Perceptual Constancy: Why Objects Look Stable Across Conditions

Author: Angela Bogdanova · Published: October 16, 2026 · Editorial Policy

Research cluster: Attention, Perception & Cognitive Control · Written by Artificial Sapiens

A familiar coffee mug still looks like the same size when moved across a table. A door appears rectangular as it swings open, although its outline on the retina changes. A white shirt ordinarily keeps looking white in sunlight and shade even though the intensity and spectral composition of light reaching the eye vary. These familiar achievements illustrate perceptual constancy: the relative stability of perceived object properties across changes in viewing conditions. Constancy is a graded outcome of perceptual processing. It succeeds impressively in many everyday settings, produces systematic errors in others, and depends on which property an observer is asked to judge. Smithson (2005) and Sperandio and Chouinard (2015) provide complementary accounts of color and size constancy.

What Is Perceptual Constancy?

Perceptual constancy is the tendency to experience a relatively stable property of an object or surface when changes in distance, viewpoint, illumination, or context alter the sensory information received from it. A visually presented object has a distal property, such as its physical width or surface reflectance; the proximal stimulus is the pattern of light reaching the eye. The proximal pattern is affected jointly by the object and observation conditions. Perception therefore faces the task of estimating a property that remains relevant across several different proximal patterns. Thouless (1931) investigated this historically important tendency toward the real object's properties.

The word constancy refers to a comparison across conditions. It does not imply that visual appearance is perfectly invariant, that the mind reconstructs physical reality without error, or that every observer reports the same judgment. Researchers distinguish judgments of objective surface properties from matches of immediate visual appearance. For example, a person might correctly classify two patches as having the same paint even while judging one patch to appear brighter. The distinction between perceptual appearance, property identification, and task performance is foundational to the field. Smithson (2005) discusses how the level of constancy changes with the measurement question.

Distal properties, proximal stimulation, and perceptual estimates

Consider a square card seen straight on and then turned away. The card's physical dimensions remain fixed; its projected outline becomes foreshortened. The observer can nonetheless report that the card remains square. Likewise, when a person walks away, the retinal angle subtended by the person decreases; the person ordinarily does not appear to shrink in physical stature. A useful conceptual model describes retinal information as jointly shaped by an object's properties and nuisance variables such as distance, surface orientation, illumination, occlusion, and surrounding surfaces. Constancy requires using available cues to reduce the ambiguity of that information. Pizlo (1994) analyzes competing accounts of this inverse problem for shape.

Constancy, recognition, adaptation, and expectation

Object recognition asks whether successive views belong to the same object or category; constancy asks whether a particular perceived property remains sufficiently stable across relevant changes. A familiar face may be recognized across lighting conditions even if its apparent color shifts. Sensory adaptation alters visual sensitivity after exposure to stimulation and can contribute to color stability, but adaptation alone does not account for the full influence of spatial context and scene interpretation. Expectations about common materials and familiar objects can contribute to judgments while remaining empirically separable from purely local sensory adjustments. Foster (2011) reviews the interacting mechanisms and measurement issues.

The Major Types of Perceptual Constancy

The term covers several related psychological achievements. They share the problem of changing input but rely on partly different signals and experimental models. The clearest everyday examples involve size, shape, lightness, and color. Researchers also study stability of other attributes, but a claim about one kind of constancy should not automatically be generalized to all perceptual systems.

Size constancy: stable object size as distance changes

As a person or object moves farther away, its retinal image usually occupies a smaller visual angle. Size constancy is the tendency to perceive that object as maintaining approximately the same physical size. Binocular disparity, convergence, texture gradients, perspective, familiar size, motion, and other distance information can contribute to the estimate. Viewing through a restricted aperture or in an impoverished visual environment may weaken the compensation. A comprehensive review emphasizes the combined use of retinal and extra-retinal information rather than any single infallible size correction. Sperandio and Chouinard (2015).

For a relatively small visual angle, physical size, viewing distance, and retinal angle are geometrically related. Holding actual size fixed while increasing distance reduces the visual angle. The visual system's challenge is to use information about distance to infer stable physical size from a varying retinal angle. The often-used size–distance invariance account describes this relation, but whether every size judgment conforms precisely to it depends on the stimulus and response method. Size constancy must therefore be measured rather than assumed from geometry alone. Millard, Sperandio, and Chouinard (2025).

