/
Writing / Full Essay / English

Drawing the City Before Designing the Building

Representation, Art and Architectural Thinking

Representation, Art and Architectural Thinking

Before architecture, drawing

Before I understood architecture through models, zoning, area schedules or technical coordination, I understood things by drawing them. Drawing was not only a way of producing an image. It was a way of slowing down enough to look.

That relationship with visual work started before architecture school. Later, at the Universidad Central de Venezuela, representation became tied to architecture: plans, sections, collages, diagrams, models and images were no longer separate from the design process. They were the way I tested what a project could become.

I still think that way. The tools have changed enormously, but I continue to understand design by moving between different forms of representation. I draw, model, compare, measure, overlay, render and redraw. Each method reveals something that the others do not.

Robin Evans described architecture as a discipline that works through an intervening medium: the architect rarely works directly with the building, but through drawings, models and other forms of representation (Evans, 1997). That idea describes my own experience. The drawing is not outside the project. It is one of the places where the project begins to exist.

Representation is not the final image

Architecture often presents representation as something that happens at the end: the drawing or rendering that explains a project once the design has been resolved. That has never been how I experience it.

For me, representation is most useful when the project is still uncertain. A crude diagram can expose a relationship that a polished rendering hides. A plan can reveal a circulation problem that is invisible in a perspective. A section can show that two spaces that look close in plan are actually disconnected vertically. A massing model can make the consequences of height and setback immediate.

The representation does not simply describe the project. It argues with it.

Juhani Pallasmaa describes drawing as an act of discovery rather than a neutral record. For him, sketching joins perception, thought and mental image; it is a way of thinking through the hand (Pallasmaa, 2009). I recognize that condition. A sketch often tells me something before I can explain it in words.

Drawing the city

The more my work moved toward urban design and master planning, the more important representation became. At that scale, the problem is not only the form of one building. It is the relationship among many things: streets, blocks, towers, parks, transit, servicing, topography, existing buildings and future phases.

A master plan cannot be understood from one drawing. I often need to see the same condition several ways: an aerial massing view to understand the district, a site plan to understand connections, a section to understand height and infrastructure, an area diagram to understand development capacity, and a ground floor plan to understand what a person actually encounters at the street.

This is one reason I became interested in diagrams. A good diagram removes information, but it has to remove the right information. Simplification is not the same thing as making something simple. It means deciding what the drawing is trying to say.

James Corner’s work on mapping is useful here. He argues that a map is not only a neutral description of existing conditions; it can reveal relationships, uncover possibilities and become an instrument of design (Corner, 1999). In master planning, I have found that to be true. The diagram does not come after the idea. Sometimes it produces the idea.

From the hand to the model

My tools now move constantly between analogue and digital thinking. In professional work I have used AutoCAD, Revit, Rhino, Photoshop, Illustrator, InDesign and Lumion across drawing, modelling, documentation and visualization. I have also repeatedly looked for better ways to move information between Rhino and Revit because the design model and the documentation model do not always want to behave as the same object.

That friction is useful. Rhino allows a certain freedom when testing form and geometry. Revit demands more explicit relationships, information and coordination. Photoshop or Illustrator can isolate an idea quickly. A sketch can ignore almost everything except a proportion or a connection. None of these tools is complete by itself.

The danger begins when the software starts determining the architecture because it is easier to model one solution than another.

Precision and ambiguity

I have learned to value both precision and ambiguity, but at different moments.

Early design needs enough ambiguity to allow alternatives to exist. A diagram or sketch can remain open. Later, architecture requires commitment. Dimensions matter. Areas matter. Clearances matter. The model has to coordinate with other disciplines. Information has to become transferable.

The skill is not choosing between a loose drawing and a precise model. It is knowing when the project needs one and when it needs the other.

The plan is still radical

For all the technology available to architecture, I still think the plan is one of our most radical instruments. A plan can show ownership, access, programme, circulation, adjacency, efficiency and public space at the same time. It can make visible who gets what part of the ground.

This became increasingly important in my Toronto work. On large mixed use projects, the plan is where many ambitions meet reality. Public realm, residential lobbies, retail, loading, garbage, ramps, bike access, structure and landscape all compete for the same ground floor. A rendering can postpone those conflicts. A plan cannot.

That is why I often return to the plan even when the design discussion begins with massing or image.

