Dear Tomas. I don't mean to be disrespectful, but the claim, based on your status as a physicist, that the urban growth model is deterministic "as in physics" (which isn't true for all of physics either, and I can say this with complete confidence, even though I'm a biologist, not a physicist, and have been working in territorial planning for 16 years), and that you can make such an outrageous claim based on internet data alone, turns this medium into just another digital outlet promoting misinformation. Saying this, when there are very serious studies based on archaeological and demographic evidence, which appeal to "complexity" precisely because deterministic science is insufficient to understand urban processes of the past, or those of the present, is disingenuous. This isn't the first time you've written on a variety of topics, including many "urban" ones, with simplistic assertions. I recommend reading "Complexity" by Lewin, or "Complex Environmental Systems" by Rolando García. There are also numerous online research papers in the journals Urban Planning, Landscape Planning, and npj Urban Sustainability, among others. This would allow you to be a little more humble and, even with the same data, introduce probability and compare your 30 or perhaps slightly more city cases. This would show you a minimal margin of error, which you ignore, but which justifies an analysis based more on function than on parts. This, like any error, also confirms the rule, but expresses the probabilities that lead to understanding why urban processes are so varied throughout history. And even more so today. As in gas physics or thermodynamics, for example.
Thanks! I think you might have taken “deterministic” in a way that is too literal. I didn’t mean it as in math, more as in “there are some very strong rules that really push some spots in certain directions.”
To illustrate: if we went back into history 10k years ago and let history unfold 1 million times to see what cities appear where, and then take the averages of these simulations, we would have a distribution of potential city placements and their size. My guess is that distribution would seem eerily similar than the actual cities we see today. Paris and London would definitely be in the same place as they are today, and would be also the biggest cities in their area. Moscow might not end up exactly there, but it would end up close by. Etc
Dear Tomás. Forgive my vehemence, but this is a topic I'm passionate about, and a complicated one at that. I'd love to provide more information, and I'll try to do so to enrich the debate. But for now, I'll leave aside a fundamental popular text on complexity, which begins with a case study of a pre-Hispanic American city by Roger Lewin:
In general I disavow arguments by reference: If you have an argument to make, make it, don't refer to something else. Especially true for books: I can't possibly read every book I'm recommended.
I did just skim the 2nd link (article) and I'm underwhelmed:
• It's purely theoretical. It doesn't make any specific claim.
• It doesn't generate concrete, testable predictions.
• It's full of jargon, frequently used to obfuscate lack of insight.
• It has no empirical evidence.
• None of its claims are against mine! It claims cities are unpredictable in detail (I agree!) not as a whole.
To this last point, the article uses a metaphor with thermodynamics (which it doesn't support well, and uses one of its concepts, entropy, which I've studied in grad school, and I don't think it means what they think it means: Cities are a clear example of local reduction in entropy, by extracting energy from its surroundings).
We can use another thermodynamics metaphor to illustrate the difference between my claims and those of the article: You can't predict the movements of a single metaphor, much less the state of its atoms. They are complex systems. But as a whole, when you take billions of molecules, you can absolutely tell how they will behave. You can know how their temperature and pressure will evolve in different situations, with fairly simple rules that are quite deterministic.
This is what I mean in my article: There are absolutely rules that can define where cities will emerge, how big they will be as a whole, and the types of structures that are more likely to be optimal in them. This doesn't mean you can determine which building will go up where, or things like that, which is what your 2nd link is focused on.
In any case, I picked a random paragraph in your article:
"La inercia de las estructuras relacionales –característica intrínseca de la identidad– conlleva la existencia de patrones que perviven en el tiempo y que se asocian al fenómeno de las persistencias. Las estructuras subyacen y permanecen y permiten identificar esta capacidad de reorganización propia de la ciudad compleja, a menos que realmente se opte por la tabula rasa (Ruiz Sánchez, 2001, p. 32-35). Es en las transformaciones directas sobre la estructura relacional en las que la identidad y unicidad de la ciudad se puede ver realmente afectada, provocando cambios que alteran su identidad."
Can you actually extract anything valuable from this type of paragraph? This sounds like BS to me.
