Semiotica 200 (2014), 1-20.
In this study, we construct a general semiotic model representing the physiological and mental processes involved in visual perception. The model has the form of an ordered typology of perceptual processes, where the ordering is determined by the Peircean categories of thought. This method of semiotic modeling is then applied to the specific mental processes involved in the perception of camouflage and visual illusions. The typology of visual illusions includes a unified explanation for the perception of static visual illusions that is based on semiotic principles.
Keywords: visual signs; Peircean categories; semiotic union; visual phenomenology; visual illusions; camouflage.
In this study, we construct a conceptual model of visual perception that is based on semiotic principles. Visual perception is a mental process that involves the formation and interpretation of mental images. Physical, physiological and behavioral correlates of this process are the objects of scientific study. The experience of visual perception is the object of study in semiotic phenomenology (“On phenomenology,” EP2: 147). We define a semiotic model to be a mental process that is represented in language and that depends upon the action of mental signs (Cantor 2009, on semiotic modeling). Following Charles S. Peirce, we define a mental sign to be a triadic relation among a visual image, a recollected image and an interpretation based on learned conventions, i.e., a representamen, an object and an interpretant (cf. Cantor 2011). We make use of the fact that any number of signs with representamina in a visual scene may be combined to form a new sign by the mental operation of semiotic union (Cantor 2003). The process of semiotic union produces a compound sign for which the representamen consists of coexistent images, the object consists of concurrent recollections and the interpretant is the sentential conjunction of the interpretants of its component signs. Our semiotic model is in the form of an ordered typology of perceptual processes, where the ordering is determined by the Peircean categories of thought (“On a new list of categories” EP1:1-10). We interpret the Peircean categories of thought to be Attribution (1st Category), Opposition (2nd Category) and Mediation (3rd Category). The categorically ordered phases in the process of visual perception are taken to be Presentation, Representation and Valuation, where Presentation precedes Representation and Representation precedes Valuation. After constructing a general model of visual perception from these processes, we present specific models for the perceptual phenomena of camouflage (visual deception) and visual illusions. Empirical information on the process of visual perception may be found in Palmer (1999). The vision science that is involved is presented from a semiotic perspective in Cantor (2012). Phenomena that are explained by our semiotic model of camouflage are listed in Behrens (2009). Phenomena that are explained by our semiotic model of illusions are described in Gregory (2009). In the next section, we begin the construction of a conceptual model of visual perception with the process of visual presentation.
The preretinal process of visual presentation may be subdivided into categorical phases of Reflection, Selection and Projection. Light reflected from the surfaces of objects is a carrier of information about their physical attributes. Light reflected from objects is collected by the eyes following orientation of the body, head and eyes. These movements establish an opposition between a selected region of the visual world and the retinal surfaces of the viewer. After its selection, ambient light is processed by the cornea/lens system to form a preretinal optic image within the chamber of the eye that is presented to the retina as a perspective projection. This virtual image mediates between the selected region of the visual world and the brain of the viewer.
The initial event in the representation process is the formation of a retinal image in each eye. A retinal image is a two-dimensional distribution of electrochemical signals produced by the retinal photoreceptors. These physiological images generated in both eyes are subsequently processed, combined and perceived as a single stereoscopic image of world, referred to as the visual scene. The perceptual process of visual representation may be subdivided into three categorical phases which are referred to as Segmentation, Figuration and Organization.
In natural vision, color perception is obligatory. Hence, the visual scene is perceived as subdivided into colored regions with boundaries determined by color contrast. This perceptual process is termed segmentation. A categorical subdivision of the process of segmentation entails the attribution of color to regions in the visual scene, the opposition of contiguous colored regions and mediation between contiguous regions by a common boundary.
Figure/ground segregation or figuration is the process by which a colored region in a visual scene is distinguished as a figure that represents a unitary object in the visual world. Since figuration establishes an opposition between a figure and its background, which is interpreted as an opposition between a physical object and its spatial complement, it is a 2nd Category mental process. Categorical principles of figuration are Saliency, Surroundedness and Familiarity (cf. Palmer 1999:282).
