BRAIN AND LANGUAGE 17, 146-165 (1982) Deep Agraphia DANIEL BUB AND ANDREW KERTESZ University of Western Ontario A case of agraphia is documented which resembles in every respect the pattern of results obtained in deep dyslexia. Pronounceable pseudowords could not be written to dictation while concrete nouns were more accurately transcribed than abstract nouns. Verbs and functions words were written very poorly and oral spelling was impaired. In addition, semantic paragraphias were present. In contrast to the deficit observed in writing, no such impairment was found for reading. It is concluded that writing can occur without phoneme-grapheme conversion and that codes generated for reading are functionally distinct from codes generated for writing. INTRODUCTION Studies of acquired dyslexia in terms of modem information processing concepts have revealed the existence of two reading processes which can be separately affected by brain damage. Lexical processing can take place without accessing phonological codes and is compromised in instances of surface dyslexia (Marshall & Newcombe, 1973; Shallice & Warrington, 1980). As a result, orthographically irregular words are read incorrectly while greater accuracy occurs for pronunciation of nonwords and orthographically regular words. A second process involves grapheme-phoneme conversion and when impaired results in an almost total inability to read pronounceable nonwords. In its pure form this particular deficit is the major feature of phonological dyslexia (Beauvois & Derouesne, 1979), but cases have also been reported of “deep” dyslexia, which besides showing impaired grapheme-phoneme conversion also reveal a complex array of additional deficits for single-word reading. Specifically, these patients (Marshall & Newcombe, 1973; Patterson & Marcel, 1977; SaIfran & Marin, 1977) produce semantic paraphasias when reading (e.g., TABLE is read as CHAIR), are considerably more accurate when reading concrete as opposed to abstract nouns, and read function words and verbs very poorly. Address reprint requests to Daniel N. Bub, Department of Clinical Neurological Sciences, St. Joseph’s Hospital, London, Ontario N6A 4V2, Canada. 146 0093-934X/82/050146-20$02.00/O Copyright All rights 0 1982 by Academic Press, Inc. of reproduction in any form reserved. DEEP AGRAPHIA 147 While the existence of surface and deep or phonological dyslexia supports the existence of a two-process reading model, the organization of the writing system is obscure. The importance of phonology in writing has often been stressed, in contrast to the lesser role such processing usually plays in reading. Indeed, Wernicke (1874) maintained that writing without the mediation of sound images was impossible since a “direct path from the concept formed by sensory images to the motor writing center” did not exist. Marie (1897) and Dejerine (1926) concurred with this viewpoint, suggesting that writing was inextricably linked to oral language. Some visual-orthographic processing must occur for writing, however, since irregularly spelled words are normally correctly written. A purely phonological approach would result in errors. The nature of such processing and its relationship to phonological writing is unknown, however. While orthographic processing must in principle take place for writing irregularly spelled words, it is not clear whether regularly spelled words are also on occasion processed nonphonologically. It may be the case for example, that visual-orthographic information is only used for unusually spelled words and that phonological writing occurs in all other instances. Recent work with normals has indeed supported the strong dependence of writing on phonology (Frith, 1979, 1981; Hotopf, 1981) and does not show convincing evidence for frequent use of an orthographic route. This in itself does not demonstrate that a purely orthographic form of writing cannot occur, however. The data obtained with normal subjects may be due to the fact that phonological writing is the preferred mode in most instances or alternatively, that testing methods are at present insufficiently sensitive to measure the role of visual processing. A number of studies using brain-damaged populations have in fact supported the existence of orthographic writing in patients with impaired phonological processing. Hier and Mohr (1977) report a case study of a Wernicke’s aphasic who, in spite of very poor oral naming, was nevertheless capable of spontaneously producing the written form of many concrete nouns. These authors conclude that writing occurred “lexicographically,” without reference to the sound of