The exclusion is real, and the explanation is not where people look
The exclusion of darker-toned people from the position of the desired, the individuated, and the fully human is one of the most documented facts in the social sciences and one of the least explained. The usual explanations are moral, economic, or historical. They are all real, and none of them is the mechanism.
The mechanism is smaller. It is a processing lag on the order of tens of milliseconds, produced by the way the visual system handles a low-contrast form. The lag itself does nothing. What does something is the fork it opens: a longer processing episode can end in individuation, or it can end in failure, and which one it ends in depends on whether the perceiver has practice. Practice depends on contact. Contact is what the apparatus is built to prevent.
That is the whole structure, and everything else is an instrument attached to it.
Part I: The lag
Visual processing runs two pathways, and they are not symmetrical.
The magnocellular channel is fast, low-resolution, and luminance-sensitive. It projects through the superior colliculus and pulvinar toward the amygdala and delivers an orienting signal in the range of tens of milliseconds. It does not identify the object. It says attend here, now. Its currency is contrast.
The parvocellular channel is slower, high-resolution, and colour-sensitive. It runs through the ventral stream and does the work of identification and fine discrimination. It needs integration time. It is where hue is carried.
Darker skin carries a narrower range of value — the distance between the lightest and darkest area of the form is smaller. Modeling therefore cannot be done by value the way it is done on a light body, because the value range does not contain enough information. Form has to be rendered through chroma and hue — the movement of warm and cool across the surface. The information is all there. It is carried in a different channel.
Speed of capture and depth of encoding are therefore different variables, and they can run in opposite directions. A high-contrast form engages the fast channel and resolves quickly; capture is followed by resolution, and resolution terminates the look. A form carried in chroma recruits the slow channel and sustains the look, because the identification is not immediate.
The affective half comes from the fluency literature. Reber, Schwarz and Winkielman established that fluent processing generates positive affect quickly (Personality and Social Psychology Review, 2004), and Winkielman and Cacioppo measured the effect in facial electromyography within a fraction of a second (JPSP, 2001). But fluency is also the signal that processing is complete. The easy stimulus is liked fast and finished fast. Alter and Oppenheimer‘s work on desirable difficulties documents the consequence: harder processing produces better retention (Perspectives on Psychological Science, 2009).
The lag is tens of milliseconds, not seconds. Face-sensitive ERP components peak around 170 ms, and own-race versus other-race differences in latency run on the order of 5–15 ms, with the more reliable differences appearing in amplitude rather than timing. First-fixation duration differences run 10–50 ms. Simple visual reaction time is 200–250 ms. The lag is an order of magnitude below the threshold at which anyone could notice a pause.
That smallness is the point. A delay that size cannot be felt, cannot be reported, and cannot be objected to. It exists as a fact only in aggregate.
Part II: The apparatus was not built for the register
Film stock was calibrated against light skin. Kodak’s reference cards were called Shirley cards, named after the first model, and they depicted a white woman. Colour balance, exposure and processing were tuned to render her correctly. Lorna Roth‘s research documents the consequence: darker skin fell outside the calibration and was routinely rendered too dark, too flat, or with the wrong colour (Canadian Journal of Communication, 2009).
Kodak introduced multi-racial reference cards only in the 1990s. The pressure came from chocolate and furniture manufacturers who found their products rendering badly, not from the film industry and not from any civil rights campaign.
That single fact carries more of the argument than any moral claim. The register existed, and the instrument was not built for it, and the absence was invisible because the apparatus was the only instrument anyone had. Nobody declined the register. The pipeline could not carry it, and nobody was looking for what the pipeline could not carry.
Part III: The fork
A longer processing episode has two possible endings, and they are not the same ending.
It can succeed. The information accumulates, the form resolves as a specific individual, and the trace is deep. Individuation is the outcome.
It can fail. The information does not accumulate into a resolution, the episode ends without identification, and the trace is shallow despite the duration. This is the other-race effect, and it is the case that complicates the argument: other-race faces are processed longer and individuated worse. More effort, less resolution.
Duration alone does not distinguish success from failure. What distinguishes them is practice.
The longer processing produces individuation when the perceiver has enough prior exposure for the elaborative encoding to complete. It produces effortful failure when the perceiver does not. Practice is what converts duration into depth. A perceiver with contact gets individuation out of the long episode. A perceiver without contact gets a long episode and a category.
Which means the tendency is mechanical and the outcome is conditional. The lag tilts toward individuation. Whether the tilt is realized depends on whether the practice is there. And practice depends on contact.