3.2.1. Saliency. In this study, saliency refers to a significant difference between the visual qualities of a figure and its ground, e.g., a difference in color, size, shape or motion. Such percepts depend upon prevailing viewing conditions as well as perceptual judgments of the viewer. Specific signs of saliency that affect figure/ground segregation include: color contrast between a figure and its ground, relative size with figure smaller than ground, relative shape with local convexity of the figure and concavity of its contiguous ground and movement of a figure relative to its ground (cf. Palmer 1999:282). Since saliency entails attribution, it is a 1st Category principle of figuration.
3.2.2. Surroundedness. Visual figures are produced by light reflected from the surfaces of opaque objects that are nearest to the viewer. Hence, surroundedness is produced by obstruction (occlusion) of the line of sight. Surroundedness of a figure relative to the contiguous region of its ground may be complete or incomplete. Since surroundedness entails an opposition between the outside and inside of a figure, it is a 2nd Category principle of figuration.
3.2.3. Familiarity. By familiarity, we mean a visual memory produced by either a single salient visual experience or by repetition of a visual experience. The interpretation of signs of figuration depends upon familiarity of the viewer with corresponding objects of similar shape, orientation or movement. Familiar features of figures that signify their biological origin, and therefore have particular significance for humans, are bilateral symmetry, vertical or horizontal orientation, parallelism and movement (cf. Palmer 1999:282). Since familiarity mediates between present and past awareness, it is a 3rd Category principle of figuration.
Visual organization is the perception of a unitary figure that is determined by the perception of relations between subordinate figures in a visual scene. The unitary figure resulting from visual organization is termed a configuration or Gestalt. The process of visual organization is also referred to as grouping. Visual configurations may be perceived as static or moving. The categorical principles of visual organization are the same as the categorical principles of mental association, i.e., Similarity, Contiguity and Conventionality (see Cantor 2011, on the categorical principles of mental association).
3.3.1. Similarity. Visual organization that is based on the principle of similarity entails a commonality of attributes, e.g., a sameness of color, size, shape, orientation or motion of the figures that constitute a group. Similarity in this sense includes the classical Gestalt principles of ‘similarity,’ and ‘parallelism’ for static configurations and ‘common fate’ for moving configurations (cf. Palmer 1999:258). The classical Gestalt principles of ‘symmetry,’ ‘continuation’ and ‘completion’ (cf. Palmer 1999:259) are based on self-similarity. Since similarity entails a sameness of attributes, it is a 1st Category principle of visual organization.
3.3.2. Contiguity. Visual organization that is based on the principle of contiguity entails a commonality of location in space and time of the figures that constitute a group, e.g., spatial contact or proximity for static configurations and temporal coincidence or succession for moving configurations. Contiguity in this sense includes the classical Gestalt principle of ‘proximity’ in space and time (cf. Palmer 1999:258). Since contiguity entails an opposition of place, it is a 2nd Category principle of visual organization.
3.3.3. Conventionality. Visual organization that is based on the principle of conventionality entails a commonality of attributes or location of the figures in a group as determined by a convention, custom or habit. Commonality in this sense includes the classical Gestalt principles of ‘interposition,’ ‘connection’ and ‘circumscription’ (cf. Wertheimer 1938). Since conventionality in this sense entails mediation, it is a 3rd Category principle of visual organization.
