Microsoft Word - Brysbaert Nazir final

VISUAL CONSTRAINTS IN WRITTEN WORD RECOGNITION:

VIDENCE FROM THE OPTIMAL VIEWING POSITION EFFECT

Marc Brysbaert

Tatjana Nazir

Royal Holloway, University of London

CNRS, Université Lyon 1

address : Marc Brysbaert

Royal Holloway, University of London

Department of Psychology

Egham TW20 OEX

United Kingdom

marc.brysbaert@rhul.ac.uk

BSTRACT

In this paper we review the literature on visual constraints in written word processing.

We notice that not all letters are equally visible to the reader. The letter that is most

visible is the letter that is fixated. The visibility of the other letters depends on the

distance between the letters and the fixation location, whether the letters are outer or

inner letters of the word, and whether the letters lie to the left or to the right of the

fixation location. Because of these three factors, word recognition depends on the

viewing position. In languages read from left to right, the optimal viewing position is

situated between the beginning and the middle of the word. This optimal viewing

position is the result of an interplay of four variables: the distance between the

viewing position and the farthest letter, the fact that the word beginning is usually

more informative than the word end, the fact that during reading words have been

recognised a lot of times after fixation on this letter position, and the fact that stimuli

in the right visual field have direct access to the left cerebral hemisphere. For

languages read from right to left, the first three variables pull the optimal viewing

position towards the right side of the word (which is the word beginning), but the

fourth variable counteracts these forces to some extent. Therefore, the asymmetry of

the OVP curve is less clear in Hebrew and Arabic than in French and Dutch.

VISUAL CONSTRAINTS IN WRITTEN WORD RECOGNITION:

VIDENCE FROM THE OPTIMAL VIEWING POSITION EFFECT

It is now abundantly clear that reading and visual word recognition are not simply

based on orthographic information but involve the activation of phonological codes.

This has been shown both at the level of individual word processing (e.g., Drieghe &

Brysbaert, 2002; Harm & Seidenberg, 2004) and at the level of sentence and

discourse understanding (e.g., Brysbaert, Grondelaers, & Ratinckx, 2000; Inhoff,

Connine, Eiter, Radach, & Heller, 2004). Also, children with deficient phonological

awareness (i.e., awareness that spoken words consist of sequences of sounds,

phonemes) are at risk for not acquiring good reading skills (e.g., Schatschneider,

Fletcher, Francis, Carlson, & Foorman, 2004).

Unfortunately, the recent emphasis on phonological coding in reading has

overshadowed the fact that a written or printed word is a visual stimulus in the first

place and that limitations of the human visual system put strong constraints on the

speed and the accuracy with which words can be recognised. To redress the balance,

we will review these constraints in the present paper.

1. THE DROP OF VISUAL ACUITY OUTSIDE THE FIXATION LOCATION

The most important variable that limits visual word recognition is the steep drop of

visual acuity outside the centre of fixation. Even at an eccentricity of 1 degree, there

is already a reduction in visual acuity to about 60% of maximum (Wertheim, 1894).

This makes that humans find it difficult to recognise words presented a few letter

positions to the left or to the right of the fixation location (unless these words are large

enough). In such situations, participants have a strong tendency to move their eyes so

that the stimulus word becomes fixated.

Starting from the observation that in reading participants mainly fixate words between

the beginning and the middle of the word, O’Regan (1981) wondered whether the

drop of acuity outside the fixation location not only had implications for the

processing of parafoveally presented words (i.e., words presented a few letter

positions away from the fixation location) but also for the processing of foveally

presented words. His reasoning was that if the drop of visual acuity is important

enough, it should be easier to recognise words after fixation on the middle letter than

after fixation on the first or the last letter. Central fixation makes maximally use of the

high acuity region around the fixation location, whereas fixation on the first letter

makes the last letters fall more than 1 degree away from the fixation location (under

usual reading conditions there are some 3 to 4 letters per degree of visual angle). The

same is true for fixations on the last letter: They make the first letters fall in

parafoveal vision.

