PSYCHOLOGY OF READING
PSYCHOLOGY OF READING PSY 338K
Popular in Course
Popular in Psychlogy
This 6 page Class Notes was uploaded by Marco Wolf on Monday September 7, 2015. The Class Notes belongs to PSY 338K at University of Texas at Austin taught by Staff in Fall. Since its upload, it has received 57 views. For similar materials see /class/181808/psy-338k-university-of-texas-at-austin in Psychlogy at University of Texas at Austin.
Reviews for PSYCHOLOGY OF READING
Report this Material
What is Karma?
Karma is the currency of StudySoup.
You can buy or earn more Karma at anytime and redeem it for class notes, study guides, flashcards, and more!
Date Created: 09/07/15
Chinese Phoenicians and the Orthographic Cipher of English Philip B Gough Margaret A Walsh University of Texas at Austin To read English one must know two things One must know the English language and one must know the orthography of English Put otherwise in order to read one must be able to recognize the words on a page and then one must understand those words We call this the Simple View of reading Gough amp Tunmer 1986 Hoover amp Gough 1990 It is disavowed by most scholars probably because they believe that the processes of recognizing words and comprehending text are inextricably intertwined But there is evidence that we can think of them as separate skills For one thing the two processes can be dissociated Gough amp Tunmer 1986 The average 5year old can understand a story yet cannot read and the dyslexic can exhibit superior listening comprehension accompanied by decidedly inferior decoding Vellutino 1979 On the other hand many of us can like Milton s daughters adequately decode a language we do not understand and the hyperlexic can evidently decode his native language quite skillfully yet understand it poorly Healy 1982 Moreover there is evidence that even within the normal range of reading ability the two processes make separate contributions to reading ability The Southwest Educational Development Laboratory measured the decoding listening and reading skills of 210 bilingual first graders and took the same measures in the second third and fourth grades Hoover and Gough 1990 have analyzed these data to show that at each grade level the product of the child s decoding and listening scores correlates 84 or above with reading So we hold that reading ability R equals the product of decoding ability D and listening ability C R D x C We know little about listening ability we guess that it is made of up knowledge of the language combined with 199 200 GOUGH AND WALSH intelligence But we ha 39 39 least in EngliSh ve some ideas about the nature of decoding ability at To begin with we hold that an im portant part of decodin an al h b t39 aghpgraphy COHSISts of knowledge of its lettersound correipondtlzncaese I a we call the orthographic cipher of the language its cipher for short THE ciPHER hat is thei cipher The cipher of English is alphabetic its printed characters respon to the phonemes that constitute its spoken words That is what mags it a ipher rather than a code Kahn 1967 e cip er is very complex English letters are ma 39 pped onto E p orgemels in a muchmore elaborate way than the orthographies of langxiliilgse vlvoildf sht0r Finnish Mario Pei 1955 wrote that English spelling is the as awesome mess p 310 Lauri K rt 39 39 doelsh not even have a word for spelliiig a lumen tom us that meSh e correspondences clearly involve strings of letters 39 ma ed ont t f phonemes Perhaps the only letter that always and everysvrhere mpssroliitg Single phoneme is the letter v The correspondences of all other 1 tt depend on their context e m Consider for example the initial let 39 ter c Before a u and o it ma s t gq beforee i and y it corresponds to s Followed by ie together Shh to sometimes c and other times k and that choice is determined by the olyying letters sometimes as many as three compare chore and chord If e t in of it as a system of rules they must be largeunit rules to use the rrnino ogy of Baron 1979 and there must be hundreds or erha s thousands of them p p even Nevertheless skilled readers mana 39 39 ge to internalize the ci her Th readily pronounce nonsense like knifer or psarth or chrisml3 and thy 2 convsstently spell pseudowords like Isunt or glek or zao y Ciphe d 13 liiigqsvthloyvthey do so we do not know the mental form of the a i is a system of rules others that it 39 39 mechanism still others that it consists of co 39 39ls an analoglcal nnections In this context w d not care We do care that it is an abilit that 39 c o hildren ha b do not namely the abilit some c39 ve t Others y translate a string of letters into h 39 p onolo icalf to qecode or better decipher print This ability or the lackgof itfhrgs pro ound consequences for the child s reading and spelling Correlates of the Cipher Children who h 39 39 have Ot ave internalized the Cipher read differently than children who In the first place they read better Gough Juel and RoperoSchneider 15 ORTHOGRAPHIC CIPHER OF ENGLISH 201 1983 measured knowledge of the cipher by asking children to read a list of pseudowords and then a story They found a significant correlation between the number of pseudowords that the child could read and the speed and accuracy of reading the story But they also read differently An important difference relates to word frequency Word familiarity matters much less to children who have the cipher than to children who don t In the study just cited Gough et a1 computed the frequency with which each child had previously encountered