Hermann von Helmholtz hypothesized that visual perception involves an unconscious process of inference (Helmholtz 1925:2, 4; Palmer 1999:56-57). Charles S. Peirce claimed that awareness of perceptual facts is given by perceptual judgments. For Peirce, perceptual judgments are not under conscious control, are accompanied by feelings of belief or doubt and may be asserted as propositions. (“On phenomenology” EP2:155; “What pragmatism is” EP2:336). In our understanding, the final phase in the process of visual perception is valuation. A visual valuation is the nonconscious acceptance or rejection of a visual representation, which results in its being either seen or not seen. Presumably, acceptance of a representation is due to the assimilation of an immediate visual experience with visual memory while rejection of a representation indicates a failure to assimilate an immediate experience with memory. This failure of assimilation may be due to the presence of a significant difference between images formed in present and past visual experience. Hence, a negative perceptual valuation may be modeled as a binary relation between images in visual awareness and images in visual memory, where one relate is marked by a significant difference. In previous work, it has been proposed that three categorical types of difference are determined by three categorical types of markedness and that these types of markedness determine three categorical types of valuation (Cantor 2006a, 2011). Hence, there are three categorical types of negative or rejective valuation that apply to relations between perceived and recalled images:
Hence, there are three categorical types of negative valuation, which may result in a clearly visible representation not being perceived.
e process of visual perception is experienced as occurring in three categorical phases that we refer to as Detection, Localization and Identification (Cantor 2002). Detection involves the attribution of visual qualities to regions of the visual scene. Hence, detection depends upon visual segmentation (section 3.1) and is a 1st Category mental process. Localization is an awareness that certain regions in the visual scene represent objects that occupy specific places in the visual world. Hence, localization depends upon visual figuration (section 3.2) and is a 2nd Category mental process. Identification is the recognition of a specific identity or class membership for objects in the visual world that are represented in a visual scene. Hence, identification depends upon visual organization (section 3.3) and is a 3rd Category mental process. Note that visual identification entails visual localization and visual localization entails visual detection. This is a manifestation of the categorical inclusion rule by which mediation entails opposition and opposition entails attribution (cf. Liszka 1996:46).
A figure in a visual scene is the representamen of a triadic sign that refers to an object in the physical world in a way given by an interpretant, where the interpretant depends upon personal visual experiences and learned conventions. In fact, the entire visual scene is the representamen of a global sign that may be considered as the semiotic union of innumerable regional and local signs (cf. Cantor 2002). Visual perception is veridical when it is both personally univocal and interpersonally consistent. In this way, veridical perception is said to represent a real world that is independent of the will of the perceiver. However, all perceivers are fallible and may be deceived by appearances in any phase of the perceptual process, i.e., in the presentation, representation or valuation of a percept. Nonveridical perception is involved in camouflage and visual illusions.
From a pragmatic perspective, visual signs are used for the detection, localization and identification of objects and events in the visual world (cf. Cantor 2002). Camouflage confounds these mental processes. Camouflage may be thought of as a deceptive presentation of objects or events in the biological world. It may be an evolutionary outcome as in protective animal coloration or an intentional artifact as in defensive military concealment. In general, it involves alterations of appearance or presentations that prevent or delay the detection, localization or identification of objects or events. Such presentations or alterations may affect the color, size, shape or motion of objects. There are three categorical types of camouflage that confound visual detection, localization and identification. Examples of these types of camouflage are given in Behrens (2009).
7.1. Detection camouflage
Detection camouflage may involve:
Since detection camouflage confounds the perception of attributes, it is a 1st Category process.
7.2. Localization camouflage
Localization camouflage may involve:
Since localization camouflage confounds relative depth perception, it is a 2nd Category process.
7.3. Identification camouflage
Identification camouflage may involve:
Since identification camouflage confounds recognition and recognition mediates between present and past visual experience, identification camouflage is a 3rd Category process.
Richard Gregory (2009: Table 2) has devised a comprehensive classification of visual illusions that is based on their causes (reception, perception and conception) and their kinds (blindness, ambiguity, instability, distortion, fiction and paradox). In this section, we present a typology of visual illusions that has been derived from the Gregory classification but is based on semiotic principles. Visual illusions may be classified as either natural or artifactual. Natural illusions are produced by physical or biological processes that are independent of mind, while artifactual illusions are produced either directly or indirectly by mental processes. Artifactual illusions may be produced by pictorial images or spatial constructions that represent imagined objects or scenes. In fact, all pictorial images may be considered as artifactual illusions, if veridical perception is defined to be the result of nonmediated natural vision. Pictorial examples of the illusions described in the following sections are given in Gregory (2009) or Palmer (1999). Visual illusions may be produced in the presentation, representation or valuation phases of visual perception. In each phase, illusions may originate in the processes of detection, localization or identification.