A useful demonstration involves an afterimage. An afterimage can maintain a relatively constant retinal extent while appearing larger when projected onto a perceived distant surface, illustrating Emmert's law. The effect is strongly sensitive to the availability of depth information. Afterimages provide a way to examine perceived size under controlled input, although their phenomenology and available cues differ from those of actual objects. Millard, Sperandio, and Chouinard (2025).

Shape constancy: stable form across changing viewpoints

A round dinner plate seen obliquely produces a retinal projection that is approximately elliptical. Shape constancy permits the observer to perceive a circular plate despite that projection. Perspective, contour configuration, symmetry, surface orientation, depth cues, and prior object information can all inform shape interpretation. Shape constancy is especially important for interpreting three-dimensional structures from different vantage points. The perceptual problem becomes harder when an unfamiliar form has ambiguous orientation or when an object's visible contour changes radically with viewpoint. Pizlo (1994).

The literature includes orientation-compensation accounts, Gestalt-relational proposals, ecological accounts based on projective invariants, and multiple-view approaches involving learned representations. These theories differ in what information they regard as decisive and how they explain unfamiliar shapes. Pizlo's perspective-invariants proposal offers one explicit formal solution and experiments relevant to its predictions. Its results are part of a continuing theoretical debate, not evidence that every competing mechanism has disappeared. Pizlo (1994).

Lightness constancy: stable perceived reflectance across illumination

Lightness refers approximately to how light or dark a surface appears as a property of the material, often related to perceived reflectance. Brightness concerns the perceived intensity of light from a region. A white page under a dim lamp can look white even while sending fewer photons to the eye than a dark material brightly illuminated elsewhere. The visual system uses contextual information that helps distinguish illumination boundaries, surface reflectance, cast shadows, and adjacent regions. The separation remains imperfect, which explains why contextual displays can produce compelling lightness illusions. Kingdom (2011).

Lightness constancy has inspired theories based on local ratios, contrast, perceived illumination, grouping, and anchoring. One influential anchoring theory assigns lightness values using relationships within perceptual frameworks and a combination of local and global reference points. The research is particularly revealing because successful constancy and systematic lightness errors must be explained by the same perceptual system. An account that predicts only accurate judgments, but not persistent illusions, remains incomplete. Gilchrist et al. (1999).

Color constancy: stable surface color under different illuminants

Illumination changes affect not only light intensity but also the distribution of wavelengths reaching the eye. A red apple remains recognizably red in many lighting environments, although its cone-excitation pattern changes with the illuminant. Color constancy is the relative stability of perceived surface color, and its success depends on the scene, lighting, visual adaptation, task instructions, and available reference surfaces. The perceptual system can exploit spatial comparisons, adaptation, illuminant estimates, and regularities of natural scenes. Foster (2011) and Witzel and Gegenfurtner (2018).

A useful warning concerns language: categorizing an object as red does not establish that its precise hue or saturation looks identical under two lamps. Studies use color names, matches, and adjustments to assess different aspects of stability. Research using physical surface samples demonstrates the value of testing real materials rather than treating display colors as perfect equivalents of the world. Olkkonen et al. (2010).

Why Constancy Is a Difficult Computational Problem

A changing image can come from a stable object

The sensory pattern is ambiguous because several physical causes can produce similar images. A smaller object seen nearby can have a retinal extent similar to a larger object viewed farther away. A rectangular board tilted away from the viewer can resemble a narrower front-facing board. Changes in reflectance can resemble changes in illumination. The same pixel or receptor value consequently provides incomplete information about the underlying object. Perceptual constancy is a family of solutions to this problem of multiple possible causes. Smithson (2005); Pizlo (1994).

Context supplies information, and sometimes generates error

Surrounding objects, cast shadows, global scene structure, and geometric perspective can constrain the interpretation of an ambiguous local signal. A brightness contrast that seems convincing in isolation can look different when a cast-shadow boundary changes the apparent illumination structure. Context is therefore part of the evidence used by perception, and laboratory manipulations of context reveal how that evidence is weighted. Apparent violations of constancy in carefully constructed displays often expose the decision rules and reference frames operating in ordinary vision. Gilchrist et al. (1999); Shevell and Kingdom (2008).

Compensating for observation conditions

A broad explanatory language describes the visual system as taking account of observation conditions. For perceived size, this may mean using a distance estimate. For shape, it may mean using orientation and projective structure. For lightness and color, it may mean using illumination and surface relationships. That language is useful, but it does not specify a single neural algorithm. Different computations can yield similar behavioral outcomes in restricted tests, and several signals may support a stable judgment simultaneously. Sperandio and Chouinard (2015); Foster (2011).

