15, 197-237 (1983)
Conscious and Unconscious
and Word Recognition
Unit, MRC Applied Cm&ridge, England
sciousness and visual word processing.
field was followed by a pattern mask. Subjects had to make one of three decisions:
Did anything precede the mask? To which of two probe words was what preceded
the mask more similar graphically? To which of two probe words was it more
similar semantically? As word-mask stimulus onset asynchrony
duced, subjects reached chance performance
semantic decisions in that order. In Experiment
which of two words was more similar either graphically or semantically to a nonde-
tectable masked word, but the forced-choice
graphic and semantic similarity. Subjects were now unable to choose selectively
on each dimension, suggesting that their ability to choose in Experiment
passively rather than intentionally mediated.
make manual identification responses to color patches which were either accom-
panied or preceded
by words masked to prevent
words facilitated reaction time (RT), color-incongruent
periment 4 used a lexical decision task where a trial consisted of the critical letter
string following another not requiring a response. When both were words they
were either semantically associated or not. The first letter string was either left
unmasked, energy masked monoptically,
vent awareness. The effect of association
masked cases, but absent with energy masking. In Experiment
word-plus-mask (where the SOA precluded
increased the association effect on a subsequent lexical decision, but had no effect
on (b) detectability or(c) the semantic relatedness of forced guesses of the masked
word. It is proposed that central pattern masking has little effect on visual pro-
cessing itself (while peripheral energy masking does), but affects availability
records of the results of those processes to consciousness.
itself is unconscious and automatically proceeds to all levels of analysis and rede-
scription available to the perceiver. The general importance
cast doubt on the paradigm assumption that representations
analysis are identical to and directly reflected by phenomenal
are presented which explore the relation of masking to con-
In Experiment 1 a single word or blank
(SOA) was re-
graphic, on the detection,
2, subjects again had to choose
stimuli now covaried
In Experiment 3 subjects had to
words delayed RT. Ex-
or pattern masked dichoptically
was equal in the unmasked and pattern
5 repeating a
detection) from 1 to 20 times (a)
of these findings is to
yielded by perceptual
previous draft of this paper by Betty Ann Levy and Earl Hunt.
1, 3, and 4 were presented at the meeting of the Experimental
Scotland in July 1974. The author thanks Paul Rajan, Howard
David Nicholls, and Jeanette Bye for their help in running and analyzing
Helpful discussion of the work was provided by Michael Turvey and on a
All rights of reproduction
1983 by Academic Press. Inc.
in any form reserved.
ANTHONY J. MARCEL
The purpose of this paper is to reassess the role of visual pattern mask-
ing. In doing so it challenges certain aspects of recent information-
processing approaches to perception. This paper is primarily experimen-
tal and general discussion is limited to some immediate and general impli-
cations of the findings; a further paper follows wherein a general approach
to consciousness will be proposed and various phenomena will be dis-
cussed in terms of the differences between conscious and nonconscious
processes. It is necessary first to set the general theoretical context of the
Scientific paradigms, in the Kuhnian sense (Kuhn, 1970) carry with
them assumptions, often implicit, according to which investigations are
carried out and data are interpreted. One paradigm assumption central to
psychophysical and information-processing
which is the focus of the present paper is what will be referred to as the
Identity Assumption. The representations which constitute conscious ex-
perience are assumed to be the very same ones that are derived and used
in sensory and motor processes. Characteristics of intentional responses
or perceptual report are often assumed to directly reflect percep-
tual-cognitive processing. That is, (a) representations which result from
analysis or processing of an event or aspect of it and which can influence
behavior often fail to be distinguished from (b) representations which can
be consciously reflected upon or reported or serve as the basis for inten-
tional choices. Another paradigm assumption, more explicit, is that of
Perceptual Microgenesis. This assumption postulates the nonimmediacy
of percepts and nonunity of their aspects. In essence, the course of per-
ceptual processing is held to be linear, sequential, and hierarchical. (In-
teractive models, where top-down and bottom-up processes are com-
bined, do not in fact violate the essential logic.) Haber (1969) and Posner
(1969) provide good examples of these assumptions. The linear, sequen-
tial aspect amounts to conceiving of different kinds of representations as
being derived one from another in a particular structural and temporal
order. The hierarchical aspect has conceived of this order either as syn-
thetic, “higher level” information being derived from “lower level” infor-
mation, or analytic, where perception proceeds from the general to the
specific. These particular paradigm assumptions have had important con-
sequences. For example, in holding to linearity and the Identity assump-
tion, interpretations of Reicher’s (1969) and Wheeler’s (1970) results on
the superiority of letter identification in the context of a word have pro-
posed the analytic hierarchic notion that somehow the “wordness” of a
word is processed before its component letters. Similar inferences are
drawn from studies of visual search (Brand, 1971; Ingling, 1972) that the
category of a character can be analyzed before its identity. An example of
approaches to perception,
CONSCIOUSNESS, MASKING, AND WORD RECOGNITION
the synthetic hierarchic notion is the assumption that if “higher level”
information is reportable or voluntarily usable, then all “lower level”
information must also be. The converse of this is that a higher level of
representation may be interfered with or prevented while leaving intact
lower levels or earlier stages of representation. It is on this assumption
that backward masking has often been used and interpreted, i.e., that if
processing of a visual stimulus is sufficiently interfered with at a stage of
precategorical representation, descriptions derived from that representa-
tion cannot be achieved (Haber, 1969; Sperling, 1967; Turvey, 1973).