each of the words in the story in his or her basal reader and then calculated the slope of the regression of error on that frequency Errors tend to decrease with frequency so these slopes are negative and the greater the decrease the greater the slope They found that this slope correlates 69 with knowledge of the cipher The two kinds of children also make different kinds of errors If the cipher reader misapplies the cipherthis can result in a pseudoword a neologism the cipher reader will make nonword errors But the reader without the cipher cannot do this all he or she can do is guess another word So Gough et al predicted and observed a correlation between knowledge of the cipher and the probability that an error will be a nonword Only the cipher reader will make nonword errors But every child will make word errors Some of these will result from an overreliance on context we believe that the use of context is independent of knowledge of the cipher But there are other sources of word errors ie word substitutions or paralexias The cipher reader s misapplication of the cipher may result in a word as in misreading dead as deed But the child without the cipher will make another kind of word error he or she will try to guess what the word might be and the guesses will tend to be drawn from the words that were previously encountered in the text Gough Juel and RoperSchneider found that the proportion of word errors that are drawn from words in the child s basal reader varies inversely with knowledge of the cipher The cipher also has consequences for the child s spelling The child who does not have the cipher can only spell from memory relying on associations and they are selective and idiosyncratic Gough amp Juel 1989 But the child who has internalized the cipher can make use of its correspondences in spelling So for one thing spelling errors will sound like the target word whereas those of the child without the cipher will not In support of this Juel Griffith and Gough 1984 found that 44 of the spelling errors of firstgrad er cipher readers ie children who could read many pseudowords were homophonous with the target compared to only 04 of those committed by children who read no pseudowords at all Children who spell from memory may have a complete image of the word But if that image is incomplete then he may have to insert something to fill it out For example a child may remember only that camel is the one with humps in the middle In consequence children without the cipher should 202 GOUGH AND WALSH insert more random letters in a misspelling than cipher readers Juel et al found that the proportion of inserted letters in errors eg misspelling come as cuame was more than twice as great among children who do not have the cipher compared to those who do Some striking evidence comes from other insertions that some of these children make Children who do not have the cipher evidently do not understand that letters represent phonemes that the parts of printed words correspond to parts of spoken words So for them the characters on the page are only arbitrarily related to the spoken word and any character a digit an ampersand an exclamation point might do In fact Juel et al found that half the children without the cipher inserted numbers or other symbols into their spellings whereas only one in 20 cipher readers ever made such a mistake The sum of all these differences is a huge difference in spelling ability J uel et al found that when asked to spell 69 words drawn from their basals cipher readers spelled 77 of them correctly whereas children without the cipher averaged 19 Reading without the Cipher Evidently mastery of the cipher is an important component in the develop ment of reading and spelling ability But could one develop reading and spelling proficiency without it There are scholars who seem to think so One reason is that students of acquired dyslexia have found that there are individuals who can read words without the ability to read pseudowords Some of them are called deep dyslexics Coltheart Patterson amp Marshall 1980 others are we think strangely called phonological dyslexics F unnell 1983 Patterson 1982 Many eg Coltheart 1978 take this fact to support the idea that there are two separate mechanisms for word recognition one phonologically mediated ie employing the cipher the other direct If there are two distinct mechanisms operating in parallel either should suffice and we should be able to find individuals who have developed skill in the one without the other Deep and phonological dyslexics do not quite fit the bill for even though they lack the cipher and can still read many words they are by no means skilled readers they are after all dyslexic Are there any skilled readers who can read without the cipher One attempt to find them was initiated by Baron and Strawson 1976 Chinese and Phoenician Baron and Strawson 1976 proposed that there are two kinds of pro cient readers One which relies on the rules they called Phoenician after the creators of the alphabet The other which relies instead on a different word a b 4 m 15 ORTHOGRAPHIC CIPHER OF ENGLISH 203 specific mechanism for word recognition they called Chinese after readers skilled with a logography an orthography lacking an alphabetic cipher In their initial study Baron and Strawson tested the ability of 60 under graduates to decipher nonsense words as indicated by their ability to decide whether the item was homophonic with a word like 3017 or not like magor They also tried to measure their ability to use