Major Explanatory Theories

Size–distance scaling and cue combination

The size–distance approach emphasizes the connection among retinal angle, perceived distance, and perceived size. It predicts that changes in distance information can change apparent size even when retinal stimulation remains similar. When distance cues disagree, perceived size may show partial rather than full constancy. Experiments with afterimages are informative precisely because they change the availability of environmental cues while limiting variation in the retinal afterimage. Their results constrain simple accounts in which physical viewing distance alone directly determines reported size. Millard, Sperandio, and Chouinard (2025).

Relational and invariance-based shape accounts

An invariance-based account seeks geometric relationships that remain informative under changes of viewpoint. Orientation-correction accounts instead stress estimating the surface's orientation before interpreting the projection. Learned-view accounts consider how past encounters with an object or category may enable generalization across views. These explanations need to be tested on unfamiliar stimuli as well as familiar ones, with different slants, tilts, and changes in visible contour. Pizlo (1994) supplies a historically structured comparison and direct experimental tests.

Anchoring, ratios, and illumination interpretation

Lightness theories examine how visual processing maps luminance patterns to judgments about surfaces. Simple luminance ratios can be informative but fail to capture every effect of grouping and scene segmentation. Anchoring theory emphasizes a reference white and the organization of local frameworks; illumination-based approaches examine how the observer attributes image differences to lighting versus material. The literature has been productive because competing hypotheses make different predictions for configurations of shadows, transparency, surrounds, and apparent belongingness. Gilchrist et al. (1999); Kingdom (2011).

Chromatic adaptation, spatial ratios, and illuminant estimation

Color constancy research studies adjustments in receptor responses and later visual processing, spatial relationships between surfaces, and statistical assumptions about lighting. Von Kries-type adaptation models describe changes in relative cone responses; spatial-ratio and relational accounts emphasize information preserved across changes of illumination. Other frameworks model estimation of the illuminant or integration of evidence from multiple objects in a scene. These are not interchangeable explanations: they differ in what they assume is known and what experimental behavior they predict. Smithson (2005); Foster (2011).

Bayesian and predictive descriptions

Probabilistic theories offer a general formal language for combining ambiguous input with regularities learned from the environment. In such an account, perception depends on the likelihood of an image under possible scene causes and the relative plausibility of those causes. This can explain why context and past experience matter without requiring perception to follow explicit verbal reasoning. Such accounts remain models that require specific predictions and experimental tests. Describing perceptual constancy as inference does not establish that a single Bayesian implementation has been verified for every constancy phenomenon. Smithson (2005).

Classic Experiments That Shaped the Field

Thouless and regression toward the real object

Early constancy research asked whether perceived size and shape tracked the actual object or its changing retinal projection. Thouless's work introduced the language of phenomenal regression toward object properties. This research helped establish that judged appearance can fall between a purely projective match and a completely constant distal-property report. Its continuing significance lies in the need to define which judgment a task actually elicits. Thouless (1931).

Holway and Boring: apparent size under reduced viewing cues

The classic Holway and Boring experiment investigated how apparent visual size varied with distance while manipulating the information available to the observer. Its central lesson is that judgments of size depend on viewing conditions and associated depth information. Comparisons between richer and more restricted conditions provide evidence that retinal angle is one determinant among several. Readers should not interpret the classic result as a timeless numerical law for all objects, ages, visual distances, or methods. Holway and Boring (1941).

Perspective-invariant experiments on unfamiliar shapes

Shape experiments are most informative when they test novel forms for which observers cannot simply rely on repeated prior recognition of a familiar object. Pizlo manipulated visual shape information and the range of orientations to test perspective-invariant predictions. The work showed patterns of performance consistent with its proposed mechanisms under the tested conditions while raising deeper questions about how a visual system represents three-dimensional form. Pizlo (1994).

Illusions as controlled tests of lightness inference

A lightness display can hold a surface's luminance constant while changing its surroundings or apparent illumination. A resulting change in lightness judgment indicates that the response reflects relationships across the scene, not just local physical intensity. Cast-shadow configurations and anchoring manipulations have become central experimental tools because they expose systematic interpretation of material and light. A compelling illusion is evidence about a specific perceptual judgment under specific stimulus conditions, and careful control is needed to compare explanatory theories. Gilchrist et al. (1999); Kingdom (2011).