Another example of the synthetic assumption has been the interpretation
of reaction time data from same-different judgments in terms of the linear
hierarchy of stages (e.g., physical, name, category). This has relied upon
the Identity Assumption in supposing that a subject’s response can be
based on a particular stage of processing uncontaminated by any further
stage of processing. Indeed the “Levels of Processing” approach in the
hands of Craik and Lockhart (1972) even holds explicitly that the upper
limit of perceptual processing is under subjects’ conscious voluntary
control, insofar as they may choose to concentrate their processing at a
particular stage in a synthetic hierarchy.
Recently dissatisfaction has been expressed with certain aspects of the
above assumptions. Most particularly,
distinction between conscious and nonconscious states and processes
(Dixon, 1971; Posner & Snyder, 1975; Shallice, 1972) and between auto-
matic processes and those under strategic control (Anderson & Bower,
1973; Posner & Snyder, 1975; Shiffrin, 1975). However, the paradigm
assumptions mentioned above have remained largely intact. This paper
seeks to concentrate mainly on that of the Identity of perceptual process-
ing with conscious representation and strategic control. There are several
reasons to question this assumption.
First, an enormous amount of visual processing is necessarily carried
out automatically and without awareness. The aspects of visual percep-
tion emphasized by Gibson (1950, 1966) have been largely ignored by
cognitive theorists who, for the most part, have used measures based on
conscious manipulation or judgment or on memory. Not only do the
aspects of vision stressed by Gibson support activities such as balance,
locomotion, and orientation, but might well be the basis of articulation of
the visual field for object perception (Man-, 1976; Turvey, 1975; Fox,
1978). Indeed focal attention could not be guided as it is, either visually as
in eye movements, or auditorily as in attending to speech streams if unat-
tended information was not analyzed to high levels of significance.
Second, information-processing theorists have paid little or no attention
to the phenomena of “subliminal” perception. This may be partially ex-
plained by all the doubts raised as to the alternative explanations of the
attention has focused upon the
ANTHONY J. MARCEL
studies carried out in the 1950s directed at motivational aspects of per-
ception (e.g., criticisms such as experimental artifacts, word-frequency
effects, experimenter effects). However, this neglect is hardly justified in
view of the number of replicable demonstrations of perception without
awareness reviewed by Dixon (1971). Certainly in the area of selective
attention many recent studies demonstrate, less problematically,
while people are unable to comment on the nature of unattended stimuli,
they affect both the general state of subjects (Corteen & Wood, 1972; von
Wright, Anderson, & Stenman, 1975) and their responses to attended
stimuli (Lewis, 1970; Mackay, 1973).
