deliberately

that

e nha nce me mory pe rforma nc e,

an d

this must b e a ck no wl ed ge d a s a factor in poor

readers

inferior

recall. However,

t he re a re

also

studies

demonstrating

that

when

good an d poor readers

ar e equated

on the

use

of particular strategies, differences in

performance

level ar e still present

Cermak,

Goldberg,

Cermak

Drake, 1980;

Cohen

Netley, 1981;

Torgesen Houck,

1980 .



52

r dy

problem for

speech

was present. for all

language

t.asks but.

w as d iscern ib le o nl y

if

th e

t.ask

were

sufficiently difficult.

To test

t.his line

of

reasoning,

it was necessary to

de te rmine w het her differ-

e nc es i n p er ce pt io n c an b e d em on st ra t. ed even

under

clear list.ening condit.ions. In t.he previous

st.udy with

both

groups

at near

ceiling performance levels on

th e

noise free condition, t.he

task

m ay not

have been sensitive

enough.

reading

group

differences

in percept.ual processing

efficiency

a re p re se nt

for

clear

list.ening

let s c on si de r h ow

such

differences might occur:

poor

r ea de rs m ig ht

be

slower at

encoding

a

phonetic item

a nd/or

th e q ua li ty o f th e

phonetic

representation

might

be

less fully

accurate.

Wit.h

relat.ively

e as y p ho ne ti c material and no

t.ime

limitations, poor readers

could

possibly

perform

a de qu at el y w it h e it he r or both o f t he se

processing

limitations.

With

these

quest.ions in

mind, Brady,

Poggie

an d

Merlo,

1986

examined third-grade good

an d

poor r ea de rs o n

a

speech

perception

task designed to discover

whether

reading

group

differ-

ences a re e vide nt , under s om e c ir cu m st an ce s, w he n s ti mu li a re

presented

clearly. Three kinds of

stimuli were presented:

monosyllabic

w or ds, mul ti sy ll ab ic w or ds,

and

pseudowords.

In

this way

t.he phonological

d em a nd s o f th e task

were va ri ed i n c as e

monosyllabic

words previously

tested)

were

no t

sufficiently difficult

to process to

reveal

potential group

differences.

Therefore the length

of

t.he

stimuli

wa s

i ncre as ed i n

th e

multisyllabic

condition,

a nd t he

familiarity

was dec re as ed by

using

novel

constructions that

.could

occur

in English) in

th e

pseudoword

condition.

Both of

t.hese

ar e

known

to increase processing demands.

Reaction

time

measures were collected

t o

assess

processing

speed

a nd th e responses

were scored for accuracy.

When t he p er ce p tu a l

skills

of good an d poor

readers were examined, we

found that the

poor

readers were significantly less

accurate

on

the m ore

difficult multisyllabic

and

pseudoword

s tim ul i. W hi le

there were group

differences

in

accuracy,

no

difference

b et we en g oo d

an d

poor

readers was obtained

for

speed of re spondi ng. The se findings

suggest

that

th e

critical

differences in

phonological

processing

be tw e en good

an d poor

r ea de rs a re r el at ed to

th e

a c curac y of formula t ing

phonetic representations,

no t

wit.h

th e

rate o f p ro cessi ng

them. This

re sult c ompl eme nt s

and

clarifies t.he

earlier finding

of

inferior speech

percept.ion

by

poor

readers

wit.h

speech-in-noise.

In

sum,

it seems when th e speech items a re s om ew ha t

difficult

to perceive

for

a ny r ea so n,

either

b ec au se o f th e signal-to-noise ratio, or

th e

phonological complexity or

novelty

children with

r ea di ng p ro bl em s d o significantly

less well.

Thus when

th e

task

is

demanding,

th e

lesser

ability

of

poor

readers t o e xt ra ct

an d process th e

phone t ic i nformat i on from

th e

s t imulus i t ems bec ome s

apparent.

Phonetic

encoding nd ST

Ou r

motive for looking at percept.ual skills was t o

address

whet.her

th e

poor r ea de rs h av e e nc od in g deficits that

ma y

c o nt ri bu te t o

their

characteristic lim

it.at.ion in

verbal

STM. Having demonst.rat.ed phonological deficits in

poor r ea de rs o n

percept.ual

t.asks

th e

next step was to

see ho w closely

perception

and

memory

performance

correspond

for

children

wh o

va ry in

reading

ab ili ty. We must.

return

at

this

point

to

th e

developmental

study

by

Merlo

an d

B ra dy to

which we

referred earlier.

As

part of

that

study, t wo additional groups

of 8t-year-olds were t es te d: g oo d r ea de rs and p o or r ea de rs .

