BRAIN

AND

LANGUAGE

15,

340-352 (1982)

Spelling Dyslexia: A Form of C[oss-Cuing
LYNN
UniversiQ

J. SPEEDIE, LESLIE J. ROTHI, AND KENNETH
of Florida

College

of Medicine

and

Veterans

Administration

M. HEILMAN
Medical

Center

A patient with spelling dyslexia maintained the ability to write, spell, and
pronounce spelled nonsense words and words that require knowledge of orthographic rules of language. He was unable to perform grapheme-phoneme conversions except after naming the letter involved. We propose that he uses a
letter-naming strategy to circumvent the disconnection of visual areas from the
area of visual word images and that his letter-naming strategy represents a
compensatory cross-cuing strategy.

INTRODUCTION
Dejerine (1891) described a patient who was unable to read or write
(alexia with agraphia) and had a left-side angular gyrus lesion. Dejerine
thought that the angular gyrus contained visual word images (engrams)
and that destruction of these engrams induced the alexia with agraphia.
Dejerine (1892) and others (see review by Benson and Geschwind, 1969)
also described patients who could not read but could write. This has
been termed “alexia without agraphia” (Benson & Geschwind, 1969).
These patients usually have a left occipital lesion which may include the
splenium of the corpus callosum. One of the postulated mechanisms
underlying alexia without agraphia is a disconnection between the functioning visual areas and the intact left-hemisphere
language areas, including the area of visual word images. A disconnection of the visual
word images from primary visual areas would produce alexia because
visual stimuli could not gain access to the left-hemisphere language area.
However, because the area of visual word images was intact, the patient
could write.
HCcaen and Albert (1978) described several types of alexia without
agraphia. In one of these, verbal alexia, the patient reads by letterThis work was supported by National Institutes of Health Grant NS 15229 and by the
Medical Research Service of the Veterans Administration. Reprint requests should be
addressed to Kenneth M. Heilman, M.D., Department of Neurology, Box J-236, College
of Medicine, University of Florida, Gainesville, FL 32610. Presented at the annual meeting
of the International Neuropsychological Society, Atlanta, Georgia, February 1981.
340
0093-934x/82/020340-13$02.00/0
Copyright
All rights

Q 1982 by Academic
Press. Inc.
of reproduction
in any form reserved.

CROSS-CUING

IN

SPELLING

DYSLEXIA

341

naming. Reading letter by letter has also been termed “spelling dyslexia”
by Kinsbourne and Warrington (1962).
At least three mechanisms have been proposed to account for spelling
dyslexia. Wolpert (1924) first described a syndrome he termed “simultanagnosia.” In this syndrome he included an alexia in which the patient
could read words only by spelling them letter by letter. Wolpert felt that
the letter-by-letter
phenomenon was another manifestation
of the patient’s inability to appreciate the totality of a visual composition in the
face of spared ability to identify its discrete elements.
Kinsbourne and Warrington (1962) in a series of investigations demonstrated that the deficits in simultanagnosia
were not a consequence
of general intellectual deficit and that recognition of discrete visual entities, regardless of linguistic or symbolic import, was qualitatively similar
to that of normals. Simultaneous and sequential presentation of stimuli
suggested a deficit in which elements are perceived only one at a time
and that a time interval is required before the patient can process another
entity.
Subsequently,
Warrington
and Rabin (1971) studied visual span of
apprehension and visual short-term memory in another group of patients
with simultanagnosia.
Many of these patients were as impaired as those
in the 1962 study in terms of visual span of apprehension and short-term
memory, but they did not exhibit spelling dyslexia. These observations
suggest that a defect in the visual span of apprehension or visual shortterm memory is insufficient to account for the letter-by-letter
reading
characteristic of spelling dyslexia.
In an additional study, Warrington and Shallice (1980) further demonstrate that in two spelling dyslexic patients, visual short-term memory
and letter span were equivalent to those of the patients in 1971 who did
not have spelling dyslexia. Warrington and Shallice also reject the disconnection hypothesis and instead propose that the letter-by-letter
reading of the spelling dyslexic arises from damage to the system through
which a visual word-form is attained. The visual word-form system is
defined as that which parses letter strings into ordered and familiar units
and categorizes these units visually. They suggest that formation of a
whole word unit or word-form is necessary before phonological or semantic analysis.
Warrington
and Shallice also noted that in their patients, concrete
words were no less vulnerable to disruption than any other type of word
and suggested, therefore, that the right hemisphere’s capacity to contribute to the residual reading process is negligible, at least in the spelling
dyslexic.
A summary of the view of Warrington and colleagues is that: (1)
reading is a unitary process in which the visual word-form system may
process letters, letter strings, and words and does so before further