Third, certain phenomena from the clinical field appear to imply that
adequate perceptual and cognitive analysis may not be reflected directly
by people’s responses. Patients with an acquired reading impairment
which has been termed Deep Dyslexia (Coltheart, Patterson, & Marshall,
1980) make responses which are semantically, but neither phonologically
nor graphemically, related to a target word presented singly and with
unlimited viewing time (e.g., “buy”
same is true for another type of patient when attempting to repeat single
spoken words (Goldstein, 1948; Morton, 1980). In neither case can the
errors be completely explained by a word-finding problem in spontaneous
speech. This suggests that words have been read or heard correctly, in-
asmuch as their appropriate lexical or semantic representations have been
accessed, but that the patients are unable to recover their identity in their
Interpretations of experiments in the information processing framework
have largely rested upon the lack of a distinction between perceptual
processing and the ability to voluntarily utilize the results of that pro-
cessing or verbalize about it. The phenomena mentioned above illustrate a
dissociation between the two. Perhaps the most dramatic illustrations are
cases of “Blindsight.” The patient reported by Weiskrantz, Warrington,
Sanders, and Marshall (1974) was blind in part of the visual field due to a
lesion in one occipital lobe, i.e., he was not aware of any stimulus. Yet in
the hemianopic field, when forced to, he could reliably make certain
shape discriminations and reach accurately for small light sources. The
patient denied seeing anything and claimed he was guessing. Apart from
the implications for the dissociation of conscious awareness, the phenom-
enon has been interpreted in terms of the two visual systems hypothesis
(Humphrey, 1972), which is an instance of distributed as opposed to linear
processing of different aspects of visual stimuli.
for debt, “swear” for curse). The
Some years ago the present author was conducting some investigations
of reading in children and adults. One experiment (Marcel, Katz, &
CONSCIOUSNESS, MASKING, AND WORD RECOGNITION
Smith, 1974) consisted of single words being briefly exposed followed by a
pattern mask. The subjects’ task was to report whatever words or letters
they were able to. A small but significant proportion of erroneous word
responses, while showing little graphic or phonological relation to the
stimulus, bore a striking semantic relationship to it. Thus green led to
responses such as “blue” and “yellow,”
light to “dark,”
happy to “joy,”
“table.” These responses are noteworthy for several reasons. First, there
was little delay before a response, and therefore it is hard to argue that it is
the result of some memory effect. Second, there was no semantic context
to bias responses. Third, in reporting from tachistoscopic presentations
subjects are usually reluctant to violate in their response any phenomenal
visual impressions that they have. That is, subjects either report letters or
try to generate words which conform to partial graphic or orthographic
Thus, unless it was purely by chance, the subjects appeared to be
exhibiting some knowledge of the stimulus at a lexical or semantic level
without being able to report any other characteristics of the word giving
rise to such knowledge. Unfortunately, Ellis and Marshall’s (1978) crit-
icism of Allport’s (1977) paper, which actually arose from the experiments
reported here, suggests that some or all of these responses may well have
been on a chance basis. In Allport’s study, semantic errors similar to
those in the Marcel, Katz, and Smith data, were found in the responses to
pattern-masked words. Ellis and Marshall estimated the proportion of
randomly paired stimuli and responses from Allport’s data that are seen as
semantically similar by judges and found that the proportion actually
obtained by Allport fell within those limits. The same procedure as Ellis
and Marshall’s was threrefore used (Marcel, 1980a) to estimate the valid-
ity of the semantic errors in the Marcel, Katz, and Smith study. The mean
chance estimate for semantically related errors for that stimulus and re-
sponse sample was found to be 3.4%. The actual proportion of semantic
errors found in the original experiment, discounting derivational and
graphically similar errors (“grass” for green, “long” for large) was 6.43%
of whole-word error responses. Even allowing for some conservatism, the
obtained proportion is considerably higher than the chance estimate,
which leads one to believe that at least some of the semantic errors were
This observation of the independence of the availability of a word’s
meaning and its identity or physical characteristics was reminiscent of at
least two other sets of phenomena mentioned so far. One is the
paraphasias and paralexias noted in acquired aphasia and dyslexia and
discussed by Goldstein (1948), Werner (1956), and Marshall and New-
combe (1973). The other is the recent literature on perception without
queen to “king,”
clock to “time,”
ANTHONY J. MARCEL
awareness (Dixon, 1971). One study in the latter domain which seemed
particularly pertinent was reported by Wickens (1972). He presented
subjects with a word for 50,60,70, or 80 msec followed by a broken-letter
mask for 1.5 sec. Subjects were then presented with a word for 5.0 set
which they had to judge as similar or not to the “unseen” word. Similarity
was defined on poles of Semantic Differential dimensions. On two of these
dimensions, subjects performed above chance while being apparently un-
able to report the first word.