wordspecific information by giving them a dictated spelling test and then a twoalternative forcedchoice spelling test over the same words reasoning that any improvement from the first test to the second must result from the use of wordspecific information in the latter Students were then classified on the basis of a combined score their number of errors on the pseudoword test added to the difference between their scores on the forcedchoice test and their dictated spelling score Those with high scores there were 8 were labelled Chinese those with low scores 11 Phoenician Baron and Strawson compared the Chinese and Phoenician readers performance in naming regular words and exception words They found that the Phoenicians took longer to name exception words than regular words whereas the Chinese did not Here then is supposed to be evidence that there are pro cient readers the Chinese who can read English without using the cipher or at least using it poorly presumably because they have another skill But we think that it is important to note that the Chinese readers identi ed by Baron and StraWSon were identified not by virtue of exhibiting some kind of skilled performance but rather by two kinds of error errors in pronouncing pseudowords and errors in spelling In spelling Baron and Strawson emphasized that the Chinese readers could make use of wordspecific information to improve their spellings But the only way one could show improvement was to misspell the word in the first place Thus Chinese readers were characterized not by having a skill but rather by having two deficiencies In their subsequent work on adults Baron Treiman Wilf and Kellman 1980 abandoned the effort to find readers skilled at using wordspecific information In fact they abandoned the effort to measure wordspecific knowledge for they concluded that It is in fact hard to find a good measure of knowled ge of wordspecific associations for adults Instead they chose to choose readers only by virtue of their ability to read and spell by rules They assumed that the ability to use rules is largely uncorrelated with the ability to use wordspecific associations p 163 So their subsequent work took a curious turn Chinese readers were no longer identified by what they had but rather by what they lacked ie the cipher No evidence was presented that these readers were proficient at anything Yet they were still described not simply as poor readers poor Phoenicians but as different kinds of readers differing along a dimension simply presumed to be orthogonal to reading ability 204 GOUGH AND WALSH As we see it Baron Treiman Wilf and Kellman s results show only that there are differences between readers good and poor at the cipher They prov1de no eVidence that there are skilled readers without the cipher or that there is another mechanism that might replace the cipher The Growth of Two Skills Baron turned next to children for evidence that there are two mechanisms that make separate contributions to word recognition skill Baron 1979 asked fourth graders to read lists of regular words exception words and pseudowords and examined the intercorrelations of these abilities He reasoned that the child could rely on the orthographic ie wordspecific or leXical mechanism to read both regular words R and exceptions words E and on the rule mechanism to use our term the cipher to read both regular grds and pseudowords N So there should be substantial correlations m tvieten R agd an lfifetween R and E But exception words and pseudowords us e rea usmg i erent mechanisms 5 th 39 N Shoald be lower 0 e correlation between E and Baron and his coworkers have consistent 39 39 I y found such eVidence in children Combining four different groups of subjects ranging from first to Sixth graders Baron 1979 obtained a correlation of 65 between R and E and one of 84 between and N whereas the correlation of E and N was only 42 In a group of 45 third and fourth graders Treiman 1984 found R and E canelated 75 and R and N 81 but N and E only 55 e agree with Baron and Treiman that these results su 39 I D I 39 pport the idea that word recognition skill in English requ1res two kinds of information and thus two distinct subskills But we wondered whether they really support the idea that these result from two independent mechanisms One fact that disturbed us is that although the correlation between nonsense and exception words is lower than the other two it is still very substantial In their adult work Baron Treiman Wilf and Kellman 1980 had explicitly assumed that the ability to use rules is largely uncorrelated With the ability to use wordspecific associations p 163 If children s ability is cfredad 5101186356 words epends on rules and if their ability to read exception 8 e n s blatantly paelse on wet speCIfic associations their expliCit assumption is The most startling data given their assum 39 ption come from Baron and Treiman 1980 They asked 13 first graders 30 second graders 12 third graders and 17 fourth graders to read regular exception and pseudowords Of thes 72 sulbiiects they threw out half who had either high or low total scores onet e ess t ey re rt a correlat39 f exception words po ion 0 90 between nonsense and Baron 1979 suggests that correlations between N and E represent the 15 ORTHOGRAPHIC CIPHER OF ENGLISH 205 variance in all measures due to general ability and instructionquot p 63 We are skeptical about this proposal We agree that both intelligence and instruction contribute to the child s ability to decode words of all kinds we can readily