the word. Wapner and Gardner (1979) have suggested that anterior aphasics show only weak evidence for preserved phoneme-grapheme conversion when writing and that they instead appeared to retain a memory of the visual characteristics of the word. Wapner and Gardner (1979) speculate that both lexical and phonological routes exist in writing and that these processes are to some extent dissociable in aphasics. A similar conclusion has also been reached by Langmore (1979) and Marcie and Hecaen (1979). The presence of two writing processes still remains a strong possibility therefore and deserves further investigation. An additional question con- 148 BUB AND KERTESZ terns the relationship between the lexical and phonological codes generated for reading and writing. If common lexical and phonological codes are involved in both processes, then a particular type of agraphia would coexist with an identical deficit in reading. Both would be due to functional impairment affecting a single processing pathway. Thus, for example, writing performance in deep dyslexia usually mirrors reading performance (Newcombe & Marshall, 1980) and can be characterized in terms of a general impairment in transcoding information phonologically. Both phoneme-grapheme and grapheme-phoneme conversion are affected. It is possible however, that phonological and lexical codes generated for reading and writing are functionally distinct, at least to some extent. If this proved correct, then phonological or lexical agraphia could occur in the absence of a corresponding deficit in reading performance. Such a functional disconnection would provide considerable understanding of the organization of input and output systems in reading and writing. Recently, Beauvois and Derouesne (1981) have documented a patient with surface or lexical agraphia, who although capable of correctly writing pronounceable nonwords and orthographically regular words, made many errors when writing irregular words. In contrast, the same patient showed evidence of phonological dyslexia when reading, i.e., pseudowords could not be read while both orthographically regular and irregular words were read without difficulty. Beauvois and Derouesne (1981) conclude that lexical processing in reading is functionally distinct from lexical processing in writing. In this paper, a patient is documented with agraphia, showing a pattern of deficits identical to that produced by deep dyslexics when reading. The ability to transcribe pronounceable nonwords to dictation was totally absent, the patient instead producing confabulatory responses or converting the stimuli into words that sounded similar to the targets. In addition, semantic paragraphias were present, abstract nouns were less accurately written than concrete nouns and performance for verbs and function words was very poor indeed. In marked contrast, reading of both words and pronounceable nonwords was extremely accurate with no evidence of semantic paralexia or difference in accuracy due to word class. The presence of deep agraphia without deep dyslexia suggests that phonological processing for reading is functionally separate from phonological processing for writing and that writing can occur even when the ability to perform phoneme-grapheme conversion is grossly impaired.’ ’ It is worth noting that this written analog of deep dyslexia proved to be transient in the present case. Six months post onset, a portion of the test stimuli were readministered to J. C. for writing to dictation. Performance for pseudowords had improved dramatically (8/10) and no difference in accuracy was found between concrete (IO/lo) and abstract (9/10) words. Recovery of function clearly rules out any explanation of the deficit based on developmental difficulty in spelling. DEEP AGRAPHIA 149 CASE REPORT J. C., a 21-year-old female, was admitted to hospital in February 1981 to manage the lack of intrauterine growth during 4 months of pregnancy. Four months prior to this she suffered a stroke with aphasia. During hospitalization, she continued to have multiple small episodes of transient neurological deficit. Examination failed to reveal evidence of fetal heart activity and induction of labor was carried out resulting in stillbirth. Neurological testing showed a minor drift of the right arm and slight facial weakness, with normal reflexes and intact sensation. No visual field defects were evident. Her stroke was thought to be embolic in origin and she was anticoagulated. Lesion Localization The patient had a C.T. with radiopaque enhancement 4 months after her stroke. This showed an atypical lesion in the left middle cerebral