Wickens’ method seemed to promise an experimental grasp on the
phenomenon. However, his experiment is subject to at least two crit-
icisms. First, when backward pattern masking is employed there is a wide
interindividual variance in the critical
word-mask stimulus onset asynchrony (SOA). Examination of the liter-
ature cited by Turvey (1973) shows a range much wider than Wickens’
50-80 msec. Therefore some individuals may have had a different quality
of information from others. Second, the fact that the subject cannot report
a word does not indicate that sufficient visual information has not been
analyzed. There may well be an influence of response criterion. As a
matter of fact, Wickens gives no indication whether or not subjects could
report the first word or any part of it. He merely states that the exposure
duration was “typically too short to result in target identification.”
The initial serendipitous observation and Wickens’ experiment are po-
tentially of great significance. The currently held interpretation of mask-
ing (Sperling, 1967; Turvey, 1973) is that it disrupts a relatively raw repre-
sentation of visual input (iconic memory), without which input cannot be
processed to achieve semantic or phonological coding. If report is impos-
sible due to masking then it is supposedly because the icon has been
interrupted and semantic features of the stimulus should not be repre-
sented. It is thus of considerable importance to establish the validity of
Wickens’ findings. The first experiment was an attempt to investigate the
phenomenon more closely, specifically to examine subjects’ knowledge of
visual and semantic features of the stimulus with respect to its detectabil-
ity as masking is made more severe.
interstimulus interval, or
The object of the first study was to obtain comparative estimates of the
availability and usability of three aspects of word stimuli over a range of
stimulus-mask onset asynchronies. The three aspects were presence vs
absence, graphic characteristics, and semantic characteristics. The rea-
son these three aspects were chosen was that there is a necessary logical
order to their processing according to most approaches to perceptual
microgenesis. However, if processing is dissociated from the recovery of
information in responses it is an open issue as to the relative effect of
CONSCIOUSNESS, MASKING, AND WORD RECOGNITION
masking on the latter aspect at critical target-mask onset asynchronies.
The method adopted was to require judgments of presence or of the
graphic or semantic similarity of succeeding stimuli to the test stimulus.
two sets of word association norms (Bousfield, Cohen, Whitmarsh, & Kincaid, 1961; Post-
man & Keppel, 1970). The words ranged from four to eight letters in length. The words were
used for both the Graphic and Semantic similarity conditions. Half of them were used in the
For the purposes of Graphic Similarity judgments, a pair of words was chosen to be judged
against each of the stimulus words selected from the above-mentioned norms. Neither word
appeared in the norms as an associate of the stimulus word. The words were chosen so that
one had a high rating of graphic similarity, the other a low rating. For this, Weber’s Index of
Graphic Similarity (Weber, 1970) was used with one modification. Since the words were
presented in lower case, graphic similarity was felt to include word shape. The nearest
approximation to this was to include a score for ascenders (h, t) and descenders (g, p).’ Low
graphic similarity was counted as beneath 60, high was counted as above 200.
For the purposes of Semantic Similarity judgments, another pair of words was chosen for
each of the stimuli. One of these was the primary associate given in the association norms.
The second was a word equated with the associate for graphic similarity (?50), which was
not associated in any obvious manner with the stimulus word. For the two sets of word
pairs, three independent judges were unanimous in each choice of a word on the basis of its
graphic and semantic similarity to the stimulus word.
Each word was drawn in black ink in the center of a white 6 by 4-in. card using a UN0
lower case stencil, No. 2.101. Letters measured approximately 0.1 by 0.1 in. The words
subtended from 1.6 to 3.4 degrees of visual angle when viewed in an Electronic Devel-
opments 3-Field Tachistoscope. The word-pair choices were presented one on top of the
other. Half had the “correct” word on top, half beneath. One hundred cards had the words
“present” above “absent,” 100 vice versa.
In addition there was a fixation point and a mask field. The fixation point was a black disc
subtending just under 0.2 degrees. The mask field was composed of parts of letters from the
same stencil, printed in random orientations, over an area of 2.5 in. wide by 0.5 in. high.
The subjects were 24 undergraduates at the University of Sussex.