imagine that the dull child or the poorly instructed child might decode all words i e regular words exception words and pseudowords less well than the child blessed in both respects But neither intelligence nor instruction can pronounce a pseudoword or an exception word The effects of those variables must be mediated by the child s word recognition mecha nisms If one s abilities depend on those mechanisms then if those abilities are strongly correlated not largely uncorrelated this suggests to us that the mechanisms themselves are connected The other thing that bothered us was the very idea that children could learn to read many exception words without mastering the cipher We concede that children can and do learn to read sight words well before they master the cipher But we subscribe to the view Gough amp Hillinger 1980 Gough amp Juel 1989 that these first words are l arned through a process of selective association The child selects an arbitrary feature or aspect of the printed word eg a single letter and associates the spoken word only with that The child can learn dozens perhaps even hundreds of words in this way but there are two problems with it asabasis for acquiringword specific information First each word is harder than the last so that it would be very difficult to acquire anything but a small sight word vocabulary in this way Second the specific information that the child learns about the word in selective association is by definition incomplete it is selective What one would need to recognize many exception words in the absence of the cipher is surely nonselective it is complete information about those specific words Given these reservations we wished to further examine the relationships between these variables So we asked 93 first second and third graders in an Austin public school to read the 36item regular exception and nonsense lists devised by Baron 1979 and examined the three intercorrelations Like Baron and Treiman we found the correlations between regular and exception words 80 and regular and nonsense words 76 to be higher than that between nonsense and exception words 66 As BarOn and Treiman have argued these results seem to indicate that the recognition of regular and exception words depends on one kind of information and that of regular and nonsense words on another so that the recognition of exception and nonsense words depend on different sources of information It certainly seems possible that they spring from two parallel mechanisms But an inspection of the scatterplots of the relationships suggests an alternative account If we examine the scatterplot of exception word reading against nonsense word reading what we see is that the ability to read exception words is not simply correlated with the ability to read pseudowords it is instead very systematically related to that ability that is the ability to 205 GOUGH AND WALSH 40 3539 I a II I A I lll 30 39 ii A I All Q A I 25 H g A 3 39 3 20 15quot A I A l A I II A II I I ll 10 A 5 A O I l I I l I I o 5 10 15 20 25 30 35 40 Nonsense FIG 151 Exception word naming as a function of nonsense word naming r 662 read exception words does not simply increase with the ability to read pseudowords it increases in a particular way What the scatterplot exhibits is that if children can read many pseudowords they may or may not read many exception words But if they can read few pseudowords then they can read few exception words Alogician might describe the relationship as implication If a child can read exception words E then the child can read pseudowords N or E implies N That is to say the ability to read many exception words is sufficient but not necessary for the cipher whereas the cipher is necessary but not sufficient for the ability to read many exception words What this suggests to us is a model very different from that proposed by Baron and Treiman It has in common with their model the assumption that two sorts of information are required for skilled word recognition the one we call the cipher the other wordspecific information But we propose that these two are neither acquired independently nor utilized in different 15 ORTHOGRAPHIC CIPHER OF ENGLISH 207 mechanisms Insteadwe posit that the wordspecific information can only be acquired with the aid of the cipher Instead of parallel mechanisms residing side by side in the mind we hold that the wordspeci c mechanism is a superstructure built upon the cipher as foundation A direct test of this hypothesis seemed possible to us We wondered how knowledge of the cipher would relate to the learning of exception words It seemed possible that knowing the cipher might interfere with this learning or be totally unrelated to it But if our hypothesis were correct what we should see is that the child s ability to read pseudowords would be positively related to the learning of new exception words The Cipher and Exception Words To test this hypothesis we administered a REN test of our devising to 49 spring first graders The test consisted of 25 regular words of decreasing word frequency beginning with will and ending with musk 25 exception words yoked to the regulars in length and frequency beginningwith said and ending with womb and 25 pseudowords yoked to the exception words by virtue of being homophonous with them sedd to woom Using their responses we thenidentified 8 exception words blood aisle busy sweat broad tomb cough and choir that almost none of the children identified correctly We typed these words on ashcards and proceeded to teach them to 34 of the same