Monday, May 2, 2011

New Urbanism on Narrow Streets: Rhetoric and Reality

"If a street is to provide the sense of enclosure that pedestrians desire — if it is to feel like a room — then it cannot be too wide. ... This ... should come as no surprise to observant travelers, many of whom have no doubt used the words narrow and charming in the same sentence to describe the famously walkable streets of our older cities. Whether it be Montreal's Old Town, Boston's Beacon Hill, or Elfreth's Alley in Philadelphia, the narrowest streets are typically the ones most cherished by tourists and residents alike."
Andres Duany et al., Suburban Nation, p. 78.
Author's photo
Beacon Hill Street
(Acorn St.)
Circa 1820s
"Boston's most photographed byway"
Building Height: 3 stories

Roadway: 7 feet
Sidewalk: 5 feet
Landscaped Buffer: None
Setback: None

Total ROW: 12 feet*

Elfreth's Alley
Circa 1730s
"America's oldest residential street"
Building Height: 2 stories

Roadway: 6 feet
Sidewalk: 10 feet
Landscaped Buffer: None
Setback: None

Total ROW: 16 feet
Old Town Montreal Street
(Rue Saint-Paul)
Circa 1670s
"Montreal's oldest street"
Building Height: 3-6 stories

Roadway: 18 feet
Sidewalk: 12 feet
Landscaped Buffer: None
Setback: None

Total ROW: 30 feet*

*Measurements are estimates

Minimum Requirements of the Smart Code (Author: Duany, Plater-Zyberk & Company) for Urban Rights-of-Way

T4: "General Urban Zone"
Maximum Building Height: 3 stories

Minimum ROW Dimensions:
Roadbed: 18 feet
Sidewalk: 10 feet
Landscaped Buffer: 10 feet
Setback: 12 feet

Total ROW: 38 feet
Distance Between Facing Buildings: 50 feet
T5: "Urban Center"
Maximum Building Height: 5 stories

Minimum ROW Dimensions:
Roadbed: 28 feet
Sidewalk: 21 feet
Landscaped Buffer: None
Setback: 2 feet

Total ROW: 49 feet
Distance Between Facing Buildings: 51 feet

T6: "Urban Core"
Maximum Building Height: 8 stories

Minimum ROW Dimensions:
Roadbed: 36 feet
Sidewalk: 21 feet
Landscaped Buffer: None
Setback: 2 feet

Total ROW: 57 feet
Distance Between Facing Buildings: 59 feet


  1. Most New Urbanist stuff is really an attempt to create what I call 19th Century Hypertrophism. Actually, it is often a sort of Suburban Hell/19th Century Hypertrophism hybrid, with the addition of setbacks, Green Space and parking.

    Of course this is a recipe for failure.

    I would note, in your description, that a proper Really Narrow Street does not have a sidewalk. It is a pedestrian street from building to building, so it does not have a division between an automobile roadway and a pedestrian sidewalk. You could say the entire ROW is the "sidewalk." The second photo, of Elfreth's Alley, is a good example of what I mean. The first photo, of Beacon Hill Street, has what you could call a vestigal sidewalk. This may have been added later, during the 20th century, or it may be a reflection of 19th Century Hypertrophic design (wide streets with sidewalks) creeping into what is mostly a Traditional City design.

    It is hard to build right to the edge of a property, so often there is a small setback of 12-24 inches. Also, a small ridge, sort of like a curb but maybe two or three inches high, can be useful for drainage purposes. So, you see what sort of looks like a curb or sidewalk (building setback) in Traditional City Really Narrow Streets. However, basically they don't have sidewalks.

    The street in Montreal is more of what I call an "arterial street." This is also a part of the Traditional City, but paired with lots of Really Narrow Streets.

  2. Nathan -- thanks for the comment. Your piece on the New Urbanists was what inspired me to put together this post. Considering the praise that Duany lavishes on really narrow streets, helpfully citing one specific example, I thought it was noteworthy that his own code actually forbids the construction of streets anywhere near those dimensions (except for the rear alleys, as you point out, which are reserved for cars only). So what's going on here?