342

SPEEDIE,

ROTHI,

AND

HEILMAN

phonological
or semantic analysis; (2) letter-naming is a direct result of
a breakdown of the visual word-form system; and (3) the right hemisphere
contributes negligibly to whatever residual reading may occur in the
spelling dyslexic.
Alternatively,
Rothi and Heilman (1981) proposed that patients may
use a spelling strategy to cross-cue. In comparison with Warrington and
colleagues, Rothi and Heilman believe that (1) reading is a complex
process in which graphic information is acted on by multiple and discrete
mechanisms, (2) letter-naming is not a direct result of the breakdown in
the processing systems per se but rather a compensatory cross-cuing
strategy that allows entry to phonologic processing; and (3) the right
hemisphere is capable of treating letters as objects and naming them
accordingly.
According to this view, the area of visual word images
contains the orthographic rules that allow graphemic stimuli to gain access to phonological,
graphomotor, and semantic systems. As postulated
by Dejerine (1892), in alexia without agraphia these systems are intact
but disconnected from visual input. Thus the typical procedure of grapheme-phoneme
transcoding cannot take place. However, this model
allows for inclusion of alternative or compensatory strategies, the need
for which has been emphasized by Gibson and Levin (1975); and in the
case of spelling dyslexia, object-naming
is the proposed compensatory
mechanism.
A letter or set of letters may be viewed as graphemic symbols of a
phoneme or phonemic sequence. Alternatively,
a letter may also be
viewed as an object and named as such. As an example, in the Public
Broadcasting System’s program “Sesame Street,” “B” becomes Mr. B
and is described as an object with a certain shape. When asked to name
objects, patients with alexia without agraphia usually show only mild
defects (Albert, 1979). These patients are also able to spell, write, and
understand words spelled to them. When words are presented visually,
by naming the letters, the patients may be spelling to themselves, thereby
cross-cuing and gaining access to the intact phonological,
graphomotor,
and semantic systems of the left hemisphere.
We recently had the opportunity to examine a patient who had spelling
dyslexia, and we performed some tests that support the cross-cuing
hypothesis of spelling dyslexia.
CASE REPORT
An 84-year-old right-handed man was admitted to the hospital with the
chief complaint of sudden onset of vertical diplopia 3 days before admission. The patient was a retired farmer and had 1 year of high-school
education.
His past medical history and family history were unremarkable.
Examination of mental status revealed that the patient did not know the

CROSS-CUING

IN SPELLING

DYSLEXIA

343

current year. He showed some reduction of remote memory. (He did
not know the president during World War II but could remember the
date and place of his own birth.) Recent memory was minimally impaired.
(He recalled two of three objects in 5 min.) He could repeat six digits
forward and four in reverse order. Spontaneous speech, comprehension,
naming, and repetition were all normal. Calculations, mapping, praxis,
right-left orientation,
and finger-naming were all within normal limits.
The patient could recognize and sing familiar tunes. He could write
spontaneously and to dictation but could not read what he had written
after a brief delay.
Examination
of cranial nerves was significant for a right homonymous
hemianopia with relative sparing of the paracentral right superior quadrant. The patient also had mild anisocoria, the left pupil being larger than
the right. His hearing was somewhat decreased bilaterally.
Motor examination revealed normal strength except for mild weakness
of the left hamstring and a mild intentional tremor in both hands. Sensory
examination revealed some diminished sensation to vibration distally in
the lower extremities.
The remainder of his neurological examination
was unremarkable.
Computer tomography disclosed a hypodense area consistent with a
left occipital infarction (Fig. l), which did not appear to include the
splenium of the corpus callosum. An electroencephalogram
showed decreased alpha activity and and decreased driving to photic stimulation
in the left occipital area, compatible with a destructive lesion in the
occipital region.
During 10 days of evaluation, he showed the constellation of problems
typically present in patients with verbal alexia or spelling dyslexia, including intact writing, spelling, and comprehension of spelled words with
impaired letter-by-letter
reading. On a short form of the Boston Naming
Test (Goodglass & Kaplan, 1972) he made 7/20 errors of a circumlocution
type (e.g., “Eskimo house” for “igloo”), which indicated a mild naming
disturbance. Prompting, regardless of type, did not improve his naming
ability. The patient also had marked difficulty in naming colors and in
supplying the color of common objects like bananas, but he could match
colored chips and place them in order according to hue.
Reading

and Writing

Tests

We asked the patient to perform several tasks of reading, writing, and
spelling. When one reads with a letter-naming
strategy, long words are
typically more taxing to memory systems. If the letter-naming strategy
were the sole method by which our patient was able to read, we would
expect that he should have greater difficulty with long words than with
short words. In reading aloud, words varying from 3 to 11 letters long,