Recall

that

in tha t

study children

were t.ested

on

both

phonetic and memory

t.asks.

Th e

p ho ne ti c t as ks

were

w ord re pe t it i on

for

monosyllabic

and

multisyllabic words, repetition

of

t ongue t wi st ers, and words s po ke n w he n

a

nonspeech tone was

heard.

Speed

an d

accuracy measures were

collected.

The

memory t as ks w er e

word st.rings

a nd the

nonverbal Corsi test.



Shod term Memory P1 Ocessing and

Reading

Ability

53

As expected, t.he reading groups differed significantly on verbal memory,

bu t

did not differ on

nonverbal STM. On

t.he

phonetic tasks , the results

replicat.ed

th e

pattern

in t.he st .udy

described

just. prior

t o

this Brady

et

a1.,

1986 .

Poor

readers

were significantly less accurate in phonetic

processes, and this was

again

apparent. on t.he more demanding t.asks. Once again the reading

groups did

not. differ

on speed of processing. Therefore

we have a consist.ent.

result t hat

poor

readers,

t.hough not. slower, have less accurate

phonetic coding, and

verbal

short-term

memory

deficit.s.

is

not.eworthy t.hat significant

correlations

were

obtained

for

the

reading

groups

between

t he accuracy

scores on

the more demanding measures of phonetic

skill

and the verbal STM

performance: multisyllabic errors and VSTM, = - ..

52;

tongue twister errors and VSTM, = -

.40.

These correlations imply that.

th e

reading group differences in verbal short .- term memory are

associat.ed wit.h differences

in phonetic encoding processes common

t o

percept.ion

and memory

t.asks.

Concluding

Remarks

To summarize, in this paper we referred t o ext.ensive evidence t.hat poor readers have verbal

short.-term memory

deficits.

I n trying

to invest.igate t.he basis

of

t.he

memory

deficit, we have

adopted

a

cognitive framework

for viewing

short.-term

memory

as

a

limited

capacit.y syst.em.

Wit .h in t .hat syst.em, i f initial

encoding

is less efficient,

the u lt imate

resources for

memory and

comprehension

will be

reduced. In the

research present.ed

here,

t.his

approach

was

supported

in

a

developmental study showing

a close link

between th e

efficiency

of phonetic

skills

and

t.he

available

storage

capacit.y in

short-term memory.

Turning

t o

children with

reading

problems, we

t.hen

demonstrated

t.hat. poor readers a re relatively less efficient at

phonetic

processes as seen in

the accuracy measures , and that this performance corresponds with their reduced memory space

as seen in

the

correlations

between accuracy and memory span .

The

identification of less accurat.e

phonetic encoding

by chidren who ar e poo r readers pro

vides a

parsimonious account of their

difficulties

on

a

number of language

tasks involving

memory

and/

or perception,

including

sentence comprehension

see Shankweiler

Crain,

1986, for elab

orat.ion

of

t.his

idea .

Furt.her,

th e

present.

research

adds

t o

t.he

burgeoning

evidence

that

poor

readers have

difficulty in

the phonological domain. The problems in phonetic processing, here

observed

in

verbal short-term

memory

and

in speech percept.ion, have also

been noted

in

other

language tasks: poor re aders are general ly lacking in metal inguis tic awareness

of

phonological

structure, produce more errors when naming objec ts , and are

less accurat.e

at

comprehension of

spoken

sent.ences. As

Liberman and

Shankweiler 198.5 discuss in a review

of

these findings, all

of t.he difficulties

appear

to st.em from deficiencies involving

th e

phonological component. of the

language. To

build

on

t.he current.

knowledge

about. t.he origin

of reading

disabilit.y,

it

would

be

valuable:

1

t o

explore t.he int.errelat.ionship of

these

diverse abilit.ies in

normal

development. and

in children encount.ering difficulty in reading; and 2 to

understand

better t.he underlying phonetic

and memory processes.

Practical

Implications

Practit.ioners working t o alleviat.e reading

problems

in children will reasonably be asking if

t.he

current

research,

though short of

complete knowledge, has any

practical application.

We

have

seen

t.hat. poor

readers

are less efficient

at

phonetic

coding and have memory

deficit.s.

The

development.alliterature suggests

t.hat

with experience encoding

operat.ions

increase

in efficiency

and memory

span

increases. Thus, it may generally be

useful

to employ more memory games and



54

r dy

t.asks in school t o

provide

great.er practice (cf. Mann Liberman, 1984). For t.hose individuals

wit.h memory deficit.s,

it

becomes

particularly important in reading

acquisition (as

Perfetti

Lesgold, 1979, and

Perfetti,

1985, have st.ressed)

to

overlearn decoding skills t o increase encoding

speed.

