Colour Illusions
AIC COLOUR '97 - KYOTO
Osvaldo da Pos
The paper deals with a
series of problems concerning general theories of perception and various
concepts of illusions. In studying perception, explications are initially
considered as formal correlations between variables, and this first step is
believed to be the basis of further theoretical interpretations in terms of
mechanisms or processes. One main distinction is postulated between physical /
neuro-physiological variables on the one hand and phenomenological variables on
the other hand, and on this distinction visual illusions are then defined and
classified. A series of colour effects is examined to support the assumption
that phenomenological illusions are
easier to be detected, more evident and surprising than psychophysical illusions, which in turn should not be properly
considered illusions.
1. Introduction.
Perceptual illusions can
be considered as icebergs in the ocean of all perceptual phenomena: they alert
the researcher that there is something of importance to be studied and they
impose a series of questions which hardly can be avoided. The interest for
studying illusions rises in different scientific and artistic fields, for
different reasons: in some cases the most general problems of perception are
the focus of the investigation, while in other cases people are concerned about
how a particular illusion can be practically used in specific situations.
In this study I try to
show that perceptual illusions can have a single very general definition, but
also that a further subdivision can be made which helps the researcher to more
easily recognize them, to more clearly deal with the theoretical implications,
to more effectively use them in practical applications.
Speaking of perceptual
illusions, my argument necessarily starts from considering the phenomenal world
as it appears to all individual observers. The traditional psychological
problem of perception is usually so formulated: "why the world appears as
it appears", and the psychologist's task has been recognized as finding an
explanation, a solution of the problem.
At a very first general
level explanation consists in identifying some regular relationships between
different elements which are correlated with the event to be explained. At this
level explanation has mainly an informational value: the knowledge of a series
of relations is sufficient for knowing the characteristics of the corresponding
event. Therefore one can easily understand why often scientific definitions and
laws insist on the functional dependency of an event on a series of conditions.
The procedure for
explanation requires two logical steps: first the isolation of the relevant
variables, secondly the determination of an appropriate mathematical
relationship between the variables.
There are two different
realms from which relevant variables can be derived: the physical world and the
phenomenal world. When physical or neuro-physiological variables are correlated
with phenomenal variables, one might speak of physical or neuro-physiological
theories; when correlations are detected within phenomenological variables, one
speaks of phenomenological theories. For instance, to explain why some surfaces
look more or less yellowish than others, one can apply either to variations in
the power and frequency of electromagnetic waves impinging the eye, or to
variations in some purely perceptual dimensions like figural belongingness or
colour similarity.
There is no apriori
principle imposing the choice: from the point of view of information, the goal
of an explanation is the description of the necessary and sufficient conditions
for the event, i.e. the formalization of the functions relating the variables
to the event. The best explanation is reached when it is possible to derive a
full knowledge of the event from the knowledge of the hypothetical relevant
variables (in the form: "a perceptual event of a specified kind is given if and only if a series of conditions is
fulfilled in the relevant variable"). Strictly speaking, only the
informational foundation of the knowledge is here involved, and therefore
mathematical models do not presuppose philosophical (ontological) inferences,
like causal relationships. When the correspondence between variables is one to
one the explanation is complete. But often the correspondence is many to one,
and at least in one direction the information is fully determined; if the
correspondence is many to many, although it is still rich in information as it
excludes what is irrelevant, the predictive power of this knowledge is
relatively low. It seems that explanations in psychology need a much more
complex description of the relevant correlated variables than in other fields
of science.
Perception may be
understood to mean just the process (e.g. a cognitive or a neurological
process) or it may be understood to mean the end result of the process, i.e.
the "phenomenal world", the world as we experience it. Mathematical
models are often derived from the phenomenal world but may be used to make
inferences about the possible processes. The use that is made of the models
depends on the theory. There are many reasons for considering mechanisms and
causality relationships as very useful constructs for explaining perceptual
events. Nevertheless, although science has recently achieved gratifying
successes, it seems that the relationships between neuro-physiological
variables and the characteristics of the perceived events are still limited to
rather simple aspects of perception, if any at all (Westheimer, 1990 [1]; Uttal
1996 [2]). Zrenner et al. [3]
p. 204 put it strongly: "Despite much effort, a direct correlation between
classical color metrics and neuro-physiological mechanisms is not yet
possible".