artery territory. The inferior portion of this infarct appeared to be continuous in undercutting the pars opercularis of the inferior frontal gyrus, involving precentral and postcentral gyri and the superior temporal gyri, the temporal operculum of the insula, the supramarginal and angular gyri partially, and the superior portion of the precentral and middle frontal gyri anteriorly (Figs. l-5). The lesion resembles in some respects the CT scans obtained in deep dyslexic patients. In the cases discussed by Marin (1980), the lesions were all extensive, perisylvian, somewhat larger than expected for the degree of language impairment and in all cases the supramarginal gyrus region, insula, and subcortex were involved. Language Testing Testing on the Western Aphasia Battery (Kertesz, 1979) 4 months after stroke (see Table 1) disclosed halting, telegraphic speech typical of Broca’s aphasia. Auditory comprehension of yes/no questions (19/20) and single words (19/20) was excellent but some difficulty was experienced in executing sequential commands (Ull). Repetition of single words was accurate (6/6) but when short phrases and sentences were used as stimuli, the response was halting and nonfluent with word-finding difficulty and phonemic paraphasia. Naming of common objects was relatively good (15/20). Sentence completion and single-word responses to questions were performed flawlessly. Written comprehension of single words was found to be intact within the sample of items tested. J. C. correctly matched simple written words to both real objects (6/6) and visual representations of these objects (6/6). She was also capable of reading single words (6/6) and letters (IO/lo) aloud. Comprehension of written sentences was adequate (6/S), though difficulty occurred when 150 BUB AND KERTESZ FIG. 1. The lowermost cut shows two lesions in the inferior temporal lobe marked by the arrows. The pars opercularis of the inferior frontal gyrus may also be involved. 152 BUB AND KERTESZ FIG. 3. The arrows outline the extent of the lesion which is mostly in the superior temporal gyrus and the postcentral gyrus. The exact extent of the lesion is difficult to determine. The two linear shadows beside the top arrow and above it represent dilated fissures. DEEP AGRAPHIA 153 FIG. 4. The lesion between the arrows involves the supramarginal and angular gyri 154 BUB AND KERTESZ FIG. 5. This cut shows two lesions marked by arrows, the anterior one is probably in the second frontal convolution (“Exner’s center of writing”). The posterior lesion is in or near the angular gyrus. 155 DEEP AGRAPHIA TABLE 1 SUBSCORES ON THE WESTERN APHASIA Comprehension Yes/no questions Auditory word recognition Sequential commands Repetition Naming Object naming Word fluency Sentence completion Responsive speech Reading Comprehension of single words Comprehension of sentences Comprehension of written commands Spelling Recognition of orally spelled words Oral spelling of words Writing Name and address Sentence to dictation Dictated words Spontaneous sentences Copying Praxis Construction Drawing Block design Calculation BATTERY” 57160 58/60 56180 66/100 51160 12120 lO/lO lo/lo 22122 26140 14120 l/6 l/6 616 2/10 8/10 l/34 1000 57160 30/30 919 20124 a Scores are scaled so as to yield an Aphasia Quotient of 100 and are consequently different from those reported in the text. length and grammatical complexity increased. Oral reading of these sentences was markedly telegraphic, paralleling spontaneous output. Recognition of orally spelled words was very poor (l/6), as was oral spelling of dictated words (l/6). Transcribing dictated concrete words was good (5/6). Copying of written sentences was excellent (lo/lo) as was writing individual letters to dictation (9110). Writing of dictated sentences and spontaneous writing of sentences were both impaired. J. C. correctly wrote 2 words of a IO-word dictated sentence even though the sentence was repeated once and divided into smaller clusters of words. When asked to write a few short sentences in response to a picture of a picnic scene, J. C. only wrote “A boy.” Praxis, drawing, block design and calculation were all adequately performed. Further Tests of Reading and Writing During standard language testing it became clear that J. C.