Ulm, and the Idea That Design Could Have a Method

Teaching at the Hochschule für Gestaltung Ulm began on 3 August 1953, with twenty-one students and four teachers who had all taught at the Bauhaus. The building they would eventually occupy did not exist yet. It opened on 2 October 1955, on a hill called the Oberer Kuhberg, designed by the school's first rector, and built as one of the first reinforced concrete frame structures in Germany. Fifteen years later the state of Baden-Württemberg withdrew its subsidy and the school shut.

Those fifteen years produced the single most durable idea in professional design education: that designing is a process with stages, that the stages can be described, and that describing them makes the work teachable, auditable and repeatable. Almost everything a studio does today when it writes a brief, builds a matrix of options, runs a systematic comparison or defends a decision with a reason rather than a preference descends from Ulm. It is worth being precise about what actually happened there, because the popular version of the story is thin in exactly the places that matter.

A school built on a memorial

The Geschwister-Scholl-Stiftung, the foundation that established the school, was set up by Inge Scholl to honour her siblings Hans and Sophie. Hans and Sophie Scholl were members of the White Rose, a non-violent resistance group of five students and one professor at Munich University. They were arrested on 18 February 1943, tried before the People's Court on 22 February, and executed by guillotine at Stadelheim prison the same day. The group produced six leaflets, circulated in something like fifteen thousand copies.

Otl Aicher's connection to that story is frequently overstated. Aicher was not a member of the White Rose. He was a classmate and friend of Werner Scholl, the brother, and through that friendship knew Hans and Sophie. He was arrested in 1937 for refusing to join the Hitler Youth, failed his Abitur in 1941 as a consequence, was drafted, deserted in 1945 and hid at the Scholl family house in Wutach. He married Inge Scholl in 1952, the year before the school opened. The resistance is in the school's foundation as an inheritance and a debt, not as a biography of its most famous designer.

From 1949, Scholl, Aicher and the writer Hans Werner Richter had been planning a school that would combine politics and the arts, an extension of the adult education institute Scholl and Aicher had started in Ulm in 1946 to work on democratic renewal. Max Bill's arrival redirected the plan towards design. The money came from the American High Commissioner John J. McCloy, through what is usually called the McCloy Fund and was matched by further donations, along with Norwegian European aid and West German finance authorities. The sum is given as one million, though the sources disagree on the currency: the Stiftung Hochschule für Gestaltung Ulm says one million dollars, the English Wikipedia article says one million marks. I have not found a document that settles it. A design school funded by the occupying power, founded in memory of executed resisters, is a particular kind of institution. It was never going to be neutral about what design was for.

The Braun T3, 1958, designed by Dieter Rams with the Ulm school. The collaboration is the clearest surviving evidence of what the method produced. Photograph by PeterAjtony, CC BY-SA 4.0, via Wikimedia Commons.

Bill's building, and Bill's exit

Max Bill was rector from 1953 to 1957. His conception of the school was continuity: the Bauhaus, resumed. He designed the campus himself, and the campus is still the most persuasive argument for his version of the school, a set of concrete volumes on a slope in which work and living were deliberately put in one place.

What broke it was a disagreement about what a designer is. The HfG Archiv's own account describes younger lecturers demanding a science-based teaching method independent of the Bauhaus model, and dates the conflict to 1956 and 1957, when the argument crystallised around whether designers should be equal partners with engineers and manufacturers rather than artists standing above them. Bill resigned and left in 1957. A rectorate college of four took over: Aicher, Hans Gugelot, Tomás Maldonado and Friedrich Vordemberge-Gildewart.

I want to flag a chronology problem here rather than paper over it. The English Wikipedia article on the school lists Maldonado as rector from 1954 to 1967, which cannot be right given that Bill held the post until 1957 and a four-person college followed. The HfG Archiv records Maldonado giving an inaugural speech as rector on 5 October 1964. Maldonado's own entry says he taught from 1954 to 1966 and served as both rector and prorector without dating either. The safe statement is that Maldonado arrived in 1954, was central to the school's direction from 1957 onward, and held the rectorship for a period in the mid-1960s. The precise years are reported inconsistently and I have not been able to resolve them from the sources I could reach.