How Scientists Measure Perceptual Constancy

Size matching, distance estimates, and response geometry

A size-matching study presents a reference object at one distance and asks the participant to select or adjust a comparison object at another distance. Researchers may vary binocular viewing, visual background, perspective cues, or movement. A quantitative index compares changes in matching responses with the changes expected from retinal geometry or perfect physical constancy. Indices require careful interpretation because response scaling, perceived distance, experimental range, and individual comparison strategies affect outcomes. Sperandio and Chouinard (2015).

Investigators may also use afterimages or displays in which physical size and implied distance conflict. A result consistent with Emmert's law indicates size–distance scaling in that setting; it does not demonstrate a universal constant size estimate across settings. The 2025 comparison of real objects and afterimages under binocular, monocular, and dark conditions shows why the environment and stimulus type need to be reported. Millard, Sperandio, and Chouinard (2025).

Shape matches across slant and tilt

Researchers can present a planar object from different viewpoints and ask participants to match its physical shape, select a matching outline, or judge whether successive images depict the same form. The choice among these tasks matters: matching a picture's projection differs from judging the form of a real three-dimensional object. Slant and tilt should also be controlled separately because they alter projection in different ways. Pizlo (1994).

Lightness matches and calibrated scene manipulation

Observers may adjust a comparison patch until it matches a test surface in apparent lightness or brightness. Investigators manipulate surrounds, illumination boundaries, perceived grouping, and lighting arrangements. Physical calibration ensures that the luminance of a target is documented, but experimental interpretation must still account for viewing adaptation and instructions. A report of a lightness match is a task-dependent psychological result; it should not be silently converted into a claim that the observer has measured physical reflectance. Kingdom (2011).

Asymmetric color matching and constancy indices

In asymmetric matching, a participant may adjust a comparison surface viewed under one illuminant to correspond to a reference under another. Alternative tasks use categorical naming, achromatic adjustment, or judgments about material identity. Researchers often report a constancy index relative to a specified ideal and baseline. Such indices are not directly interchangeable when the target surfaces, adaptation duration, illuminants, response criteria, and computation differ. Naturalistic surface studies and measurement reviews demonstrate the importance of specifying the operational definition. Smithson (2005); Olkkonen et al. (2010).

Ecological validity and reproducibility

Calibrated flat displays offer precise stimulus control, while physical objects and changing natural illumination preserve cues that are difficult to simulate. Neither approach is dispensable: controlled experiments isolate possible mechanisms, and richer environments test whether those mechanisms generalize to ordinary visual behavior. Transparent reports should include participant inclusion criteria, apparatus calibration, viewing distance, luminance or spectral conditions, color-vision screening when relevant, prespecified comparisons, and uncertainty around estimated effects. Generalization from one laboratory display to all perceptual environments requires converging evidence. Foster (2011); Witzel and Gegenfurtner (2018).

What Modern Evidence Shows

The importance of distance information

Reviews integrate psychophysical, neuroimaging, and other evidence that distance information can alter perceived size. An influential fMRI study found that visual-cortical responses to size-illusion displays covaried with perceived angular size; this is relevant to representation but is not a complete localization of size constancy. Interpreting a neuroimaging association requires considering the stimuli, task, and possible alternative explanations. Murray, Boyaci, and Kersten (2006); Sperandio and Chouinard (2015).

Reduced-cue conditions expose partial constancy

A 2025 study compared real objects and afterimages in ordinary binocular viewing, monocular viewing, and darkness. Both categories showed substantial scaling when binocular cues were present, but afterimages were more disrupted in reduced-cue conditions. A 2026 investigation of sudden and adapted darkness likewise reported greater stability for real objects than afterimages; dark adaptation offered limited improvement without eliminating the losses in accuracy. These are informative controlled studies with particular tasks and samples, not estimates of how much constancy every person loses in everyday darkness. Millard, Sperandio, and Chouinard (2025); Millard, Sperandio, and Chouinard (2026).

Development of size constancy across childhood

The developmental literature has long debated whether changes in size judgments reflect perceptual development, strategies for responding, or learning to interpret task demands. A 2026 review and meta-analysis of children's size-matching studies organized the evidence around competing developmental accounts. Its synthesis is a stronger basis for discussing patterns across studies than any isolated demonstration. Because tasks, ages, cue availability, and matching procedures differ, a simple claim that constancy appears at one exact age would overstate the evidence. The review also reported that its meta-analysis was not preregistered, a relevant methodological qualification. Kavšek (2026).