344

SPEEDIE,

ROTHI,

AND

HEILMAN

he could read the shorter words, albeit laboriously, with 97% accuracy
(Table 1). Typically he named each letter aloud, sometimes repeatedly,
before saying the word (e.g., for the word “cup” he said “c, u, pcup”). He had more difficulty with the longer words (17% correct), which
suggested that his letter-naming
strategy was less effective for the longer
word that taxed immediate memory.
To determine whether he might have some other reading strategy
available (ideographic or grapheme-to-phoneme
conversion), we limited
the time of presentation. When time of presentation was limited to less
than 3 set, the patient was able to read very few 3- to 4-letter highly
familiar words (15% correct). Typically, he guessed what the word might
be, based on the initial letter or letters (e.g., for fish he said “f, i,” and
then guessed “fist”), or he complained that he did not have enough time.
His inability to read rapidly suggests that these other more rapid strategies may not have been available to him.
For patients who must rely on a letter-naming
strategy, transcoding
from script to print should be more difficult than transcoding from print
to script because the letters themselves are less visually distinct in script
than in print and thus more difficult to treat as an object to be named.
Our patient was able to transcode four words and three brief sentences
from print to script in 7 min and 23 set but required 15 min and 33 set
to transcode script of the same material presented in counterbalanced
fashion over multiple days. In both situations, he used the letter-naming
strategy. That he was slow on both tasks but required twice as long to
transcode from script to print as from print to script and required the
letter-naming
approach in both tasks is consistent with the cross-cuing
hypothesis.
The ability to transcode a single grapheme to a phoneme or vice versa,
should not be affected by either simultanagonosia
or a breakdown of the
visual word form because no word form is involved. We asked the patient
to transcode grapheme to phoneme and vice versa. When asked to select
a grapheme from an array of graphemes when its phoneme was pronounced, he did so with 63% accuracy. When asked to provide the
phoneme for a particular grapheme, he did so with 11% accuracy. In
neither case did he select a response without first saying the name of
the letter involved. These results not only are incompatible
with the
simultanagnosia
and word-form hypotheses of spelling dyslexia but also
demonstrate that the patient could not perform grapheme-to-phoneme
conversion without first naming letters. These results again suggest that
the letter-naming
may reflect cross-cuing.
According to the cross-cuing hypothesis, the patient’s morphophonological conversion rules should be intact. To demonstrate this, we asked
him to discriminate words for which correct pronounciation
and meaning

CROSS-CUING

IN SPELLING

DYSLEXIA

345

346

SPEEDIE,

ROTHI,

AND

FIG. 1 .-Continued

HEILMAN

CROSS-CUING

IN SPELLING

FIG.

1 .-Continued.

DYSLEXIA

347

348

SPEEDIE,

ROTHI,

AND

FIG. l-Continued.

HEILMAN

CROSS-CUING

IN SPELLING

349

DYSLEXIA

TABLE 1
SUBJECT’SPERFORMANCE ON GRAPHIC TESTS
Task
Read short words (3-6 letters)
Read long words (7- 11 letters)
Read short words with limited time
Transcribe phoneme to grapheme
Transcribe grapheme to phoneme
Pronounce rule-governed words
Pronounce nonsense words
Spell nonsense words

Correct/items

Correct (%)