8.1 Presentational illusions
Presentational illusions differ from representational and valuational illusions in that they are present in the virtual images that precede the formation of retinal images. Hence, preretinal or presentational illusions may be reproduced photographically.
8.1.1. Presentational illusions of detection. Blindness due to opacification of the cornea or lens and double vision due to binocular malalignment are natural illusions, whereas effects of detection camouflage may be regarded as either natural or artifactual illusions (cf. section 7.1).
8.1.2. Presentational illusions of localization. Depth localization of objects in the three-dimensional visual world is based upon the interpretation of depth signs that present in two-dimensional visual scenes. The fundamental effects of perspective projection that indicate depth are:
Hence, figural contraction, convergence of parallels and visual occlusion are natural depth signs. In natural depth signs, the figure that refers to a thing constitutes the representamen, the thing that is referred to constitutes the object and the perceived relation of depth between object and viewer is the interpretant. Depth signs that refer to different objects indicate their relative size when their depth is known or assumed to be the same and indicate their relative depth when their size is known or assumed to be the same. Furthermore, depth signs that refer to different surfaces of a familiar object indicate its orientation relative to the viewer. Hence, perception of the relative size or depth of two objects or of the orientation of a single object in the visual world depends upon the availability of relative depth signs as well as collateral knowledge. Artifactual illusions of depth localization may be produced by simulators of natural depth signs or by ambiguous depth signs. A simulator of a natural depth sign is a sign whose representamen is similar to that of a natural depth sign but whose object is different. When the representamen of an artifactual depth sign is a faithful copy of the representamen of a natural depth sign that has the same object, the artifactual sign is a true depth sign. When the representamen of an artifactual depth sign appears from only one point of view to be the same as that of a natural depth sign but with a different object, the simulator is a false depth sign. An ambiguous depth sign is produced by the semiotic union of true and false depth signs or the semiotic union of true depth signs with inconsistent or contradictory interpretants.
Examples of natural illusions of localization include the effects of refraction at interfaces between transparent media of different densities, e.g., mirages at air-air interfaces and illusory displacements at air-water interfaces, the effects of mirror reflection that produce horizontal and depth reversals and the effects of perspective projection that produce contraction of figures and convergence of parallels in depth. Artifactual illusions of depth localization are produced by ambiguous depth signs. The Ames Room illusion (Gregory 2009: Plate 4) is produced by the semiotic union of true and false depth signs and knowledge of the relative sizes of familiar objects. The Impossible Triangle illusion (Gregory 2009: Fig. 44) is produced by the semiotic union of true and false depth signs. The painting "Carte Blanche" by Magritte (Gregory 2009: Plate 5) presents illusions as the semiotic union of pictorial depth signs with inconsistent figure-ground interpretants. The engraving The Fisherman by Hogarth (Gregory 2009: Fig. 22) presents illusions as the semiotic union of pictorial depth signs with inconsistent interpretants of relative size and depth. Furthermore, the effects of localization camouflage are all artifactual illusions.
8.1.3. Presentational illusions of identification. Natural similarities between things create natural illusions of identification. The portrait paintings by Archimboldo, where parts of the figure are represented by fruit, vegetables, flowers, etc (Gregory 2009: Fig. 20), the effects of identification camouflage (section 7.3) and the perception of Impossible Objects (Palmer 1999: Fig. 1.1.8) are all presentational illusions of identification.