"This is what I mean in my article: There are absolute rules that can define where cities will emerge, how big they will be as a whole, and the types of structures that are most likely to be optimal in them. This doesn't mean you can determine which building will go up where, or things like that, which is what your second link is focused on." And you're right. I'll try, take some time, and gather more specific and appropriate bibliography, but I insist that the history and evolution of cities cannot be addressed in a deterministic way, because by definition in a complex system, the object of study is not given, and is constructed by the nature of its variables, which are also multifunctional. I understand that you'll say it's a theorization (I'm also a professor of science methodology), but I will strive to show with data why a city cannot be addressed with deterministic models.
Who says the author is making “an outrageous claim based on internet data alone?”
If memory serves correct, I think that “city patterns are based on physics” is exactly the same argument that Geoffrey West made in his book “Scale.” Indeed, I believe that West is a leading proponent of Complexity theory, so I am not sure why you consider Complexity to be incompatible with physics.
Dear Michael, I'm going to get and read the book you suggested, for the simple reason that it argues what I maintain. It's not deterministic physics that explains the phenomenon, but rather its complexity. And the author makes this clear when he refers to "scale," at which various "rules" are produced that explain some structures that emerge from the function beyond the type of system being studied. Scale is "the exception" that confirms the rule, but precisely the one that establishes why its trajectories are not deterministic, because what at certain scales are processes are structures at others. For example: a wetland in a city at a certain scale is the structure of the urban watershed, but at the urban scale it is a floodable area due to the organization of its drainage network. For planners, this is a problem because they must resolve its "multifunctional dimension" and predict its evolution simply by being affected by what, in terms of complexity, defines error: exogenous variables. We can continue talking about problems in "nondeterministic" physics, such as the three-body problem and the probability of "determining" the position of a body when it is affected by two others. No urban process can be reduced to two variables!
My main concern was the “vengeance” that permeated your comment. You could have phrased things in much better way. It was, after all, a very good article. I know something about villages and cities in history, and Pueyo was largely correct.
As for the book, the author believes that cities, biological organisms, and all other complex system can be explained by “ Quarter-power scaling laws” where positive consequences outweighs negative consequences as a thing grows in size.
Tomas, it might be interesting to study how these patterns apply to Amish farmers in the US. Their 30 minute radius is determined by the speed of a horse-drawn buggy and they have certainly been require to expand to new villages as their population has grown. From beginnings in eastern Pennsylvania they have expanded to new communities in Ohio, Maryland, Indiana, West Virginia and further west.
Leonardo, you say you “don’t mean to be disrespectful,” but then call the article “misinformation,” “outrageous,” “disingenuous,” and “simplistic.” That feels more dismissive than constructive.
I agree complexity matters, and of course no single model explains everything. “Deterministic” is a strong word and does need nuance, but I didn’t read the article as trying to replace archaeology or demography..It is pointing out that clear regularities show up in urban data and making those patterns visible for a wider audience. That’s not misinformation, it’s sparking curiosity and debate.
Calling it “misinformation” just shuts down a useful conversation. The references you shared look great, and adding probability and more cases is a solid next step. But if we’re talking humility, maybe that reminder cuts both ways. We’ll get further by building on these ideas instead of writing them off.
You're right, Patricia. I didn't mean to be offensive, but unfortunately, those of us who deal with territorial planning face simplistic analyses on a daily basis (I swear I don't mean to be offensive) that reduce urban rivers to piped drains, draining surfaces to dry plazas, and turn any land surface into a business opportunity, even when there's evidence that the soil may be unstable or unhealthy. I can assure you that the history of a city's development doesn't obey deterministic laws, and Tomás exposes some of this in the explanation, but not in the data (surely due to the nature of his article). Of course, there are patterns that define trajectories, especially throughout history, but urban science is moving in a different direction and is laboriously working on approaches diametrically opposed to what Tomás proposes. I insist, a quick review of the scientific literature on the subject over the last 30 years reveals a different approach based more on function than on parts.
Yes, I would be as well, because I think (with a few minor exceptions) you got it exactly right.
I think the commenters reaction was more about keeping people from outside self-identified urban design experts quiet. As a professor, I saw this attitude all the time when outsiders suggested a new approach or theory. But the best ideas often come from outsiders.
Thanks for explaining, Leonardo. I get where you’re coming from, and I see how frustrating it can be when complexity gets reduced to oversimplified plans. What I took from Tomás piece, though, was less about dismissing that complexity and more about showing the broad regularities that stand out.
And I think there’s room for both, the detailed work you mention and the broader patterns that make these ideas accessible to people like me. That is exactly what I enjoy about this newsletter, different angles coming together and making the conversation richer (at least that's how I learn!). Appreciate you taking the time to add to your initial comment.