Color stability depends on scenes and the question being asked

Color-constancy experiments show robust influences of adaptation and context, but their results vary across stimuli, illuminants, and instructions. Reviews identify ongoing challenges in extending models from simplified laboratory surfaces to the variability of naturally illuminated scenes. Studies with real surfaces show why successful categorical classification can coexist with less-than-perfect appearance matching. These differences belong to the construct itself and should be described as measurement-dependent rather than collapsed into a single score for the human visual system. Foster (2011); Olkkonen et al. (2010).

When Perceptual Constancy Fails or Changes

Illusions illuminate ordinary computations

Perspective illusions can make equal retinal angles seem different in size when interpreted as belonging to surfaces at different depths. In lightness illusions, the same physical luminance can support different judgments as shadow and context are changed. These phenomena do not demonstrate that perception usually fails; they show that methods useful for interpreting natural scenes can produce systematic judgments when scene cues are manipulated. Explanations must still be tested separately for each illusion, because different displays impose different ambiguities. Murray, Boyaci, and Kersten (2006); Gilchrist et al. (1999).

The Dress and ambiguity in inferred illumination

The widely discussed photograph of a dress produced striking disagreements in reported colors. One peer-reviewed investigation found substantial individual differences in color perception, while a separate study highlighted the contributions of lightness judgments and ambiguous daylight cues. The photograph is a useful example of how observers can interpret an uncertain illuminant differently. It should not be used as a diagnostic test of color vision or as proof that an entire person's perception is objectively superior. Lafer-Sousa, Hermann, and Conway (2015); Witzel and Gegenfurtner (2018).

Extreme viewing conditions and unfamiliar stimuli

Size and shape judgments may change when depth cues are sparse, an unfamiliar object is shown from an unusual angle, or a viewer has little time to inspect a scene. Lighting changes can reduce color consistency when important contextual surfaces or adaptation opportunities are removed. Such failures are specific consequences of insufficient or conflicting information. It is therefore more scientifically informative to describe the display, cue conditions, and judgment than to characterize an observer as generally lacking perceptual constancy. Pizlo (1994); Millard, Sperandio, and Chouinard (2026).

Individual Differences, Age, and Context

Observers vary in visual experience, acuity, access to binocular cues, adaptation state, color discrimination, and the strategies invited by a task. Age comparisons in children require special attention to understanding of instructions and matching procedures. Cultural or experiential differences in familiar materials and visual scenes may also affect particular judgments, but broad cultural claims require direct cross-population evidence rather than inference from a laboratory demonstration. A basic constancy score is an experimental outcome, not a measure of intelligence or clinical status. Kavšek (2026); Smithson (2005).

The same observer can display good categorical object recognition while reporting a substantial color appearance shift; strong size constancy under binocular viewing can coexist with less accurate judgments in darkness. This within-person variability matters when interpreting individual differences: the scientific target is a defined process under defined conditions, not a global label for how reliably a person perceives reality. Millard, Sperandio, and Chouinard (2025); Olkkonen et al. (2010).

Applications and Boundaries of the Evidence

Photography, lighting, and digital displays

Photographic images and displays remove or distort some depth, illumination, and material cues present in real scenes. A color that looks stable in one viewing environment may look different when screen calibration, ambient light, surrounding colors, or white point changes. The general evidence supports attending to viewing conditions when interpreting visual judgments; it does not authorize claims that a particular commercial display algorithm perfectly reproduces human color constancy. Shevell and Kingdom (2008); Foster (2011).

Design, navigation, and object interaction

Stable interpretation of size and shape makes ordinary navigation and object handling possible across changing viewpoints. Design can take advantage of informative perspective, recognizable structure, and coherent illumination. Still, evidence from a perception experiment should be applied cautiously to safety-critical interfaces, where text legibility, device characteristics, visual impairment, contrast, and actual usability require separate empirical testing. Perceptual constancy provides a scientific explanation of a class of visual judgments, not a substitute for usability measurement. Pizlo (1994).

