- seCtIon tHRee: BoDY AnD WoRLD
Despite the confusion, in the 1970s and ’80s these objections prompted
progress in both research methods and theory. The basic requirement was to
demonstrate a dissociation between two measures: a ‘direct’ measure, taken
to indicate conscious perception, and an ‘indirect’ measure, to indicate uncon-
scious perception. British psychologist Tony Marcel (1983) adapted the method
of semantic priming, in which one word (the prime) influences the response to
a second word (the target). For example, if the prime and the target word are
semantically related (e.g. doctor and nurse), recognition of the target is faster.
Marcel made such primes undetectable by flashing a visual mask immediately
after them, yet semantic priming still occurred. This seemed to mean that peo-
ple’s word recognition was affected by primes they did not see: when asked to
say whether a word had been presented before the target one, or to guess the
masked word, their verbal responses made clear that they were unaware of the
words’ presence. Other kinds of masked priming were also used, but contro-
versy ensued because although some people successfully replicated the effects,
others failed to.
The controversy was resolved when Canadian psychologists James Cheesman
and Philip Merikle (1984, 1986) proposed a distinction between what they called
the ‘objective threshold’ and the ‘subjective threshold’. The objective threshold
is defined as ‘the detection level at which perceptual information is actually dis-
criminated at a chance level’, whereas the subjective threshold is ‘the detection
level at which subjects claim not to be able to discriminate perceptual informa-
tion at better than a chance level’ (1984, p. 391). The latter, naturally, is higher than
the former.
Cheesman and Merikle used a Stroop-priming task in which participants had to
name a colour after being primed with a colour word for different lengths of time.
Congruent colour words reduce reaction time, but incongruent words increase it
(the Stroop effect). The question was whether primes presented so briefly as to
be undetectable would affect reaction times. Cheesman and Merikle measured
participants’ objective threshold using a reliable four-alternative forced-choice
procedure, and their subjective threshold by asking them to judge their own
ability to discriminate the words. They found a priming effect (i.e. evidence for
unconscious perception) when the length of time between the prime and the
mask was below the subjective threshold, but none at all when it was below the
objective threshold.
Their conclusion was that unconscious perception occurs primarily when infor-
mation is presented below the subjective threshold but above the objective
threshold. They were then able to show that previous experiments had confused
the two, with some measuring one and some the other. From this we can con-
clude that the objective threshold really is the level below which stimuli have no
effect of any kind, but there is a level above that at which a stimulus can have an
effect even though the person denies being conscious of it.
Many of these experiments used words as stimuli and implied the possibility of
unconscious semantic analysis. This possibility has been debated for a century
or more, with many arguments about just how much meaning someone can
extract from a stimulus they deny seeing. Psychologist John Kihlstrom concludes
that ‘With respect to subliminal stimulation, the general rule seems to be that