8.2. Representational illusions
The process of visual representation may be thought of as occurring in three phases: Registration, Recollection and Interpretation. Hence, the perception of a visual sign involves registration in forming its representamen, recollection in forming its object and interpretation in forming its interpretant. Again, the phenomenology of visual perception entails the detection, localization and identification of objects by means of visual signs.
8.2.1. Representational illusions of detection. An illusion of detection may be produced by either a nonveridical absence or a nonveridical presence of sensation. An absence of sensation is produced by lesions of the brain that interrupt the visual pathways extending from the retina to visual cortex. A transient blindness is produced by continuous visual stimulation that results in neuronal fatigue. A nonveridical presence of sensation may be perceived as an illusory brightness, color, contour or motion. Such illusions of detection include the formation of after-images (Gregory 2009:168), the effects of contrast on the perception of brightness and color (Gregory 2009: Fig. 13 and Plate 1), the illusory contours of the Kanizsa Triangle (Gregory 2009: Fig. 41) and the illusion of movement created by static repeated patterns (Gregory 2009: Fig. 24).
8.2.2. Representational illusions of localization. Representational illusions of visual depth are created in the perception of certain configurations of lines. The retinal image of a physical line or object-line is a linear distribution of electrochemical signals that is mapped topographically to the visual cortex and perceived as an image-line (cf. Cantor, 2012). Note that physical lines (or stripes) that are drawn or painted on the ground surface have length, width and orientation.
In perspective projection, both the length and width of physical lines appear to decrease with increasing distance from the viewer. Recall that the idealized lines of Euclidean geometry, which are assumed to have length without width, are represented or drawn as physical lines that have length and uniform but negligible width. Similarly, representational depth illusions are produced by drawings consisting of lines of uniform width that are unconsciously interpreted as representing lines on the ground in front of the viewer. The illusions are based on the expected effects of perspective projection on the length and orientation of physical lines, while its effect on the width of lines is neglected. The following observations may be confirmed by the reader:
Hence, we may distinguish four elementary frontal depth signs. The representamina of three of these depth signs consist of vertical, vertical-oblique and vertically convergent line segments that may or may not intersect. The representamen of one frontal depth sign consists of a horizontal line segment. Since the same perspective effects are produced when the line of sight is rotated ninety degrees, we may distinguish lateral depth signs with corresponding representamina consisting of horizontal, horizontal-oblique, horizontally convergent and vertical line segments.
Illusions of depth localization involve nonveridical relations between perceived length and depth. In configurations of line segments, the perception of relative length is based on two elementary equivalence signs. The representamina of these signs consist of two parallel line segments of equal length and of a line segment that is divided by a point into two parts of equal length. Representational illusions of depth localization may be created by induced depth signs. Induced depth signs are compound depth signs produced by the semiotic union of elementary depth signs and elementary equivalence signs, where the interpretant of the compound sign is the conjunction of the interpretants of its components. In the following sections, we construct semiotic models of the classical illusions of expansion, contraction, convergence/divergence and displacement. Furthermore, we demonstrate that a hypothetical process of semiotic induction provides a unified explanation for the phenomena of representational depth illusions.