The stimuli were 240 words selected from the stimulus terms in
The first part of a session was concerned with finding the approximate stimulus onset
asynchrony between word and mask (SOA) at which the subject began to have difficulty in
deciding whether or not a word had appeared. This consisted of a crude “hunting” in which
only presence-absence judgments were required. A trial consisted of the following se-
quence (i) the central fixation point lasting 500 msec, (ii) a word or blank field for a variable
duration, (iii) the mask field lasting 500 msec. The experimenter informed the subject that on
50% of trials a word would be presented, on 50% a blank card. When a SOA was found
where the subject first made errors of detection the experimental trials were begun.
The SOAs used ranged from 5 msec above the point where the subject first showed any
’ For ascenders and descenders, the term +z was added to Weber’s formula, where z is
calculated by counting 2 for each ascender/descender in equivalent (-t 1) positions from the
beginning of the word and subtracting 2 each time an ascender/descender appears in a
position more than two letters away from where one exists in the stimulus word.
ANTHONY J. MARCEL
ments 3, 4, and 5 do not appear to be SO,~ and the latter three studies are
more valid anyway in that they indirectly assess the representation of
information. Should these studies be regarded as necessarily related to
consciousness? One response to this is to propose that the presentation
conditions referred to as nonconscious are merely a matter of the sub-
ject’s criterion for reporting awareness. Perhaps so. But this is to advance
a theory of what it is to be aware rather than to deny the present relevance
of consciousness. Most important for deciding how to interpret the pres-
ent findings is another study using the same techniques as reported here
(Marcel, 1980b). That study compared the effects of polysemous words
(PALM), when masked or not, on the processing of a subsequent word
related to one of its meanings (WRIST). When it was not masked, PALM
only facilitated processing of WRIST when itself preceded by HAND;
when preceded by TREE, it delayed processing of WRIST, compared for
example to CLOCK PALM WRIST, When it was masked, however,
PALM facilitated processing of WRIST, irrespective of what preceded it.
Effects that are qualitatively different but at least as strong suggest a
qualitative difference in the state of the internal representation of the
critical word. This encourages the view that the appropriate interpretation
of the present experiments is indeed in terms of consciousness.
The most satisfactory account of the present experiments, then, is that
pattern masking does not impede visual processing but rather impedes the
availability to consciousness of what is masked. Discussion of how it does
so and of what is involved in being conscious of an event is left to a
theoretical paper which follows this. However, there are certain meth-
odological implications which follow directly from the present experi-
There are several especially important points raised by the findings
reported here. First, masking has been considered a particularly valuable
4 Since this and the subsequent theoretical paper were originally written, various repli-
cations have been attempted. Experiments 1 and 4 have been successfully replicated by
Fowler, Wolford, Slade and Tassinary (1981). While they were able to find a possible artifact
in Experiment 1 they were unable to do so for Experiment 4. Experiment 3 has been
replicated with almost identical latencies by Ellis (unpublished manuscript, University of
Bangor, Wales). However, an attempt to replicate Experiment 4 by Creighton (unpublished
data, Applied Psychology Unit, Cambridge) yielded very small and variable nonconscious
priming effects, and an attempt by Evett (University of Bristol) was unsuccessful. The point
of this is that the phenomena reported here may depend on experimental factors as yet
unexplored. Chambers (Deakin University, Australia; manuscript in preparation) reports
that dichoptic masking is critical. This is highly plausible, since it is the only way to ensure
that pattern masking does not operate at a peripheral level at low SOAs.
CONSCIOUSNESS, MASKING, AND WORD RECOGNITION
technique for revealing early stages of perceptual processing. As Haber
(1969) pointed out it is the one technique whereby one can supposedly halt
stimulus processing at a certain stage without relying on the subject’s
ability to address him/herself to that stage and remain uncontaminated by
any further automatic processing. Masking can no longer be considered to
fulfill this role, since it does not appear to impede such processing. Indeed
our whole conception of what has been referred to as Iconic memory has
to be reconsidered. This will be broached in the theoretical paper which
follows. More generally what the present studies reveal to have been
misleading in studies of masking, and indeed of early perceptual process-
ing, is the response measure used, that of report. Report may reveal some
aspects of conscious experience and its limits, but hardly anything of
perceptual processes themselves. In fact, there is good reason to believe
that it may be extremely misleading in several ways to rely on conscious
percepts to tell us about perceptual processing. First the languages of
representations derived at a nonconscious and a conscious level may be
quite different, a point to be explored in the paper to follow. Second, as
Nisbett and Wilson (1977) suggest, reports, even of tachistoscopic stimuli
or of one’s own sensations, probably tell us more about people’s beliefs
about sensation and cognition than about those processes themselves.