29130
116
3120
lo/16
2118
24130
10112
9110

97
17
15
63
11
80
83
90

cannot be gained from strict letter-to-sound
transcoding without the
knowledge of morphophonological
conversion rules (such as the difference in pronounciation
between “tap” and “tape”).
He correctly pronounced 80% of the items. He also reasonably approximated the pronunciation
of 83% of printed nonsense words such
as “arpistan.”
Finally, we asked him to spell aloud nonsense words such
as “mishapool.”
He offered reasonable spelling on 90% of these items.
His ability to pronounce and to spell aloud nonsense words and to discriminate between words that require knowledge of morphophonological
conversion rules suggests that his orthographic code or area of visual
word images was intact.
DISCUSSION
This patient suddenly developed visual difficulties and was found to
have a lesion in the left occipital region and resultant characteristics of
alexia without agraphia. Like previously described patients, such as one
in a report by Denkla and Bowen (1973), our patient could, however,
read some words using a letter-naming strategy (verbal or spelling dyslexia). We explored his reading behaviors in the hope of clarifying the
nature of the underlying mechanisms.
We propose that this patient’s alexic difficulties arose because visual
input of graphic information was isolated from an intact area of visual
word images containing the morphophonological
rules of spoken language. He was unable to approach any of our tasks without verbalizing
the letter names that constituted the grapheme sequences, and his ability
to perform in any fashion declined drastically when presentation time
was limited. Consequently, we believe that verbalization of letters may
substitute for visual input into the area of visual word images. His reliance
on letter-naming, both in grapheme-phoneme conversion and in print-script
transcoding provides further evidence that the patient used object-naming
of letters as a compensatory
cross-cuing strategy. He performed very

350

SPEEDIE,

ROTHI,

AND

HEILMAN

slowly in transcoding script to print and print to script, but required
twice as long to transcode script to print. This fact may reflect the relative
difficulty in treating script letters as independent objects, since they are
less visually discrete than printed letters.
For the naming to succeed, the verbalizations must gain access to an
intact left-hemisphere
orthrographic
code (area of word images). That
the left-hemisphere
word images were indeed intact was reflected in the
patient’s ability to spell aloud nonsense words and pronounce written
nonsense words as well as words that required knowledge of morphophonological
rules of conversion in order to be spoken or understood.
Two issues need to be addressed. As HCcaen and Kremin (1977) noted,
how can objects be named at all in patients whose functioning visual
areas are disconnected from the left hemisphere? What role may the
right hemisphere play in such a naming process?
Oxbury, Oxbury, and Humphrey (1969) suggested that degraded visual
information
traverses the corpus callosum sufficiently to allow gross
visual recognition for objects but not the fine visual recognition required
for reading. In contrast, Geschwind (1965) suggested that visual input
within the right hemisphere may invoke sensory associations other than
visual ones and that derived information may cross the corpus callosum
anteriorly and thus be recognized, named, or in some fashion used.
However, the right hemisphere itself may also have the capacity for
naming objects, Findings from patients with split-brain syndrome suggest
that the right hemisphere is capable of certain linguistic processing usually conducted to better advantage by the left hemisphere, even in the
split-brain populace. Specifically, Levy and Trevarthen (1977) demonstrated that the right hemisphere can match visually presented words
and can select an appropriate picture for simple words. This confirmed
a similar finding of Gazzaniga and Sperry (1967), that the right hemisphere
is capable of reading simple words. Zaidel (1976) reported that, given
sufficient time, the right hemisphere of certain patients can process words
of all grammatical classes. Furthermore, the right hemisphere may initiate
articulations characterized by single syllables or highly habitual phrases
(Levy & Trevarthen,
1977). Thus, letter-naming
in certain patients may
reflect the capacity of the right hemisphere to treat the graphemes as
objects and to name them.
Further evidence for right-hemisphere
contributions
to language processing arises from residual skills of aphasic patients as well. Heilman,
Rothi, Campanella, and Wolfson (1979) described a patient with global
aphasia who selected entrees from a cookbook to indicate his preferences
for dinner and who also learned Amerind, an ideographic gestural system.
After a subsequent right-hemisphere
stroke, the patient lost both skills.
In Japanese alexic and aphasic patients (Sasanuma, 1975; Sasanuma,
Itoh, Mori, and Kobayashi, 1976; Yamadori, 1975) their relatively pre-