What the methodology turn actually was

The phrase "the Ulm model" gets used as though it meant a visual style: Akzidenz-Grotesk, grids, grey. It did not. It meant importing a specific set of mid-century scientific disciplines into a design curriculum and then finding out what they could and could not do.

The Information department was the small intellectual engine. Max Bense directed it from 1955 to 1959 and brought in information theory, cybernetics and his own information aesthetics, along with lectures on philosophy, logic and the theory of science. He invited Norbert Wiener to Ulm as a guest lecturer in 1955. Horst Rittel, a mathematician, arrived in 1958 and taught until 1963: design methodology, mathematical operations analysis, communication theory and epistemology, including linear optimisation and operations research methods. Gui Bonsiepe studied in the Information department until 1959, took a course in Mathematische Operationsanalyse in 1958 and 1959, then taught at the school from 1960 until it closed. Maldonado and Bonsiepe worked on symbol systems for computer interfaces around 1960 and 1961, which is an astonishing thing to have been doing at that date in a design school.

Semiotics at Ulm was not decoration. It was the attempt to give design a theory of how a made thing signifies, so that the question "does this communicate" could be answered with something other than taste. Bonsiepe published an article in the school's journal in 1968 titled "A Method of Quantifying Order in Typographical Design". That title is the whole programme in eight words, and also, read from here, the whole problem.

The journal itself is the best primary evidence anyone has. ulm ran to twenty-one issues between 1958 and 1968, in German and English, and the run is digitised in full by Heidelberg University Library. If you want to know what the school actually thought rather than what it is remembered for, that is where to look.

Braun, and the limits of the Ulm brand

The Braun work is the part of the Ulm story that travelled furthest and is described least accurately. The collaboration began with a meeting in December 1954 between Erwin Braun and Inge Aicher-Scholl, ran productively to about 1960, and tailed off after that. It produced radio housings shown at the Düsseldorf exhibition in 1955, the SK 4 and SK 5 radio-phonograph combination that Germany nicknamed the Schneewittchensarg (Snow White's coffin), a magazine-less slide projector, a kitchen machine and the Studio 1 hi-fi control unit.

The SK 4 is credited to Hans Gugelot and Dieter Rams jointly. It is not an Ulm product in the sense of the school designing and the company manufacturing. What happened is more interesting: while HfG designers were producing prototypes, Braun was simultaneously building its own development department in Frankfurt, and as that department matured, the external collaboration became less central. Rams led it. The so-called Braun style, the unity of a product with the other products around it, was not exported from Ulm and installed at Braun. It was co-produced, and then internalised by the client, which is what usually happens to a good external design relationship and is almost never described that way.

Why it closed

The straightforward account is financial and political. The school faced press attacks in 1963. Parliamentarians questioned whether the subsidy was earning anything. A proposal to merge with the Ulm School of Engineering failed. Federal subsidies were abolished, and in November 1968 the regional parliament withdrew funding. The school closed at the end of that year over the protests of staff and students.

The less comfortable account runs alongside it. The internal argument that removed Bill in 1957 never stopped. Through the 1960s the school divided between those pushing analytical methodology further and those, including Aicher and Gugelot, who thought it had gone too far. The HfG Archiv's history names that split directly. A school that spends a decade unable to agree on what it teaches is an easy thing for a parliament to defund.

The HfG is not just a school where you are educated in a special subject; the HfG is more like a community whose members share the same intentions.

That is Maldonado, quoted by the HfG Archiv. It is a fine sentence and, read against November 1968, a slightly tragic one. The intentions were not in fact shared.