There is a quite animated
debate about the relationships between the physical and the perceptual worlds.
As previously mentioned, causality seems to be more a construct (Masin, 1993
[4]) than a "real" (ontological) process, and therefore it is
difficult to defend the claim that our perceptions are reproductions or
representations of the physical world because they are caused by it. As far as
we know, one can only say that some characteristics of our perception can be
correlated with some variables used by physical science to describe some aspects
of the world; and moreover further linking propositions are needed for
interpreting the formal symbolization (Teller, 1984 [5]). For instance,
relationships between different tones in auditory perception (for instance one
octave) can be correlated with corresponding relationships between frequencies
of molecular vibrations; and relationships between different hues, perceived in
specific conditions, can be correlated with some relationships between
wavelengths of electromagnetic (e.m.) radiations. Nevertheless it would be
wrong to conclude that colours are simply the perceptual aspects of e.m.
radiations, or that sounds are correspondingly the perceptual aspects of
material vibrations (inside a range of frequencies). The famous dilemma about
whether a falling tree in the forest makes any noise (when there is nobody to
hear it) receives a negative answer, and the same happens for the fire in the
same forest which makes no light when nobody is there to see it. As sensations
certainly depend on the particular cortex area activated by the sensory nerves,
if the optic tracts and the auditory nerve were interchanged somewhere in the
brain before reaching the cortex, we would see lights when the ears are
stimulated by molecular vibrations and hear sounds when the eyes are stimulated
by e.m. radiations. Lights and sounds still would keep some structural
invariant relationships with the physical stimuli (for instance two colours
might be correlated with a particular relationship between molecular
vibrations, and correspondingly sounds would be correlated with relationships
between e.m. waves), allowing thus the organism to correctly behave in the
environment ("veridicality of perception", Logvinenko, 1996 [6]). In
this case the falling tree would produce lights and not noises, while the fire
would produce noises instead of light. So the phenomenal characteristics which
distinguish lights from sounds are not derived by physical aspects of the world
but only depend on the anatomical structure of our sensory systems. Lights and
sounds are not so subjective to have nothing to do with the physical aspects of
the world: the two realms are linked by a functional / relational dependency
properly described by a set of formal relations. It remains to verify how
useful as an explanation this description can actually be, especially if it is
compared with other kinds of descriptions, for instance the phenomenological
ones. An example of the priority of phenomenological variables is given by
Rabin et al. (1992 [7]): an illusory Necker cube presented to stimulate S cones
(only chromatic), is still perceived as long as its visibility allows:
according to the authors, phenomenal visibility is the relevant variable,
rather than the particular pathway traversed.
One may notice that the
visual system appears more complex than the auditory system: one reason could
be that e.m. radiations are more effective in signalling environmental
variations than molecular vibrations, especially because of the speed and
directionality of the e.m. waves, and therefore they are predominantly used for
quick and precise motor response allowing the organisms to fit the environment
better. We conclude then that the characteristics of the perceptual world are
characterised by our biological make up rather than by the physical world; they
are not reducible to physical aspects, cannot be questioned by theories and
only phenomenological descriptions are appropriate for them. There are still
relevant relationships between some variables in the first realm and other variables
in the second one, which psychophysics and neuro-physiology try to describe.
According to experimental
phenomenology, a rather old discipline directly derived from Gestalt theory,
our phenomenal world can be described in terms of phenomenal concepts, i.e. of
concepts that can be used to describe the content of our perception. For
instance redness, lightness, shadow, and so on, are perceivable characteristics
used to describe and communicate aspects of our phenomenal world. We know that
unique red, like unique green, blue and yellow, can be defined only on a
perceptual basis: R G B Y are then concepts referring to specific colour
characteristics and how they appear in our phenomenal world. Unfortunately the
same words and symbols (R G B, for instance) are also used to describe
particular e.m. waves or combination of waves. It seems to be a common practice
to use terms referring to one set of variables instead of the other when they
are correlated (of course an ontological implication is often assumed when a
formal correlation is found, presupposing causal relationships, which would
justify the interchange of the two terms, the cause and the effect). In
additive color mixing R G B stand for the corresponding radiations: in this
case, moreover, R G B are not radiations but constructs, i.e. they mean all the
sets of radiations which are correlated with the same apparent colour.