In

t.his way th e individual

can

compensat .e for a furt.her

limitation

in STM, which is

already,

by

it s

nature,

brief

and limited

in capacit.y.

While

t.hese recommendations are perhaps not

curative,

underst.anding th e underlying

prob

lems

for children with reading difficulty can serve as an important aid in avoiding inappropriat.e

diagnoses and inappropriate t.reat.ment.. There are

some

practical implications in

this

direction

from

the cognitive

research.

In closing, these will

be

briefly described:

1. Evidence is steadily mounting t.hat t.he problem poor readers have is specific to language.

Many prereading programs incorporate pract ice in labeling environmental

sounds.

Our research

indicat.es

this

is a separate, unrelated skill:

that

children with reading problems are not

having

a difficulty with

nonspeech

sound

categorization,

thus suggest.ing a teacher would

be

better off

spending time on relevant language tasks

such

as auditory training on phonological

awareness

(see Liberman, Shankweiler, Blachman, Camp,

Werfelman, 1980, for

examples.)

2.

the

errors poor readers made on STM tasks provided evidence that these children were

using some other strategy, such as v isual or semantic coding, one might conclude t.hey should

be t.aught. by a different.

method.

But. evidence that. poor readers use deviant. coding st.rategies

has not

been

obt.ained in t.he research surveyed here.

Error

analyses confirm t.hat. good and poor

readers use t.he same st.rategies, but poor readers are less accurate.

3. Sometimes the well-intended mot.ive to teach to the child s st.rength goes off-track.

Obviously, a practioner needs to deal pragmatically and sensitively with th e child s self-esteem

and self-confidence, but what

has

been observed is that the poor reader needs help in phonolog

ical processes. The memory

research,

and t.hat.

on

linguist.ic awareness, point.s directly to this.

Avoidance of

the

problem will

not improve the

essential phonological skills.

4. A similar comment. can be

made

about t.he recommendat.ion t o teach to th e right hemi

sphere. From what. is known about language processing, it is evident t.hat

when

humans are

presented with a letter or word, whether by eye or by ear, a phonetic represent.at.ion is formed

in short.-t.erm memory. Neurological st.udies confirm

that

this involves regions of th e left cerebral

hemisphere (Milner, 1974). Going to th e whole

word

met.hod

of

instruction in no way bypasses

linguistic processing: the child is simply left with the more difficult task of figuring out. phonic

relat.ionships and

decoding

skills without. helpful instruction,

but

the

left-hemispheric

linguist.ic

processes involved

have not

been eliminated

.

.5 Anot.her lesson from the research on phonological deficit.s pert.ains t o t.he issue

of

visual

learners

vs. audit.ory

learners, visual

and

audit.ory dyslexics, etc.

On

t.he visual

and auditory

short-t.erm

memory tasks for

verbal

mat.erial, poor readers were

found

to

have

a gener l phonetic

coding

deficit. The modalit.y

of

presentation

did

not. matter.

Why

might.

the

misunderst.anding about. visual

and

audit.ory l ea rn ing have come about.?

t.he verbal material has been presented visually (either letters, or words, o r, maybe especially,

pictures), t.he person tes ting the sub ject may focus on t.he visualnat.ure

of

th e task

and

interpret

a performance deficit. as visual in origin. But

again,

from

th e

studies that

have

been reviewed, it

is known that

the

child will encode

the presented

material phonetically. So linguistic processing



Short term

Memory

Processing an d Reading Ability

155

is central to this v isual task. t is

important

to recognize the language requirements of visual

tasks,

so as

not

to

draw

the wrong conclusion and possibly

recommend

an inappropriate activity

for remediation.

6

In the perception studies,

the

necessity of having sufficiently sensitive tasks for identifying

deficits ha s b ee n demonst ra ted. In three different studies, poor r ea de rs have bee n obser ve d to

have

perceptual problems, but

these

were o nly d iscern ib le o n

the

more demanding

t as ks . Ma ny

of the standard

tests

for s cr ee ni ng c hi ld re n ha ve rather easy tasks that

may demarcate

large

developmental differences,

but

no t be sensitive

to

important individual differences. Therefore, it

is

appropriate to consider whether a task has

been

sufficiently difficult. before concluding that a

child has adequate skills i n a given area.

References

Baddeley, A. 1966).

Short-term

memory for w or d sequ enc es as a f un ct io n

of

acoustic, semantic,

and formal similarity. Quarterly Journal of Experimental Psychology

18,

362-365.

