Places need signs
Information design, architecture and making buildings readable
Sign designers are convinced that architects don’t like them. They complain of being called in at the last minute, when it is too late to integrate the signs into the building, or of finding there is no suitable place to which they can actually attach the signs. Or they find themselves being briefed to develop signs that are effectively invisible – small grey ones, or even transparent ones. Tellingly, the sign outside the Royal Institute of British Architects headquarters in London is made from glass, for fear of obscuring the building beyond. This theme is reflected throughout Edo Smitshuijzen’s 2007 Signage Design Manual: ‘[Architects] perceive signage as an assault on the aesthetics of their creation and as an insult to the self-evidence of their spatial design. A lot of them carry an almost sacred but entirely unfounded belief in the functionality of their “wordless” buildings.’
In Learning from Las Vegas (1977), Robert Venturi and his co-authors attribute this preciousness to Modernism: ‘The integration of the arts in Modernist architecture has always been called a good thing. But one did not paint on Mies. Painted panels were floated independently of the structure … The diminutive signs in most Modern buildings contained only the most necessary messages, like LADIES, minor accents begrudgingly applied.’
There is merit in the idea of transparency. It would be good news if buildings or sites were more self-explanatory and so required fewer signs, and less cognitive effort on the part of the user. But the public building that truly works without signs is elusive – there is probably a limited set of building types that can genuinely self-explain. Their function will be restricted largely to insiders, with few visitors; they will be highly conventional in the way space is organised; the cost of failure must be low (you can afford to fail to find the entrance to a museum but you had better find the hospital emergency room quickly); they will probably have a simple memorable form; and the outside shape will reflect the internal organisation of space. I think I’ve just described a church.
Even if they do perfectly communicate their function, most public buildings get reconfigured and repurposed during their life – a challenge that document designers don’t face. In his classic book on this theme, How Buildings Learn (1994), Stewart Brand says: ‘In wider use, “architecture” always means “unchanging deep structure”. It is an illusion. New usages persistently retire or reshape buildings.’ So even if form does follow function, function changes, and, as Brand remarks, ‘Often information … is a cheaper fix than physical correction … The countermanding sign is an example of the way most problems are handled in buildings once they’re occupied.’
That words, whether in signs or captions, are an admission of failure seems to be quite a common belief – the utopian vision of universal communication is largely pictorial (viz. numerous discussions of data visualisation). But language is the supreme human ability. I am happy to see words on the door if it avoids the humiliation of trying to push when I should have pulled. And I would rather see the word ‘Paris’ in a film if
it averts the accordion music.
Reading a building
What can information designers learn from the way we navigate physical places, whether Las Vegas or anywhere else? And what can architects learn from the way we read text? They frequently use metaphors from each other’s domains: information designers speak of user journeys, and navigating a document or a website, as if it were a spatial environment, while architects speak of reading a building, as if it were a message.
When, following Kevin Lynch’s seminal The Image of the City (1960), we speak about legible environments, we’re looking at something altogether more significant and strategic than when we speak of legible books. Legible environments have clear pathways, areas and edges that explain a city and enable to us to form a coherent mental model to help us navigate it. But in print and on screen, legibility is just how easily we can read at the level of the word and the line. It is a hygiene factor – the equivalent of an architect getting the right pitch for a flight of steps, or the right height for a doorway. Unlike architects and planners, we have no word for the ease with which we can see typographic structures: the way topics are diagrammed through layout, the distinctions between different kinds of content, the clear and natural flow between pictures and text. It’s a pity.
The concept of reading a building can perhaps be traced to Ruskin’s distinction between ‘three great branches of architectural virtue’ that require of a building ‘1. that it act well’, ‘2. that it speak well’ and ‘3. that it look well’. He immediately opted out of developing further thoughts about the second branch, on the grounds that ‘it is evident that there are no universal laws’, and also because the modern observer of a historical building cannot know the context in which the intended audience would have read it. This is a pity, as a superficial view of what we are left with (functionality and aesthetics) appears to ignore communication as an explicit goal of the architect.
It is a small step, though, to a broader definition of functionality that includes clear communication. These days, from speech act theory, we understand that to speak is, in fact, to act: fire exit signs are part of the act of protection, along with the physical provision of doors; entrance signs are part of the act of welcoming. As well as the physical work of walking, opening doors and climbing stairs, users of buildings also do perceptual and cognitive work – looking, matching, asking, learning, thinking, inferring. Reading, in fact.
A naïve view of reading is that it is a linear processing of streams of written words, in the order that the author wrote them down. While that may be true of some reading for entertainment, almost all functional reading is an active process in which readers search for answers to questions, and seek to understand the whole structure before selecting where to pay attention. Reading is goal-directed – even where readers think they are reading compliantly, they are acting out the part of an intended reader, seeking goals the writer has planted. Even where the topic may be unfamiliar, or key information is missing, readers do their best to imagine possible worlds in which the writer’s words make sense. So sense-making in text environments depends on contributions from both the writer and the reader.