To test the idea, O’Regan, Lévy-Schoen, Pynte, and Brugaillère (1984) systematically

manipulated the participants’ initial fixation location within a word by displaying

words in such a way that they were shifted horizontally relative to an imposed fixation

location. More specifically, participants had to fixate a gap between two vertically

aligned fixation lines placed just above and below the horizontal position where the

words were presented. Then, a word appeared shifted so that a different letter position

fell between the two fixation lines (see Figure 1).

Figure 1: Example of how the initial fixation position is manipulated in the

optimal viewing position paradigm. The participant is asked to fixate the gap

between two vertically aligned lines. Words are presented in such a way that the

participants initially look at different letter positions within the words.

An impressive series of experiments has established a consistent pattern of results due

to the fixation position manipulation. This pattern is present in word naming, lexical

decision, and perceptual identification. In addition, it has been observed in many

different languages (French, Dutch, Hebrew, Arabic, Hebrew, Japanese). Figure 2

shows this pattern, which has been called the Optimal Viewing Position (OVP) effect,

for words of 5 and 7 letters in three different tasks: Word naming, lexical decision,

and perceptual identification. In all tasks, word identification was best when the

words were fixated between the beginning and the middle, and performance dropped

when participants were forced to fixate on the extreme letters of the words. The

processing cost was larger for fixations on the end letters than for fixations on the

beginning letters.

Figure 2 : OVP effect for 5-letter (squares) and 7-letter (triangles) words. Left

panel: word naming; middle panel: lexical decision (data from Brysbaert, 1992,

1994). Notice that for these two panels, not all letter positions within the 7-letter

words were tested. Right panel: perceptual identification (data from Stevens &

Grainger, 2003).

watch

Perceptual identification

0.30

0.40

0.50

0.60

0.70

0.80

0.90

-3 -2 -1 0 1 2 3

letter initially fixated relative to word

centre

identification probability

To directly test the idea that the drop of visual acuity is equally important for foveally

presented and parafoveally presented words, Brysbaert et al. (1996) investigated

perceptual identification for 5-letter words presented so that the observers looked

either four character spaces in front of the word, two character spaces in front of the

word, on the first letter of the word, on the middle letter of the word, on the last letter

of the word, two character spaces after the word, or four character spaces after the

word. Presentation duration of the words varied from 14 ms to 70 ms. Figure 3 shows

the results. As can be seen, there was no discontinuity between the performance to

foveally and parafoveally presented words. Performance levels at all fixation

positions were well captured by a Gaussian curve that was shifted slightly to the left

(accounting for the fact that performance was better when words were presented in

the right visual field than in the left visual field; see below).

Figure 3 : The continuity of the drop of performance for foveally presented and

parafoveally presented 5-letter words (data from Brysbaert et al., 1996). Five

different presentation durations were used. The data for each duration fell on a

Gaussian distribution that was slightly shifted to the left of the word centre.

2. MEASURING THE DROP OF PERCEPTIBILITY FOR EMBEDDED LETTERS

To get a more precise measure of the drop of letter perceptibility as a function of

stimulus eccentricity, Nazir, O’Regan, and Jacobs (1991) inserted lower-case target

letters in tachistoscopically presented strings of 8 k’s (e.g. kkkkckkkk) and asked

participants to identify the target letter. The strings were presented in such a way that

participants either fixated on the first or the last k. Target letters could be presented at

each location of the k-string. Figure 4 shows the findings of the study, as a function of

fixation position (beginning vs. end) and position of the target letter.

Figure 4 : Recognition probability of letters presented within a sequence of ks,

either fixated on the first letter (open symbols) or on the last letter (closed

symbols; Nazir et al., 1991).

Nazir et al. (1991) observed three important effects. First, identification of the most

extreme letter (i.e., the last letter when fixated at the beginning or the first letter when

fixated at the end) was better than identification of the nearby inner letters. This is due

to the well-known phenomenon of lateral inhibition (Bouma, 1970): Letters are more

difficult to recognise when they are embedded within other letters than when they are

presented against an empty background. Second, when the most extreme letter was

discarded the drop of performance could be approximated rather well with a linear

regression (notice that a large part of the slopes of the normal distributions in Figure 3

can also be approximated by a linear regression line). Finally, Nazir et al. observed

that the drop of perceptibility was stronger in the left visual field (when participants

fixated on the last k) than in the right visual field (when participants fixated on the

first k). The slope of the linear regression line was 1.8 times steeper in the left visual

field than in the right visual field.