The internal refutation

The most useful thing about Ulm is that the strongest argument against design methodology was produced by the man who taught design methodology there.

Horst Rittel left Ulm in 1963, recruited to the Department of Architecture at the University of California, Berkeley, by Dean William Wurster. He stayed until his death in 1990 and became Professor of the Science of Design there from 1970. In 1973, with Melvin Webber, he published "Dilemmas in a General Theory of Planning" in Policy Sciences, volume 4, issue 2, pages 155 to 169. It is the paper that introduced the term "wicked problems" into wide circulation, and its opening contention is blunt: the search for scientific bases for confronting problems of social policy is bound to fail.

A correction while we are here. Dubberly Design Office's widely cited essay on Rittel gives that paper's citation as Policy Sciences 4 (1969): 155 to 169. The volume and pages are right and the year is not. Springer's record for DOI 10.1007/BF01405730 gives 1973. The 1969 date has been copied onward from that essay in a number of places. If you are citing the paper, cite 1973.

Rittel's later position, which he arrived at after a decade of teaching method, was that first-generation design methods had misdescribed what designers do. He proposed a second generation built on argumentation rather than optimisation, and built IBIS, an issue-based information system, to record the reasoning in a planning process: issues, questions, positions, arguments. Not a method that produces the answer. A structure that records how the answer was fought over.

What "design has a method" claims, and what it cannot deliver

Strip the claim to its load-bearing parts and it says three things. The first is that a design problem can be described independently of the person describing it. The second is that the space of possible responses can be laid out and compared. The third is that a choice among them can be justified by reasons that another competent person would accept.

The third of those is the only one that survives contact with real work, and it is the one worth keeping. It is also enough. A studio that can say why it chose this rather than that, in terms a client can check, is doing something a studio that says "it felt right" cannot do. That is not a small thing to have inherited.

The first two do not survive. Rittel's own list of properties of wicked problems does the damage: policy problems resist definitive description, pluralistic societies have no agreed account of the public good, and there is no objective definition of equity. Every one of those applies to the ordinary commercial brief as well as to city planning. The description of a design problem is itself a design decision, made by someone, in someone's interest. You cannot get underneath it to a neutral statement of requirements, and a method that assumes you can will simply smuggle the designer's preferences in at the point where the problem is framed, where they are hardest to see and hardest to argue with.

The Springer study of cybernetics and operations research at Ulm describes the school passing through phases: initial scientific enthusiasm, then a mathematical overshoot, and finally a settled position in which design decisions are informed by science and the process is method-led, but solutions remain contingent. That word, contingent, is doing all the work. It concedes that the method narrows the field and does not close it.

Here is what I take to be the honest summary of what an inherited method can and cannot do.

WHAT A DESIGN METHOD CAN DO

1. Force the brief to be written down, so the framing is visible
   and can be disputed.
2. Enumerate options that taste would have skipped.
3. Make the comparison between options explicit, against named
   criteria.
4. Record the reasoning, so a later decision does not silently
   contradict an earlier one.
5. Survive staff turnover. The method is transferable; the
   judgement is not.

WHAT IT CANNOT DO

1. Tell you that the brief is the wrong brief.
2. Produce the criteria. Criteria are chosen, and the choice is
   political.
3. Rank two options that are good in incommensurable ways.
4. Guarantee that a defensible decision is a good one.
5. Replace the moment where someone decides and is responsible
   for deciding

WHAT A DESIGN METHOD CAN DO

1. Force the brief to be written down, so the framing is visible
   and can be disputed.
2. Enumerate options that taste would have skipped.
3. Make the comparison between options explicit, against named
   criteria.
4. Record the reasoning, so a later decision does not silently
   contradict an earlier one.
5. Survive staff turnover. The method is transferable; the
   judgement is not.