‘s writing of dictated words contained semantic paragraphias despite the absence 156 BUB AND KERTESZ TABLE 2 REPETITION, READING, AND WRITING TO DICTATION Percentage Nonwords N Reading Writing Repetition Single-syllable four-letter stimuli Polysyllabic eight-letter stimuli First-order approximations Second-order approximations Third-order approximations Abstract nouns (Orth. Reg.) Concrete nouns (Orth. Reg.) Abstract nouns (Orth. hr.) Concrete nouns (Orth. hr.) Verbs Function words 20 85 5 100 6 6 5 20 20 1.5 15 20 20 83 100 100 100 100 100 100 0 100 100 100 100 100 100 100 100 100 0 0 45 75 33 80 3.5 30 of similar errors in oral reading. On the assumption that these errors indicated the presence of deep agraphia without corresponding deep dyslexia, a series of word lists were compiled and administered to J. C. for writing, oral repetition, and reading. Stimuli were held constant between tasks. Test I: Single-syllable nonwords. Writing: Twenty single-syllable fourletter pronounceable nonwords were dictated individually. J. C. correctly transcribed l/20 (see Table 2). The remainder were either omissions (N = 6), attempts to convert the nonword into a similar sounding word (N = 6) or confabulations bearing no resemblance to the stimulus (N = 6). The remaining response appeared to be the result of a semantic paragraphia for a word generated by J. C. to a nonword. Thus, BLOM was written as FLOWER. The etiology of the response is hypothesized to be BLOOM generated for BLOM and then FLOWER written for BLOOM. Reading: Stimuli were printed in capital letters on white cards and were adminstered individually. Seventeen of twenty nonwords were correctly read. Three stimuli were converted into words that sounded similar to the target. Thus, SHAD was read as “SHADE without the E,” DOSH as GOSH and CRED as SHRED. Repetition: All nonwords were correctly repeated. Test 2: Polysyllabic nonwords. Writing: An additional list of polysyllabic eight-letter nonwords were compiled from norms published by Hirata and Bryden (1971) which varied in the degree of approximation to English. Six fourth-order, six third-order, and five second-order approximations were dictated. These stimuli had no letter appearing more than once and were constructed so that second-, third-, and fourth-order approximations reflected digram, trigram, and tetragram frequencies of 157 DEEP AGRAPHIA English, respectively. Higher-order approximations more closely correspond to conventional English. Part of each fourth-order approximation was converted into a similar sounding word when written to dictation. Thus, RINTOMBE was written as RUNTOM, CLIMPANY as BALCONY, EPONSTER as EPUMONSTER, etc. For third-order approximations, fewer words were written and these bore less resemblance to the actual stimuli. One response was omitted while one bore no relationship to the dictated stimulus. An interesting error again suggests a semantic paragraphia for a word sounding similar to a nonword. Thus, LOBIGNER was written as OYSTER. The etiology of the word is presumed to be LOBIGNER = LOBSTER = OYSTER. For second-order approximations only two responses bore any resemblance to the dictated stimuli, one being TAIL written for TAYORGUC and another ATHYBUE written for ATHIBUNG. The percentage of letters correctly written for each approximation, without regard to correct serial order of the letters is illustrated in Fig. 6. Performance is compared to that of a neurologically intact control subject matched with J. C. for sex, age, and educational level. Thus, both J. C. and the control subject were 21 years of age and had successfully completed high school up to and including grade 9. In computing the percentages, any letters that could substitute for the stimulus elements (e.g., I instead of Y) were counted as correct. Single vowels that correctly depicted the sound pattern conveyed by two vowels (e.g., I instead of IE) were counted as two letters correctly written. Reading: Only fourth-order approximations were administered for reading. J. C. read 5/6 correctly. One nonword was converted into a word (CLIMPANY was read as CLIMAX). Repetition: All stimuli in each group of approximations were correctly repeated. 100 O-1 Matched Control 80 Percent Letters correct 60 40 : \-: J.C. 4 Order of 3 2 Approximation FIG. 6. Percentage letters correct for polysyllabic nonwords written to dictation. Higher-order approximations more closely correspond to English. 