# Chapter 4: Learning from Precedent When the New Urbanists began their revolution in the 1980s, they did something radical: they looked around. While academic planners debated theories and traffic engineers calculated level-of-service ratios, this small group of architects and planners simply asked: _which places do people love? What patterns make neighborhoods walkable, memorable, and resilient?_ Then they measured them. This empirical approach—learning from what works rather than theorizing about what should work—has a long history in design. It's how humans have always built, at least until recently. The Roman architect Vitruvius didn't invent the column orders; he documented what Greek builders had refined over centuries. Medieval cathedral builders didn't calculate flying buttresses from first principles; they learned from the successes and failures of previous cathedrals. Precedent served every role at once. Somewhere in the 20th century, we decided we were smarter than history. We could engineer optimal cities from abstract principles. The results speak for themselves: the places we love were mostly built before 1945, while many of the places built after leave us wanting. It's time to be humble again — to acknowledge that thousands of years of city-building left us something worth knowing, and to measure it. ## The Empirical Turn The approach underlying this book is simple: study great places, measure their dimensions, identify patterns. Not anyone's personal favorite places, not theoretically optimal places, but places that independent observers have identified as successful. Places that have endured. Places people choose to visit, to live in, to preserve. This empirical approach has three key advantages over theoretical planning: First, it's been tested by time. A neighborhood that's thrived for centuries has proven it can adapt to technological change, economic cycles, and shifting demographics. That's worth more than any theory. Second, it's been tested by people. Real humans have chosen these places — moving to them, investing in them, defending them. Tourist guides recommend them. Retirees settle in them. Property values reflect their desirability — centuries of real-world evidence. Third, it reveals universal patterns. When neighborhoods in Barcelona, Charleston, and Kyoto share similar dimensions despite being built centuries apart by different cultures, we're seeing something fundamental about human scale and urban form. ## Selecting the Data Set The first challenge was choosing which places to study. Personal favorites would introduce too much bias. The solution was to rely on sources that had already done the hard work of identifying great places. The American Planning Association proved invaluable. Since 2007, they've annually designated "_Great Neighborhoods,_" "_Great Streets_," and "_Great Public Spaces_" across America. Their criteria include: - Architectural features and accessibility - Functionality and memorability - Community involvement and vision - Environmental sustainability - Economic vitality This provided a solid core of American examples. But great urbanism isn't uniquely American, so the search expanded internationally. Travel guides like Fodor's and Lonely Planet provided examples from around the world—the neighborhoods they consistently recommend to visitors. The AARP's list of great places to retire added another dimension, focusing on places that work for people across their entire lives. The study also included places repeatedly cited in urban planning literature: Savannah's squares, Barcelona's Eixample, Paris's boulevards. These canonical examples might be overstudied, but they're studied for a reason. The final data set included over 100 neighborhoods from around the world, ranging from ancient Roman settlements to 19th-century American towns to medieval European cities. Small towns with populations under 500 and major capitals with millions. Hot climates and cold. Rich and poor. The diversity was intentional—if patterns emerged across such different contexts, they'd likely be universal. ## The Measurement Protocol With places selected, the next step was developing a consistent analytical method. Geographic information systems (GIS) and Google Earth transformed what would have been quite difficult a generation ago. But technology was just a tool—the real work was deciding what to measure and how. The process follows these steps: Start with an aerial photograph (or GIS data if available) of a neighborhood. Overlay trace paper (or create a new digital layer). Draw only the public rights-of-way—streets and alleys. All the visual noise disappears: no buildings, no trees, no cars. Just the permanent framework. Label each right-of-way and measure its width. Count how many of each width exist. Calculate the average, the median, the distribution. Create a frequency chart showing which widths are most common. For blocks, measure each dimension. Calculate perimeters. Note the ratios of length to width. Identify the median block (half are bigger, half are smaller) and use it as representative. Create a scatter plot with short dimension on one axis, long dimension on the other. Each neighborhood becomes a dot on the chart. This process [^1] revealed something unexpected: how much precision went into these places. Streets weren't random widths—they clustered around specific dimensions. Blocks weren't arbitrary shapes—they followed clear geometric patterns. Even in medieval cities that looked organic from above, the underlying order was remarkable. ## Why Precedent Beats Theory Thomas Huxley, 19th-century English biologist, once quipped: _"The great tragedy of science—the slaying of a beautiful hypothesis by an ugly fact."