220.127.116.11. Illusory expansion. The phenomenon of illusory expansion in apparent depth has been termed the size-distance paradox (Palmer 1999:323). The Horizontal-Vertical illusion (Gregory 2009: Fig. 40) is due to the effect of perspective projection by which a line segment that is directed away from the viewer in the line of sight is perceived as foreshortened. The sign is produced by the semiotic union of a depth sign and an equivalence sign with representamina consisting of a vertical line and a bisected horizontal line, where the two lines are the same length and the lower end of the vertical line coincides with the midpoint of the horizontal line. Hence, the upper end of the vertical line is unconsciously inferred to be farther from the viewer than the lower end. Consequently, absence of the expected foreshortening in depth of the vertical segment is unconsciously perceived as a lengthening or expansion. The Ponzo illusion (Gregory 2009: Fib. 36) is based on the effect of perspective projection by which parallel lines directed away from the viewer appear to converge in depth. The illusion is produced by the semiotic union of an equivalence sign and a depth sign, with representamina consisting of two parallel horizontal line segments interposed between a symmetrically arranged pair of convergent vertical-oblique lines. In this compound sign, the elementary depth sign induces a direction of depth in the interpretant of the equivalence sign. Hence, the higher horizontal segment of the equivalence sign is unconsciously inferred to be farther from the viewer than the lower segment. Since the parallel lines are of equal length, absence of the expected contraction in depth of the higher segment is unconsciously interpreted and perceived as an expansion. The Table Top illusion (Gregory 2009: Fig. 39) is also based on the principle of semiotic induction. The illusion is produced by a drawing that resembles a table with a square or rectangular top, in a false perspective projection, where opposite edges of the ‘tabletop’ are drawn as parallel lines of equal length. It follows that absence of the expected convergence of parallels is unconsciously interpreted and perceived as divergence of parallels, i.e., expansion in depth. The Müller-Lyer illusion (Gregory 2009: Fig. 37) is based on the effects of perspective projection by which parallel lines appear to converge as they recede from the viewer and diverge as they approach the viewer. Such configurations of lines may be seen at the farthest edges of a rectangular room as viewed from the inside and at the nearest edges of a rectangular structure as viewed from the outside. The illusion is produced by the semiotic union of two compound depth signs with the two components of a parallel equivalence sign. In one presentation, each compound depth sign is the semiotic union of two horizontal depth signs that converge to a point such that
By semiotic induction, one of the parallel lines is unconsciously inferred to be farther from the viewer than the other and is expected to be shortened, although the two parallel lines have the same length. Hence, the line that appears to be farthest from the viewer is perceived as lengthened.
18.104.22.168. Illusory contraction. The Bisected Müller-Lyer illusion (Gregory 2009: Fig. 37b) is based on the effect of perspective projection by which a line segment that is directed away from the viewer is perceived as foreshortened. The illusion is produced by the semiotic union of a compound depth sign and a bisected equivalence sign, where the depth sign has two components that converge to a point with vertical orientation and the bisected equivalence sign is vertically oriented and connects the apices of the components of the compound depth sign. In this illusion, the interpretant of the compound depth sign induces a direction of depth in the interpretant of the equivalence sign. Hence, the upper end of the vertical segment is unconsciously interpreted as being farther from the viewer than the lower end and the upper half of the segment is perceived as foreshortened.
22.214.171.124. Illusory convergence/divergence. Illusions of vergence are based on the effects of perspective projection by which the distance between parallel lines appears to contract as they recede from the viewer and to expand as they approach the viewer. In its simplest form, the Zöllner illusion (cf. Palmer: 2009 Fig 7.3.5) is drawn as a symmetric configuration consisting of two vertically oriented parallel lines and two columns of convergent vertical-oblique line segments that are superimposed on the parallel lines. The illusion is produced by the semiotic union of a compound depth sign and a parallel equivalence sign, where the compound depth sign is an iterated semiotic union of equally spaced and superiorly oriented convergent depth signs and the parallel equivalence sign is vertically oriented. The illusion is created when the compound depth sign induces a superior direction of depth in the interpretant of the equivalence sign. Since the vertical lines are parallel, absence of convergence in the induced direction of depth is unconsciously interpreted and perceived as divergence. The Zöllner illusion is usually presented in an augmented form by increasing the number of parallel lines that constitute the equivalence sign and increasing the corresponding number of columns in the compound depth sign. In this augmented illusion, adjacent vertical line segments appear to converge and diverge in succession.
