Person. indirld. LX’ Vol. IO. No. 2, pp. 185-189. 1989 0191-8869s89 53.00 +O.

OO
Pnntcd in Great Bntain. All nghts reserved Copyright C 1989 Pergamon Press plc

PRIMING, WORD RECOGNITION AND

PSYCHOTIC TENDENCIES

M. FISHERand J. WEINMAN
Unit of Psychology, United Medical Schools of Guy’s and St Thomas’s Hospital, London Bridge,
London SE1 9RT. England

(Receiced 15 Ocrober 1987; received for publication 25 May 1988)

Summary-Frith (1979) proposed a model for the cognitive deficit underlying schizophrenia, suggesting
a defect in inhibitory selective mechanisms which limit the contents of consciousness. Using a semantic
priming paradigm to measure the operation of inhibitory mechanisms, it was predicted that, if psychosis
is conceptualised as a pathological exaggeration of the features measured in scales of ‘psychoticism’ or
‘schizotypy’, then high scorers on these scales would show a reduced inhibition effect. No such correlation
was obtained; a result in apparent opposition to that of Bullen and Hemsley (1984). though an alternative
explanation for the disparity is suggested. This data may be of relevance in assessing the validity of
schizoid tendency scales, adding weight to the assertion of Raine (Br. J. clin. Psychol. 26, 305-310, 1977)
that such scales measure different components of personality rather than the same underlying dimension.

INTRODUCTION

Recent theories of the nature of the cognitive functioning in schizophrenia have provided particular
predictions about deficits in word processing. For example, Frith (1979) has proposed that
schizophrenics may be more likely to be aware of the multiple meanings of words particularly if
these are ambiguous. Frith (1979) has argued that in schizophrenics there is a problem with the
mechanism which controls and limits the contents of consciousness and that there is an increased
awareness of products of normally pre-conscious processes. Thus he has proposed that schizo-
phrenics “should fail to inhibit alternative meanings of ambiguous words”.
A large proportion of words in the English language are ambiguous and studies have shown that
it takes longer to process sentences containing these words (Schvaneveldt, Meyer and Becker, 1976).
Moreover the recognition of single words can be either facilitated or inhibited by altering the
context of an ambiguous prime, presented immediately beforehand (Marcel, 1980). Thus ‘river’ is
recognised more rapidly when preceded by the words ‘shore-bank’ whereas it takes longer if it
follows ‘money-bank’.
Thus, if Frith’s hypothesis is correct and if schizophrenics fail to inhibit alternative meanings
of ambiguous words, then the recognition of ‘river’ should not be slowed if it is preceded by
‘money-bank’. Limited support for this has been provided by Bullen and Hemsley (1987) who
found a non-significant weaker inhibition effect in schizophrenics as compared with normal and
psychiatric controls.
In an earlier study along similar lines Bullen and Hemsley (1984) provided further support for
Frith’s hypothesis by relating the inhbition effect to psychoticism as measured by the Eysenck
Personality Questionnaire (Eysenck and Eysenck, 1975) in a normal population. This study was
premised on the notion that psychoticism is a dimension and that questionnaires can measure the
degree of psychoticism in a normal population (e.g. Eysenck and Eysenck, 1975; Claridge and
Broks, 1984). Support for the validity of such scales has emerged from both psychophysiological
and cognitive measures, in addition to evidence demonstrating elevated scores in schizophrenic
patients and antisocial individuals ‘bordering on schizophrenia’ (see Raine, 1987, for a review). For
example, Beech and Claridge (1987) provide evidence of a reduced inhibition effect on the Stroop
test in individuals who were high scorers on the schizotypy scale of the Schizotypal Personality
Questionnaire (Claridge and Broks, 1984).
The present study aims to replicate and extend Bullen and Hemsley’s (1984) study since this only
used a dimensional measure of psychoticism in a small sample and also failed to demonstrate a
facilitation effect in word priming. Thus this study incorporates two different measures of psychotic
tendencies in a larger normal sample and relates scores on these to the degree of the inhibition effect
I85
186 M. FISHER and J. WEISSIAS

on the word-recognition task developed by Marcel (1980) and used by Bullen and Hemsley (1984).
but modified to deal with the methodological problems inherent in their study.

METHOD
Subjects
Twenty-two subjects were tested-10 female. 12 male: age range 19-22 yr, with a mean of 20.6 yr.
They were an unselected group of medical, dental, optical and nursing students. with no more than
a basic education in psychology.