158 BUB AND KERTESZ Test 3: Words. One hundred ten words were dictated to J. C. for writing and also for repetition. Subsequently, J. C. was required to read the words typed in capital letters on plain white paper. Since reading and repetition were flawless, only the results of the writing test will be discussed in detail. Results are shown in Table 2. Orthographically regular abstract and concrete nouns. Twenty abstract and twenty concrete nouns were dictated. Mean frequency (Kucera & Francis, 1967) for concrete nouns was 111 (SD = 126) and for abstract nouns was 280 (SD = 353). Word length was approximately equal for the two groups. J. C. correctly wrote 17120concrete nouns and 9/20 abstract nouns. Semantic paragraphias were evident. For example, TIME was written as CLOCK, SKY as SUN, DESK as CHAIR, CHAIR as TABLE. J. C. appeared unaware of these errors while writing took place. Immediately after each paragraphic error, however, the discrepancy between the dictated stimulus and written output was noted, since J. C. was capable of reading the response. Orthographically irregular abstract and concrete nouns. An additional 15 abstract and 15 concrete orthographically irregular nouns were dictated. Frequency was approximately equated between the two word groups with concrete nouns having a mean frequency of 124 (SD = 144) and abstract nouns having a mean frequency of 137 (SD = 157) per million. Twelve of fifteen concrete nouns and 5/15 abstract nouns were correctly written. It can be seen from these results that orthographically irregular nouns are not written more poorly than orthographically regular nouns (x2 = 0.19, p > .l). In contrast, concreteness exerted a significant effect on accuracy when word groups are collapsed across regular and irregular spellings (x2 = 5.23, p < .025). Verbs. Twenty pure verbs (i.e., verbs without noun forms) were dictated. All words were high frequency (greater than 55 per million). Seven were correctly transcribed. Function words. Twenty function words were dictated. Six were correctly written. The remaining errors were omissions (i.e., no responses) or were incorrect functors (e.g., OUR was written as MY, THEY as THEIR). This bears a striking resemblance to the reading performance of deep dyslexics, who typically also preserve word class when incorrectly pronouncing written functors. Overall, writing performance for function words and verbs together was significantly poorer than for nouns (x2 = 4.22, p < .OS). DISCUSSION The high accuracy obtained for oral repetition of all dictated stimuli rules out the possibility that impaired perception is producing the errors in written performance. Instead, results indicate that for J. C., phoneme-grapheme conversion is impossible and that lexical and semantic processing is required for written output to occur. DEEP AGRAPHIA 159 The dependence of writing on lexical processing is illustrated by the results obtained for mono- and polysyllabic nonwords. For monosyllabic nonwords, responses were either omissions and confabulations or were an attempt to convert the stimuli into similar-sounding words. This provides evidence that a lexical-orthographic writing route is operating and mediates performance when words are written to nonwords. Writing via a phonological mechanism, however, is severely impaired since nonword responses (e.g., SEP, MOOY, CASD) did not resemble stimuli in any meaningful way. This is further indicated by the results obtained with polysyllabic nonwords (see Fig. 6). For stimuli closely resembling English, a high percentage of responses consisted of single words or two words combined that corresponded quite closely to the sound pattern of the target syllables. As the resemblance between stimuli and English words decreased, response accuracy for writing deteriorated and omissions or confabulations were evident. In addition to the inability to correctly transcribe nonwords, written performance for words was influenced by semantic aspects of the stimuli. Concrete nouns were written more accurately than abstract nouns, regardless of orthographic regularity, while performance for function words and verbs was very poor. Semantic paragraphias were present. APPEAR is written as MAGIC, TIME as CLOCK, GIVE as TAKE, etc. This pattern of results corresponds closely to that obtained in deep dyslexia and provides strong support for writing via higher-order semantic processes without recourse to phonology. In contrast to these results, reading of all words tested was performed flawlessly, while reading of pseudowords, though slightly impaired, was vastly superior to written performance for the same stimuli. This functional disconnection between grapheme-phoneme and phoneme-grapheme conversion processes extends the distinction offered by Beauvois and Derouesne (1981). Both phonological and orthographic systems are at least partially dissociable in reading and writing. Furthermore, the fact that for J. C., deep agraphia occurs without deep dyslexia indicates that two functional “lesions” must have occurred in cases showing parallel deficits in both modalities, consistent with the model offered by Newcombe and Marshall (1980). Writing without Phonology The nature