Albers Ran Experiments, Not Poster Campaigns
Interaction of Color was published in 1963 as a box of 150 silkscreen plates weighing about twenty-two pounds. It is a set of experiments with results, not a book of pictures.
Date
Category
Writer
The first edition of Interaction of Color weighed about twenty-two pounds. Yale University Press issued it on 12 June 1963, which was Anni Albers's birthday, in an edition of two thousand copies, and announced at the time that there would never be any more. It was not a book in the ordinary sense. It was two volumes, text and commentary, and more than eighty folders holding 150 silkscreen plates, the whole thing housed in a hinged portfolio with a slipcase. You did not read it on a train.
The price is the first thing I could not pin down. Yale's own news office puts it at $200, roughly $2,000 in current money. The Otis College special collections entry gives $181.20, roughly $1,500 in 2020 dollars. Both are institutional sources. I suspect one is the list price and the other the subscription or pre-publication price, but I have not found a document that says so, and I am not going to pretend the discrepancy away.
What matters about the format is that it was not a format chosen for luxury. Josef Albers spent thirty years running a colour course, first at the Bauhaus, then at Black Mountain College from November 1933, then at Yale, where he was appointed in 1950 to head the Department of Design. The course did not consist of lectures with illustrations. It consisted of students making things out of cut paper and then being asked whether the thing they had made did what they thought it did. The book had to reproduce the conditions of that test, which meant the colours had to be printed as flat opaque areas at reliable sizes, on separate sheets, with flaps that could be lifted. Otis's description of the plates mentions exactly this: a flap down and a flap up, the same inner colour, two different impressions. A bound book of halftones cannot do that. So Yale printed a box.
What the book actually claims
The opening sentence of the text is the entire argument, and the Albers Foundation quotes it: in visual perception a colour is almost never seen as it really is, as it physically is. He follows it immediately with the consequence, that this fact makes colour the most relative medium in art. Elsewhere he puts it as a flat statement: colour deceives continually. Late in life, on camera for Yale, he was blunter still.
All color perception is really illusional. That's the theme of my book.
That is a scientific claim about the visual system, not an aesthetic preference, and it has consequences that most people who cite Albers do not actually accept in practice. If colour appearance is a function of context, then no colour has a fixed identity that can be inspected on its own, and a specification that names a colour without naming its surroundings has not specified an appearance.
The demonstrations he built to establish this are stated as challenges rather than as propositions. The structure of the course, in the order the exercises run, looks roughly like this:
Demos 1 and 2 are the famous pair, and people repeat them as slogans. They are not slogans. They are tasks with a pass condition. In Demo 2 you are given two colours and you must find two grounds that make them converge. If you cannot find them, you have failed the exercise, and you have failed it in a way that is visible to everyone in the room. That is the part that does not survive being quoted.
Studies for Homage to the Square. The squares are a fixed armature so that only the colour varies. Photograph by Selena N. B. H., CC BY 2.0, via Wikimedia Commons.
The exercise almost nobody mentions
The demonstration I find most useful is the one named after the Weber-Fechner law, which sits near the end of the sequence. The exercise asks the student to produce a series of steps that look evenly spaced in intensity. The finding is that even-looking steps do not come from adding equal amounts of pigment. They come from doubling: each step requires a multiplication of the quantity, not an addition of it, because perception responds to ratio rather than to difference.
Anyone who has built a tonal scale for an interface and found the middle steps bunching up has rediscovered this the hard way. It is a well-established property of the visual system and it is the reason a linear ramp between two hex values produces a scale that looks wrong in the middle, and it is also the reason that fixing it by eye works and fixing it by arithmetic does not, unless the arithmetic is the right arithmetic.
The other late exercises are equally concrete. Vibrating boundaries: place two colours of similar intensity but contrasting hue against a sharp edge and the edge appears to move. Vanishing boundaries: match the light intensity of two adjacent colours closely enough and the border between them dissolves. The Bezold effect: change one colour in a pattern and the whole pattern shifts, which is a property of the pattern and not of the changed colour. Optical mixture, which explains both pointillism and halftone printing. Transparence and space illusion, made by adjusting the softness or hardness of boundaries where two areas overlap. Film colour and volume colour, which is what happens to intensity across distance and through a transparent body.