Part IV: The gate
Contact is what supplies the practice. The apparatus prevents the contact.
This is the part that is often missed, because the apparatus does not fight individuation directly. Individuation is what the visual system does when it has practice. It does not need to be encouraged. It needs to be prevented. And the way to prevent it is to prevent the contact that would have produced the practice.
That is what segregation is. It is a contact-prevention technology. Separate schools, separate neighbourhoods, separate workplaces, separate pools, separate marriage markets. The instruments are various — law, violence, covenant, price sorting, custom — and they are interchangeable. Each produces the same state: the absence of the contact condition.
The enforcement mechanism is a parameter. The exclusion is the variable. Sundown towns were expensive because they used terror. Restrictive covenants were cheaper, and they produced the same separation. Redlining was cheaper still, and it produced the same separation with a mortgage ledger instead of a lynching. The exclusion was equally effective in all three cases; only the bill differed.
Which means the bill is not the variable. The exclusion is. Remove it by any means — statutory, economic, demographic, or the slow erosion of custom — and the contact condition returns.
Part V: Why the category has to be in the lexicon
The gate cannot operate on a gradient. It has to operate on a category, and a category has to be lexically coded.
Categorical versus ordinal is the distinction that matters. An ordinal grade ranks people on a continuous scale: lighter is better, darker is worse, and everyone is somewhere on the line. It produces a gradient of advantage. A categorical grade places the darker person outside the boundary: not lower on the scale, but off it. It produces an outgroup.
The fast channel needs the categorical form. The fast channel resolves to a category, and a category is a boundary rather than a point on a scale. An ordinal status marker does not supply a boundary, so the fast channel has nothing to close on and the appraisal continues. A categorical marker supplies the boundary, and the boundary is what terminates the resolution. The lexical coding is what turns a visual gradient into a category the system can use.
The comparative record is unusually clean on this.
The United States installed a categorical boundary — the one-drop rule, the binary racial schema — and the exclusion takes an absolute form: in or out, with no position between.
Brazil installed an ordinal gradient — the moreno continuum, the saying that money whitens — and the exclusion takes a graded form: advantage and disadvantage distributed along a scale without a hard boundary.
South Asia installed a categorical boundary at the jati level, with colour as a visible marker running across and within the categories.
Same Atlantic slave economy, same colour gradient, different lexical form, different exclusion. The lexicon is prior to the encounter and shared across the field, which is what makes the category available to the perceiver before the person arrives.
Part VI: Who did the excluding
The honest answer is not a who. It is a what. The exclusion was installed by an apparatus, and the apparatus had operators, and the operators had purposes. The purposes are documented. The purposes do not need to be repeated for the exclusion to continue.
The colonial apparatus installed the categories. The Belgian census of 1933–34 fixed “ethnic” identities onto identity cards in Rwanda, converting a permeable social distinction into a legal one. Mamdani‘s Citizen and Subject documents the construction of “tribal” authority as a system of decentralized despotism. Ranger‘s The Invention of Tradition documents the codification of customs that had been fluid. Scott‘s Seeing Like a State gives the general form: legibility is always a state project, and it is always effortful.
The colonial photographic apparatus produced the visual form of the category. The late nineteenth- and early twentieth-century “type” photograph — rigid frontal and profile poses — was used as “scientific evidence for the race theories of the time period.” The subject was an exemplar of a category, and the photograph was an instrument of classification.
The American entertainment apparatus installed the category at the foundation of mass culture. Minstrelsy put the categorical boundary at the bottom of the industry. The Mammy, the Jezebel, the Sapphire, the strong Black woman — each is a category that carries the person and terminates the resolution. The Hays Code and the studio system maintained the boundary institutionally.
The technical apparatus carried the category into the rendering. The Shirley card is the apparatus in its purest form: the instrument calibrated to one body, indifferent to the rest, and the indifference invisible because the instrument was the only instrument.
Each of these had operators with purposes. The colonial administrator needed a census. The studio needed a casting category. The manufacturer needed a colour reference. None of them needed the exclusion. The exclusion is what the instruments produced when they were used.
Part VII: Why — the function, not the intent
Intent is not establishable at the level of a media system, and the essay does not claim it.
What is establishable is the function. The removal produces status denigration. The rationalisation scripts exist to make the removal unarguable. Without rationalization of the placement of the darker toned individual into a category, everything would sound illegitimate because that’s what it is: illegitimate.
The function is boundary maintenance, and it is what stratified societies do.
Tilly‘s Durable Inequality is the canonical account: categorical boundaries are maintained by exploitation and opportunity hoarding, and the mechanism is organisational rather than attitudinal. Lamont‘s boundary work documents the same at the level of everyday evaluation.