children using the standard procedures of pairedassociate learning with anticipation and correction Instruction continued until the child correctly anticipated all 8 words on 2 successive trials or reached 15 trials Then the child was asked to spell the 8 words We found that the number of trials to criterion decreased with score on the nonword test r 56 p lt 01 children with the cipher learned the eight exception words faster than children without it The presence of the cipher evidently facilitated the acquisition of the Wordspecific information neces sary to read the irregular words Moreover the children with the cipher could also use that information to spell the words that are exceptions to that cipher The number of letters spelled correctly per child correlated 67 p lt 01 with the child s score on the nonword reading test children who know the cipher learn to spell words that are exceptions to that cipher better than children who do not Our results then support a modelin which wordspeci c information does not reside in a mechanism separate and apart from the cipher but instead is accumulated on top of the cipher and cannot be otherwise acquired A clear implication of this model is that there can be no such thing as proficient Chinese or Phoenician readers To be a skilled reader you need both the foundation and the superstructure you need the cipher and wordspecific information The skilled reader is neither Chinese nor Phoenician he or she is if you will both 208 GOUGH AND WALSH But our data suggest an even stronger conclusion Although there are no skilled Chinese or Phoenicians there are semiskilled readers who might be called Phoenician because they know the cipher but lack wordspecific information But there is no Chinese counterpart There are no readers of English or we suspect of any other alphabetic orthography who are even semiskilled without knowing the cipher If a leader does not have the cipher he can only be a poor reader We conclude then that the cipher is the basis of reading ability in English It is not one of two equivalent mechanisms for word recognition two equal partners in the practice of reading The cipher is the basic mechanism the senior partner To be sure it is not enough it must have help By itself the cipher will fail to recognize many perhaps a majority of the words in English 80 further information wordspeci c information must also be acquired But that information must be added to the cipher and the cipher supports the whole endeavor Reading ability in our language like its orthography begins in Phoenicia REFERENCES Baron J 1979 Orthographic and wordspeci c knowledge in children39s reading of words Child Development 50 6072 Baron J amp Strawson C 1976 Use of orthographic and wordspeci c knowledge in reading words aloud Journal of Experimental Psychology Human Perception and Performance 2 386 393 Baron J amp Trelman R 1980 Use of orthography in reading and learning to read In J F Kavanagh amp R L Venezky Eds Orthography reading and dyslexia Baltimore Uni versity Park Press Baron 1 Treiman RWilf 1 amp Kellman P 1980 Reading and spelling by rules In U Frith Ed Cognitive processes in spelling New York Academic Press Coltheart M 1978 Lexical access in simple reading tasks In G Underwood Ed Strategies of information processing London Academic Press Coltheart M Patterson K amp Marshall J C 1980 Deep dyslexia London Routledge amp Kegan Paul Funnel G 1983 Phonological processes in reading New evidence from acquired dyslexia British Journal of Psychology 74159 180 Gough P Bamp HillingerM L 1980 Leaming to read An unnatural act Bulletin of the Orton Society 30179 936 Gough P 3 amp Juel C 1989 The first stages of word recognition In L Rieben a C Perietti Eds L Ippen39i quotquotquotquot quot L L 39 39 quot pedagogiques pp 85 102 Neuchatel et Paris Delachaux et Niestle Gough P 3 Inc C amp RoperSchneider D 1983 Code and cipher A twostage conception of initial reading acquisition In J A Niles amp L A Harris Eds Searches for meaning in readinglanguage processing and interaction 32nd Yearbook of the National Reading Conference Rochester NY National Reading Conference Gough P B amp Tunmer W E 1986 Decoding reading and reading disability Remedial and Special Education 7 610 Healy J M 1982 The enigma of hyperlexia Reading Research Quarterly 17 319338 15 ORTHOGRAPHIC CIPHER OF ENGLISH 209 Hoover W A amp Gough P B 1990 The simple view of reading Reading and Writing 2 127 w o Juel1 C Griffith P I amp Gough P B 1984 Reading and spelling strategies of rst grade dhildren In J A Niles amp R Lalik Eds Issues in literacy A research perspective pp 306 309 Rochester NY National Reading Conferenlsieih h D 196 The codebreakers New York Mac an I 11tt2rson K 1982 The relation between reading and phonological coding Further neuropsychologieal observations In A W Ellis Ed Normality and pathology in cognitive functions London Academic Press k L n 39 l39 h New Yor lpprnco 350 E231 v V in spelling and reading styles Journal E e 39mental Child Psychology 37 463477 Trei iiian p11quot amp Baron J 1981 Segmental analysis ability Development and relation to readihg ability In G E MaeKinnon amp T G Waller Eds Reading research Advances m theory and practice Vol 3 New York Academic Press Vellutino F 1979 Dyslexia Cambridge MA MIT Press
Are you sure you want to buy this material for
You're already Subscribed!
Looks like you've already subscribed to StudySoup, you won't need to purchase another subscription to get this material. To access this material simply click 'View Full Document'