Sometimes we devise some sets of abstract concepts which can be used to explain
our perceptions: for instance, X Y Z are neither physical entities nor directly
perceivable, but are analogous to R G B which are perceivable.
Other examples of such
constructs are: potential energy, electromagnetic field, gravitation on the
physical side, receptive fields, mechanisms,
filling-in, adaptation on the neuro-physiological side, making a hypothesis,
unconscious inference, problem solving on the cognitive side, colour /
illumination complementarity or splitting, belongingness, and figure-ground
organization on the phenomenological side. The usefulness of constructs for
making predictions can be appreciated in that a wider set of relationships are
this way embraced which lead to the formulations of new hypotheses to be
verified (Spillmann & Dresp, 1995 [8]). Sometimes experience itself can
challenge some theoretical constructs. For example there are occasion when the
unconscious inference of cognitive psychology breaks down: transitivity often
does not hold in perception. If the two extremities of a coloured stripe which
looks uniform match the extremities of another coloured stripe in a different
context, this second stripe will not necessarily look also uniform (de Grandis,
Fig. 7-17)
To distinguish between
psychophysical descriptions and phenomenal descriptions the concept of stimulus
can be useful. Stimulus strictly means what can stimulate the sense organ, and
receptors can be excited or inhibited by specific forms of physical energy. As
distance and contact senses identify two main kind of sensory systems (Kardos, 1984
[9]), also the concept of stimulus assumes two traditional, different meanings:
proximal stimulus is the form of energy which effectively reaches the receptors
and acts on them (the molecules of some external body for touch, smell and
taste receptors on the one hand, e.m. waves or vibrations for vision and
hearing receptors on the other), while distant stimulus is considered the
source where the proximal stimulus comes from (for instance a group of
molecules emitting e.m. waves). Because of the tendency of exchanging the terms
for cause and effect (for instance e.m. radiations / light; Liljefors, 1995
[10]) and of considering a formal dependency as causal relationship (Nakayama
et al., 1990 [11]), both the distant stimulus for the distance sense and the
proximal stimulus for the contact senses are frequently considered the
"object" of our perception, what we "perceive". So, on the
basis of a correlational dependency, a coloured object of our perceptual world
is often considered as identical to
the assembly of some molecules; in another view the coloured object is
considered as the representation of
the correlated molecules (Caivano, 1990 [12]. Assuming a more neutral
philosophical position, we can remain safely on a correlational approach, with
some further specifications about the possible asymmetry of the relationships.
If from one set of relationships (for instance in the realm of physical
entities) we derive more information than we do from the other set (for
instance in the realm of perception), we may make inferences about
"real" (ontological) relationships linking the two fields (Burigana,
1996 [13]). But this, although useful for making predictions, seems admissible
only on an hypothetical ground. Psychologists often find that the contrary is
true: an effect like transparency with coloured surfaces is much more
completely described with reference to perceptual variables (colour
similarities) rather than physical variables (radiations; da Pos, 1991).
We are now ready for a
basic distinction between two kinds of perceptual illusions (da Pos, 1997 [14]:
A) on the one side an illusory
perception is given when there are discrepancies between what we perceive (e.g.
redness) and the physical, not perceivable variables ( e.g. wavelength) which
are known to be correlated: I would call these psychophysical illusions. B) on the other side an illusory
perception is given when the discrepancies appear within the phenomenal world,
as, for instance, when the same perceived object appears now with some and now
with contrasting characteristics: I would call these phenomenological
illusions. Most people spontaneously refer to the first kind of illusion when
asked about a personal definition, probably because of the naive identification
of the perceived object with the stimulus. But it is rather evident that
psychophysical illusions do not appear as such unless some more or less deep
knowledge of psychophysical functions is reached. In this case, the illusory
aspect of a percept is still referable to a more complex and exhaustive set of
relationships. So a psychophysical illusion is not a mistake performed by our
perceptual system, but simply involves a relationship with the physical
variables structured in such a way that the behaviour of the animal is
impaired. Perception is then mainly functional for movement responses, or more
widely, for adaptive behaviour in the environment.