Baddeley, A.,

H it ch , G.

1974).

Working memory.

In

G. A. Bower

Ed. ,

The

psychology

of

learning an d

motivation Vol. 8). New York: A ca demi c P re ss .

Baddeley, A., Thomson, N.,

Buchanan, M. 1975).

Word length and

the

structure

of short-term

memory.

Journal of Verbal Learning an d Verbal Behavior

14 , 575-.589.

Brady , S., Mann, V.,

Schmidt, R. in press). Errors in short

term

memory for g oo d and poor

readers.

Memory

Cognition.

Brady, S., Poggie, E., Mer lo, M. 1986). An investigation of speech perception abilities in

c hi ld re n w ho differ

in reading

ability. Haskins Laboratories Status Report on Speech

Research

SR 85

23 27.

Brady, S., Shankweiler, D.,

Mann. V. 1983). Speech perception and memory coding in relation

to

reading

ability.

Journal of Experimental Child Psychology

35, 345-367.

Case,

R.,

Kurland, D.,

Goldberg, J. 1982). Operational efficiency and the growth of short-term

memory span.

Journal of

Experimental

Child Psychology

33,

386-404.

Cermak, L., Goldberg, J. , Cermak, S., D rake , C. 1980).

The

short-term

memory

abili ty o f

children

with

learning disabilities.

Journal

of

Learning Disabilities

13 , 25-29.

Chi , M. T. H. 1976). Short-term memory

limitations

in children: Capacity or processing deficits?

Memory

Cognition 4 559-572.

Cohen, R. ,

Netley, C. 1981). Short-term

memory

deficits in

reading

d isab led child ren in

the

absence of opportunity for rehearsal strategies. Intelligence 5, 69-76.

Conrad, R. 1972). Speech and reading. In J. F. Kavanaugh

G.

Mattingly Eds. ,

Langttage

by ear

and

by eye: The

relationships

between speech

an d

reading.

Cambridge, MA: MIT

Press.

Corsi, R. M. 1972). Human memory an d

the

medial

temporal region

of

the brain.

Unpublished

doctoral

dissertation,

McGill University.

Daneman, M., Carpenter, P. 1980). Individual differences in working

memory and

reading.

Jottrnal

of ~ r b l Learning and Verbal Behavior J Q 450 466.

Dempster,

F.

N

1981). Memory span: Sou rces o f ind iv id ual and development.al differences.

Psychological

Bulletin 89, 63-100.

Engle,

R.,

Marshall, K. 1983). Do d ev elop mental chang es in d ig it span result. from acquisition

strategies?

Journal of Experimental Child Psychology

36 429-436.



156

Brady

Hall, J. , Wilson, K., Humphreys, M., Tinzmann, M., Bowyer, P. 1983). Phonemic-similarity

effects in good vs.

poor

readers.

Memory

Cognition 11,520-527.

Hanson,

V. L.,

Liberman, I.

Y., Shankweiler, D. 1984). Linguistic coding by deaf chi ldren

in relation to beginning

reading

success.

Jou1 nal

of

Experimental Child Psychology 37 ,

378-393.

Hoosian, R. 1982). Correlation between

pronunciation

speed

and

digit

span

size.

Perceptual

an d

Motor Skills

55, 1128.

Hulme,

C

Thomson, N., Muir,

C

Lawrence,

1984). Speech rate and the development of

short-term memory

span. Journal

of

Experimental

Child Psychology 38, 241-253.

Huttenlocher,

J.,

Burke, D. 1976). Why does memory span increase with age?

Cognitive

Psychology 8, 1-31.

Jorm, 1983). Specific reading

retardation and

working memory: A review. Bl itish Journal

of Psychology

74, 311-342.

Katz, R., Shankweiler , D. ,

Liberman, I.

981). Memory for i tem order and phonetic recoding

in the beginning reader. Jou1 nal of

Experimental

Child Psychology

32,

474-484.

LaBerge, D., Samuels, S. 1974). Towards a

theory

of automatic

information

processing in

reading. Cognitive Psychology 6, 293-323.

Liberman,

I.

983). A language-or iented view of reading and its disabilities. In

H

Myklebust

(Ed.), Progress in

learning

disabilities ol. V). New York: Grune Stratton.

Liberman,

I.,

Mann, V. A., Shankweiler, D., Werfelman, M. 1982). Children s memory for

recurring linguistic and nonlinguistic material in relation to reading ability. Cortex 18 ,

367-376.

Liberman,

I.,

Shankweiler, D. 198.5). Phonology and

the

problems of learning to read and

write.

