This paper’s reference: Mishkin, M. & Forgays, D.G. (1952). Word recognition as a function of retinal locus. Journal of Experimental Psychology, 43, 43-48.

WORD RECOGNITION AS A FUNCTION OF RETINAL LOCUS

MORTIMER MISHKIN AND DONALD G. FORGAYS

McGill University

Stimulus conditions may be altered in a variety of ways and yet in terms of the reaction produced the situations may be equivalent. One form of equivalence is demonstrated, for example, by man's ability to recognize an object in any visual field of sufficient acuity. Evidence of this kind indicates that the same response may be elicited by the stimulation of different sensory cells, and it has been generally concluded, as in "field theory" (Kohler, 1947) and "equipotentiality" (Lashley, 1942), that the perception must therefore depend on a pattern of excitation whose locus is unimportant. Opposed to this interpretation is the theory presented by Hebb (1949) in which it is assumed that a particular perception depends on the action of particular neural cells, assembled slowly by the repeated stimulation of a specific receptor matrix. The proposal that separate parts of a receptor surface are individually trained leads to the view that equivalence is not spontaneous but is established by earlier learning. The necessary learning conditions are clearly present in vision since ordinarily retinal areas corresponding to fields of equal visual acuity receive the same amount of training. In reading English teat, however, this may not be the case. The reader is persistently presented with the neat word for recognition in the right field while he is still engaged with a word in central The purpose of the present study was to determine whether this constitutes a selective training condition as would be suggested by the finding of non-equivalence between projections of the same word on different retinal loci. The results should provide evidence related to the opposing theories.

EXPERIMENT 1

A comparison was made of the recognition of English words in the left and right visual fields together with a control comparison in the upper and lower fields.

Procedure. Forty-eight eight-letter words in two series of 24 each were presented tachistoscopically to 16 adult Ss. Each word, 2 in. long and 1/4 in. high, was penned in lower case lettering on a white card. The same card was used for presenting the word in different visual fields. [Details of tachistoscope omitted here.] For the first series of 24 words each S was instructed to fixate a small target is the center of the field. The stimulus words appeared in random order 1.5 in. above or below; or 2.0 in. to the left or right. The measurements were made to the center of the word. For the second 24-word series fixation was directed in random sequence to one of four targets in the upper, lower, left, and right parts of the exposure field, with the word always appearing in the center. The distances from fixation to the center of the word were the same as before. The second method allowed S to anticipate the position of a word by predirected attention, permitting evaluation of the effect of this variable on recognition scores.

The choice of stimulus words was determined by the possibility that words containing commonly used prefixes and suffixes might have unequal effects in the left and right fields. For example, when a word with a common suffix is presented to the left of fixation, the part that provides the best cue for recognition falls at some distance from central vision, whereas it lies close to central vision when presented to the right. The exclusive use of such words might produce superior scores in the right field, and conversely for words with common prefixes. Stimuli were carefully selected to balance and permit detection of these effects if they existed. One group of words (e.g., precious, youthful) contained commonly used suffixes; another group (prohibit, transfer) contained commonly used prefixes; while a third group (broccoli, daffodil) had neither, or else was made up of two English words of which either might provide an equally good clue to the whole word (bluebird, woodpulp).

In each method of presentation the Ss, words, and directions of words from the fixation point were arranged in a 4 X 4 latin square. Frequency of recognition in each field could then be determined apart from particular Ss or particular stimulus words.

Each S was given a short practice session in which he became familiar with the conditions of the experiment.

The Ss were observed carefully by the two Es and if eye movement occurred at the moment of exposure the response was discarded and a substitute word was added to the series.

Results. – One point was scored for a completely correct response, and one-half point for a response in which four or more letters of the word were reported in correct sequence. An analysis of the results revealed no unequal effects of the different word types; in particular, recognition scores in the right visual field were as high for words containing common prefixes as for those containing common suffixes.

Since no significant differences were found between the scores obtained in the two experimental procedures, they were combined. The result, presented in Table 1, is that recognition below the fixation point was nearly twice as good as above, and recognition to the right was more than two and a half times better than to the left. These differences between below and above, and right and left, are significant at better than the .01 level of confidence (t(15) = 5.08 and 7.79, respectively). It is clear that exposures of the same word in the left and right visual fields are not equivalent stimulus situations. However, it is possible that several factors other than a selective retinal training may have been responsible, particularly since a recognition difference was also revealed in the control comparison. Alternative interpretations are considered in the following experiments.