The author of the book that I linked gives a very persuasive argument that cities do, in fact, obey deterministic laws: Quarter-power scaling.
If “ urban science is moving in a different direction and is laboriously working on approaches diametrically opposed to what Tomás proposes” it might well be driving itself into a dead end. This would not be the first time that urban design zealots have tried to reinvent cities only to make them worse.
Understanding why cities grow organically according to certain identifiable patterns and then working within them seems to be a much better approach.
Leonardo, my reaction to this article is quite different from yours. That any regularity at all arises in village placement and structure is wondrous. Rather than denying complexity, to me it brings to mind Phil Andersen's epithet "More is different". I gather, as an urban planner you see important differences between instances of urban growth, which is another aspect of complexity. To me that doesn't diminish the possibility that human-built things like villages arise with certain patterns because all humans share certain features. In a similar way I'm reminded that while all humans are unique, and human development is an extremely complicated process, yet most humans have 2 legs, 2 arms and 2 eyes. That doesn't in any way diminish the uniqueness of each human, but there is something in the complexity that produces some regularity and similarity between most humans. "More is different" hit the vernacular just as the study of complexity became mainstream in physics. To me this article explores that kind of complexity. It is a kind of determinism in the sense that "beings with these attributes in environments with these other attributes will develop villages with these kinds of observable features". Isn't that what all of us mean by complexity? (Or maybe you are saying that Tomas' data about villages is wrong, so his analysis and discussions are pointless. I can't speak to that.)
Dear Robert, I partially agree with what you say, because I don't deny that there are unique emerging patterns of urban relations throughout its history. But the deterministic tradition of physics and chemistry still doesn't allow emerging sciences of the 20th century, such as biology, ecology, and economics (and urban sciences are no exception), to change the approach to their study systems in order to understand certain interactions. Despite this, it was a biologist, Ludwig von Bertalanfy, who proposed General Systems Theory. And that's what I want to focus on: reducing reality (systems), even when studying urban systems, is necessary, but for that today we have more adequate statistical, mathematical, and geographical approaches. And even then, the delimitation of its variables is the most difficult task because defining the object of study in complex systems is not a given (I can assure you from experience). In order to be clearer about what I want to say, I propose this paragraph by Marc Barthelemy for better understanding:
"The beginning of statistical physics can be traced back to thermodynamics in the nineteenth century. The field is still very active today, with modern problems occurring in out-of-equilibrium systems. The first problems (up to c. 1850) were to describe the exchange of heat through work and to define concepts such as temperature and entropy. A little later many studies were devoted to understanding the link between a microscopic description of a system (in terms of atoms and molecules) and a macroscopic observation (e.g., the pressure or the volume of a system). The concepts of energy and entropy could then be made more precise, leading to an important formalization of the dynamics of systems and their equilibrium properties. More recently, during the twentieth century, statistical physicists invested much time in understanding phase transitions. The typical example is a liquid that undergoes a liquid-to-solid transition when the temperature is lowered. This very common phenomenon turned out, however, to be quite complex to understand and to describe theoretically. Indeed, this type of “emergent behavior” is not easily predictable from the properties of the elementary constituents and as Anderson (1972) put it: ”… the whole becomes not only more than but very different from the sum of its parts.” In these studies, physicists understood that interactions play a critical role: without interactions there is usually no emergent behavior, since the new properties that appear at large scales result from the interactions between constituents. Even if the interaction is “simple,” the emergent behavior might be hard to predict or describe. In addition, the emergent behavior depends, not on all the details describing the system, but rather on a small number of parameters that are actually relevant at large scales (see for example Goldenfeld 1992). Statistical physics thus primarily deals with the link between microscopic rules and macroscopic emergent behavior and many techniques and concepts have been developed in order to understand this translation – among them the notion of relevant parameters, but also the idea that at each level of description of a system there is a specifically adapted set of tools and concepts."