Abstract intelligence has
a higher cognitive value than perception, and in humans it allows the detection
and overcoming of illusions, allowing a still better adaptation to the
environment for which perception is inadequate. (Some illusions are just
adaptatively neutral, e.g. brightness / lightness discrepancy, while other
illusions may be seriously dangerous, as when linear acceleration is felt as an
upward acceleration in speedy aeroplanes. Pilots have been known to crash while
making what they feel to be corrections to the aircraft's trim). According to
Boring (1942) , when such illusions are fully understood, they cannot be
considered illusions anymore, because the lack of knowledge which justified the
surprise of their discovery has been overcome. Phenomenal illusions on the
contrary always keep their illusory characteristics as far as the same event
appears (in perceptual sense) under contrasting aspects. Therefore the more
direct is the connection of contrasting aspects with the same event, the stronger is the evidence of an illusion.
2. Classification.
We can then distinguish different
kinds of illusions according to the immediacy of the perceived link between
opposing characteristics and the object which bears them.
Case A. In the visual
field there is no change which could be ascribed to physical variations: the
object in question appears alternatively with different characteristics only
because of some changes on the part of the observer: either the time is just
passing, or the fixation point has changed, or the attitude (global vs
analytic) is different.
Case B. As before the
visual field is still physically unchanged, but there are two objects (supposed
to be identical) which appear different in different contexts. One common way
of ascertaining the identity of two objects (colours, in most our cases) is to
look at them in a reduction situation, i.e. through a reduction screen (Katz,
1935) or better through a black tube. It is possible to have the same results,
at least in some cases (Masin, 1984 [15]) also assuming a proper attitude in a
free vision .
Case C. There is only one
critical object and its characteristics appear to change because either the
object itself is moving or the context has undergone some changes.
In the cases A and C the identity of the object (for
instance it is the same dog, the same stick, the same movement and so on) is
directly perceived, while in B it is known through different sources of
knowledge (I have been told, I produced it, I measured it, and so on).
Case D. The object is
recognised as the same as one perceived some time before, but its
characteristics look different. Remembering a previous perception is involved
here.
Case E. Discrepancies are
not perceived, but there are contrasts between what we know and what we see. A
set of known correlations between physical or neuro-physiological and
perceptual variables are supposed to hold while in a specific situation this
does not happen.
A list of examples
involving colours are given in the following section
Case A.
Differences
connected with spatial arrangement.
as1 - partial transparency: the same object appears
transparent in some parts of it and opaque in others (Metelli et al. 1981 [16],
da Pos 1976 [17], Kozaki et al. 1989 [18]).
as2 - paradoxical transparency: one object appears not
only opaque here and transparent there, but also of different colour (Metzger,
1955 [19])
as3 - the Mach book: surface colours and illumination
appear different as a function of perceived three-dimensional organization
(Mach, 1865)
as4 - Thiéry coloured prism: one region appears either
in shadow or of different surface colour according to the perceived 3D organization
(Gregory, 1994 [20])
as5 - Ambegujas: a four colour and six coplanar areas
display appears of distinctive surface colours and illumination as a function
of many different 3D organizations (Bergström et al., 1996)
as6 - a particular area of the visual field changes
its colour characteristics passing from surface colour to shadow according to
the 3D perceived organization of a cone (Logvinenko, 1994 [21])
as7 - Hermann grid: dark or coloured dots are not
perceived in the fixation area (Spillmann, 1994)
Differences
which appear as time passes (after-effects of all kinds)
at1 - colour desaturate after prolonged observation
Differences
connected with perceptive attitude.
ap1 - the colour and illumination of an object changes
as a function of its belongingness to different figural structure (Gilchrist,
1977)
ap2 - metacontrast
and attention: the disappearance of a first object as a consequence of the
appearance of a second is prevented by particular figural organizations which
depend on attention (Ramachandran, 1995)
ap3 - an analytical attitude reveals aspects of the
objects that contrast with what is perceived when observing them with a global
attitude:
ap3a - Kanizsa triangle and subjective contours
(Kanizsa, 1976)
ap3b - Eherenstein, van Tujil, Varin, neon effects
(Redies et al., 1984 [22], Goda, 1997 [23]).
