Oct 13 2010 (a 6 minutes read)
As with many a good discussion, this got started by chance. While I was cruising around with too many tabs open in my browser, one of them caught my attention. Stefano Bussolon (@sweetdreamerit), psychologist and information architect, was commenting on something Peter Morville (@morville) had just tweeted on structures and patterns. Stefano's tweet went like this:
Information architects organize the patterns inherent in data (@morville). Participatory IA discovers patterns inherent in data (me).
This bugged me somewhat, for reasons I'll explain, and I replied. After a few exchanges, Stefano concluded by saying:
inherent= in the head of our (potential) users: let their mental patterns go outside ;)
Now, with this in mind, let's bring in one of my favorite famously unrelated examples. The night sky is a fascinating view when you are lucky enough to see it from places where the bright lights of big cities are not there to dim the Milky way out of sight. Some of us can find the most well-known stars or visible planets, and name them. We also know that stars can be clustered together in constellations, in patterns. Like in "give them stars some structure, dammit". Some of these constellations are famous, like Ursa major, the Great Bear, or Ursa minor, the Little Dipper.
Those of us who for unfathomable reasons really paid attention in school can possibly trace out the profile of Cepheus , or Cassiopeia.
Constellations are evidently a mnemonic artifact. But is that all? No sir, it is not, as anyone interested in their horoscope could tell you. Have they been useful? Yes, definitely, and in many different wondrous ways, all throughout human history. Constellations were probably linked to initial religious beliefs, were widely used by farmers to understand the passing of the seasons way before calendars became a common tool , and we know from a thousand bad adventure movies that you can use Polaris, the North star, to navigate at sea. Constellations still have a proper scientific use today, as they are used to identify which stars astronomers are talking about  in a much quicker way than providing their latitude and longitude. Mu Cepheus is sufficiently precise for conversation.
In all, constellations certainly did their part. But then again, constellations do not really exist. They are patterns projected on the celestial sphere, that fictional globe surrounding the earth on which the stars appear to be. The key word is "appear": stars in the same constellations can be hundred of light years apart, and constellations are simply an illusion created by distance, and by our position inside our galaxy. If the Earth were to occupy a different place in space-time, we wouldn't have these constellations at all. They are not inherent, we hand-picked them, we built them.
I believe Peter's initial tweet, the one Stefano Bussolon commented and re-tweeted, taps into something Richard Saul Wurman wrote more than 30 years ago.
In 1976, Wurman chaired the national conference of the American Institute of Architects (AIA). He chose "The Architecture of Information" as the conference's theme. According to what he said to Dan Klyn in a recent series of interviews , he had no plan in mind to start anything, he was just trying to "find patterns for himself". I love that sentence, the designer in me frolics and cartwheels around. Patterns for himself.
For that conference, Wurman came up with a three-headed definition. The information architect is:
While I still like point #2, and believe #3 is a bit too fuzzy or broad for my tastes, I always had an issue with #1: not with the general idea, that's a very good definition of an IA in my book, but with the part which calls the patterns in data 'inherent'. The dictionary says that inherent means "existing as an essential constituent or characteristic; intrinsic". So, 'patterns inherent in data' does not mean 'in the head of (the) potential users', as Stefano suggested in his follow-up, but rather intrinsic, built-in into the data itself. Remember our constellations.
I need to quote Wurman once more: Presenting information for a purpose is an architectural task, and nothing in architecture is preordained. I could stop here, as that pretty much sums it up, but unfortunately I can't condense my examples like Andrew Hinton does . I need to add one more.
One of my favorite novels, and one that features prominently even in the upcoming Pervasive Information Architecture book I'm co-writing with Luca Rosati, is The Name of the Rose, by Umberto Eco. It's a complex, multi-layered story, which can be read at so many different levels. It's also an incredibly captivating book, which has to be read in its original Italian  to be appreciated in full.
Anyway, the novel can be read as a whodunnit kind-of-story, where people get killed and a detective, Brother William, investigates. In the course of a week William manages to understand quite a lot of what is going on, including that an homicide is actually a suicide, that heresy has nothing to with the crimes, and that everything revolves around a book which should be long lost and forgotten.
But at the end, when the plot gets untangled, everything is lost, and William is defeated, we finally understand. As William says:
"I arrived at (the villain) through an apocalyptic pattern that seemed to underlie all the crimes, and yet it was accidental. I arrived at (the villain) seeking one criminal for all the crimes and we discovered that each crime was committed by a different person, or by no one. I arrived at (the villain) pursuing the plan of a perverse and rational mind, and there was no plan, or, rather, (the villain) himself was overcome by his own initial design and there began a sequence of causes, and concauses, and of causes contradicting one another, which proceeded on their own, creating relations that did not stem from any plan.”
Contemplating the furious fire which is ravaging the abbey William finally comments: “There was no plot, and I discovered it by mistake.”, to which Adso, his scribe and disciple, replies: “But in imagining an erroneous order you still found something.”
That's it, that's what we do: we build structures and patterns, we imagine orders.
These are not there, in the data, in re, but are a part of our individual or collective imaginary, of our professional expertise. As such they can be fantastically "wrong" but still produce meaning, and good design.