3. SIMULATING OVP CURVES ON THE BASIS OF LETTER PECEPTIBILITY

MEASURES

Subsequently, Nazir et al. (1991) examined whether they could predict the OVP

curves obtained in perceptual identification, on the basis of the perceptibility of the

individual letters and by assuming that a word is recognised only when all its

constituent letters are recognised. Table 1 illustrates the reasoning. The first row

shows the probability of identifying each letter when the participant fixates on the

first letter. The probability of identifying the first letter when the eyes are on this letter

is 1.00. The probability of identifying the second letter is .94 (a drop of .06). The

probability of identifying the third letter is .88 (another drop of .06), and so on.

Likewise, when the eyes are looking at the last letter (the last line of Table 1), the

probability of identifying this letter is 1.00. The probability of identifying the second

last letter is .89. This is a drop of .11, due to the fact that the perceptibility of letters

decreases 1.8 times more rapidly as a function of eccentricity in the left visual field

than in the right visual field. Chances of identifying the third last letter when the eyes

are looking at the last letter, are .78 (i.e., another drop of .11) and so on.

Position of

fixated letter

Probability of letter identification

Probability of

recognizing a

chain of 5

letters

1 2 3 4 5

1 1 .94 .88 .82 .76 .52

2 .89 1 .94 .88 .82 .60

3 .78 .89 1 .94 .88 .57

4 .68 .78 .89 1 .94 .44

5 .57 .68 .78 .89 1 .27

Table 1 : Probability of identifying a word under the assumptions (1) that all

letters must be identified, (2) that the chances of perceiving a letter drop linearly

as a function of the distance from the fixation location, and (3) that the drop of

perceptibility is 1.8 times higher in the left visual field than in the right visual

field.

Figure 5 shows the results of the simulation and compares them with empirical data

obtained for 5-letter words. As can be seen, the simulation captures the asymmetry in

the OVP curve very well, but underestimates the overall recognition probability of the

words. Further simulations indicated that the model predicts a strong word length

effect (poorer performance for long words than for short words), which is not

observed in the empirical data of adults either (although it is present in the data of

beginning readers; Aghababian, & Nazir, 2000).

Figure 5 : Theoretical and empirical OVP curves (theoretical data based on the

model presented in Table 1; empirical data from Nazir et al. (1991).

There are two reasons why the Nazir et al.’s (1991) simulation underestimated the

overall word recognition probability. The first is that it did not take into account the

higher perceptibility of the first and the last letter (see the rightmost data points in

Figure 4). The second is that the model assumed that all letters must be recognised

before a word can be identified. The latter assumption is not realistic, given that many

words can be guessed on the basis of a subset of the letters that make the word (e.g.,

tabl-, h-us-).

Stevens and Grainger (2003) examined whether it was possible to predict the

empirical OVP curves on the basis of individual letter recognition probabilities by

taking into account the probabilities of recognising words on the basis of incomplete

information. They were able to do so, if they used a coding scheme in which not the

absolute letter positions were used (1

, 2

, 3

, …) but relative letter positions (first,

last, or middle letter). The latter observation agrees with the fact that words can be

recognised rather easily when the positions of middle letters have been swapped

(Grainger & Whitney, 2004; Perea & Lupker, 2004). Ben Boutayab (2004) recently

also showed that it is possible to simulate empirical OVP curves on the basis of

individual letter recognition probabilities and the probability of producing the target

word on the basis of partial input.

4. ACCOUNTING FOR THE ASYMMETRY IN THE OVP EFFECT

In the previous sections, we have seen that because of the drop of visual acuity

outside the fixation location and because of the existence of lateral inhibition not all

letters of a word are equally visible. This makes that a word is more rapidly

recognised when readers fixate on the centre letters than when readers fixate on the

outer letters. Two factors are involved in the word processing quality (either accuracy

or speed): (1) the perceptibility of the individual letters as a function of the fixation

location, and (2) the extent to which the most visible letters isolate the target word

from its competitors.