Third, paying attention to stimuli, which is required by most studies in-
volving report, may well induce a state quite inimical to revealing auto-
matic nonconscious processes. That is, people attempt to base their be-
havior on notions of “rationality.” If the task is normally one of reading
or reporting, one does not report a word, even if it “comes to mind,” if one
is not conscious of having seen one. Further, paying attention to the outer
world (i.e., to specific stimuli) or being prepared to act are likely to
narrow one’s attention and preclude awareness of what has only an inner
Psychologists have not realized what they are doing. Indeed, whenever
they suspect that their studies are “contaminated”
processes (postcategorical representations, beliefs, strategies), all too
often their first response is to try to prevent their presence or influence. In
fact, these “problems” present the most truly psychological issues and
those domains worthy of infinitely more study than they are at present
However, how can perceptual processes themselves be studied at a
psychological level? First, indirect rather than direct measures should be
used. If we are interested in the processing of a stimulus at a certain level
or in certain terms, without contamination of later coding or response
processes, then we should look at its effects at that level or in those terms
on the processing of another stimulus. The time course and structure of
by “higher level”
ANTHONY J. MARCEL
processing can be tapped by varying the time between the influencing and
the to-be-influenced stimulus.
But perhaps even indirect measures are not enough. As suggested
above, allowing an event to reach a state of conscious representation at all
may produce misleading effects. Not only may different dimensions of
coding obtain for conscious and nonconscious representations, but some
of what is represented of an event prior to consciousness may be inhibited
by that event’s access to consciousness (see Marcel, 1980b). To the extent
that an investigator is interested only in early nonconscious perceptual
processes, it follows that consciousness of the relevant stimulus ought to
be prevented by masking or some functionally equivalent procedure. It
would at least be interesting to compare the results of direct measures
involving conscious perception and indirect measures where conscious-
ness is precluded. Experiments which approach this via unattended
stimuli are open to criticisms of strategic sampling by the subject. In
studies which rely only on inability to report, but not on inability to
detect, subjects can always be said to have partial information available.
Awareness itself needs manipulation.
Allport, D. A. On knowing the meaning of words we are unable to report: The effects of
visual masking. In S. Dornic (Ed.), Attention
Academic Press, 1977. Vol. 6.
Anderson, J. R., & Bower, G. H. Human associative
Bousetield, W. A., Cohen, B. H., Whitmarsh, G. A., & Kincaid, W. D. The Connecticuf
free associational norms. University of Connecticut, Studies on the Mediation of Ver-
bal Behavior Technical Report No. 35, 1961.
Brand, J. Classification without identification in visual search. Quarterly
1971, 23, 178- 186.
Breitmeyer, B. G., & Ganz, L. Implications of sustained and transient channels for theories
of visual pattern masking, saccadic suppression and information
Psychological Review, 1976, 83, p-36.
Coltheart, M., Patterson, K. E., & Marshall, J. C. (Eds.) Deep dyslexia. London: Routledge
& Kegan Paul, 1980.
Corteen, R. S., & Wood, B. Autonomic responses to shock-associated words in an unat-
tended channel. Journal of Experimental
Craik, F. I. M., & Lockhart, R. S. Levels of processing: A framework for memory research.
Journal of Verbal Learning and Verbal Behavior,
Davelaar, E., & Coltheart, M. Effects of interpolated items on the association effect in
lexical decision tasks. Bulletin of the Psychonomic
Dixon, N. F. Subliminal perception: The nature of a controversy.
Dyer, F. N., & Severance, L. J. Stroop interference with successive presentations of sepa-
rate incongruent words and colors. Journal
and performnnce. New York/London:
memory. Washington, D.C.: Winston,
Journal of Experi-
Psychology, 1972, 94, 308-313.
1972, 11, 671-684.
Society, 1975, 6, 269-272.
of Experimental Psychology,
CONSCIOUSNESS, MASKING, AND WORD RECOGNITION
Ellis, A. W., & Marshall, J. C. Semantic errors or statistical flukes? A note on Allport’s “On
knowing the meaning of words we are unable to report.” Quarterly
1978, 30, X9-575.
Ferguson, G. A. Nonparametric trend ana/ysis. Montreal: McGill Univ. Press, 1965.