of the processing deficits shown by J. C. is depicted in diagram form in Fig. 7. It should be noted that whereas phonological analysis, phoneme-grapheme and grapheme-phoneme conversion are separated spatially in the figure for convenience, anatomically these processes may well be subserved by closely related neural systems. As indicated, reading systems involving grapheme-phoneme conversion and lexical-orthographic processing are intact. J. C. correctly read 160 BUB AND KERTESZ SPEECH PHONOLOGICAL t WRITING FIG. 7. Input and output systems that are intact and impaired for J. C. Dotted enclosures indicate impaired processing. pseudowords and orthographically irregular words. Word class had no effect on performance. In contrast, writing via phoneme-grapheme processes is impaired, so that J. C. is forced to access lexical and semantic information to transcribe to dictation. Superior performance for concrete nouns compared to abstract nouns, function words and verbs is due to the fact that translation of semantic information into graphemes is impaired for certain items or that the semantic representations themselves are compromised. Evidence tends to support the first alternative since reading comprehension was relatively good compared to the writing of even the most elementary abstract noun. In addition, comments supplied by J. C. during writing suggest that she understood the word but could not generate a graphemic representation for the stimulus. For example, when asked to write MIND, J. C. commented, “Mind over matter, but I can’t write it.” An additional interesting aspect of J. C.‘s writing, besides the absence of phonological processing, concerns her inability to spell orally. Oral spelling, though not tested in great detail, was markedly impaired relative to writing (see Table 1). In this respect, J. C. differs from cases reported of intact oral spelling but poor writing ability (Kinsbourne & Rosenfield, 1974; Rothi & Heilman, 1981). Thus, while J. C. correctly wrote words like TELEPHONE, HAMMER, and DOCTOR to dictation, she was incapable of orally spelling words like CAT, POND, or PENCIL, or of synthesizing words from spelling patterns like D-O-G or N-O-S-E, in spite of the fact that individual letter naming was performed accurately on visual confrontation (cf. Kinsbourne & Warrington, 1965). 161 DEEP AGRAPHIA Oral spelling presumably occurs via both phoneme-grapheme conversion and by means of access to internally generated visual codes (see Fig. 7). Evidence supports the conclusion that J. C. is capable of writing even though these processes are no longer intact. Writing to dictation could still occur, however, if lexical units are converted directly into graphomotoric code (cf. Weigl & Fradis, 1977), without visually processing the individual letter components of the word. Thus, J. C.‘s performance points to the functional independence of a number of systems within the writing process. Phoneme-grapheme conversion takes place separately from the more direct route which converts semantic and lexical information into graphemic units. In addition, writing appears to involve an analytic (and probably visual) processing of letter components and a more unitary translation of lexical items into graphomotor acts. APPENDIX: EXAMPLES OF WRITING RESPONSES Single-Syllable Nonwords Oral stimulus Written response BIST CRED HUFT SOOM SNEP DOSH SHAD ROSK BESH HAFE WABE MAFE DANT FESS MUNT BUME TESH DEET FOGE BLOM BROD R FACT MOOY CASD HEIF FOWE NR BASH F WADE NR NR FESS NR PUMP SEP TOP FOG FLOWER (brsth) (krcd”) (hufth) (sum) (sncd (d$ (Sapd) (r3sk) (be;) ChefI (web) (mef) (da&) (f=) (mu&) (bjum) Q&S) (dit) @x9 (bljm) 162 BUB AND KERTESZ Polysyllable Nonwords Oral stimulus Written Fourth POLICKET RINTOMBE SURFINDO OMENTACK APONSTER CLIMPANY (p3jjkcth) (rtntomb”) (safindo) (omantrek”) (aep3nsts) (kltmpani) PERABOSI ALKINGER STABOURN LOBIGNER GANDEOPL CALFROVE (pbrabbsi) (Blkms) (&barn) (gaendgopl) (cafrov) EMICHANS ATHIBUNG UNDECHIV DOWACLIS TAYORGUC (ami&nz) (&eabirg) (ondaciv) (dauwakhs) (tt5aguk”) response order POLYAKE RUNTOM SURFTON ONETAC EPUMONSTER BALCONY Third order PAIRUBOE ALKING STOBURN OYSTER B RUNSTOP Wlgns) Second order Orthographically MARTRAS ATHYBUE BU CATCY TAIL Regular Nouns Concrete Oral Stimulus Written SKY DESK BOAT BIRD FOOD MOON TELEPHONE NOSE CHAIR TREE RIVER METAL HORSE DOCTOR GRASS GIRL CLOCK APPLE HOUSE BOOK SUN CHAIR BOAT BIRD FOOD MOON TELEPHONE NOSE TABLE TREE RIVER METAL HORSE DOCTOR GRASS GIRL CLOCK APPLE HOUSE BOOK Abstract Response Oral Stimulus TIME AMOUNT LIFE EFFORT ACT MIND HOPE CARE NAME FACT FUND GAME END PRIDE JOB EVENT METHOD CHOICE CRIME