One element that still has to be accounted for, however, is the asymmetry in the OVP

curve. Why are words recognised better after fixation on the first letters than after

fixation on the last letters? Three factors seem to be involved

The first factor is the observation that in general the initial letters of a word are more

informative about the identity of the word than the last letters. Some authors have

argued that this is due to the way spoken words are recognised (for a review, see

Shillcock, Ellison, & Monaghan, 2000). Because the production of a spoken word

takes a few hundred milliseconds (depending on the length), it is more efficient and

communicatively effective to pack the maximum possible information at the

beginning of the word. Then, spoken words can be recognised before the speaker has

ended the pronunciation, leaving more time for other (syntax and discourse-related)

processes.

The effect of the information distribution within words has been examined in a

number of papers (Brysbaert et al., 1996; Farid & Grainger, 1996; O’Regan et al.,

1984; Pynte, Kennedy, & Murray, 1991) and the results have been consistent. The

information distribution has some influence, but on its own it does not reverse the

asymmetry of the OVP effect. Figure 6 shows the results of Brysbaert et al. (1996).

Five-letter words were used that had a high informative beginning (the words had a

chance of 84% to be correctly produced by participants given the first 3 letters,

against a chance of only 9% when given the last 3 letters) or that had a high

informative end (71% correct target production on the basis of the last 3 letters vs. 8%

correct target production on the basis of the first 3 letters). The task was perceptual

identification after tachistoscopic presentation.

0.40

0.50

0.60

0.70

0.80

0.90

135

letter fixated in 5-letter word

percentage correct

Figure 6: Identification probability for five-letter words as a function of initial

fixation (first, middle, last letter) and information distribution within the word

(informative beginning [squares] vs. informative end).

As can be seen in Figure 6, the OVP pattern indeed changed as a function of the most

informative word part, but whereas for words with a high informative beginning

performance was much better after initial fixation on the first letter than after initial

fixation on the last letter, no comparable word end advantage was observed for words

with a high informative end. For these words, performance was equal after fixation on

the first letter and fixation on the last letter.

A second factor that plays a role in the left-right asymmetry of the OVP effect is the

reading direction. The OVP curve is much more symmetric for languages read from

right to left (Arabic and Hebrew) than for languages read from left to right (French

and Dutch). Figure 7 shows perceptual identification data for tachistoscopically

presented Arabic and French words (Farid & Grainger, 1996). As can be seen, the

OVP curve for Arabic has become symmetric; it has not turned into an advantage for

fixations on the rightmost (initial) letter of the word. A similar pattern has been

observed in Hebrew (Nazir, Ben-Boutayab, Decoppet, Deutsch, & Frost, 2004).

Figure 7: The OVP effect for Arabic and French seven-letter words; perceptual

identification (figure from Farid & Grainger, 1996).

The main reason why the reading direction influences the OVP effect probably has to

do with low-level perceptual learning. Nazir (2000; Nazir et al., 2004) argued that left

to right text reading results in many words being recognised in the right visual field

(either when they are in parafoveal vision or after fixation on the initial letters). There

is evidence that repeated presentation of a visual stimulus in the same region of the

visual field leads to enhanced discrimination of that stimulus at that particular region

(Nazir & O’Regan, 1990) but not at other nearby parts of the visual field. Given this

location-dependent perceptual learning, words will be more easily recognised after

fixation on the first letters than after fixation on the last letters, because during

reading the eyes land more often on the first part of a word than on the last part of a

word (see Figure 8).

12345

5-letter words

STATISTICS

Frequency in %

Landing site in the word

9-letter words

13 795

Proportion correct

PERFORMANC

9-letter words

13579

5-letter words

12345

Fixation Zones

Figure 8. Correlation between the distribution of the landing sites of the eyes

during reading and word recognition performance. Top panel. Distribution of

landing positions in 5- and 9- letter Roman words, observed during reading.

Bottom panel. Probability of correct responses for 5- to 9- letter words as a

function of the location of the eyes on the word (from Nazir, 2000).