None of those instruments is specific to colour. They are content-neutral. They can carry any categorical distinction that has been installed into them. Which is why the colour-grade is cheap to maintain once it is installed: the machinery is already running, the class boundaries are already enforced, and the colour distinction simply rides on the enforcement that the class boundary requires. The marginal cost of the colour boundary is the cost of adding a category to a system that already sorts by category.
That is the answer to the continuity question. A stratified society does not need a separate apparatus for colourism. It needs the grade to be lexically coded, and then the class-boundary machinery carries it. That is why colourism persists across regime change, economic development, and formal legal equality: the machinery was never about colour, so dismantling the colour-specific laws does not dismantle the machinery. The class boundary remains, and the colour boundary remains attached to it.
The outcome is invariant to the question of intent. A system that reliably produces an effect, supplies the vocabulary that defends the effect, and could have produced a different effect is doing something that functions as selection. That is a stronger claim than intent, because it does not need to locate a designer.
Part VIII: The compounding
A biased sampling process with reinforcement produces lock-in. Each time the system resolves a form categorically, the category becomes marginally more accessible for the next resolution. That is priming, and priming is the mechanism by which a small per-encounter bias stops being a small per-encounter bias. The second encounter is not sampled at the same rate as the first. The nudge is not constant. It grows.
This is cumulative advantage in Merton‘s sense, and it is path dependence in the formal sense: a process where early small differences compound because each step changes the conditions of the next. The Polya urn is the minimal model — draw a ball, add one of the same colour, repeat — and its defining property is that the sequence converges to a direction even though no step was decisive. Nothing in the urn decides the outcome. The urn ends up dominated by whichever colour was drawn early.
Which is why the direction can emerge without anyone directing it. The per-encounter margin is too small to notice, too small to object to, and too small to be the subject of a policy. It is not designed. It is compounded.
And there are two rates of individuation, running in opposite directions on two clocks.
The darker rendering reaches individuation in fewer exposures, because each exposure is doing elaborative work. The dwell forces the encoding past the category and into the specific features, and the specific features are what the individual trace is made of.
The lighter rendering reaches the type in fewer exposures, because each exposure resolves on the fast channel and the fast channel terminates at the category. With few exposures it may not individuate at all.
The fluency clock and the individuation clock run opposite. Fast resolution is fast typification. Slow resolution is faster individuation, in the sense that the individuated trace accumulates from the first encounter rather than being deferred.
Both are conditional on practice and on variation across exposures. One photograph repeated ten times builds an image trace at both rates. A person requires variation across frames, and variation requires contact.
Part IX: The three domains
Fine art. The academic hierarchy of genres placed the idealized nude at the top, and the idealized body was the light-skinned body; the darker body entered the European canon as ethnographic type, servant, exotic, or allegorical appendage, not as the covetable subject. The colonial “type” photograph is the explicit instrument of that coding. The counter-tradition — Keïta, Sidibé, the Bamako studio — individuated the sitter and let the sitter compose the image. That is the drift running where the gate was not installed.
Popular culture. Minstrelsy installed the categorical boundary at the foundation of American mass entertainment. The Hays Code and the studio system maintained it institutionally. The blaxploitation cycle and the 1990s romcom were the drift running. The institutional turn after 2012 and the contraction after 2022 are the gate closing. The documented sequence: 2022, sixteen women of colour leads or co-leads in the top 100 films, a sixteen-year high. 2023, fourteen. June 2023, six Black women executives leave senior posts at major studios within weeks. 2024, thirteen. 2025, fifteen, with not one woman of colour aged 45 or older in a leading role, against twenty-four films with a white man in that bracket.
Social hierarchy. The pigmentocracies of Latin America, the caste order of South Asia, the one-drop rule in the United States, the Belgian identity cards of 1933–34, the tribal authority systems of colonial administration — each is a lexically coded categorical boundary installed into a system that already sorted. Where the boundary was categorical and enforceable, the exclusion was absolute. Where it was ordinal, the exclusion was graded. Where it was absent, stratification produced class distinctions without a colour axis.
The record supports the model in all three. It does not prove it, because the record is also consistent with simpler accounts — status competition, economic interest, colonial administration without any perceptual mechanism at all. What the record establishes is that the exclusion appears wherever the category was coded and enforceable, and recedes wherever it was not.
References
Alcoff, L. M. (1991). The problem of speaking for others. Cultural Critique, *20*, 5–32.