FUN Written CLOCK AROUND LIFE NR ACT NR HOPE CARE NR FACT FUND GAME NR NR JOB NR NR NR CRIME NR Response 163 DEEP AGRAPHIA Orthographically Irregular Nouns Abstract Concrete Oral stimulus LIGHT FRUIT FLOOD POET FLOOR SCHOOL YACHT BLOOD SIGN GUEST CASTLE CIRCLE DOOR COMB KNIFE Written response LIGHT FRUIT FLOOD PE PEAT PE PA POET NR SCHOOL BOAT BLOOD SIGN NR CASTLE CIRCLE DOOR COMB KNIFE Oral stimulus HEIGHT SCIENCE GUESS DOUBT SIGHT LEAGUE EIGHT FIGHT HEALTH DEBT WORD SOUL BIRTH WORK MONTH Verbs Oral stimulus READ EAT ASK CARRY FIX TAKE SPEAK GIVE BEGIN SEND LEND LEARN LIVE CHOOSE GROW SEEK HAVE APPEAR STOP ENTER Written READ EAT NR CARRY FIX NR NR TAKE NR SAID NR NR LIFE CHOOSE NR NR NR MAGIC STOP ENTER Written NR SCIECE,SCEINE,SCEICE NR NR NR NR EIGHT FIGHT NR DEED NR SOUL BIRTH NR MONTH Function response Oral stimulus THE THAN FROM INTO AFTER THEM. WHAT AROUND THEY ME WHO WHERE AT ALSO OUR YOURS WHY HERE WHEN YOU response words Written THE NR FROM INTO NR NR NR NR THEIR ME WHO WHERE TH NR MY OUR NR HERE NR NR response 164 BUB AND KERTESZ Examples of Semantic Paragraphs Oral stimulus Written response HAPPY YACHT SKY TIME DESK CHAIR APPEAR GIVE LAUGH GIVE GIVE FUNNY BOAT SUN CLOCK CHAIR TABLE MAGIC KIND SMILE TAKE KEEP REFERENCES Beauvois, M. F., & Derouesne, J. 1979. Phonological alexia: three dissociations. Journal of Neurology, Neurosurgery and Psychiatry, 42, 1115-l 124. Beauvois, M. F., & Derouesne, J. 1981. Lexical or orthographic agraphia. Brain, 104, 21-49. Dejerine, J. 1926. Semiologie des affections du systeme nerveux. Deuxieme tirage conforme a l’edition de 1914. Paris: Masson. Frith, U. 1979. Reading by eye and writing by ear. In P. A. Kolers, M. E. Wrolstad, & H. Bouma (Eds.), Processing of visible language. New York: Plenum. Frith, U. 1981. Cognitive processes in spelling. New York/London: Academic Press. Hier, D. B., & Mohr, J. P. 1977. Incongruous oral and written naming. Brain andLanguage. 4, 115-126. Hirata, K., & Bryden, M. P. 1971. Tables of letter sequences varying in order of approximation to English. Psychonomic Science, 25, 322-324. Hotopf, N. 1981. Slips of the pen and slips of the tongue. In U. Frith (Ed.), Cognitive processes in spelling. New York/London: Academic Press. Kertesz, A. 1979. Aphasia and associated disorders: Taxonomy, localization and recovery. New York: Grune & Stratton. Kinsbourne, M. & Rosenfield, D. B. 1974. Agraphia selective for written spelling. Brain and Language, 1, 215-225. Kinsboume, M., & Warrington, E., 1965.A case showing selectively impaired oral spelling. Journal of Neurology, Neurosurgery and Psychiatry, 28, 563-566. Kucera, H., & Francis, W. N. 1967. Computational analysis of present-day American English. Providence, RI: Brown Univ. Press. Langmore, S. 1979. The written spelling deficit in Broca’s aphasics, Unpublished preliminary report of a doctoral dissertation, Northwestern University, Evanston, IL. Marcie, P., & Hecaen, H. 1979. Agraphia: writing disorders associated with unilateral cortical lesions. In K. Heilman & E. Valenstein (Eds.), Clinical neuropsychology. New York: Oxford Univ. Press. Marie, P. 1897. L’evolution du langage consideree du point de vue de l’aphasie. La Presse Medicale, 109, 397-399. Marin, 0. S. M. 1980. In M. Coltheart, K. Patterson, & J. C. Marshall (Eds.), Deep dyslexia. London: Routledge & Kegan Paul. Marshall, J. C., & Newcombe, F. 1973. Patterns of paralexia: A psycholinguistic approach. Journal of Psycholinguistic Research, 2, 175-200. DEEP AGRAPHIA 165 Newcombe, F., & Marshall, J. C. 1980. Transcoding and lexical stabilization in deep dyslexia. In M. Coltheart, K. E. Patterson, & J. C. Marshall (Eds.), Deep dyslexia. London: Routledge & Kegan Paul. Patterson, K. E., & Marcel, T. 1977. Aphasia, dyslexia and the phonological coding of written words. Quarterly Journal of Experimental Psychology, 29, 307-318. Rothi, J., & Heilman, K. E. 1981. Alexia and agraphia with spared spelling and letter recognition abilities. Brain and Language, 12, l-13. Saffran, E. M., & Marin, 0. S. M. 1977. Reading without phonology: Evidence from aphasia. Quarterly Journal of Experimental Psychology, 29, 515-525. Shallice, T., & Warrington, E. K. 1980. Single and multiple component central dyslexic syndromes. In M. Coltheart, K. E. Patterson, & J. C. Marshall (Eds.), Deep dyslexia. London: Routledge & Kegan Paul. Weigl, E., & Fradis, A. 1977. The transcoding processes in patients with agraphia to dictation. Brain and Language, 4, 11-22. Wapner, W., & Gardner, H. 1979. A study of spelling in aphasia. Brain and Language, 7, 363-374. Wernicke, C. 1874. Der aphasische symptomencomplex. Breslau: Cohn & Weigert.