Notice that the perceptual learning factor, just like the information distribution factor,

predicts that the OVP effect in a language read from right to left (Arabic, Hebrew)

should be the mirror image of the effect observed in languages read from left to right

(French, Dutch). So, one final factor must be invoked to explain why the OVP effect

is more asymmetric for languages that are read from left to right than for languages

read from right to left.

Brysbaert (1994, 2004) argued that this factor is the asymmetry of the brain for

language processing. For the vast majority of people, the left cerebral hemisphere is

more important for language processing than the right cerebral hemisphere. Because

information in the right visual half field is projected directly onto the left cerebral

hemisphere whereas information in the left visual half field requires interhemispheric

transfer to reach the left cerebral hemisphere, word recognition will be slightly easier

after fixation on the leftmost letter of a word than after fixation on the rightmost letter.

This will be true both for languages read from left to right and for languages read

from right to left (remember that the leftmost letter is the first letter of a word when

the word is read from left to right, but the last letter of the word when the word is read

from right to left).

Brysbaert (1994) tested the laterality hypothesis by comparing the OVP effect in

participants with left hemisphere language dominance and participants with atypical

language dominance

. Participants named words of three, four, five, seven, and nine

letters. Figure 9 shows the results.

Figure 9: Naming latencies for words of 3, 4, 5, 7, and 9 letters as a function of

the initial fixation location and the cerebral dominance of the participants

(figure from Brysbaert, 1994).

As can be seen in Figure 9, the word beginning advantage observed in unselected

participants (the vast majority of whom are left dominant for language) did not turn

into a word end advantage for participants with right hemisphere dominance, but the

effect of cerebral dominance on the asymmetry of the OVP curve was significant. So,

a third reason for the strong word beginning advantage in words that are read from

left to right is related to the fact that fixation on the leftmost letter makes the whole

word fall in the right visual half field which has direct connections to the dominant

left hemisphere. This finding also implies that the fovea does not project information

bilaterally to both hemispheres, as is sometimes assumed (see Brysbaert, 2004;

Lavidor & Walsh, 2004; Leff, 2004; Monaghan, Shillcock, & McDonald, 2004;

Whitney, 2004 for further discussion of this issue).

To further test the laterality hypothesis, Nazir et al. (2004) examined the letter

perceptibility as a function of eccentricity for Hebrew letters. Remember from Figure

4, that for French participants the drop of letter visibility as a function of letter

eccentricity is 1.8 times steeper in the left visual field than in the right visual field. If

perceptual learning is the only factor that contributes to this left-right asymmetry, then

for Hebrew readers and Hebrew letters, the pattern should be reversed with an 1.8

times steeper drop in letter perceptibility for letters presented in the right visual field

than for letters presented in the left visual field. On the other hand, if the right visual

Assessment procedures available at the time did not allow the author to be 100% sure that participants

were right hemisphere language dominant; some participants could have had a symmetric language

representation. Therefore, the data of figure 9 are likely to slightly underestimate the effect due to

cerebral dominance. We intend to repeat the study in the near future using brain imaging to assess the

cerebral dominance (Knecht et al., 2000).

field advantage for French participants is the sum of perceptual learning and cerebral

asymmetry, then the asymmetry should be much smaller for Hebrew participants,

because for them perceptual learning (which favours the left visual field) and cerebral

dominance (which favours the right visual field) should cancel each other out. This is

exactly the pattern Nazir et al. (2004) obtained (see Figure 10). There was no

difference between the regression lines in the left and the right visual field.

Figure 10: Recognition probability of Hebrew letters in a homogenous string of

characters, as a function of letter eccentricity and fixation location (left character

of the string or right character).

5. CONCLUSION

In this paper, we have reviewed the literature on visual constraints in written word

processing. We have seen that not all letters are equally visible to the reader. The

letter that is most visible is the letter that is fixated. The visibility of the other letters

depends on (1) the distance between the letters and the fixation location, (2) whether

the letters are outer or inner letters of the word, and (3) whether the letters lie to the

left or to the right of the fixation location. Because of these three factors, word

recognition depends on the viewing position. In languages read from left to right, the

optimal viewing position is situated between the beginning and the middle of the

word. This optimal viewing position is the result of an interplay of four variables: (1)

the distance between the viewing position and the farthest letter, (2) the fact that the

word beginning is usually more informative than the word end, (3) the fact that during

reading, words have been recognised a lot of times after fixation on this letter

position, and (4) the fact that stimuli in the right visual field have direct access to the

left cerebral hemisphere. For languages read from right to left, the first three variables

pull the optimal viewing position towards the right side of the word (which is the

word beginning), but the fourth variable counteracts these forces to some extent.