_ This captures why precedent matters more than theory in city planning. A theory can be internally consistent, mathematically elegant, and completely divorced from human reality. But a place that's worked for centuries, tested by millions of lives, thousands of businesses, hundreds of economic cycles — that's not theory, that's evidence. Consider the Radiant City of Le Corbusier—towers in a park, separation of functions, abundant green space. In theory, it solved the problems of overcrowded industrial cities. In practice, it created isolated towers that discouraged street life, vast unusable lawns, and social alienation. The theory was perfect; the reality was Pruitt-Igoe. Or consider the traffic engineering theories that gave us the street hierarchies of modern suburbs—local streets feeding collectors feeding arterials. In theory, this separates through traffic from local traffic, improving safety and flow. In practice, it concentrates traffic on a few overburdened roads, makes walking impossible, and prevents the fine-grained structure that cities need to thrive. Meanwhile, the "boring" grids of successful cities keep working century after century. Their redundant connections mean traffic has options. Their consistent dimensions mean any block can accommodate any use. Their simple geometries are perfect for the simple boxes we call home. They work not because they're theoretically optimal but because they've been optimized by use. The architect Louis Sullivan famously declared that 'form follows function.' But in city planning, we face a paradox: the function keeps changing while the form stays constant. A street built for horses now carries cars. A warehouse becomes lofts. A mansion becomes apartments. The uses—the functions—transform with each generation, but the street pattern—the form—endures. This is why precedent remains relevant: it shows us forms that have successfully accommodated countless functions over time. ## The False Promise of Connectivity Metrics In our search for better ways to measure and regulate street networks, a seductive solution emerged: _connectivity metrics_. These mathematical formulas promised to distill the complex reality of urban form into simple numbers. Planning departments embraced them. LEED-ND adopted them. Even Virginia's Department of Transportation built entire regulatory systems around them. There was just one problem: they don't work. To understand why, let's examine what these metrics claim to measure. The three most common are: _Connectivity Index_ (links divided by nodes): Supposedly tells you if a network is rural (<1.4), suburban (1.4-1.6), or compact (>1.6). _Intersection Density_ (intersections per square mile): LEED-ND awards points based on thresholds from 200 to 400+ intersections per square mile. _Street Density_ (centerline miles per square mile): Another attempt to capture network intensity through simple division. The appeal is obvious. Planners can require "a minimum connectivity index of 1.6" and feel they've ensured good urbanism. Developers can calculate their score and know they've met the standard. Clean, objective, scientific. Except when we tested these metrics on real places, the results told a different story. ### The Savannah Deception Take Savannah, Georgia—universally acknowledged as one of America's finest examples of urban planning. When we measured 25 of Oglethorpe's famous wards together, the metrics confirmed what we already knew: - Connectivity Index: 1.63 ("Compact") - Intersection Density: 755.8 per sq. mi. (Maximum LEED points) - Street Density: 46.1 mi. per sq. mi. (High score) Perfect scores for a perfect place. The metrics work! But Savannah wasn't built 25 wards at a time. It grew ward by ward, each one a complete unit. So what happens when we measure a single ward—the actual building block of Savannah's urbanism? - Connectivity Index: 1.40 ("Suburban", nearly "Rural") - Intersection Density: 1060.8 per sq. mi. (a 40% jump) - Street Density: 57.8 mi. per sq. mi. (a 25% jump) The same perfect urbanism now scores as suburban sprawl. By simply changing where we draw the study boundary, intersection density and Street Density jump while the connectivity index plummets. If Oglethorpe had faced modern connectivity requirements, he might not have been allowed to build Savannah. ### The Manhattan Manipulation The problem compounds when we realize how easy this is to game. Take a section of Manhattan's famous 1811 grid—200-by-600-foot blocks that have successfully hosted everything from farmhouses to skyscrapers. Measure four blocks (2×2) and you get: - Connectivity Index: 1.33 ("Rural") The same grid that supports some of Earth's densest urbanism rates as "rural." Expand your boundary to sixteen blocks (4×4) and suddenly: - Connectivity Index: 1.60 (Compact) Same streets. Same blocks. Same urbanism. Completely different scores. ### Why Metrics Fail The mathematical reason is simple: these metrics divide discrete elements (whole intersections, integer street segments) by continuous areas (which can be 40.000 acres or 40.001 acres depending on where you draw the line). It's like measuring the quality of a novel by dividing the number of periods by the weight of the book—the result changes based on irrelevant factors like paper thickness or margin width. The practical implications are worse. A savvy developer can manipulate scores by adjusting study boundaries. Include that cul-de-sac subdivision next door? Your connectivity drops. Draw the boundary down the middle of a street? Your intersection count changes. The metrics incentivize gaming rather than good design. But the deepest problem is what these metrics miss entirely: the actual dimensions that make urbanism work. A connectivity index of 1.5 tells you