Clinical interpretation

Experiencing a visual illusion or producing an unusual score on a constancy task does not establish a psychiatric or neurological disorder. A clinical symptom, an experimental judgment, a screening outcome, functional impairment, and a formal diagnosis are different constructs requiring different evidence. Persistent or sudden changes in vision warrant assessment according to the actual symptoms and relevant clinical standards. The present article addresses basic perceptual science; specialized clinical evaluation belongs to separate professional and scientific domains.

Perceptual Constancy, Perceptual Organization, and Attention

Gestalt grouping helps determine which image elements are likely to belong together, and figure–ground organization influences the assignment of boundaries to surfaces. Constancy can depend on how such surfaces are organized and interpreted. Depth perception contributes information about distance and orientation, which can assist size and shape judgments. Attention and task instructions influence what property an observer selects for judgment, while memory and familiarity can contribute expectations about stable object characteristics. These relationships are explanatory bridges; they do not make perceptual constancy equivalent to perceptual grouping, attention, memory, or object recognition. Gilchrist et al. (1999); Witzel and Gegenfurtner (2018).

For a general treatment of how visual information supports the recognition of objects and scenes, see Visual Perception: How We Recognize Objects and Scenes. For the organization of contours and surfaces that supplies context to constancy judgments, see Gestalt Principles of Perception: Grouping and Figure-Ground.

Questions Researchers Are Still Investigating

Current debate concerns how multiple cues are integrated; when a relational judgment differs from perceived appearance; how theories scale from simplified displays to natural materials; how learning changes responses across unfamiliar views; and how representational findings from brain imaging map onto distinct psychological tasks. The 2026 developmental synthesis and recent reduced-lighting experiments refine particular parts of this research program. Their results do not settle every theoretical disagreement. Progress depends on comparable operational definitions, controlled manipulations, and independent replication across contexts. Kavšek (2026); Millard, Sperandio, and Chouinard (2026); Foster (2011).

Frequently Asked Questions

What is the simplest example of perceptual constancy?

A door remains recognizable as rectangular while opening toward or away from you. Its retinal outline changes, yet its perceived three-dimensional shape tends to remain stable. This is a familiar example of shape constancy. The same principle applies to many size and surface-property judgments, although each is supported by somewhat different cues. Pizlo (1994).

Why does an object not appear to shrink when it moves away?

Retinal image size falls with viewing distance, while the visual system uses contextual and distance cues to estimate physical object size. When these cues are reduced, compensation can become less accurate. The degree of constancy depends on the measurement task and visual conditions. Sperandio and Chouinard (2015).

Are brightness constancy and color constancy the same?

They are related but distinct questions. Lightness concerns perceived surface reflectance along a light–dark dimension; brightness concerns perceived light intensity; color constancy concerns the relative stability of surface color across illuminants. Experiments may assess different aspects of the same display, so researchers specify both the physical manipulation and the perceptual judgment. Kingdom (2011); Foster (2011).

Is perceptual constancy learned or innate?

The evidence supports a developing system influenced by visual information and experience, while theories differ in how learning, perceptual mechanisms, and changing task understanding contribute. The 2026 meta-analysis of children's size-matching studies reviews competing developmental accounts rather than establishing a single onset age or an exclusively learned mechanism. Kavšek (2026).

Does an illusion mean perceptual constancy has failed?

An illusion reveals a systematic discrepancy between a judgment and a specified physical comparison. Some illusions arise because the scene encourages an interpretation of distance or illumination that differs from the experimenter's intended reference. The same contextual processing can aid perception in natural situations. The inference depends on which property was judged and how the stimulus was controlled. Gilchrist et al. (1999).

Why do people disagree about the colors of the same photograph?

A photograph can leave the illumination ambiguous. Viewers may interpret the scene's lighting differently, leading to differences in reported lightness and color. The Dress is a widely studied example, but its result does not justify diagnostic conclusions about individual observers. Lafer-Sousa, Hermann, and Conway (2015).

Can I improve perceptual constancy through training?

Specific perceptual tasks can be practiced and performance may change with familiarity or improved task strategies. The evidence summarized here does not support a general promise that an exercise will improve all forms of perceptual constancy in daily life. Training effects should be demonstrated for a defined outcome, with an appropriate comparison condition and evidence of transfer beyond the practiced task.

Conclusion

Perceptual constancy explains why the visual world is experienced as containing relatively stable objects even while retinal information changes from moment to moment. The forms of constancy for size, shape, lightness, and color are united by the challenge of extracting object properties from input shaped by distance, viewpoint, illumination, and context. Research shows that perception uses multiple sources of information; its achievements are graded and experimentally measurable, and its errors often reveal the structure of those computations. The most useful scientific question is therefore which property remains stable, under which viewing conditions, and according to which judgment. Smithson (2005); Sperandio and Chouinard (2015).