Apparatus and materials

A three field tachistoscope was employed for presenting the stimuli. The first two fields were set
at 40% intensity, and the third (target) field at 5% (after Bullen and Hemsley, 1987). Four timers
were used.
The stimulus cards consisted of white index cards, 6 x 4 in. Each word was positioned in the
centre of a card, letter height 2.5 mm.
The following tests were also used:

(i) Eysenck Personality Questionnaire EPQ; (Eysenck and Eysenck, 1975) which provides a
measure of psychoticism together with measures of extraversion and neuroticism.
(ii) Schizotypal Personality Questionnaire (STQ; Claridge and Broks, 1984). which provides
separate measures of schizotypy (STA) and of borderline personality (STB).
(iii) Mill HiN Vocabularly Test-Form 1 (Senior, 1977 revision)-only the synonyms test was
used to provide a measure of verbal ability.

Stimuli
A list of 36 ambiguous words was taken from those compiled by Schvaneveldt et al. (1976), which
had been selected according to the following criteria. Each word had to have two distinct and
unrelated meanings, both of which should be relatively common interpretations of the word with
the same spelling and pronunciation. Two words related to each meaning of each ambiguous word
were selected (again using Schvaneveldt et al., 1976), with each word being six letters or less, and
in common use. The choice of meaning for the target word was chosen randomly and the priming
stimuli selected accordingly (after Bullen and Hemsley. 1987).

Temporal parameters
The temporal relationship between the stimuli was designed to provide the optimal conditions
for the operation of priming effects (both facilitation and inhibition) based on previous research
in the field (e.g. Schvaneveldt et al., 1976; Neely, 1977; Bullen and Hemsley, 1987). Thus the
sequence of stimuli(s) and interstimulus intervals was as follows: Sl, 500 msec; 0; S2, 500 msec;
200msec; S3, variable; Response, 8-10 set; Sl . . . The interval between each trial was approx.
10 sec.

Design
In each trial, the subjects were presented with a sequence of three words. The first two words
were priming stimuli, the second word always being ambiguous, and were presented at fixed
durations above threshold. The third word in the sequence was designated as the ‘target’ word,
which the subject was required to name, and was presented at a level just below the subject’s
threshold range (individually determined). If the subject failed to name the target word, the
sequence of words was repeated. The number of presentations required for identification was the
main dependent variable.
The task-related independent variable was the nature of the semantic relationship between the
priming stimuli and the target stimulus. There were three conditions:
(a) Concordant (C): the first priming stimulus and the third stimulus (target) were both related
to the same meaning of the central, ambiguous word, e.g. SHORE-BANK-RIVER
(b) Discordant (D): the first priming stimulus and the target word were related to separate
meanings of the central, ambiguous word, e.g. MONEY-BANK-RIVER
 Word priming and psychoticism 187

(c) Neutral (N): the first priming stimulus consisted of a series of Xs, and the target stimulus
was related to one meaning of the central, ambiguous word, e.g. XXXXX-BANK-RIVER
Order of presentation was randomised and error data (i.e. a subject giving an incorrect word
as a response) was treated as missing data in the analysis.

Procedure
Subjects were told that the tests were designed to investigate speed-reading ability, and its
relationship to various personality dimensions. Initially, a block of practice trials was presented
to help the subject understand the nature of the task, and for the experimenter to determine the
subject’s threshold range.
The experimental trials were then presented in three blocks with a break between each. Each trial
consisted of the presentation of a triad of words as described above. The initial value of the
exposure duration of the target was set from the practice trials for each individual subject to be
just below that subject’s threshold level. The subjects were reminded at intervals to take note of
the first two words because they could act as clues to the identity of the target.
The Mill Hill, EPQ, and STQ were completed in the breaks between trials. At the end of the
experiment, the subjects were assured of anonymity and confidence of their questionnaire replies,
and were fully debriefed.

RESULTS
Priming effects
An ANOVA was used to analyse the effect of the experimental conditions on the ease of
recognition of the target words. The relevant data are presented in Table 1. The three experimental
conditions were found to have significantly different effects on the ease of recognition of target
words (F = 15.42; d.f. 2, 21; P < 0.001).
The data were split up and paired t-tests applied to determine the possible significance between
pairs of conditions. Differences between D and N (t = 2.19; P < 0.05; d.f. = 21) between N and
C (t = 3.39; P < 0.01; d.f. = 21) and between D and C (t = 5.81; P < 0.001; d.f. = 21) were all
significant.
This study has therefore demonstrated both a facilitation and an inhibition effect.