Case B.
b1 - many contrast examples: the "same
colours" or the "same coloured objects" appear different because
placed in different coloured contexts.
b1a - For instance the two identical grey triangles in
the Benáry cross appear different, despite locally identical contexts, because
the contextual difference is in figural belongingness (Benáry, 1924)
b1b - the red and blue squares in the cover of the
book by W.D. Wright (1958 [24]) appear different because they are perceived as
lying over differently coloured figures
b1c - the White's effect: almost the same as in
Wright's example but in the achromatic domain (although discovered as a
"new effect" about 20 years later, White 1979)
b1c - an analogous chromatic example displayed by
Hesselgren (1987 [25]) on the cover of his book "On Architecture".
b1d - some more recent examples by Adelson in the
achromatic range (Adelson, 1993 [26]) and by Logvinenko (1997 [27]) in the
chromatic one: in these cases the contexts of the two associated coloured
regions differ not only in surface colour but also in perceived illumination
(shadow), which makes the apparent colour difference much more evident.
b2 - many assimilation examples: as before,
differences are perceived when the "same colours" are located in
different colour contexts. One of the most famous example is taken from Bezold
in the chromatic domain (von Bezold, 1874), and from Helson (1959 [28]) in the
achromatic domain, but there are interesting examples also from Kanizsa (1957
[29]).
Case C.
The
object moves
co1 - subjective colours: since Fechner, Benham, and
Bidwell either black and white displays or coloured disks are known to change
their colour depending on whether or not they move, and on the direction of
their movement (Spillmann, 1990). Intermittent lights produce analogous effects
(Violetteffekt, Welpe, 1978)
co2 - colour from motion: a couple of effects are
referred to here which show a colour spreading dependent on the perception of
motion (Cicerone et al. 1995 [30]; Zanforlin, 1996 [31]) . A common feature is
the appearance of 3D organization.
co3 - fluttering heart (Grünau von, 1976) and MacKay (1958) illusory
displacement: a figure appears moving in relation to the background on which it
is fixed when both are moved
co3 - red-blue moving slit illusion (Mollon &
Polden, 1975): the blue part of a slit is seen moving behind the collinear red
part when both are rigidly shifted.
The
object is constant and context is changed:
cc1 - the colour of a hole appears to change when the
surrounding surface is differently illuminated (Hess & Pretori, 1884, 1970 [32])
cc2 - illusions dependent on spatial frequency:
particular changes in the object, for instance the appearance of the neon
effect (Redies et al., 1984) and spatial colour fusion (Cornsweet, 1970), are
perceived when the observer approaches or goes away from the object.
cc3 - Purkinje effect: if the illumination of a scene
is decreased or increased, the lightnesses of some objects appear mutually
inverted (Purkinje, 1825).
cc4 - many aspects of the observed objects undergo
more or less drastic changes when seen in reduction or in isolation (i.e.
either through a holed screen or a black tube)
Case D
The identity of the object whose
characteristics are perceived conflicting is given by memory.
d1 - in the moon
illusion, the surprise about its different sizes depends on our believing that
it is the same moon.
Case E
e1 - Land retinex theory: perceived surface colours do
not always correspond to the e.m. radiations reaching the eyes from the
corresponding areas (Land, 1977 [33]).
e2 - Abney effect: changes in colorimetric purity are
correlated with changes not only in saturation but also in hue (Abney, 1910)
e3 - Bezold-Brücke shift (Purdy, 1931): keeping
constant the spectral composition of some radiations and changing only their
power, observers notice changes not only in the intensity of the light, but
also in its hue.
e4 - Helmholtz-Boswell-Kohlrausch effect: the
brightness of saturated colours appears overestimated as compared with
expectations based on their luminance (Nayatani, 1994)
e5 - Helson-Judd effect: under monochromatic
illumination neutral surfaces appear either of the illuminant or of the
complementary colour depending on their reflectance (Evans, 1948)
One illusory phenomenon can belong in
different categories. In agreement with Boring's suggestions, the effects of case
E should not be considered illusions: after we have improved our knowledge
about them, while illusions of cases A-C hold their deceptive appearance.