Fowler, C. A., Wolford, G., Slade, R., & Tassinary, L. Lexical access with and without
awareness. Journal of Experimental Psychology:
Fox, .I. Continuity, concealment and visual attention. In Cl. Underwood (Ed.), Strategies of
New York/London: Academic Press, 1978.
Fromkin, V. A. (Ed.) Speech errors as linguistic
Gibson, .I. J. The perception of the visual world. Boston: Houghton-Mifflin,
Gibson, J. J. The senses considered as perceptual
Goldstein, K. Language and language disturbances.
Haber, R. N. Repetition as a determinant of perceptual recognition processes. In W.
Wathen-Dunn, J. Mott-Smith, H. Blum, & P. Lieberman (Eds.), Models for the percep-
tion of speech and visual form. Cambridge, MA: MIT Press, 1967.
Haber, R. N. Introduction to R. N. Haber (Ed.), Information
New York: Holt, Rinehart & Winston, 1969.
Humphrey, N. K. Seeing and nothingness. New Scientist,
Ingling, N. Categorization: A mechanism for rapid information processing. Journal of Ex-
perimental Psychology, 1972, 94, 239-243.
Keele, S. W. Attention demands of memory retrieval. Journal of Experimental
1972, 93, 245-248.
Kucera, H., & Francis, W. N. Computational
Providence, RI: Brown Univ. Press, 1967.
Kuhn, T. S. The structure of scientific revolutions.
Laner, S., Morris, P., Oldfield, R. C. A random pattern screen. Quarterly
perimental Psychology, 1957, 9, 105- 108.
Lewis, J. L. Semantic processing of unattended messages under dichotic listening. Journal
of Experimental Psychology,
1970, 85, 225-228.
Mackay, D. G. Aspects of the theory of comprehension, memory and attention. Quarterly
Journal of Experimental Psychology,
1973, 25, 22-40.
Marcel, A. J. Conscious and unconscious perception: Word recognition, visual masking and
an approach to consciousness. Unpublished manuscript, University
Marcel, A. J. Conscious and preconscious recognition of polysemous words: Locating the
selective effects of prior verbal context. In R. S. Nickerson (Ed.), Attention
formance VIII, Hillsdale, NJ: Erlbaum, 1980(b).
Marcel, A. J., & Forrin, B. Naming latency and the repetition of stimulus categories. Jour-
nal of Experimental Psychology, 1974, 103,450-460.
Marcel, A. J., Katz, L., & Smith, M. Laterality and reading proficiency. Neuropsychologia,
1974, 12, 131-139.
Marcel, A. J., & Patterson, K. E. Word recognition and production: Reciprocity in clinical
and normal studies. In J. Requin (Ed.), Attention
Erlbaum, 1978. Vol. 7.
Marr, D. Early processing of visual information. Philosophical
Society, B, 1976, 275, 483-524.
Marshall, J. C., & Newcombe, F. Patterns of paralexia: A psycholinguistic approach. Jour-
nal of Psycholinguistic Research, 1973, 2, 175- 199.
Meyer, D. E., & Schvaneveldt, R. W. Facilitation in recognizing pairs of words: Evidence
of a dependence between retrieval operations. Journal
1971, 90, 227-234.
Journal of Experi-
1981, 110, 341-362.
evidence. The Hague: Mouton, 1973.
systems. Boston: Houghton-Mifflin,
New York: Grune & Stratton, 1948.
processing approaches to
1972, 30th March, 682-684.
analysis of present-day American English.
Chicago: Univ. of Chicago Press, 1970.
Journal of Ex-
Transactions of the Royal
of Experimental Psychology.
ANTHONY J. MARCEL
Meyer, D. E., Schvaneveldt, R. W., & Ruddy, M. G. Activation of lexical memory. Paper
presented at the meeting of the Psychonomic Society, St. Louis, MI, November 1972.
Morton, J. Grammar and computation in language behavior. In J. C. Catford (Ed.), Studies
in Language and Language Behaviour. Center for Research in Language and Language
Behavior, Progress Report, No. VI, Ann Arbor: University of Michigan, 1968.
Morton, J. Categories of interference: Verbal mediation and conflict in card sorting. British
Journal of Psychology,
1969, 60, 329-346.
Morton, J. A Functional Model for Memory. In D. A. Norman (Ed.), Models of memory.