Correlations between the test scores and the word processing measures
The mean scores on the Mill Hill, EPQ and STQ are shown in Table 2. A correlation matrix
was derived to ascertain whether there was any relation between these scores and the inhibition
effect on the word recognition task. The determination of this inhibition effect raised a problem.
Bullen and Hemsley (1984) used the difference between the mean scores in discordant and neutral
trials (i.e. D - N) as their measure of the inhibition effect. This choice was presumably influenced
by their failure to demonstrate a facilitation effect. The labelling of the third experimental condition
as ‘neutral’ might be open to criticism, since the use of ‘XXX’ rather than a word prime could be

Table I. Means (with standard deviations) for the number of

presentations required for accurate recognition in each experimental
condition
Experimental condition
Concordant (C) Neutral (N) Discordant (D)
Mean 1.71 2.33 2.82
SD 0.61 0.72 0.91

Table 2. Means and standard deviations for scores obtained on the Mill Hill. the EPQ and the STQ

EPQ srQ
Mill E N L P STA STB
Hill (Extraversion) (Neuroticism) (“Lie” Scale) (Psychoticism) Schizotypy Borderline

R Il.2 13.5 II.8 3.9 2.7 12.6 5.8

SD 2.6 3.7 5.3 3.9 2.4 6.0 4.3
188 M. FISHER and J. WEISMAN

Table 3. A correlation matrix to determme anv relationshm between ouestionnaire ratines_ and onmine
. _ effects
~~--
Mill Hill I
Extraversion (E) -0.31 I
Neuroticism (N) 0.0 I -0.18 I
“Lx” score (L) -0.3 0 I6 -0.2 I
Psychoticism (P) -0.13 0.14 -0.05 0.01 I
Schizotypy (STA) 0.03 0.23 0.64*** -0.32 - 0.26 I
Borderline (STB) 0.12 0.08 0.65”. -0.43. 0.01 0.6 I I
D-C 0.06 -0 I9 0.05 -0.33 0.03 -0.03 0.05 I
D-N 0.1 I -0.18 0.05 -0.44. -0.06 -0.04 0.0 0.64’** I
N-C -0.03 0.0 -0.07 0.29 0.09 .- 0.05 0.01 0.27 -0.55*** I
lf c: 0.05; l*P < 0.01; l**p < 0.005.

considered incomparable. Therefore, perhaps the difference between the mean scores in concordant
and discordant trials (i.e. D - C) would be a more valid measure of the inhibition effect. However,
this could well be confounding the facilitation and inhibition effects, and could therefore be open
to criticism. Consequently it was decided that both measures would be used in the matrix and the
results are shown in Table 3.
From the correlation matrix it can be seen that no significant correlation was found between
the Mill Hill and either of the two measures of priming. Moreover none of the three scales
measuring psychotic tendency (i.e. P, STA, STB) showed a significant correlation with the priming
measures. Only the L scale of the EPQ correlated significantly with one of the priming measures
(D - N: r = 0.44; P < 0.05).
As expected, a significant correlation was found between the STA and STB scales. The P score
was uncorrelated with the STQ scores whereas the N score was significantly correlated with both
STA and STB.