Therefore, the asymmetry of the OVP curve is less clear in Hebrew and Arabic than in

French and Dutch.

Our review has concentrated entirely on the recognition of individual words. This

raises the question to what extent the OVP phenomenon has implications for text

reading, given that in this situation several words are presented on a line of text. We

venture that the implications of the OVP phenomenon will be particularly strong for

beginning readers, because for quite some time they read text materials word by word.

The situation is slightly more complicated for proficient readers, because they pick up

information from the parafoveal word n+1 while they are still fixating on word n.

This can be concluded from the fact that the reading rate slows down when letter

information from the parafoveal word is denied (e.g., because the word remains

masked until the eyes land on it) and also from the fact that skilled adult readers skip

about one third of the English words (predominantly the short ones; Brysbaert,

Drieghe, & Vitu, in press). This means that in text reading proficient readers quite

often have rudimentary information about the word, in particular the word beginning,

when they land on a word. This advance knowledge is likely to attenuate the OVP

effect in text reading relative to isolated word recognition, as has indeed been

observed by Vitu, O’Regan, & Mittau (1990). However, it seems unlikely that the

advance knowledge could nullify the visual constraints discussed in this paper.

Indeed, the research group of McDonald and Shillcock recently showed that several

of these constraints are needed for a good understanding of eye movement control

both in normal (McDonald & Shillcock, 2005) and in dyslexic readers (Kelly, Jones,

McDonald, & Shillcock, 2004).

References

Aghababian, V., & Nazir, T.A. (2000). Developing normal reading skills: Aspects of

the visual processes underlying word recognition. Journal of Experimental Child

Psychology, 76, 123-150.

Ben Boutayab, N. (2004). Interactions des facteurs visuels et lexicaux au cours de la

reconnaissance des mots écrits. University of Lyon : Unpublished PhD thesis.

Bouma, H., (1970). Interaction effects in parafoveal letter recognition. Nature, 226,

177-178.

Brysbaert, M. (1992). Interhemispheric transfer in reading. University of Leuven:

Unpublished PhD Thesis.

Brysbaert, M. (1994). Interhemispheric transfer and the processing of foveally

presented stimuli. Behavioural Brain Research, 64, 151-161.

Brysbaert, M. (2004). The importance of interhemispheric transfer for foveal vision:

A factor that has been overlooked in theories of visual word recognition and

object perception. Brain and Language, 88, 259-267.

Brysbaert, M., Drieghe, D., & Vitu, F. (in press). Word skipping: Implications for

theories of eye movement control in reading. In G. Underwood (Ed.), Cognitive

Processes in Eye Guidance. Oxford: Oxford University Press.

Brysbaert, M., Grondelaers, S., & Ratinckx, E. (2000). Sentence reading: Do we make

use of orthographic cues in homophones? Acta Psychologica, 105, 31-56.

Brysbaert, M., Vitu, F., & Schroyens, W. (1996). The right visual field advantage and

the optimal viewing position effect: On the relation between foveal and parafoveal

word recognition. Neuropsychology, 10, 385-395.

Drieghe, D., & Brysbaert, M. (2002). Strategic effects in associative priming with

words, homophones, and pseudohomophones. Journal of Experimental

Psychology: Learning, Memory, and Cognition, 28, 951-961.

Farid, M., & Grainger, J. (1996). How initial fixation position influences visual word

recognition: A comparison of French and Arabic. Brain and Language, 53, 351-

368.

Grainger, J., & Whitney, C. (2004). Does the huamn mnid raed wrods as a wlohe?

Trends in Cognitive Science, 8, 58-59.

Harm, M.W., & Seidenberg, M.S. (2004). Computing the meaning of words in

reading: Cooperative division of labor between visual and phonological

processes. Psychological Review, 111, 662-720.