The Café Wall illusion (Gregory 2009: Fig. 31) is produced by a configuration of lines and colored areas. The simplest form of this illusion is drawn by arranging three horizontal bands of the same width so that the common edges of contiguous bands form parallel lines, coloring the bands by vertical alternating black and white stripes of the same width and translating the middle band to the right or left by a half stripe width. The seams between contiguous bands are horizontal lines that constitute the representamen of a parallel equivalence sign. Perception of the periodic pattern produced by lateral translation of the middle band induces a direction of depth (the direction of translation) in the interpretant of the equivalence sign. As in the Zöllner illusion, the parallel seams are interpreted and perceived as divergent in the induced direction of depth. This illusion is usually presented in a configuration consisting of more than three horizontal bands with edges that appear to converge and diverge toward each side in succession.
126.96.36.199. Illusory displacement. The Poggendorff illusion (Palmer 1999: Fig. 1.1.4 D) is based on the effect of perspective projection by which a line that is directed away from the viewer on the ground surface is perceived as interrupted when the line of sight is occluded by a barrier. The illusion is created by a drawing consisting of two separate vertical-oblique line segments of equal length that are collinear, i.e., that are parts of an oblique transversal that intersects two vertically oriented parallel lines of equal length. The points of intersection divide the parallel lines into unequal parts, where corresponding segments on opposite sides of the transversal are of equal length. Hence, the illusion is produced by the semiotic union of a compound depth sign and a parallel equivalence sign, where the compound depth sign is the semiotic union of two oblique depth signs with representamina that are parts of an interrupted transversal and the representamen of the equivalence sign consists of the parallel line segments of equal length that are above and below the transversal. Hence, the upper component of the compound depth sign induces a vertical direction of depth in the interpretant of the contiguous upper component of the equivalence sign. Consequently, the vertical segment above the transversal is unconsciously interpreted as being farther from the viewer than the corresponding segment that is below the transversal and therefore it is perceived as foreshortened. This results in an apparent superior displacement of the contiguous oblique segment. A similar illusory displacement is perceived when the entire configuration is rotated ninety degrees, where the parallel lines are horizontal and the interrupted transversal is horizontal-oblique.
The Moon illusion (Gregory 2009:200 and Palmer 1999:322) is a natural representational illusion that may also be explained by the action of an induced depth sign. The illusion is created when the Moon and the horizon are projected within the same field of view. The illusion is produced by the semiotic union of a depth sign and a nonlocalizing sign, where the representamen of the depth sign is an image of the horizon alone that creates the impression of depth and the representamen of the nonlocalizing sign is an image of the Moon alone. The expected apparent size of the Moon is determined by recollection of its appearance when it is high above the horizon or isolated in the field of view. In the compound depth sign, the depth component induces a sense of depth in the interpretant of the nonlocalizing component. Since the true size of the Moon does not change with its position in the sky, absence of the expected contraction in depth is unconsciously interpreted and perceived as an expansion. With reinterpretation, the Moon appears to be closer than expected.
Emmert’s Law (Gregory 2009:196-197) is another natural representational illusion that may be explained by semiotic induction. By Emmert’s Law, an after-image (produced by a brief exposure to a focused and intense bright light) that is superimposed upon the retinal image of an object appears to expand as the distance of the object from the viewer increases. This phenomenon is in contrast with the expected contraction in depth. As in previous examples, an interpretation of increasing depth in the object-image induces an interpretation of increasing depth in the after-image. Since the size of the after-image is constant, absence of the expected contraction in depth is unconsciously interpreted and perceived as an expansion.
8.2.3. Representational illusions of identification. Visual identification involves the recognition of a particular object or a type of object, where recognition entails both recollection and interpretation. Hence, illusions of identification may be due to a failure of recollection or a false interpretation. Failure of recollection may result from inattention during an earlier presentation, e.g., attentional blindness or change blindness (Palmer 1999:536, 538; Gregory 2009:106). False interpretations of identity are commonly due to perceptual judgments that are based only on similarity.
8.3. Valuational illusions
After Helmholtz, we have assumed that the formation of interpretants in visual perception is determined by a nonconscious process of inference. Furthermore, we assume that the acceptance or rejection of an interpretant is determined by a nonconscious decision process or valuation. In other words, a perceptual valuation is a nonconscious decision process that either allows or disallows the awareness of a perception. The decision to reject an interpretant may be influenced by emotions that direct attention away from an object of perception or direct attention toward another object. Again, valuational illusions may be produced in the detection, localization or identification phases of visual perception.
