I'd like to make a few comments about some great recent discussion between Dan Hill of City of Sound and Peter Tesugen.

One of Peter's posts raised Gehry's Guggenheim Bilbao and Kevin Lynch's idea of imageability in the same context, and I immediately thought of a concept that Daniel Dennett illustrated in a lecture a few years back (entitled "Problems and prospects for memes in explanations of human culture" if my notes are accurate). He described a cognitive psychology experiment based upon a variation of the old "Operator" game, wherein one person thinks of a message, whispers it to their neighbor, who in turn whispers what they heard to their neighbor, and so on throughout a large group of people. The "punchline" if you will is the last person revealing to the group the message they heard, which is sometimes very different from the original message due to the compounded transmission errors.

In the experimental version, what was being transmitted from person to person was a simple sketch, rather than a whispered message. There were two sketches used: one was a simple five-pointed star, the other was a tangled, "gestural" line (not unlike many of Gehry's small sketches). The first subject is given one of the sketches to copy, then passes their sketch to the next subject, who copies it, and so on. The experiment found that the star sketch could be reproduced through an almost indefinitely long chain and still remain true to the original, while the more casual sketch morphed in fundamental ways away from the original as it was copied. In the case of the star sketch, each subject would recognize the image they received as a star, map it to a relatively fixed conceptual image of a star in their head (which they shared with others), then reproduce the image based on that concept. There was no underlying cognitive "type" for the tangled sketch, so when people reproduced it they had no concept to guide them and inevitably strayed from the original. Dennett referred to the communication errors in this latter case as "thinkos" as opposed to "typos". It wasn't just a mistake that caused the error, it was a fundamental cognitive gap. His claim is that high-fidelity reproduction depends strongly on an underlying cognitive "alphabet". I mention this story because the idea of high-fidelity reproduction is quite similar to the idea of imageability, and I think the same cognitive factors are at play. I have found myself wondering whether any casual observer would be able to reproduce the image of Bilbao (which I think is a beautiful building, for the record), and I had suspected the answer was no. I'm not sure to what extent Dan Hill's sketches contradict this, but certainly Bilbao is far less sketchable than, say, the Eiffel Tower, and I would suggest posesses far lower imageability. I whole-heartedly agree with Dan's recommendation that The Image of the City be re-read every few years, and I'm afraid I'm due for a revisit myself and don't remember what Lynch might have suggested about the impact of legibilty at the scale of a single building instead of the larger portions of urban fabric he focused on. With regard to the legibility of software architecture (which is currently my trade), I think these observations support the software patterns movement inspired by Christopher Alexander's work, in that patterns can provide a shared vocabulary and a set of stable concepts across the wide body of software practitioners.

I haven't yet had the opportunity to read the writings and interviews by Nikos Salingaros, so I won't comment on those, but I will react to some of Dan Hill's thoughts on Salingaros' criticism. I attended a fascinating discussion with Peter Calthorpe, founder of the Congress for the New Urbanism, a few weeks ago, and for what it's worth I think he would object to holding up places like Seaside and Celebration as emblematic of the whole New Urbanist movement, as was suggested by Dan's recent post. He said that the style of neotraditionalism that has been common in New Urbanist developments must be kept entirely distinct from the real ideas about urbanism that he believes New Urbanism champions, which seem entirely in line with Jacobs' ideas about heterogeneity, mixed use, etc. It may have been hyperbole, but he claimed further that he didn't really even CARE what the buildings looked like, as long as their layout and structure improved the urbanity of the location. This was not met with enthusiasm from the architects in the crowd :) . The closest thing my notes provide to an exact quote is:

A place like Seaside does not seem likely to foster unexpected encounters, although I have not been there. I'm no historian of New Urbanism, but Calthorpe suggested that the movement had really broken into two coastal schools (West and East), with himself and the West Coast school focusing on issues of urban planning, and the East Coast school (Duany/Plater Zybek) focusing on a neotraditionalist style. New Urbanism may not be as monolithic a movement as is sometimes thought.