Inhoff, A.W., Connine, C., Eiter, B., Radach, R., & Heller, D. (2004). Phonological

representation of words in working memory during sentence reading.

Psychonomic Bulletin & Review, 11, 320-325.

Kelly, M.L., Jones, M.W., McDonald, S.A., & Shillcock, R.C. (2004). Dyslexics’ eye

fixations may accommodate to hemispheric desynchronization. Neuroreport, 15,

2629-2632.

Knecht, S., Drager, B., Deppe, M., Bobe, L., Lohmann, H., Floel, A., Ringelstein,

E.G., & Henningsen, H. (2000). Handedness and hemispheric language

dominance in healthy humans. Brain, 123, 2512-2518.

Lavidor, M., & Walsh, V. (2004). Magnetic stimulation studies of foveal vision.

Brain and Language, 88, 331-338.

Leff, A. (2004). A historical review of the representation of the visual field in primary

visual cortex with special reference to the neural mechanisms underlying macular

sparing. Brain and Language, 88, 268-278.

McDonald, S.A., & Shillcock, R.C. (2005). The implications of foveal splitting for

saccade planning in reading. Vision Research, 45, 801-820.

Monaghan, P., Shillcock, R., & McDonald, S. (2004). Hemispheric asymmetries in

the split-fovea model of semantic processing. Brain and Language, 88, 339-354.

Nazir, T.A. (2000). Traces of print along the visual pathway. In A. Kennedy, R.

Radach, D. Heller, & J. Pynte (Eds.), Reading as a perceptual process (pp. 3-22).

Oxford: Elsevier.

Nazir, T.A., Ben-Boutayab, N., Decoppet, N., Deutsch, A., & Frost, R. (2004).

Reading habits, perceptual learning, and recognition of printed words. Brain and

Language, 88, 294-311.

Nazir, T.A., & O’Regan, J.K. (1990). Some results on the translation invariance in the

human visual system. Spatial Vision, 3, 81-100.

Nazir, T.A., O’Regan, J.K., & Jacobs, A.M. (1991). On words and their letters.

Bulletin of the Psychonomic Society, 29, 171-174.

O’Regan, J.K. (1981). The convenient viewing position hypothesis. In D.F. Fisher,

R.A. Monty, & J.W. Senders (Eds.), Eye movements, cognition, and visual

perception (pp. 289-298). Hillsdale, NJ: Erlbaum.

O’Regan, J.K., Lévy-Schoen, A., Pynte, J., & Brugaillère, B. (1984). Convenient

fixation location within isolated words of different length and structure. Journal

of Experimental Psychology: Human Perception and Performance, 10, 250-257.

Perea, M., & Lupker, S.J. (2004). Can CANISO activate CASINO? Transposed letter

similarity effects with nonadjacent letter positions. Journal of Memory and

Language, 51, 231-246.

Pynte, J., Kennedy, A., & Murray, W. (1991). Within-word inspection strategies in

continuous reading: Time course of perceptual, lexical and contextual processes.

Journal of Experimental Psychology: Human Perception and Performance, 17,

458-470.

Schatschneider, D., Fletcher, J.M., Francis, D.J., Carlson, C.D., & Foorman, B.R.

(2004). Kindergarten prediction of reading skills: A longitudinal comparative

analysis. Journal of Educational Psychology, 96, 265-282.

Shillcock, R., Ellison, T.M., Monaghan, P. (2000). Eye-fixation behavior, lexical

storage, and visual word recognition in a split processing model. Psychological

Review, 107, 824-851.

Stevens, M., & Grainger, J. (2003). Letter visibility and the viewing position effect in

visual word recognition. Perception & Psychophysics, 65, 133-151.

Vitu, F., O’Regan, J.K., & Mittau, M. (1990). Optimal landing position in reading

isolated words and continuous text. Perception & Psychophysics, 47, 583-600.

Wertheim, T. (1894). Uber die indirekte Sehschärfe. Zeitschrift für Psychologie, 7,

172.

Whitney, C. (2004). Hemiphere-specific effects in word recognition do not require

hemisphere-specific modes of access. Brain and Language, 88, 279-293.