3. Concluding remarks.
What have I
proposed that is new? Where is its usefulness? First of all, not very much is
new in the arguments which have been put forward to support the idea that
illusions primarily deal with phenomenal discrepancies (Reynolds, 1988), and
only secondarily with physical / psychological conflicts. In fact the history
of psychology (Henle, 1976 [34]) shows that a consistent contribution has
always been given in this direction, since Hering's phenomenological
theorization on colours, Gestalt Psychology, and experimental phenomenology
(Bozzi, 1989 [35]; Burigana, 1996 [13]).
The main point has been to stress the
elementary and logical importance of correlations between variables as tools
for explaining our phenomenal world; and then the legitimation of the
precedence of the phenomenal variables over the physical and neuro-physiological
ones, especially from the point of view of informational and explicative power.
I can see two reasons for the
usefulness of the phenomenological approach to illusions: as surprise is
considered to be strictly connected with the discovery of discrepancies, the
more direct the perception of these discrepancies the more joyful is the
resulting surprise (except for dangerous situations!). This might be the reason
why illusions can be used effectively in art. On the other hand, once an
illusion has been perceptually detected, further explanations are urged to
establish a larger and more consistent network of relationships within which
the event can be arranged, and this is why illusions are so widely treated by
psychologists.
4. Bibliography.
For reason of space,
references not listed here can be found in: da Pos O., Zambianchi E., Visual Illusions and Effects. A Collection. Illusioni ed effetti visivi. Una raccolta. (English - Italian), Guerini Studio,
Milano, 1996.
[1] Westheimer G., Relating
neural mechanisms to visual perception. Historical and philosophical
considerations. In: L. Spillmann and J.S. Werner (Eds.), Visual perception.
The neuro-physiological foundations. Academic Press, San Diego CA, 1990, 5-9.
[2] Uttal W.R., Do
bridges exist between psychophysics and neurophysiology? In: Masin S.C.
(Ed.), Fechner Day 96, Cleup, Padua, 1996, 1-21.
[3] Zrenner E., Abramov I., Akita M., Cowey A.,
Livingstone M., Valberg A., Color
perception. Retina to cortex. In: L. Spillmann and J.S. Werner (Eds.), Visual perception. The
neuro-physiological foundations. Academic Press, San Diego CA, 1990, 163-204..
[4] Masin S.C., Some
philosophical observations on perceptual science, In: Masin S.C. (Ed.),
Foundation of Perceptual Theory, Elsevier Science Publisher, 1993, 43-73.
[5] Teller D.Y., Linking
propositions. Vision Research, 1984, 24, 1233-1246.
[6] Logvinenko A., Invariant
relationships between achromatic colour, apparent illumination and shape of
surface: implications for the colour constancy. In: C.Dickinson, I.Murray
and D.Carden (Eds.), John Dalton's colour vision legacy, Taylor &
Francis,1996, 591-597.
[7] Rabin J., Adams A.J., Switkes E., Perceptual ambiguity and the short
wavelength sensitive visual path. Vision Research, 1992, 32, 399-401.
[8] Spillmann L., Dresp B., Phenomena of illusory form: can we bridge the gap between levels of
explanation?, Perception, 1995, 24, 1333-1364.
[9] Kardos L., The
origin of neuro-psychological information, Akadémiai Kiadó, Budapest, 1984.
[10] Liljefors A.,
Redefine the term "Light"!, Proceedings of the 23rd Session of
the CIE, New Delhi, India, Vol. 1, 1995.
[11] Nakayama K., Shimojo S., Ramachandran V.S., Transparency: relation to depth, subjective
contours, luminance, and neon color spreading. Perception, 1990, 19, 497-513.
[12] Caivano J.L., Symbolicity in elementary visual signs. In: Haworth K., Deely J.,
Prewitt T. (Eds.), "Semiotics", with Bernard J., Deely J., Voigt V.,
Withalm G. (Eds.), "Symbolicity", University Press of America,
Lanham, 1990, 46-55.