Why coloured paper, and why that decision is the real method
An early chapter of the book exists only to justify the material: why coloured paper instead of pigment and paint. It is worth taking seriously because it is the methodological core disguised as a practical note.
Paint introduces a confound. If a student mixes a colour and the result is wrong, there are two possible explanations, that the judgement was wrong or that the mixing was wrong, and no way to separate them. Cut paper removes the second explanation entirely. The colour is given, fixed, identical for everyone in the room, and the only variable left under the student's control is placement. Which means every result is attributable. Albers had, without saying so in these terms, isolated a single independent variable.
The Otis entry puts the effect of this well: the work is not a theory of colour, it creates a laboratory for the study of colour as a visual phenomenon. Albers dedicated the book to his students and wrote that his students in colour had taught him more colour than books about colour had. He meant it institutionally. The plates in the box are, in large part, the students' results: the material came out of the courses he taught at Black Mountain College and at Yale, and he credited the students with visualising and discovering new problems, new solutions and new presentations. The book is a published dataset with a method attached.
Experiment versus artefact
Here is the distinction I want to draw, and it is the reason I keep returning to this book rather than to the paintings.
Property | An experiment | An artefact |
|---|---|---|
What is fixed | Everything except one variable | Nothing in particular |
What counts as a result | A difference attributable to that variable | The thing looking good |
How failure appears | The predicted effect does not occur, visibly, in front of witnesses | Someone senior does not like it |
What transfers to the next job | A relationship that holds whenever the conditions recur | The artefact, which will not fit |
Homage to the Square, which Albers began in 1949 and continued for decades, is usually described as a painting series. It is more accurately a fixed apparatus: the same nested-square format held constant, so that the main thing changing between one canvas and the next is the colour. Hold the geometry still for that long and every canvas becomes a reading of the colours in it. That is the design of a controlled study, executed in oil.
Why a screenshot of a palette proves nothing
The standard artefact of contemporary colour work is a row of swatches with hex values under them, posted as an image. Against Albers' first sentence, that image is close to information-free about the thing it claims to document.
It fixes no context, so it cannot show appearance, only specification. Every one of those colours will shift when placed on a real background, and Demo 1 says the shift can be large enough that a single colour reads as two. It sets no task, so nothing about it can fail. And it is normally judged at a size and viewing distance that has nothing to do with the sizes the colours will actually occupy, which matters because area changes the reading: a hue in a twenty-pixel label and the same hue across a full-bleed panel are not the same event for the eye.
The Albers equivalent of a palette review would be a set of pass-or-fail demonstrations run in the real context: place the two accent colours that are supposed to be distinguishable on the two backgrounds they will actually sit on and check whether they still diverge, or whether the grounds have subtracted the difference. Take the tonal scale and check whether the steps are even by eye rather than even in numbers. Take the border between the two most similar surfaces in the interface and see whether it vanishes at the intensity you have chosen. These are ten-minute tests that produce answers, and they produce answers that can be wrong, which is the property that a screenshot lacks.
Anoka Faruqee, teaching at Yale, described what students go through when they first meet the exercises as confusion or mistrust before they recognise how colour actually behaves. Mistrust is the correct response. A demonstration that does not initially seem wrong is not telling you anything you did not already know.
Corrections and open questions
Three things I want to flag rather than bury.
First, the book is frequently described as colour theory. It positions itself as the opposite. Its final movement is an argument that theory follows practice and that direct observation takes precedence over systems, and Albers' rejection of the colour wheels and colour solids of nineteenth-century theory is not incidental to the book, it is the point of it. Calling it a theory book inverts its thesis.
Second, Albers' Yale dates are reported inconsistently. He was appointed in 1950 to lead the Department of Design. Wikipedia and Yale's news office both say he retired from teaching in 1958, at seventy. The Metropolitan Museum's index gives him as head of the department from 1950 to 1960. I have not found a source that reconciles those, and it may be a distinction between the chairmanship and active teaching.