TABLE 1

RECOGNITION SCORES FOR WORDS PRESENTED Above, BELOW,

LEFT, AND RIGHT OF FIXATION, FOR EACH OF Two METHODS

SEPARATELY AND COMBINED (N = 16)

Experiment 2

It is unlikely. that differences in visual acuity could account for the non-equivalence, since monocular acuity, in itself, is nearly symmetrical for the lateral fields (Ruch, 1947) and the slight discrepancy that does exist is cancelled out in binocular vision. Even if all Ss had used only their right eyes, the small difference in acuity in favor of the temporal field could hardly have accounted for the marked superiority in recognition to the right of fixation.

However, several other explanations may be offered: Ss may have been attending selectively to the right perceptual field; an anisotropy of visual space (Koffka, 1935, pp. 275-280) may result in a greater clarity or distinctiveness for patterns in the right field than in the left; the left occipital cortex may be "dominant" for vision.

To determine whether these or similar unidirectional factors could account for the results, a comparison of left and right field recognition of English words was made concurrently with that of Yiddish words, in which the letters run in the reverse order. If the explanations are correct, any differences in recognition clearly should be in the same direction for both languages.

Procedure. – Twenty-four English and 24 Yiddish words were presented tachistoscopically to 19 bilingual Ss, whose ages ranged from 12 to 24. The English words used for all Ss and the Yiddish words used for 10 Ss were five letters in length and were hand printed. Yiddish words for the remaining nine Ss were from three to five letters in length and were cut out of an elementary reader. The words, placed on small white cards, were exposed in random order 2 in. to the right or to the left of the central fixation target, so that S did not know on which side the next word would fall nor whether it would-be English or Yiddish. Presentation was arranged with words, Ss, and visual fields in a 2 X 2 latin square.

Results. – Credit was given only for a completely correct 'response. As seen in Table 2, recognition was 44% greater for English words to the right than to the left and 25% greater for Yiddish words to the left than to the right of the fixation point.

TABLE 2

RECOGNITION SCORES FOR ENGLISH AND YIDDISH WORDS

PRESENTED TO THE LEFT AND RIGHT OF FIXATION (N = 19)

The English difference is highly significant statistically (t(18) = 3.54, p < .01), the Yiddish difference is not (t(18) = 1.78, p = .10). This may be due to the fact that the Ss, whose native tongue was English, did not have a comparable reading facility in both languages, as indicated by their consistently lower recognition of Yiddish. Nevertheless, it is clear that whatever facilitated recognition of English in the right visual field did not similarly affect the recognition of Yiddish, and it may be concluded that unidirectional factors cannot account for the differential recognition.

[Skipping two additional experiments that rule out alternative explanations of findings.]

CONCLUSIONS

It has been demonstrated that accuracy of word perception is largely dependent on the locus of retinal stimulation. Differences in visual acuity, a lateral cerebral "dominance," selective attention, the anisotropy of visual space, and the disproportionate significance of parts of a word, have been ruled out as explanations in control experiments. The results support the hypothesis that reading trains limited regions of the left hemi retina selectively. They are inconsistent with the theory of a general equipotentiality in vision since the learning involved in word recognition is not subject to complete transfer. Since there is an indication that English and Yiddish words are more accurately perceived in different visual fields, it appears that a more effective neural organization is developed in the corresponding cerebral hemispheres (left for English, right for Yiddish) as a result of training processes that are specific to the reading of those languages. (It is suggested that a factor in the training may be. the neural equivalent of a selective visual attention, although the data have indicated that when learning is complete this factor may no longer be operative.

SUMMARY

1. Four experiments were performed to investigate the accuracy of tachistoscopic recognition of words placed in the left and right peripheral fields of vision.

2. Experiments 1 and 4 demonstrated that Ss recognized significantly more words placed in certain parts of the right visual field than in corresponding parts of the left, confirming the hypothesis of a selective retinal training arising from the reading situation.

3. Several alternative interpretations of this phenomenon were ruled out in Experiments 2 and 3.

4. The results support the theory that a particular perception depends on the excitation of particular nerve cells, and are inconsistent with the theory of a general equipotentiality in vision.

(Manuscript received March 31, 1951)

REFERENCES

Hebb, D. O. (1949) The organization of behavior. New York: Wiley.

Koffka, K. (1935). Principles of Gestalt psychology. New York: Harcourt, Brace.

Kohler, W. (1947). Gestalt psychology. New York: Liveright.

Lashley, K. S. (1942). The problem of cerebral organization in vision. Biological Symposia, 7, 301-322.

Ruch, T. C. (1947). In J. F. Fulton (Ed.), Howell's textbook of physiology. Philadelphia: Saunders.