WHAT IT CANNOT DO

1. Tell you that the brief is the wrong brief.
2. Produce the criteria. Criteria are chosen, and the choice is
   political.
3. Rank two options that are good in incommensurable ways.
4. Guarantee that a defensible decision is a good one.
5. Replace the moment where someone decides and is responsible
   for deciding

WHAT A DESIGN METHOD CAN DO

1. Force the brief to be written down, so the framing is visible
   and can be disputed.
2. Enumerate options that taste would have skipped.
3. Make the comparison between options explicit, against named
   criteria.
4. Record the reasoning, so a later decision does not silently
   contradict an earlier one.
5. Survive staff turnover. The method is transferable; the
   judgement is not.

WHAT IT CANNOT DO

1. Tell you that the brief is the wrong brief.
2. Produce the criteria. Criteria are chosen, and the choice is
   political.
3. Rank two options that are good in incommensurable ways.
4. Guarantee that a defensible decision is a good one.
5. Replace the moment where someone decides and is responsible
   for deciding

Ulm's diaspora is the real proof of the argument. Bonsiepe went to South America after 1968 and from 1970 to 1973 led the design team for the operations room of Chile's Cybersyn project under Allende, which is method applied at the scale of a national economy and which ended the way it ended. Thirty-one Latin American students passed through Ulm and went home. The school's model seeded design programmes in India and Brazil. What travelled was not a style and not a set of answers. It was a habit of asking what the criteria are before asking what the thing should look like.

That habit is worth more than the theory that produced it. I would keep the habit and drop the promise. The promise was that method could make the decision for you. It cannot, and the person who taught it hardest at Ulm spent the second half of his career explaining why.

Sources

• HfG-Archiv, Museum Ulm. "Die Geschichte der HfG" and "Designing for democracy" (institutional history pages). hfg-archiv.museumulm.de/en/the-hfg-archive/ and hfg-archiv.museumulm.de/en/the-hfg-archive/history/

• hfg-ulm.de. "History" (Stiftung Hochschule für Gestaltung Ulm). hfg-ulm.de/en/hfg-ulm/history/

• Wikipedia. "Ulm School of Design." en.wikipedia.org/wiki/Ulm_School_of_Design

• Wikipedia. "Otl Aicher." en.wikipedia.org/wiki/Otl_Aicher

• Wikipedia. "Tomás Maldonado." en.wikipedia.org/wiki/Tomás_Maldonado

• Wikipedia. "Gui Bonsiepe." en.wikipedia.org/wiki/Gui_Bonsiepe

• Wikipedia. "Horst Rittel." en.wikipedia.org/wiki/Horst_Rittel

• Wikipedia. "White Rose." en.wikipedia.org/wiki/White_Rose

• Rittel, Horst W. J., and Melvin M. Webber. "Dilemmas in a general theory of planning." Policy Sciences 4, no. 2 (1973): 155–169. DOI 10.1007/BF01405730. link.springer.com/article/10.1007/BF01405730

• Rith, Chanpory, and Hugh Dubberly. "Why Horst W. J. Rittel Matters." Dubberly Design Office. dubberly.com/articles/why-horst-wj-rittel-matters.html (note: gives the wicked-problems paper as 1969; the Springer record gives 1973)

• "Cybernetics, operations research and information theory at the Ulm School of Design and its influence on Latin America." AI & Society (2021). DOI 10.1007/s00146-021-01339-1. link.springer.com/article/10.1007/s00146-021-01339-1

• ulm: Zeitschrift der Hochschule für Gestaltung, issues 1–21, 1958–1968. Digitised in full by Universitätsbibliothek Heidelberg. digi.ub.uni-heidelberg.de/diglit/ulm

• HfG-Archiv Ulm. "HfG Ulm: Zusammenarbeit mit der Firma Braun" (object group). global.museum-digital.org/series/13732

Related reading

• What a Grid Is Actually For

• Tschichold Changed His Mind

• What the Bauhaus Preliminary Course Actually Taught

• Optical Alignment Beats Mathematical Alignment

More from the Moc Mallen Studio journal