TALAMANA · THE AI LITERACY MAP FOR ARCHITECTURE AND DESIGN · Foundations · AGE 16—18 · POSITIONAL · HELD
Pattern is not precedent
Studying a case builds judgment, and looking at averages builds only familiarity.
Our position
A pattern of what courtyards tend to look like is not a precedent. A precedent is a traceable case: a named building with a site, a climate, a structure, a budget, a client, and consequences you can ask about — who designed it, under what constraints, what was built, what changed, what happened afterwards. A model's courtyard pattern is a likeness with no address. Studying a case builds judgment. Looking at averages builds familiarity, and familiarity can pass for judgment in a jury presentation.
Why we hold it
Precedent study is where an architect learns why: why this depth, this shade, this threshold, in this place. The why lives in the provenance and the context — the specific conditions — and an averaged image has removed exactly those.
The strongest objection
Precedents are already abstractions. Architects rarely visit them; they meet them through photographs, plans, sections, criticism and selective histories, and they legitimately extract patterns from them. A precedent is not made a precedent by having an address, and a pattern drawn from fifty cases may teach more than one case visited. If the Lab's real line is provenance and context, then pattern-versus-precedent is the weaker version of its own argument — and a retrieval-wired model that returns the named building with its context has already crossed it.
What would make us revise it
Evidence that pattern-first tools, in real teaching, increase rather than replace engagement with traceable cases — students opening, drawing and questioning the sources the tools surfaced — or evidence that students taught from well-provenanced patterns build judgment as reliably as students taught from cases. If the average becomes a door to cases rather than a stopping point, this card softens. We check it against studio evidence every edition.
Try it
Generate "a courtyard house in a hot climate." Then study one real courtyard house from your region for the same thirty minutes. Write down one thing the real building taught that the image could not contain.
Take it to crit
Present one design move of yours with its precedent. Be ready for the question: did you study a building, or an average of buildings, and can you tell the difference in your own work?
What this idea builds on
What this idea opens up
Sources
- architectural-thinking doctrine — pattern ≠ studied precedent
- The Seduction of the Machine
Open this idea on the map · The complete map · Logika · RBDS AI Lab, India · revised every edition.