nothing about whether blocks are walkable, whether streets create useful frontage, whether the pattern can adapt over time. Manhattan's 200-by-600-foot blocks work brilliantly. So do Savannah's wards. So do Charleston's 300-by-400-foot blocks. But the index can't tell them apart. ### Back to Reality This is why precedent beats metrics just as surely as it beats theory. The push for connectivity metrics came from good intentions—trying to ensure connected, walkable places. But like so many theoretical approaches, the abstraction lost touch with reality. Real places have actual dimensions—blocks 300 feet long, streets 60 feet wide. These dimensions either work for human habitation or they don't. They either create value through frontage or they waste land. They either allow adaptation or they lock in disfunction. No amount of mathematical manipulation changes these physical realities. The metrics are appealing because they promise certainty through calculation. But urbanism isn't resolved through ratios. It's resolved through form—specific, measurable, physical form that has proven itself through use. We measured real dimensions and found real patterns. The rules in this book are based on those findings — on form, not formulas. The numbers that matter aren't abstract ratios. They're the 300 feet you walk to the corner store, the 60 feet between building faces, the 20 feet of alley that hides the dumpster. These dimensions don't change when you redraw a study boundary. They don't depend on mathematical gymnastics. They simply work. ## The Archaeological Approach There's something archaeological about measuring old street patterns. You're uncovering decisions made by people long dead, in contexts you can only partially understand. Why did they make this street 60 feet wide instead of 50? Why are these blocks 300 feet long instead of 400? Sometimes explicit answers emerge. The _Laws of the Indies_, which governed Spanish colonial city planning, specified exact dimensions for plazas based on the number of residents. The _U.S. Land Ordinance of 1785_ established the 640-acre section as the basic unit of Western expansion. These were conscious decisions that left paper trails. More often, the logic is implicit, embedded in the pattern itself. Medieval cities that seem chaotic from above reveal, on measurement, remarkably consistent block sizes. The curves and angles (limited public space being eroded by individuals' power over centuries) that look random turn out to follow property lines, topography, or defensive needs. The apparent disorder masks underlying order. But collecting measurements was only the first step. Like archaeologists who must piece together shards to understand a whole civilization, we needed to synthesize thousands of individual observations into coherent patterns. The real discovery would come not from any single measurement, but from what emerged when all the data was viewed together. ## The Neighborhood as Unit of Analysis There was another crucial decision in our methodology: choosing to study neighborhoods rather than entire cities or random collections of blocks. This wasn't arbitrary—it reflects something fundamental about how cities actually work and how we experience them. Neighborhoods sit at the right scale for understanding urban patterns. They're large enough to function as complete systems with their own internal logic, yet small enough to perceive as coherent places with distinct characters. When people describe where they live or visit, they typically reference neighborhoods: "I'm staying in the French Quarter," not "I'm staying in New Orleans." This intuition has deep roots in planning theory. Clarence Perry's influential 1929 concept of the _neighborhood unit_ argued that cities should be organized around walkable neighborhoods scaled to daily life—particularly a child's walk to school. Perry calculated this as roughly a half-mile square, containing housing, schools, parks, and shopping within a five-minute walk of most residents. Our analysis confirmed that real neighborhoods tend to cluster around a similar scale, though often slightly larger than Perry's ideal. From Savannah's _Historic District_ to Vienna's _Innere Stadt_ to New Orleans' _Garden District_, successful neighborhoods typically span 0.6 to 0.8 miles on a side—walkable in 10-15 minutes, large enough for variety, small enough for coherence. This scale makes practical sense for analysis. Cities are essentially agglomerations of neighborhoods, stitched together by major infrastructure but maintaining distinct identities. By studying neighborhoods, we capture the repeating unit of urbanism—the module that, when successful, can be adapted and repeated to build larger cities. Focusing on neighborhoods also avoided two methodological traps. Analyzing entire cities would have introduced too many variables—industrial districts, highways, geographic barriers—that obscure the patterns of everyday urbanism. Selecting random blocks scattered across cities would have missed the systematic relationships between blocks that create functional neighborhoods. The neighborhood scale captured just enough context to see how urban elements work together while maintaining analytical clarity. ## Conclusion The database was complete. Hundreds of neighborhoods and thousands of dimensions had been measured and recorded. Those measurements revealed patterns that challenge modern planning assumptions. --- Continue to [[Chapter 5 - Universal Patterns]] → [^1]: The measurement protocol developed for this research has since been formalized into the _**Urban Form Atlas**_, a reference collection of neighborhood "x-rays" showing urban form at consistent scale. See dougalleninstitute.org/urban-form-atlas.