CROSS-CUING

IN SPELLING

351

DYSLEXIA

served capacity to process Kanji as compared with Kuna also lends
some credence to possible right hemisphere mechanisms for ideographic
processing of linguistic symbols.
Finally, it is not clear why the patient we describe here could treat
letters as objects but could not treat words as objects (ideographic reading). Rothi and Heilman (1981) postulated that the alternative strategies
available to a patient depend in part on premorbid reading abilities. We
suspect that this man before his illness relied on a nonideographic reading
strategy such as grapheme-to-phoneme
conversion; therefore, after his
stroke he was unable to use an ideographic strategy. We suspect that
after the left-side lesion he began to use a right hemisphere mechanism
for letter-naming
and then employed a verbal-auditory
letter-naming
code to access the area of visual word images, whereupon he applied
orthographic morphophonemic
rules of conversion.
REFERENCES
Albert, M. L. 1979. Alexia. In K. M. Heilman & E. Valenstein (Eds.), Clinical neuropsychology. New York: Oxford Univ. Press. Pp. 59-91.
Benson, D. F., & Geschwind, N. 1969. The alexias. In P. J. Vinken & G. W. Bruyn
(Eds.), Handbook
of clinical neurology:
Disorders
of speech, perception,
and symbolic
behavior.
New York: American Elsevier. Pp. 112-140.
Dejerine, J. 1891. Sur un cas de cecitt verbal avec agraphie suivi d’autopsie. Comptes
Rendus

Hebdomadaires

des

Seances

et Memoires

de la SociCte’

de Biologic,

3,

197-201.
Derjerine, J. 1892. Contribution a l’etude anatomopathologique et clinique des differentes
varittts de &cite verbale. Comptes Rendus Hebdomadaires
des Seances et Memoires
de la Socie’te de Biologic,
4, 61-90.
Denkla, M. B., & Bowen, F. P. 1973. Dyslexia after left occipitotemporal lobectomy. A
case report. Cortex, 9, 321-328.
Gazzaniga, M. S., & Sperry, R. W. 1967. Language after section of the cerebral commissures. Brain, 90, 131-148.
Geschwind, N. 1965. Disconnexion syndromes in animals and man. Brain, 88, 237-294,
585-644.
Gibson, E. J., & Levin, H. 1975. The psychology
ofreading.
Cambridge, MA: MIT Press.
Goodglass, H., & Kaplan, E. 1972. The assessment
of aphasia
and related
disorders,
Philadelphia: Lea & Febiger.
Hecaen, H., & Albert, M. L. 1978. Disorders of written language. In Human neuropsychology.
New York: Wiley. Pp. 54-66.
HCcaen, H., & Kremin, H. 1977. Reading disorders resulting from left hemisphere lesions:
Aphasic and pure alexias. In H. Whitaker & H. Whitaker (Eds.), Studies in neurolinguistics.
New York: Academic Press. Vol. 2, pp. 269-329.
Heilman, K. M., Rothi, L., Campanella, D., & Wolfson, S. 1979. Wemicke’s and global
aphasia without alexia. Archives of Neurology,
36, 129-133.
Kinsbourne, M., & Warrington, E. K. 1962. A disorder of simultaneous form perception.
Brain, 85, 461-486.
Levy, J., & Trevarthen, C. 1977. Perceptual, semantic and phonetic aspects of elementary
language processes in split-brain patients. Brain, 100, 105-l 18.
Oxbury, J. M., Oxbury, S., & Humphrey, N. K. 1969. Varieties of colour anomia. Brain,
92, 847-860.

352

SPEEDIE,

ROTHI, AND HEILMAN

Rothi, L. J., & Heilman, K. M. 1981. Alexia and agraphia with spared spelling and letter
recognition abilities. Bruin and Language,
12, l-13.
Sasanuma, S. 1975. Kana and Kanji processing in Japanese aphasics. Brain and Language,
2, 369-383.
Sasanuma, S., Itoh, M., Mori, K., & Kobayashi, Y. 1976. Tachistosopic recognition of
Kana and Kanji words. Neuropsychologia,
14, 547-553.
Warrington, E. K., L Rabin, P. 1971. Visual span of apprehension in patients with unilateral
cerebral lesions. Quarterly
Journal
of Experimental
Psychology,
23, 423-43 1.
Warrington, E. K., & Shallice, T. 1980. Word-form dyslexia. Bruin, 103, 99-112.
Wolpert, I. 1924. Die Simultanagnosie-Storung
der Gesamtauffassung. Zeitschrift
fuer die
gesamte

Neurologie

und Psychiutrie,

93, 397-415.

Yamadori, A. 1975. Ideogram reading in alexia. Brain, 98, 231-238.
Zaidel, E. 1976. Auditory language comprehension in the right hemisphere following cerebral commissurotomy and hemispherectomy: A comparison with child language and
aphasia. In E. Zurif & A. Caramazza (Eds.), The acquisition
and breakdown
of
language:
Parallels
and divergencies.
Baltimore: Johns Hopkins Press.