    As for the streets -- Montreal actually does not have that much in the way of really narrow ones, so I did the best I could. I almost wonder if Duany meant Quebec, which does appear to have a few. The odder thing to me is the almost complete lack of any non-North American examples here and elsewhere in the foundational New Urbanist literature. The Suburban Nation book has dozens of photos, only one of which is of a non-American place -- and it's a ten-lane Haussman boulevard in Paris. The failure to show (or even admit the existence of)European/Asian/Latin American small towns allows the claim to stand that the 19th century American small town is "traditional."

    Now it may be that Duany et. al. for whatever reason did not believe that those towns were viable models for early 21st century America(which I suspect is true), but if that was the case, he owed readers an explanation for why he reached that decision, rather than pretending that the question didn't exist. Interested to hear your thoughts.

  3. fwiw, those hyper narrow streets aren't common within those towns, they are more like alleys with housing, and alleys with housing are ways to monetize land that through the normal system of blocks and streets, would be on the interior of a block and less valuable and harder to access.

  4. I haven't read any of the New Urbanist "texts." Basically I got it all filtered through Kunstler. I had a sense that they had little to offer. Remember, I was living in Tokyo at the time.

    I would say that Duany and others were probably motivated by an urge to create "Small Town America." This is of course the most common motivation over the last 150 years, so nothing new there. Kunstler, for example, was born and raised in Manhattan, and now lives in Saratoga Springs. You don't hear much rah-rah for the 19th Century Hypertrophic City, large city version, from him.

    I think you do a wonderful job of showing the immense failures and blind spots among the U.S. "New Urbanists" today.

    I'll tell you what's more creepy, though. Why have the Europeans and Asians gone along with this stuff? Do you get anything better from Leon Krier?

  5. Nathan -- Duany's model in his writings is the "American small town" (his early 90s essay "The Second Coming of the American Small Town" is a good primer), but if you look at his actual developments, they owe more to the 1920s Garden City designs of John Nolen and Raymond Unwin than to the strictly gridded hypertrophic towns of the 19th century. These designs were basically a low-density streetcar suburb typology with more non-private green space and attention paid to the public realm (see e.g. Venice, FL).

    My hunch (and this is just speculation based on what I've read) is that Duany perceived these developments of the 1920s as striking an ideal balance between the "traditional" characteristics of the city (walkability, attention to civic spaces and public realm, mixed uses) and what was assumed to be an American market preference for freestanding single-family homes and continued near-universal automobile ownership. It was this assumption, however, that led to the design choices (requirement of wide ROWs and car-only access alleys behind all blocks) that make traditional urbanism impossible under the Smart Code.

  6. Los Angeles, as many have noted, actually has a rather high density of about 9,000 people per square mile. Of course this falls well short of the 65,000 people/mi2 of central Paris, which is seven times greater. However, I would say that Los Angeles probably represents about the highest density you can comfortably attain with a single-family detached suburban layout, with a few apartment buildings, and with the personal automobile. Generally, I would say that the residential areas of Suburban Hell are reasonably pleasant, and in fact nearly identical to Small Town America although typically somewhat denser. (1/8 acre plots instead of 1/4 acre).

    The problem then lies with the commercial/retail/office areas. If you use a 19th Century Hypertrophic model, of a line of storefronts on a wide street, then you have a parking problem. Basically, there are not enough people living within walking distance of the store (due to the low density of the suburban layout) to support the store. Without parking, the store cannot attract anyone outside walking distance. And, even those within walking distance might want to get in their car and drive to the mall or big box retailer. So, even those within walking distance are not a captive customer base. This drives retailers to locate in shopping centers etc. with lots of parking. Then you have car dependency because you can't walk to the store anymore.

    If you raise the density such that there are enough people in an area to support streetfront retail, then you end up with apartment buildings -- basically Brooklyn. Now driving is very unpleasant, practically impossible (try driving in Brooklyn), and there is a serious parking problem. Basically you end up with the dysfunctional 19th Century Hypertrophic pattern, which is not so great for pedestrians, not so great for cars, and not so great for families and so forth who now live in this pit of "congestion" (i.e. too much automobile traffic) and thus start dreaming of "green space."