The reader’s contribution is to make interpretations not only from explicit information supplied but also from inferences they themselves make, from their prior knowledge, and from their existing mental models or schemata.
The user journey when you read an unfamiliar book is not unlike that when you visit a building for the first time. You go through stages: identifying the book or building; finding the starting point or way in; identifying it as one of a particular type that might conform to a known pattern or genre; planning a strategy for using the building (which may be highly goal-directed or which may start by browsing); making inferences about the location of elements; or getting help from an index, a map, a signage system or a human guide.
Conversations with place
Information designers understand that all effective communication is to some degree conversational – writers and readers co-operate in a mutual effort to create understanding. Writers try to anticipate readers’ questions, and readers try to imagine the writer’s personality and motives.
Just as in a verbal conversation, people ask questions of buildings, seeking answers through a combination of evidence and inference, as they assess how it is organised, where the entrance is, where their specific destination is, where to get information, or where the toilets might be. They can find the evidence in maps and signs, if provided, but from what can they make inferences?
One way in which readers of documents can make inferences is through their familiarity with genres or text types. Genres are well known configurations of information that, because they are common, have acquired names – letter, textbook, newspaper, catalogue. Each is organised in a conventional way, allowing readers to use their past experience to navigate each new instance with confidence.
Genre conventions may have originated in the way they are produced, their typical content or the way they are most efficiently used. But after a time, it is enough just to imitate past solutions – readers will know where to find the index, or the sports page, or the order form. They are institutionalised and repeatable solutions to common problems. Genres work because they conform to convention, even after the convention has lost its original functionality. Online magazines still have covers, even though they no longer have to attract your attention on a newsagent’s shelf.
Some buildings work in a similar way. For example, homes built in certain eras conform to strong patterns, so we know where the kitchen or the bathroom is likely to be. Among architectural genres, suburban homes still adhere to convention (and are often built without the contribution of architects), while public buildings – theatres, hospitals, galleries or schools – are much less likely to place such value on conformity.
Christopher Alexander’s concept of ‘pattern language’ is perhaps an architectural equivalent to document genres, but less uncritically conformist. Pattern languages identify and give names to common design problems, and list optimal solutions, both at the grand scale (towns) and the humble (shelves). There are areas of information design that explicitly draw on pattern language as a methodology. Interaction design, in particular, has too short a history for many genres to have developed naturally, and the concept of pattern language has been used in that community as a way to share learning from experience and to create conventions.
Architecture and inference
Most designers will be familiar with the concept of affordance, a psychological concept that influences both architecture and information design. The word describes the ability of an environment or design feature to communicate its function and encourage a particular user response. Some genres may not have good qualities of affordance but they work because the user is literate (that is, has learned their conventions). Documents suffer from poor affordance when page breaks occur inappropriately (for example, when the topic appears to have finished but has not) or where hierarchies of headings are not distinguishable.
Buildings suffer from poor affordance when you cannot identify the door, or how to open it, when there are no sight lines to your destination, or when attached buildings connect only on certain floors.
The theory of affordance has its roots in Gestalt psychology, and the realisation that our perceptual system actively seeks form and meaning. To give their work a sense of order and good form, both information designers and architects frequently use modular structures. The layout grids used by information designers are consistent vertical, and (sometimes) horizontal, points at which information may be aligned. By restricting the designer’s options, they make layout decisions easier, and also give the document a visual consistency that can help the reader navigate. In modular buildings, each floor may follow a similar structure, so users can transfer their knowledge of their home floor to all the others. Modular structures allow users to use the logic of assembly to make inferences. Our knowledge of how buildings are made may also allow us to make further inferences that are not dependent on grids: for example, we may expect the toilets to be near the elevators, if we are aware that both require the use of the same vertical service shaft. We may, of course, come to the same conclusion without that knowledge, simply through our familiarity with a genre convention or pattern.
Architects are familiar with physical constraints: buildings must be engineered to stay up and not fall down, and, however monumental they are in scale, they are ultimately constrained by the human physique – stairs must not be too steep to climb nor doors too heavy to open; and skyscrapers were feasible only once we had the lift and the telephone. But for architecture to be truly expressive of its function, should not architects pay equal attention to the ability of users to do cognitive work? After all, mere expression is somewhat functionless without a recipient to see it and give it meaning.
This would appear to discourage innovation, since public buildings that are usable from the start must inevitably be based on existing conventions. To borrow Umberto Eco’s terms, they must largely be code-using, rather than code-extending.
The best wayfinding projects – the best motivated, the best informed, the best executed – demonstrate that, like good architects, information designers can work within constraints to innovate and delight without compromising function. And they demonstrate that, because they innovate and delight as well as inform, signs need be seen neither as a threat to the visual integrity of a building, nor as a criticism of its ability to speak through its visual form.
Just like handles on doors, signs should be there from the start.
1. I base this term on E. H. Gombrich’s ‘geometry of assembly’; see his The Sense of Order (Phaidon, 1979).
2. See, for example, some of the projects illustrated in Andreas Uebele’s Signage Systems and Information Graphics (Thames & Hudson, 2007).