What if a building could be generated by rules, the way a sentence is built from grammar? That is the idea behind the shape grammar.
A shape grammar is a design system in which shapes transform based on specific rules. You begin with an initial shape, apply rules step by step and a design emerges from the process. Instead of drawing every part directly, the designer or architect creates the logic or the rule that generates the form.
The concept was introduced by George Stiny and James Gips in their 1971 paper Shape Grammars and the Generative Specification of Painting and Sculpture. Their work showed how simple shape rules could generate entire families of paintings and sculptures which later influenced architecture too.
How does it work?
Think of shape grammar as a visual language. Just as any language has vocabulary and grammar, a shape grammar has three parts.
- A starting shape
- A set of transformation rules
- A sequence for applying those rules
A shape can be divided, extended, mirrored, rotated, or combined with other shapes according to predefined rules. As the rules are applied, the design develops. Different rule sequences can produce different outcomes while still belonging to the same design language.
Here is the simplified example- The Palladian Example
The Palladian Example
It is published by George Stiny and William Mitchell in 1978. It is a parametric shape grammar that generates the ground plans of Palladio’s villas, and it was applied to produce the plan of the Villa Malcontenta. In effect, the grammar defines the Palladian style as a set of rules rather than a list of examples.
The researchers analyzed the villas of Andrea Palladio and translated their design principles into a rule-based system. Rather than copying existing buildings, the grammar generated plans that followed the same underlying logic. This introduced a new way of understanding architectural style as a process rather than merely a visual appearance.
Villa Foscari, known as La Malcontenta. Photo by Threecharlie, via Wikimedia CommonsShape Grammar and AI
Although Shape grammar was developed decades before AI-assisted design, many of the principles behind today’s generative design systems have similar roots.
Generative design systems often explore multiple design options based on rules, constraints and objectives. AI tools can analyze patterns in existing designs and generate new alternatives. While modern technologies use advanced computation and machine learning, shape grammar focuses on something fundamental: the rules that organize design.
In a sense, shape grammar was an early attempt to encode architectural knowledge into a system that could generate new designs. Long before AI became a design buzzword, researchers were already exploring whether architectural creativity could be expressed through formal rules.
Why It Still Matters
- Design logic becomes visible. Shape grammars reveal the underlying structure behind a style.
- One system, many variations. The same rules can generate multiple design alternatives in a limited time.
- A valuable teaching tool. Students learn to think about relationships, proportions, and design systems rather than isolated forms.
- A bridge to computational thinking. Shape grammar connects traditional architectural theory with contemporary digital design methods.
Shape grammar is a fascinating example of the intersection between architecture, mathematics and creativity.
Is architecture primarily about creating forms or about defining the rules that generate them? Would you design a building form from a sketch or from a set of rules? Share your thoughts in the comments..
Also check the previous Word of the week post and come back next week for another architecture word.


Leave a Reply