DISCUSSION

The findings not only demonstrate an inhibition effect, as found by Bullen and Hemsley (1984)
and Bullen, Hemsley and Dixon (1987), but also a facilitation effect, which neither of these studies
was able to do. However, in contrast with Bullen and Hemsley (1984) no relation was found
between the inhibition effect and the measures of schizoid tendency. It could therefore be that, in
successfully obtaining a facilitation effect, the correlation was lost. If, as Bullen and Hemsley (1984)
suggest, P may measure “unconventionality”, then non-compliance by high P-scorers to paying
attention to the priming words would produce the predicted negative correlation with the inhibition
effect. However, if such subjects perceived that the primer might produce a facilitation effect, as
occurred in this study, then compliance would be rewarded, inhibition achieved, and the correlation
would be lost. Such an explanation is clearly tentative, but may be of value in attempting to align
contradictory results from two similar experiments. Further work could therefore manipulate the
design to assess compliance with instructions, for example, by altering the ratio of C to D trials
and hence making compliance more or less favourable (Bullen, personal communication).
That the relationship betwen psychoticism and inhibition is not as clear-cut as it originally
appeared from Bullen and Hemsley (1984) is further supported by other studies employing different
experimental paradigms. Beech and Claridge (1987), using a Stroop task, produced a significant
negative correlation between inhibition and the STA scale of the STQ. However, there was no such
correlation for the STB and the EPQ ‘P’ scale was in fact positively, though non-significantly,
correlated with inhibition. Bullen et al. (1987), using a matching-task paradigm, obtained an
inhibition effect negatively correlated with the Launay and Slade (1981) “predisposition to
hallucinations” score, but which did not correlate with any dimension from either the STQ or EPQ.
In the present study there was no correlation between the EPQ psychoticism scale and the two
STQ scales, even though the two subscales of the STQ showed appropriate intercorrelations. This
lack of correlation between the EPQ P scale and the STQ scales has also been reported by Rawlings
(1983) and by Claridge, Robinson and Birchall (I 983). These findings add weight to the assertion
of Bullen et al. (1987) and Raine (1987) that schizoid/psychotic tendencies are multifactorial aspects
of personality and that the different scales are measuring different components rather than the same
underlying dimension.
 Word priming and psychoticism 189

Clearly, further experimentation is required in assessing the validity of the various schizoid
tendency scales and their pertinence to perceptual and attentional anomalies before studies on
normal populations can be used to test models of the cognitive deficit in schizophrenia.

REFERENCES
Beech A. and Claridge G. (1987) Individual differences in negative priming: relations with schizotypal personality traits.
Br. J. Psychol. 78, 349-356.
Bullen J. (1982) Effects of visual priming on schizophrenics’ word recognition performance. Unpublished M.Phil. Thesis,
University of London.
Bullen J. and Hemsley D. (1984) Psychoticism and visual recognition thresholds. Person. indiuid. Drs 5, 735-739.
Bullen J. and Hemsley D. (1987) Schizophrenia: a failure to control the contents of consciousness? Br. J. clin. Psycho/. 26,
25-33.
Bullen J., Hemsley D. and Dixon N. (1987) Inhibition, unusual perceptual experiences and psychoticism. Person. individ.
Dlf. 8, 687-692.
Claridge G. and Broks P. (1984) Schizotypy and hemisphere function-l; Theoretical considerations and the measurement
of schizotypy. Person. indiuid. D# 5, 633-648.
Claridge G., Robinson D. and Birchall P. (1983) Characteristics of schizophrenics and neurotic relatives. Person. individ.
D@ 4, 651-644.
Eysenck H. and Eysenck S. (1975) Manual of the Eysenck Personality Questionnaire (Junior and Adult). Hodder &
Stoughton. London.
Frith C. (1979) Consciousness, information processing, and schizophrenia. Br. J. Psychiat. 134, 225-235.
Launay G. and Slade P. (1981) The measurement of hallucinating predisposition in male and female prisoners. Person.
individ. D# 2, 221-234.
Marcel A. (1980) Conscious and preconscious recognition of polysemos words: locating the selective effects of prior verbal
context. In Attention and Performance Vol. 8 (Edited by Nickerson R. S.). Erlbaum, Hillsdale, N.J.
Meyer D., Schvaneveldt R. and Ruddy M. (1975) Loci of contextual effects on visual word recognition. In Atfenrion and
Performance, Vol. 5 (Edited by Rabbitt P. M. A. and Dormic S.). Academic Press, New York.
Neely J. (1977) Semantic priming and retrieval from lexical memory: roles of inhibitionless spreading activation and limited
capacity attention. J. exp. Psychol.. Gen. 106, 226-254.
Posner M. and Snyder C. (1975) Facilitation and inhibition in the processing of signals. In Attention and Performance, Vol.
5 (Edited by Rabbitt P. M. A. and Domic S.). Academic Press, New York.
Raine A. (1987) Validation of schizoid personality scales using indices of schizotypal and borderline personality disorder
in a criminal population. Br. J. clin. Psycho/. 26,_305-310.
Rawlings D. (1983) An enquiry into the nature of psychoticism as a personality dimension. D. Phil. Thesis, University of
Oxford.
Schvaneveldt R., Meyer D. and Becker C. (1976) Lexical ambiguity semantic context and visual word recognition. J. exp.
Psychol., Hum. Percept. Perform. 2, 243-256.