Third, and this is a correction to something I used to repeat: the box was not a deliberate act of exclusivity. It cost what it cost because 150 silkscreen plates cost what they cost, and Albers himself approved a cheap edition as soon as one was viable. A pocket edition, about 8¼ by 6 inches, appeared in 1971 with ten colour studies he chose himself. The book has since been translated into at least eleven languages, has appeared as an iPad application in 2013 for the fiftieth anniversary, as an ebook in 2020 and as an interactive web version in 2022. The screen editions do something the 1963 box could not: they let the reader move a colour and watch it change, which is closer to the classroom than the plates ever were. Two thousand people got the box. Rather more people now get the demonstration, which was always the part that mattered.
Sources
Yale News, "The book on color that never fades", 17 February 2023. Publication on 12 June 1963 (Anni Albers's birthday), edition of 2,000 copies, 150 colour plates, 22 pounds, $200, the Press's statement that there would never be any more, Albers's line that all colour perception is really illusional, the three challenge exercises, Anoka Faruqee on student mistrust, the 2013 app, 2020 ebook and 2022 web version. news.yale.edu/2023/02/17/book-color-never-fades
Otis College of Art and Design, Special Collections guide, "Interaction of Color". Limited edition of 2,000, two volumes of text and commentary in a hinged portfolio with slipcase, 150 silkscreen plates in over 80 folders, $181.20, the work described as a laboratory rather than a theory, the dedication and the line about students teaching him more colour than books, the flap-up and flap-down plates. otis.libguides.com/special_collections/interaction_of_color
Josef and Anni Albers Foundation, "Interaction of Color". Yale University Press 1963, plates silkscreened largely by Albers's students, the opening sentence on colour almost never being seen as it physically is, "colour deceives continually", exercises framed as challenges, solutions as steps on an endless path, the 1971 pocket edition, the eleven translations. albersfoundation.org/alberses/teaching/interaction-of-color
Yale University Press, "Interaction of Color: 60 Years On", 13 February 2023. The 1963 limited silkscreen edition, 150 plates, 80 folders, nearly 22 pounds, the 1971 pocket edition at 8¼ by 6 inches with ten colour studies selected by Albers. yalebooks.yale.edu/2023/02/13/interaction-of-color-60-years-on/
UC Davis Library, "50 Features of Special Collections: Interaction of Color by Josef Albers". The oversized portfolio, over twenty exercises each introduced by explanatory text and an illustration, 80 folders of sample study plates, the flap mechanism showing identical inner colours under different surrounds. library.ucdavis.edu/news/50-features-special-collections-interaction-color-josef-albers/
Zych, Jeff, "Exercises from Interaction of Color by Josef Albers", 29 April 2020. A worked list of the exercises in order, including one colour appears as two, two different colours look alike, why colour deception, colour mixture in paper, the Bezold effect, colour intervals and transformation, the Weber-Fechner law step-doubling finding, vibrating boundaries, equal light intensity, and the closing argument that theory follows practice. jlzych.com/2020/04/29/exercises-from-interactions-of-color-by-josef-albers/
Open Library record for Interaction of Color, Yale University Press, 2013, 50th anniversary edition, xi and 193 pages, ISBN 9780300179354, first published 1963, with the chapter sequence. openlibrary.org/books/OL26888074M/Interaction_of_color
Wikipedia, "Josef Albers". Dates 19 March 1888 to 25 March 1976, Bauhaus from 1920, Black Mountain College November 1933 to 1949, Yale appointment 1950 to head the design department, retirement from teaching in 1958, Homage to the Square begun 1949, 150 silkscreen plates in the first edition, the motto "to open eyes". en.wikipedia.org/wiki/Josef_Albers
Metropolitan Museum of Art, Modern Art Index Project, "Anni and Josef Albers". Bauhaus enrolment 1920, the glass workshop, the Vorkurs, the 1933 dissolution of the Bauhaus, Black Mountain College 1933 to 1949, head of the Yale Department of Design 1950 to 1960. metmuseum.org/research-centers/leonard-a-lauder-research-center/research-resources/modern-art-index-project/anni-josef-albers
Related reading
Latest Articles.
© Moc Mallen Studio
Thoughts, ideas, and perspectives on design, simplicity, and creative process.
