discussed before the common background of information theory (Nor-
wich, 1991).
Determination of the intensity of the human sensation of taste in most
cases originates from discrimination tests or descriptive rating (by cat-
egories), scoring (on line scales) or ratio scaling (magnitude estimation).
In each of these cases it is words that have to express precisely what
sensations have been perceived. These words have to be found (if they
exist at all) and used by different assessors the same way (Thompson
and MacFie, 1983). Another possible problem arises from the fact that
if notes from complex food items are compared quantitatively, word pro-
files are compared. Possible additional effects in the overall impression
brought by a combination of the parts are left aside (Thompson and Mac-
Fie, 1983).
In “real” foods like fruits and vegetables a complex of stimuli rather
then a single stimulus is present. The effective stimulus in complex food
is not clear. Correlation of instrumental data and sensory values is only
meaningful when the instruments measure the effective stimulus. For in-
stance, in texture research there has been substantial effort to identify
the effective principles, or better, how they can be measured instru-
mentally (Szczesniak and Kahn, 1971; Kapsalis and Moskowitz, 1978;
Bourne, 1983; Brown et al., 1996).
In this context Kroeze (1990) discusses the analytic or synthetic char-
acter of the human senses. In case of mixed taste stimuli he reports ex-
amples of integration of perceptions, a breakdown to a smaller number
of sensing units. The result of these processes can be described by ad-
ditive, weighted additive, enhancing or suppressive or dominant attribute
models (McBride, 1990). In the last case one attribute above a certain
intensity suppresses the perception of another attribute—thus being dom-
inant. McBride and Anderson (1990) emphasize that a there may be no
relationship between the psychophysical functions of individual stimuli
and the mixture psychophysics.
Whatever mechanisms apply, it is not clear whether mixtures can be
analyzed backward to yield single components (or stimuli) or are per-
ceived as a whole percept, a pattern, as emphasized by Gestalt psy-
chology. Kroeze (1990) gives the examples of the drawing of a square
that does not present itself as a set of four lines or the pattern of tastes,
smells and texture that is immediately recognized as cheese. In both
cases the subjects would be able to analyze the percepts to some degree:
into four lines or into several taste, smell and texture components. But
analysis requires attentional effort, which in cases of a tightly organized
The Perception Process 183