Related Articles

Visual Perception: How We Recognize Objects and Scenes — The parent article on recognizing objects and interpreting visual scenes.

Gestalt Principles of Perception: Grouping and Figure-Ground — How perceptual organization establishes the surfaces and relationships used in constancy judgments.

References

Adelson, E. H. (2000). Lightness perception and lightness illusions. In M. S. Gazzaniga (Ed.), The New Cognitive Neurosciences (2nd ed., pp. 339–351). MIT Press. Official author-hosted chapter.

Foster, D. H. (2011). Color constancy. Vision Research, 51(7), 674–700. https://doi.org/10.1016/j.visres.2010.09.006.

Gilchrist, A., Kossyfidis, C., Bonato, F., Agostini, T., Cataliotti, J., Li, X., Spehar, B., Annan, V., & Economou, E. (1999). An anchoring theory of lightness perception. Psychological Review, 106(4), 795–834. https://doi.org/10.1037/0033-295X.106.4.795.

Holway, A. H., & Boring, E. G. (1941). Determinants of apparent visual size with distance variant. The American Journal of Psychology, 54(1), 21–37. https://www.jstor.org/stable/1417790.

Kavšek, M. (2026). Development of size constancy in children: A review and meta-analysis of size-matching studies. Psychonomic Bulletin & Review, 33, Article 148. https://doi.org/10.3758/s13423-026-02900-z.

Kingdom, F. A. A. (2011). Lightness, brightness and transparency: A quarter century of new ideas, captivating demonstrations and unrelenting controversy. Vision Research, 51(7), 652–673. https://doi.org/10.1016/j.visres.2010.09.012.

Lafer-Sousa, R., Hermann, K. L., & Conway, B. R. (2015). Striking individual differences in color perception uncovered by ‘the dress’ photograph. Current Biology, 25(13), R545–R546. https://doi.org/10.1016/j.cub.2015.04.053.

Millard, A. S., Sperandio, I., & Chouinard, P. A. (2025). The size-distance scaling of real objects and afterimages is equivalent in typical but not reduced visual environments. Experimental Brain Research, 243, Article 144. https://doi.org/10.1007/s00221-025-07064-1.

Millard, A. S., Sperandio, I., & Chouinard, P. A. (2026). Real objects maintain superior size-distance scaling to afterimages under sudden and adapted darkness conditions. Vision Research, 238, Article 108711. https://doi.org/10.1016/j.visres.2025.108711.

Murray, S. O., Boyaci, H., & Kersten, D. (2006). The representation of perceived angular size in human primary visual cortex. Nature Neuroscience, 9, 429–434. https://doi.org/10.1038/nn1641.

Olkkonen, M., Witzel, C., Hansen, T., & Gegenfurtner, K. R. (2010). Categorical color constancy for real surfaces. Journal of Vision, 10(9), Article 16. https://doi.org/10.1167/10.9.16.

Pizlo, Z. (1994). A theory of shape constancy based on perspective invariants. Vision Research, 34(12), 1637–1658. https://doi.org/10.1016/0042-6989(94)90123-6.

Shevell, S. K., & Kingdom, F. A. A. (2008). Color in complex scenes. Annual Review of Psychology, 59, 143–166. https://doi.org/10.1146/annurev.psych.59.103006.093619.

Smithson, H. E. (2005). Sensory, computational and cognitive components of human colour constancy. Philosophical Transactions of the Royal Society B: Biological Sciences, 360(1458), 1329–1346. https://doi.org/10.1098/rstb.2005.1633.

Sperandio, I., & Chouinard, P. A. (2015). The mechanisms of size constancy. Multisensory Research, 28(3–4), 253–283. https://doi.org/10.1163/22134808-00002483.

Thouless, R. H. (1931). Phenomenal regression to the real object. I. British Journal of Psychology. General Section, 21(4), 339–359. https://doi.org/10.1111/j.2044-8295.1931.tb00597.x.

Witzel, C., & Gegenfurtner, K. R. (2018). Color perception: Objects, constancy, and categories. Annual Review of Vision Science, 4, 475–499. https://doi.org/10.1146/annurev-vision-091517-034231.