8.3.1. Valuational illusions of detection. Under normal viewing conditions, the optic disk and the retinal arteries are not perceived as defects in a retinal image, although these structures are devoid of photoreceptors. This phenomenon may be thought of as a negative valuation that is expressed as a filling in of perceptual defects or a completion of the perceived image (Gregory 2009:215-216).
8.3.2. Valuational illusions of localization. Valuational illusions of localization are created by drawings that contain equivocal depth signs of figure-ground organization or perspective projection. Rubin’s Vase-Faces illusion (Gregory 2009: Fig. 15) is a symmetric line drawing with ambiguous figure-ground organization (cf. section 3.2). The ambiguity is produced by contiguous regions with familiar shapes that lack surroundedness. In this case, equally plausible figurations evoke an equivocal valuation that is experienced as ‘flipping’ ambiguity. In the Necker Cube illusion (Gregory 2009: Fig. 17), there is an absence of perspective depth signs (cf. section 8.1.2) in a configuration of parallel lines that resembles a line drawing of a cubic cage. Hence, near and far faces of the cage cannot be distinguished by contraction in depth. In this case, equally plausible orientations evoke an equivocal valuation that is experienced as ‘flipping’ ambiguity.
8.3.3. Valuational illusions of identification. . Valuational illusions of identification are created by drawings in which different contours of a figure or parts of a figure resemble different things. In Jastrow’s Duck-Rabbit illusion (Gregory 2009: Fig. 16), opposite contours of a figure resemble the face and ears of a rabbit or the head and bill of a duck. In Boring’s Young Woman-Old Woman illusion, opposite parts of a figure resemble the face and head of an old woman in lateral projection or the face and head of a young women in oblique projection. In both examples, equally plausible identifications evoke an equivocal valuation that is experienced as ‘flipping’ ambiguity.
The perception of Roentgen (X-ray) images involves the same processes of detection, localization and identification that occur in natural vision. However, Roentgen images are static pictorial images that lack the coloration of the visual scene and the occlusion of the line of sight that are characteristic of natural vision. Furthermore, natural images represent only surface features of objects while Roentgen images primarily represent interior features and provide limited information about surface features. Consequently, detection, localization and identification in the perception of Roentgen images depend upon the interpretation of Roentgen boundary signs that differ from the familiar boundary signs of natural vision (Cantor 2000).
In this study, the process of visual perception has been represented by a conceptual model. The model consists of an ordered typology that is based on the physical, physiological and mental processes involved in visual perception. The ordering of this typology is determined by the ordering of the universal categories of thought, as proposed by Charles S. Peirce. The process of visual perception is initially subdivided into the phases of Presentation, Representation and Valuation. Visual presentation involves the preretinal processes of Reflection, Selection and Projection of ambient light. Visual representation involves the categorical processes of Segmentation, Figuration and Organization of the retinal image. This is the process by which the representamina of visual signs are created. Perceptual valuation involves a nonconscious decision process that results in either the experience or absence of perception. In addition to this general model of visual perception, the specific phenomena of camouflage and visual illusions are represented as effects of the mental processes of Detection, Localization and Identification. The semiotic model of illusory depth perception is based on a hypothetical process which we have termed semiotic induction, a process by which the interpretant of one sign interacts with the interpretant of a contiguous sign. Hence, semiotic induction depends upon principles of mental association that are operative in the process of semiotic union, i.e., in the combination of signs to form new signs. This explanation of static visual illusions is consistent with Richard Gregory’s explanation of depth illusions as due to a hypothetical process termed inappropriate size scaling (cf. Gregory 2009:178). In conclusion, we have demonstrated that semiotic principles provide the basis for a comprehensive phenomenological model of visual perception.