People have been confused about neighborhood centers and edges for decades, and that confusion leads to bad decisions. "Neighborhood center" and especially "town center" are broadly used to mean commercial centers. At the beginning, when a hamlet composed of a single neighborhood in the landscape is built (like the one shown on the left), the commercial center that springs up at the crossroads around which the hamlet is constructed is the center of the neighborhood of which the hamlet is composed.
But when the hamlet grows into a village (like the one shown on the right), it reorganizes itself into multiple neighborhoods that are bounded by the main streets leading to the village center. From that point forward, as the village grows to a town and then to a city, each neighborhood has civic spaces and uses (like parks, religious buildings, and the like) at their centers and commercial uses at their corners. Confusing neighborhood centers with neighborhood corners sets a series of errors in motion:
The Single Circle Fallacy
The core error is our simplistic view of the 1/4 mile radius that most people choose to walk instead of drive. Actually, the 1/4 mile radius itself is a fallacy, as we discussed in Walk Appeal… but that's another story. Whether the walking radius is fixed at 1/4 mile or a variable, I propose two overlapping layers of circles, which combine to weave the city together. The first layer is the neighborhood, which is bounded by the busiest thoroughfares, not centered on them. If you doubt that, talk to any mother of a small child. The child is inevitably forbidden to cross the busy thoroughfares, but is given free rein over larger areas of the neighborhood within as they grow. To ignore this fact is to miss something that is fundamentally understood by hundreds of millions of non-planners.
Double Circles
So what's the other layer? The second layer of 1/4 mile radii offsets 1/2 neighborhood in each direction to center on the retail at the intersections of the busiest thoroughfares. In other words, each quarter of a neighborhood probably shops at different corners of the neighborhood. In doing so, they meet people from quarters of neighboring neighborhoods. Those people, in turn, know others from their neighborhood centers, who know others from their retail intersections... weaving a web of relationships clear across the city.
The Ghetto Mistake
The "commercial center as neighborhood center" mistake is actually the prescription for a ghetto. There, a neighborhood is self-contained with little reason to enter or leave. The history of ghettos is spotty at best, of course. I maintain that the two-layer model is actually much more similar to the way that most traditional American towns once worked, and should be the strongly preferred model for a legion of reasons.
The Transect of Risk
Does this mean that the arterial can be left uncivilized? Not at all. I agree that context-sensitive arterials are the preferred model. But even if thoroughly civilized, let's face the fact that the risk to a pedestrian on an arterial will always be somewhat more than their risk on an alley or lane. Is it not possible to see the city as an ocean of waves of risk, undulating between higher risk and lower risk? Must we force everything to the same level of risk? Isn't that exceptionally boring? What I'm really talking about here is a transect of risk, which ought to be a natural part of life. Let's just make sure that the waves peak at an acceptable level.
~Steve Mouzon
You'll receive an email from me with the subject line "Mouzon Design: Please Confirm Subscription." Click Yes to confirm your subscription for Walk Appeal book updates.
Legacy Comments
Ann Daigle · Works at CityBuilding Exchange
Steve, I remember when we discovered this. Both this and the Walk Appeal posts are excellent. I do have one quibble... The "general" 1/4 Mile walking radius is a real factor and is evident in neighborhoods everywhere (and this is key) that were developed pre-auto dominance and according to human physical limitations of time and space and within their context. That's primarly why the "overlap" is so effective, as it increases daily opportunities. Of course people will walk further if the walk is nice. However, if one had no other option but to walk, it would be as routine as hopping in a car. Those who have spent lots of time in a walkable environments know that despite the pleasure factor, distance to daily needs (which equals time) is a real determining factor giving schedules of everyday life.
Dan Bartman · Senior Planner at City of Somerville (Official)
I'd like to see these concepts diagramed over real places.
Nicolas Derome
In Tokyo and much of Japan, much of the commerce is on narrow streets near rapid transit stations. http://goo.gl/maps/5xX13
Sometimes more bustling neighbourhoods can support retail on all the arterials plus more minor streets in American style cities too.
http://goo.gl/maps/rTiaa
I think this could become relatively common in American cities with a sufficient degree of revitalization/intensification. For instance, Philadelphia's core has few arterial roads so much of the retail could be on smaller streets, as is already somewhat the case for South Street or 9th street.
Low risk commercial streets are common in many Europe city centres, and even beyond them sometimes.
http://goo.gl/maps/Wvw1D
http://goo.gl/maps/46DpI
I agree that what you describe is true for most American inner cities though. I think retail should generally be near transit, and surface transit should usually go in more or less straight lines, so usually along long streets that because of their length also end up being arterials for cars. At that point, it makes sense to place amenities for children like schools and playgrounds away from the arterials (one thing Canadian planning has gotten right, even post-WWII). http://goo.gl/maps/zIMzB
If there's already retail on these arterials though, additional retail needed to support intensification could go on smaller low risk streets. Here in Toronto, examples would be Argyle, Robinson, Niagara, Logan, Wallace, Sorauren, Hillcrest and Northland.
588+