[13]
Burigana L., Singolaritŕ della visione,
Upsel Domeneghini Editore, Padova, 1996.
[14] da Pos O., Visual
illusions and effects. In: Sivik L., Colour and Psychology, Scandinavian
Colour Institute, Stokholm, 1997.
[15] Masin S.C., Quarta A., Experimental demonstration that observers produce unbiased estimates of
reduction lightness in transparent surfaces. Bulletin of the Psychonomic
Society, 1984, 22, 529-530.
[16] Metelli
F., Masin S.C., Manganelli A.M., Partial
transparency, Atti dell'Accademia Patavina di Scienze Lettere ed Arti,
1981, 92, 115-169.
[17] da Pos
O., Contributo teorico-sperimentale alla
percezione della trasparenza parziale con colori. Analisi di un effetto
paradossale osservato da W.Metzger. Atti dell'Istituto Veneto di Scienze
Lettere ed Arti, 1977, 135, 47-70.
[18] Kozaki T., Fukuda M., Nakano Y., Masuda N., Phenomenal transparency and other related
phenomena, Hiyoshi Review of Natural Science (Keio University), 1989, 6,
68-81.
[19] Metzger W. Über Durchsichtigkeits-Erscheinungen
(Vorlaüfige Mitteilung). Rivista di
Psicologia, Fascicolo Giubilare, 1955, 49, 187-189.
[20] Gregory R.L., Even
odder perceptions. Routledge, New York, 1994.
[21] Logvinenko A., Menshikova G., Trade-off between achromatic colour and
perceived illumination as revealed by the use of psuedoscopic inversion of
apparent depth. Perception, 1994, 23, 1007-1023.
[22] Redies C., Spillmann L., Kunz K., Coloured neon flanks and line gap
enhancement. Vision Research, 1984, 24, 1301-1309.
[23] Goda N., Ejima Y. Additive effect of luminance and color cues in generation of neon color
spreading, Vision Research, 1997, 37, 291-305.
[24] Wright W.D., The
measurement of colour. Hilger and Watts, London, 1958.
[25] Hesselgren S.,
On Architecture. Studentlitteratur, Chartwell-Bratt, Bromley, 1987.
[26] Adelson E.H., Perceptual
organization and the judgment of brightness, Science, 1993, 262, 2042-2044.
[27] Logvinenko, Effect
of perceptual context on colour appearance, paper presented at the AIC
COLOR '97 8th Congress, Kyoto, 1997.
[28] Helson H., Rohles H. jr., A quantitative study of the reversal of classical lightness-contrast.
American Journal of Psychology, 1959, 72, 530-538.
[29] Kanizsa G., Gradient
marginal et perception chromatique. In: I.Meyerson (Ed.), Problčmes de la
couleur, Paris, SEVPEN, 1957, 107-114.
[30] Cicerone C.M., Hoffman D.D., Gowdy P.D.,
Kim J.S., The perception of color from
motion. Perception &
Psychophysics, 1995, 57, 761-777.
[31] Zanforlin
M., Trasparenza anomala nei fenomeni
stereocinetici (Anomalous transparency in stereokinetic phenomena). In:
Bortolatto L., da Pos O. (Eds.), Effetto Trasparenza, Le Venezie, Treviso,
1996, 40-41.
[32] Hess C., Pretori H., Quantitative investigation of the lawfulness of simultaneous brightness
contrast (Flocks H., Tenny J.H., trans.), Perceptual and Motor Skills,
1970, 31, 947-969. (Original work published 1884).
[33] Land E.H., The
retinex theory of color vision. Scientific American , 1977, 237, 108-128.
[34] Henle M., Why
study the history of psychology?, Annals of the New York Academy of
Sciences, 1976, 270, 14-20.
[35]
Bozzi P., Fenomenologia Sperimentale,
Il Mulino, Bologna, 1989.
Osvaldo DA POS
Department of General
Psychology
Via Venezia 8 - 35131 Padova Italy
E-mail: dapos@psico.unipd.it
Tel: +39 49 8276680 - Fax: 8276600