    I don't see Duany's model as a solution for any of these issues. It is basically the suburbs just as they exist today, with some minor aesthetic tweaks. The attempts to recreate the storefront commercial street would basically be a failure, for the same reason that existing legacy storefront streets in hundreds of existing legacy 19th Century Hypertrophic cities and towns have been wasting away for thirty years. Actually, the typical shopping center is not so much different than the old-fashioned storefront street, just surrounding a giant parking lot rather than stretched along "Main Street."

    With all that said, I wonder if it is possible to build what amounts to a suburban residential district, using Traditional City methods, but with universal automobile ownership. Basically, you would have to build it and try it out. What I am thinking is ROW of 15 feet, no sidewalks, no on-street parking, attached (or nearly so) townhouses and apartment buildings with garages, no setbacks, backyards/courtyards optional, local mini-parks, a little bit of neighborhood commercial (bar, restaurant, small grocery, etc.). I suspect that there could be little enough auto traffic that you could get away with it. Basically, it would be one of these Tokyo single-family-detached neighborhoods that I show (see Jiyugaoka or Seijo). It would be even better if you built near a train station, that would eliminate half the cars and car trips if people commuted by train.

  7. Nathan,

    Long time reader, first time caller. To your point (idea?) of building a traditional city neighborhood but with car ownership, I would point you to three examples from my city. Look up Porter Street, Richard's Alley or Hancock Place in Wilmington Delaware. All three streets are narrow, two of the three feature garage parking under townhomes, and are vehicle accessible. Porter street is pre-car, and in the ghetto, but if you look at it solely from a planning/urban design perspective, its not bad. I think in all three instances, you'd have to go from 2 car households to one, but that's not such a bad thing.

    One could build an entire city of such streets (with arterials every 6 to 8 blocks), support some car ownership, and still have the required density to support walking and transit. This level of density (assuming most people have jobs) would probably support diverse retail. The problem with retail/commercial in Wilmington is that it has a high unemployment rate and all the post-industrial problems of a mid-Atlantic City.

    I work in the city, and generally commute by foot. I live within walking distance of a grocery store, pharmacy, bars and parks. I could live without a car, but if I want to leave the city to visit friends/places in the county, I have to drive. I think this problem may be the one thing keeping many Americans tied to their car; in principle they could (would?) abandon the car, but so much of the built environment is suburban. If you have business or pleasure outside of the city, (where 85% of New Castle county lives) you have to drive. That's why I haven't sold my truck. That and its a real beater that I couldn't get much $$$ for...

  8. I also have to wonder how much of the smart code is based on feedback from the Fire Code issue that CNU has encountered of late. I'm not super familiar with both but it might be of interest to look into.

  9. I just wanted to lamely say that I've been enjoying reading this blog immensely--am learning a lot! Although am still pretty tied to my car...can't imagine my life without it, sadly. KBG

  10. Pantograph: Thanks for the link. The Smart Code refers to its proposed street arrangements as being in conformity with traffic engineering requirements (which in turn are guided by fire codes), so I think you're right and that this reflects a pragmatic accommodation that gives the Code a fighting chance of being adopted in a given municipality. I do know the New Urbanists have been pushing to relax fire code-related street requirements for years -- Duany's book has memorable anecdotes on this. All that said, the setback and buffer requirements do not seem to be mandated by these same concerns, as far as I can tell.

  11. The Wilmington example may be worthwhile, but I would like to see some pictures. Probably, there is a ROW in the 19th Century Hyptertrophic format, of four lanes of automobile traffic in the middle (two lanes with parking on either side), sidewalks and so forth, giving a 60-100 foot ROW.

    As for fire codes, I was just talking to a person two days ago who is trying to build an "alpine village" type development at a ski area. In his talks with city code people, they decided that a 20 foot ROW building-to-building was acceptable. The justification was that "you had to allow two ambulances to pass each other." Actually, you could probably go down to about 14 feet and still meet this requirement, even 10 feet or so if you use more compact ambulances, but 20 feet is definitely something we could work with.

  12. I would say one thing for those people who "can't imagine" (they actually use this term!) living without their automobile.