Thanks Leonardo. That's an interesting paragraph. I've sort of lost track of where this discussion is going, but in response to: "deterministic tradition of physics and chemistry still doesn't allow emerging sciences of the 20th century, such as biology, ecology, and economics ... to change the approach to their study systems in order to understand certain interactions." I recommend you check out Jim Sethna's Stat Mech book and his class notes. https://sethna.lassp.cornell.edu/Teaching/562/
Tomas, I appreciate your clearly reasoned approach to complex subjects. I was intrigued with your dialogue with Leonardo in the comments about theories of complexity. To me it illuminated how what you’re trying to do here differs from the practices of academia. It seems to me that in these articles you take a fresh approach to complex human phenomena by asking basic practical questions which any intelligent, college-level educated person might engage in, and you show how asking the right questions can lead to greater clarity about our amazing multilayered world. You don’t make outsized claims, you don’t labor under ideological agendas, and you’re open to new information and evaluating alternative analyses. Academic institutions, on the other hand, are often invested in hierarchies of status and performance which do little to support the advancement of insight.
Fascinating subject that I wish more people new existed. Thanks for the great breakdown here, and looking forward to reading more of the rest of the series!
I highly recommend The Dawn of Everything by the anthropologist David Graeber and the archaeologist David Wengrow. It is a total revision of what we thought we knew about how cities and communities emerged.
I read it! I disagree a lot with it! I think he had an agenda and overfit it to the evidence quite dramatically, even though the evidence itself is fascinating!
One extra point. The density of population is directly tied to land productivity (which is not the same as labor productivity).
Unlike most other crops, irrigated rice has an amazing ability to stay productive on smaller and smaller plots of land. This is because the nutrients are in the water, not in the soil. Productivity scales with human labor in rice production, so it is possible to have a tiny plot of land growing enough rice to support a farm household, as long as they live close enough to their field to tend it regularly. So the result is that villages in rice irrigated regions can be much bigger than other agricultural villages.
As far as I know other crops are more tied to land productivity and added more human labor does little to improve yields (beyond a certain point).
I cannot prove it, but I think this accounts for East/Southeast/South Asian settlement patterns of very dense populations in rural area combined with low urbanization rates. And also the enormous populations of India and China. This was very different from Europe and other regions.
Yes, cities tend to grow on crossroads or waterways because their main purpose is long-distance transportation (I.e. between cities and the local villages). Villages must be near farms as the farmer must grow his own food. So most cannot be on rivers or crossroads (at least not large ones).
You are correct that not all farmers cluster in villages. Some live in isolated farmhouses or hamlets (smaller than a village). My guess is that crop type, geography and culture determine how dense a settlement they prefer.
Dear Tomas. I don't mean to be disrespectful, but the claim, based on your status as a physicist, that the urban growth model is deterministic "as in physics" (which isn't true for all of physics either, and I can say this with complete confidence, even though I'm a biologist, not a physicist, and have been working in territorial planning for 16 years), and that you can make such an outrageous claim based on internet data alone, turns this medium into just another digital outlet promoting misinformation. Saying this, when there are very serious studies based on archaeological and demographic evidence, which appeal to "complexity" precisely because deterministic science is insufficient to understand urban processes of the past, or those of the present, is disingenuous. This isn't the first time you've written on a variety of topics, including many "urban" ones, with simplistic assertions. I recommend reading "Complexity" by Lewin, or "Complex Environmental Systems" by Rolando García. There are also numerous online research papers in the journals Urban Planning, Landscape Planning, and npj Urban Sustainability, among others. This would allow you to be a little more humble and, even with the same data, introduce probability and compare your 30 or perhaps slightly more city cases. This would show you a minimal margin of error, which you ignore, but which justifies an analysis based more on function than on parts. This, like any error, also confirms the rule, but expresses the probabilities that lead to understanding why urban processes are so varied throughout history. And even more so today. As in gas physics or thermodynamics, for example.
Thanks! I think you might have taken “deterministic” in a way that is too literal. I didn’t mean it as in math, more as in “there are some very strong rules that really push some spots in certain directions.”
To illustrate: if we went back into history 10k years ago and let history unfold 1 million times to see what cities appear where, and then take the averages of these simulations, we would have a distribution of potential city placements and their size. My guess is that distribution would seem eerily similar than the actual cities we see today. Paris and London would definitely be in the same place as they are today, and would be also the biggest cities in their area. Moscow might not end up exactly there, but it would end up close by. Etc
Dear Tomás. Forgive my vehemence, but this is a topic I'm passionate about, and a complicated one at that. I'd love to provide more information, and I'll try to do so to enrich the debate. But for now, I'll leave aside a fundamental popular text on complexity, which begins with a case study of a pre-Hispanic American city by Roger Lewin:
https://www.planetadelibros.com.ar/libro-complejidad/131291
This article is very good and recent:
https://revistainvi.uchile.cl/index.php/INVI/article/view/63471/67082
And this book is essential. Although I only read part of it, I think it provides a good summary of the complexity approach in urban sciences:
https://link.springer.com/article/10.1057/s41289-018-0072-1?fromPaywallRec=true
Thanks Leonardo!