New York: Academic Press, 1970.
Morton, J. Word recognition. In J. Morton & J. C. Marshall (Eds.), Psycholinguistic
London: Elek, 1979. Vol. 2. (a)
Morton, J. Some experiments on facilitation in word and picture recognition and their
relevance for the evolution of a theoretical position. In P. A. Kolers, M. E. Wrolstad, &
H. Bouma (Eds.), Processing
Morton, J. Two auditory parallels to deep dyslexia. In M. Coltheart, K. E. Patterson, &
J. C. Marshall (Eds.), Deep dyslexia.
London: Routledge & Kegan Paul, 1980.
Morton, J., & Chambers, S. M. Selective attention to words and colours. Quarter/y
of Experimental Psychology, 1973, 25, 387-397.
Motley, M. T., & Baars, B. J. Semantic bias effects on the outcomes of verbal slips. Cogni-
tion, 1976, 4, 177- 187.
Nisbett, R. E., & Wilson, T. De C. Telling more than we can know: Verbal reports on
mental processes. Psychological Review,
Posner, M. I. Abstraction and the process of recognition. In G. H. Bower & J. T. Spence
(Eds.), The psychology of learning and motivation:
New York: Academic Press, 1969. Vol. 3.
Posner, M. I., & Snyder, C. R. R. Attention and cognitive control. In R. L. Solso (Ed.),
Information processing and cognition: The Loyola symposium. Hillsdale, NJ: Erlbaum,
Postman, L., & Keppel, G. Norms of word association.
Reicher, G. M. Perceptual recognition as a function of stimulus material. Journal of Experi-
1969, 21, 275-280.
Schvaneveldt, R. W., & Meyer, D. E. Retrieval and comparison processes in semantic
memory. In S. Kornblum (Ed.), Attention
Press, 1973. Vol. 4.
Severance, L. J., & Dyer, F. N. Failure of subliminal word presentations to generate inter-
ference to color naming. Journal of Experimental
Seymour, P. H. K. Semantic equivalence of verbal and pictorial displays. In A. Kennedy &
A. Wilkes (Eds.), Studies in long term memory. London: Wiley, 1975.
Shallice, T. Dual functions of consciousness. Psychological
Shiffrin, R. M. The locus and role of attention in memory systems. In P. M. A. Rabbitt & S.
Domic (Eds.), Aftention andperformance.
Sperling, G. Successive approximations to a model for short-term memory. Acta
1967, 27, 285-292.
Thorndike, E. L., & Lorge, I. A teacher’s word book of 30,000 words. New York Teachers
College, Columbia University, 1944.
Turvey, M. T. On peripheral and central processes in vision: Inferences from an infor-
mation-processing analysis of masking with patterned stimuli. Psychological
1973, 80, l-52.
Turvey, M. T. Perspectives in vision: Conception or perception? In D. Duane & M. Rawson
(Eds.), Reading, perception and language.
von Wright, J. M., Anderson, K., & Stenman, U. Generalization of conditioned GSR in
of visible language, I. New York: Plenum, 1979 (b).
1977, 84, 231-259.
Advances in research and theory.
New York/London: Academic
New York: Academic
1973, 101, 186- 189.
Review, 1972, 79, 383-393.
New York: Academic Press, 1975. Vol. 5.
Baltimore, MD: York, 1975.
CONSCIOUSNESS, MASKING, AND WORD RECOGNITION
dichotic listening. In P. M. A. Rabbitt (Ed.), Attention
London: Academic Press, 1975. Vol. 5.
Weber, R-M. First graders’ use of grammatical context in reading. In H. Levin & J. P.
Williams (Eds.), Basic studies on reading. New York: Basic Books, 1970.
Weiskrantz, L., Warrington, E. K., Sanders, M. D., & Marshall, J. Visual capacity in the
hemianopic field following a restricted occipital ablation. Brain,
Werner, H. Microgenesis and aphasia. Journal of Abnormal
Wheeler, D. D. Processes in word recognition. Cognitive
Wickens, D. D. Characteristics of word encoding. In A. W. Melton & E. Martin (Eds.),
Coding processes in human memory. Washington, D.C.: Winston, 1972.
and performance. New York/
1974, 97, 709-728.
and Social Psychology,
1970, 1, 59-85.
(Accepted December 2, 1982)