    Living without an automobile in an automobile-dependent Suburban Hell environment is stupid. It's like trying to play golf without golf clubs.

    What I encourage you to imagine, instead, is living in a place where an automobile is not necessary, or at the very least, is not necessary for daily trips to work, shopping, school, etc. You can have an automobile for the occasional weekend trip.

    I would consider this -- the ability to imagine a no-car lifestyle in a no-car environment -- to be the first and most important step to actually making it happen.

    Here is some help in that regard:


    If that link doesn't work, search "narrow streets of Wilmington" in google maps, I've made it a public map. Most of the streets in city are 19th century, and wider than they need to be. That said, I think the mid-Atlantic model provides a good place to start for Americans to conceptualize a car free (or at least a car-lite) city. The colonial section of New Castle (also in DE) is another example.

    I think getting most Americans to imagine a car free city, they need a native, tangible example of how Americans lived before the auto. Remember that Americans are fairly provincial people, and suspicious of foreign examples of anything, especially urban design. It may not be logical, but its the world (and mindset) we have to deal with.

  14. The very narrow historical streets that you show can in fact be accommodated by the SmartCode, provided that state and local jurisdictions allow them. I'm not aware of any U.S. state or local juridictions that do allow such narrow streets and setbacks to be built in new construction, but if you can find any examples, they would be great to learn about! Alleys and pedestrian ways are a different matter.

    The SmartCode street standards are not actually required or mandated. They are found in the Complete Thoroughfare Assemblies module, which means they may be plugged into the base code or not, at the discretion of the code writers. The Complete Thoroughfare Assemblies module says,

    "Twenty-two typical assemblies are presented here for convenience. These may be added to the base SmartCode for the local calibration, and others may be created as necessary using the same template. They replicate closely the thoroughfare standards of municipal public works manuals."

    I suggest you contact Sandy Sorlien about narrow streets and SmartCode calibration. Her contact info is at

  15. Thanks, Laurence. What I was trying to do in this exercise was to take the narrowest options presented by the Smart Code and combine them to create the smallest ROW assemblies that the code appeared to comprehend, using both the Smart Code itself and the Thoroughfare Assemblies Module. If none of these outcomes even approach the historical examples Duany cites, my tentative conclusion was that the Smart Code was not really conceived with them in mind as a desirable design elements, even if with some creativity they could be accommodated in places (by designating such streets as alleys or pedestrian ways, as you've mentioned).

    I definitely will send Sandy an email on this topic, though -- thanks for the website link.

  16. Hello all,
    Laurence Aurbach is correct. The Thoroughfare Assemblies, created by several NU engineers along with DPZ, are a separate Module provided for convenience for the most common thoroughfare types needed in walkable neighborhoods, including downtowns. There is nothing in the Smartcode that prohibits narrower thoroughfares. Note: we don't use the term "roadbed" because not all thoroughfares are roads. The original authors, DPZ, were careful to analyze types of thoroughfares based on their components.
    Please see Section 3.7.3.e.ii of the v9.2 model code. You'll find your narrow streets (sic) there.
    Feel free to write off-blog, or here, if you have other questions about the code.

  17. I should add that I agree with Charlie that CATS should support the adding of narrower ROW assemblies in the Thoroughfare Module. While most very narrow thoroughfares fall into T4 or a T4-Open subzone, and the Intent for T4 (Table 1) in the SmartCode doesn't say anything about sidewalk width, it does say that there are sidewalks, which would discourage a mews or woonerf where you have shared street space.

  18. Since I live in Boston I visited Acorn Street a while ago to see what it was like in person.

    It might be picturesque, but it's not very walkable: the cobbles in the picture aren't paved, but set in concrete and are painful to walk on: I think the vestigial sidewalk is actually the vestige of the original paved roadbed and the "cobbles" were put in to keep people from driving on it.

    Based on the context of the street and the houses that surround it (as well as the history of Beacon Hill), I'm pretty sure that Acorn Street is actually a mews -- the alley, like the ones in the Suburban Follies posts, that let the wealthy Bostonians keep their horses and carriages in the back and over which the servents employed to maintain animals and equipment lived.