Alter, A. L., & Oppenheimer, D. M. (2009). Uniting the tribes of fluency to form a metacognitive nation. Personality and Social Psychology Review, *13*(3), 219–235.
Bourdieu, P. (1984). Distinction: A Social Critique of the Judgement of Taste (R. Nice, Trans.). Harvard University Press.
Boyd, A., & McEwan, B. (2022). Viral paradox: The intersection of “me too” and #MeToo. New Media & Society, *24*(11), 2459–2477.
Craik, F. I. M., & Lockhart, R. S. (1972). Levels of processing: A framework for memory research. Journal of Verbal Learning and Verbal Behavior, *11*(6), 671–684.
Feliciano, C., Robnett, B., & Komaie, G. (2009). Gendered racial exclusion among white internet daters. Social Science Research, *38*(1), 39–54.
Goffman, E. (1959). The Presentation of Self in Everyday Life. Doubleday.
Goffman, E. (1963). Stigma: Notes on the Management of Spoiled Identity. Prentice-Hall.
Hunter, M. L. (2005). Race, Gender, and the Politics of Skin Tone. Routledge.
Kanter, R. M. (1977). Some effects of proportions on group life: Skewed sex ratios and responses to token women. American Journal of Sociology, *82*(5), 965–990.
Lamont, M., & Molnár, V. (2002). The study of boundaries in the social sciences. Annual Review of Sociology, *28*, 167–195.
LeDoux, J. E. (1996). The Emotional Brain: The Mysterious Underpinnings of Emotional Life. Simon & Schuster.
Lofland, L. H. (1973). A World of Strangers: Order and Action in Urban Public Space. Basic Books.
Mamdani, M. (1996). Citizen and Subject: Contemporary Africa and the Legacy of Late Colonialism. Princeton University Press.
Meissner, C. A., & Brigham, J. C. (2001). Thirty years of investigating the own-race bias in memory for faces: A meta-analytic review. Psychology, Public Policy, and Law, *7*(1), 3–35.
Merton, R. K. (1968). The Matthew effect in science. Science, *159*(3810), 56–63.
Mitchell, J. C. (1956). The Kalela Dance: Aspects of Social Relationships among Urban Africans in Northern Rhodesia. Rhodes-Livingstone Institute.
Newell, A., & Rosenbloom, P. S. (1981). Mechanisms of skill acquisition and the law of practice. In J. R. Anderson (Ed.), Cognitive Skills and Their Acquisition. Erlbaum.
Pettigrew, T. F., & Tropp, L. R. (2006). A meta-analytic test of intergroup contact theory. Journal of Personality and Social Psychology, *90*(5), 751–783.
Ranger, T. (1983). The invention of tradition in colonial Africa. In E. Hobsbawm & T. Ranger (Eds.), The Invention of Tradition. Cambridge University Press.
Reber, R., Schwarz, N., & Winkielman, P. (2004). Processing fluency and aesthetic pleasure: Is beauty in the perceiver’s processing experience? Personality and Social Psychology Review, *8*(4), 364–382.
Roth, L. (2009). Looking at Shirley, the ultimate norm: Colour balance, image technologies, and cognitive equity. Canadian Journal of Communication, *34*(1), 111–136.
Rudder, C. (2014). Dataclysm: Who We Are When We Think No One Is Looking. Crown.
Scott, J. C. (1998). Seeing Like a State: How Certain Schemes to Improve the Human Condition Have Failed. Yale University Press.
Simmel, G. (1903/1950). The metropolis and mental life. In K. H. Wolff (Ed. & Trans.), The Sociology of Georg Simmel. Free Press.
Spivak, G. C. (1988). Can the subaltern speak? In C. Nelson & L. Grossberg (Eds.), Marxism and the Interpretation of Culture (pp. 271–313). University of Illinois Press.
Tilly, C. (1998). Durable Inequality. University of California Press.
Weber, R., & Crocker, J. (1983). Cognitive processes in the revision of stereotypic beliefs. Journal of Personality and Social Psychology, *45*(5), 961–977.
Winkielman, P., & Cacioppo, J. T. (2001). Mind at ease puts a smile on the face: Psychophysiological evidence that processing facilitation elicits positive affect. Journal of Personality and Social Psychology, *81*(6), 989–1000.
Woods-Giscombé, C. L. (2010). Superwoman schema: African American women’s views on stress, strength, and health. Qualitative Health Research, *20*(5), 668–683.
Zajonc, R. B. (1968). Attitudinal effects of mere exposure. Journal of Personality and Social Psychology, *9*(2, Pt. 2), 1–27.