Section as city

Toronto also taught me to rely more heavily on section. At projects connected to transit and concourse systems, the public city exists on several levels simultaneously. Street, lobby, retail, concourse, PATH, subway and infrastructure can occupy the same site at different elevations.

A site plan can make those systems look separate. A section reveals that they are one spatial problem.

This has changed how I think about representation. Sometimes the drawing that best explains the city is not a map. It is a cut through the ground.

Research by drawing

In Two Cities I wrote about the personal library of residential projects I have built and continue to study. Representation is central to that research. I do not only collect images. I am interested in plans, dimensions, floorplates, cores, unit relationships, massing and the way projects meet the ground.

Comparing projects visually allows patterns to emerge. You begin to see recurring dimensions, recurring mistakes and recurring strategies. The point is not to copy a precedent. It is to understand what problem another architect was solving and whether that problem has anything to do with yours.

Research becomes useful when it changes the next drawing.

Drawing as communication

Representation is also how architecture becomes collaborative. A project exists across many people before it exists as a building. Architects, planners, clients, engineers, landscape architects, contractors and public agencies may all be looking at the same problem from different positions.

The drawing is often where those positions meet.

I have become increasingly conscious that a drawing does not need to show everything it knows. It needs to show what the conversation requires. Sometimes that means a detailed Revit view. Sometimes it means a black and white diagram with three colours removed. Sometimes it means a physical markup on a printed plan.

Clarity is not cosmetic. It changes the quality of the decision that follows.

Teaching the act of looking

This is also why, when discussing architectural activities for the ACE Mentor Program, I proposed exercises around floor planning, reading construction drawings, sketching and BIM. These are different tools, but the underlying skill is similar: learning how to look at spatial information and communicate what you see.

A sketch develops proportion and observation. A plan teaches relationships. BIM reveals how information and geometry are coordinated. None replaces the others.

AI and the problem of authorship

Artificial intelligence introduces another layer to this discussion. I am interested in it because it can accelerate exploration, research, visualization and communication. But it also makes an old architectural problem more visible: the difference between generating an image and making a design decision.

An image can be convincing while the architecture behind it is wrong. Geometry can drift. A façade can change. A plan can become something that was never intended. For architectural work, that is not a small error. The drawing is evidence of a project, so changing the drawing changes the project.

Robin Evans’s question remains relevant here. If architecture already works through translation, from drawing to model to building, then AI adds another translation. The question is not only what the tool produces, but what is lost, altered or assumed in the movement between representations (Evans, 1997).

Mario Carpo argues that digital tools have shifted from instruments for making to instruments for thinking, and that machine learning introduces a new form of design intelligence (Carpo, 2017). I find that useful, but I do not confuse computational speed with architectural judgment. AI can generate options quickly. It cannot decide which option matters.

I therefore see AI as most useful when it extends architectural judgment rather than replacing it. It can help test, compare, organize and visualize, but the architect still has to decide what must remain exact and what is open to interpretation.

The faster images become, the more important that distinction becomes.

What remains constant

The tools I use today would have been almost unrecognizable when I first began drawing. But the basic act feels familiar.

Look carefully. Decide what matters. Remove what does not. Draw the relationship. Test it. Draw it again.

Technology changes the speed and resolution of that process. It does not remove the need to see.

Conclusion: representation as a way of thinking

The first essays in this series were about public life, enclosure, private development and the two cities that shaped my architectural point of view. Representation sits underneath all of them because it is how those questions became visible to me in the first place.

I do not separate drawing from architecture as easily as I once did. Drawing can be observation, analysis, research, negotiation, instruction or speculation. A model can be a design tool and a coordination system. A diagram can make a complicated urban condition understandable in a few lines.

The important thing is not whether the tool is a pencil, Rhino, Revit, Photoshop or AI.

The important thing is whether the representation helps us see the architecture more clearly.

For me, drawing the city has always been one way of learning how to design it.

References / Referencias

Carpo, M. (2017). The second digital turn: Design beyond intelligence. MIT Press.

Corner, J. (1999). The agency of mapping: Speculation, critique and invention. In D. Cosgrove (Ed.), Mappings. Reaktion Books, 213 to 252.

Evans, R. (1997). Translations from drawing to building and other essays. Architectural Association Publications.

Pallasmaa, J. (2009). The thinking hand: Existential and embodied wisdom in architecture. Wiley.