Remedial and Special Education 6,

8-17.

Liberman, I

Y., Shankweiler , D.,

Blachman,

B., Camp,

1.,

Werfelman, M. 1980).

Steps

toward

literacy.

In

P. Levinson C. H. Sloan

(Eds.), Auditory processing

an d

language:

Clinical and research

perspectives. New York:

Grune

Stratton.

Liberman,

I.,

Shankweiler, D.,

Liberman,

A. M., Fowler,

C

Fischer, F.

W.

1977). Phonetic

segmentation and

recoding in

the

beginning reader.

In

A. S.

Reber

D.

Scarbrough

(Eds.), Toward a psychology of reading: The proceedings of the CUNY

conference. Hills

dale, N

J:

Erlbaum.

Mann, V. 1984).

Reading

skills and

language

skill.

Developmental

Review

4,1-15.

Mann, V. A., Liberman, I Y. 1984). Phonological awareness and verbal short-term

memory:

Can they presage early reading problems?

Journal

of Learning Disabilities 17, .592-599.

Mann, V., Liberman,

I.,

Shankweiler, D. 1980). Children s memory for sentences and word

strings in relation to reading ability.

Memory

Cognition

8,

319-335.

Mark, L

S., Shankweiler, D.,

Liberman, I.

Y., Fowler, C. 1977).

Phonetic

recoding and

reading difficulty in beginning readers.

Memory

Cognition 5, 623-629.

Merlo, M., Brady, in preparation). Reading and

short term

memory:

The role of

phonetic

process

mg.

Milner, B. 1974). Hemispheric specialization: Scope and limits. In F. O. SdllUit.t F. G.

Worden

(Eds.), The neurosciences: Third

study

program. Cambridge, MA: MIT Press.

Olson, R. , Davidson,

R.,

Kliegl, R., Davies, S 1984). Development of phonetic memory in

disabled and

normal

readers.

Journal of E;rpcl imcntal Child Psychology 37,

187-206.

Perfet.ti, C. 1985).

Reading ability.

New York: Oxford University Press.



Short-term Memory Processing an d Reading Ability

157

Perfetti,

C.,

Lesgold,

A

1977). Discourse, comprehension, and sources of individual differences.

In M.

A

Just

P.

A

Carpenter Eds. , Cognitive processes in comprehension. Hillsdale,

NJ: Erlbaum.

Perfetti,

C.,

Lesgold,

A

1979).

Coding and

comprehension in skilled

reading

and implications

for

reading ins truc tion . In

B. Resnick

P. Weaver

Eds. , Theory

and pmctice in

early reading.

Hillsdale, NJ

:Erlbaum

Associates.

Rabbitt, P. M. A. 1968). Channel-capacity, intelligibility and immediate memory. Quarterly

Journal of Experimental Psychology

20

241 248.

Samuel, A. G. 1978).

Organizational

vs. ret rieval factors in th e deve lopment of digit span .

Journal of Experimental Child Psychology 26, 308-319.

Shankweiler, D.,

Crain,

S. 1986).

Language

mechanisms and

reading

disorders: A

modular

approach.

Cognition 24, 139-168.

Shankweiler, D.,

Liberman, 1

Y.,

Mark, L

S., Fowler, C. A.,

Fischer,

F. W.

1979). The

speech code and learning to reacl.

Jow nal

of

Experimental Psychology: Human Learning

a nd M em or y 5, 531 54 5

Stanovich, K. 1982).

Individual

differences in the cognitive processes of reading:

1

Word decod

ing. Journal of Learning

Disabilities

15 , 48.5-493.

Stanovich, K. 198.5). Explaining

th e

variance in reading abili ty in terms of psychological pro

cesses:

What

have

we learned?

Annals

of Dyslexia

35, 67-96.

Torgesen,

J. ,

Houck, D. 1980).

Processing

deficiencies

of

learning-disabled children who per

form poorly on

the

digit span test. Journal of

Educational

Psychology 72, 141-160.

Tzeng, O. J. L., Hung, D. L.,

Wang,

W. S. Y. 1977). Speech recoding in reading.

Journal

of

Experimental Psychology:

Human

Learning

and Mem ory 3,

621-630.

Vellutino, F. 1979).

Dyslexia: Theory

an d research.

Cambridge,

MA: MIT Press.

Vellutino,

F., Pruzek,

R., Steger,

J., Meshoulam,

V. 1973).

Immediate

visual recall in

poor and

normal readers as a function of orthographic-linguistic familiarity.

Cortex

9, 368-384.

phonological processing and reading ability

Documents