In general I disavow arguments by reference: If you have an argument to make, make it, don't refer to something else. Especially true for books: I can't possibly read every book I'm recommended.
I did just skim the 2nd link (article) and I'm underwhelmed:
• It's purely theoretical. It doesn't make any specific claim.
• It doesn't generate concrete, testable predictions.
• It's full of jargon, frequently used to obfuscate lack of insight.
• It has no empirical evidence.
• None of its claims are against mine! It claims cities are unpredictable in detail (I agree!) not as a whole.
To this last point, the article uses a metaphor with thermodynamics (which it doesn't support well, and uses one of its concepts, entropy, which I've studied in grad school, and I don't think it means what they think it means: Cities are a clear example of local reduction in entropy, by extracting energy from its surroundings).
We can use another thermodynamics metaphor to illustrate the difference between my claims and those of the article: You can't predict the movements of a single metaphor, much less the state of its atoms. They are complex systems. But as a whole, when you take billions of molecules, you can absolutely tell how they will behave. You can know how their temperature and pressure will evolve in different situations, with fairly simple rules that are quite deterministic.
This is what I mean in my article: There are absolutely rules that can define where cities will emerge, how big they will be as a whole, and the types of structures that are more likely to be optimal in them. This doesn't mean you can determine which building will go up where, or things like that, which is what your 2nd link is focused on.
In any case, I picked a random paragraph in your article:
"La inercia de las estructuras relacionales –característica intrínseca de la identidad– conlleva la existencia de patrones que perviven en el tiempo y que se asocian al fenómeno de las persistencias. Las estructuras subyacen y permanecen y permiten identificar esta capacidad de reorganización propia de la ciudad compleja, a menos que realmente se opte por la tabula rasa (Ruiz Sánchez, 2001, p. 32-35). Es en las transformaciones directas sobre la estructura relacional en las que la identidad y unicidad de la ciudad se puede ver realmente afectada, provocando cambios que alteran su identidad."
Can you actually extract anything valuable from this type of paragraph? This sounds like BS to me.
Tomás. I understand this point:
"This is what I mean in my article: There are absolute rules that can define where cities will emerge, how big they will be as a whole, and the types of structures that are most likely to be optimal in them. This doesn't mean you can determine which building will go up where, or things like that, which is what your second link is focused on." And you're right. I'll try, take some time, and gather more specific and appropriate bibliography, but I insist that the history and evolution of cities cannot be addressed in a deterministic way, because by definition in a complex system, the object of study is not given, and is constructed by the nature of its variables, which are also multifunctional. I understand that you'll say it's a theorization (I'm also a professor of science methodology), but I will strive to show with data why a city cannot be addressed with deterministic models.
I looks forward to it! Thanks!
LOL
“ Can you actually extract anything valuable from this type of paragraph? This sounds like BS to me.”
😂
Your 3rd article is inaccessible btw
Who says the author is making “an outrageous claim based on internet data alone?”
If memory serves correct, I think that “city patterns are based on physics” is exactly the same argument that Geoffrey West made in his book “Scale.” Indeed, I believe that West is a leading proponent of Complexity theory, so I am not sure why you consider Complexity to be incompatible with physics.
https://techratchet.com/2021/01/15/book-summary-scale-by-geoffrey-west/
I would not be so quick to dismiss an alternate opinion as “disinformation.”
To use your own words, this would allow you to be a little more humble…
Thanks for the link! I skimmed it. Hard to follow but I get the gist, and it makes sense to me!
Dear Michael, I'm going to get and read the book you suggested, for the simple reason that it argues what I maintain. It's not deterministic physics that explains the phenomenon, but rather its complexity. And the author makes this clear when he refers to "scale," at which various "rules" are produced that explain some structures that emerge from the function beyond the type of system being studied. Scale is "the exception" that confirms the rule, but precisely the one that establishes why its trajectories are not deterministic, because what at certain scales are processes are structures at others. For example: a wetland in a city at a certain scale is the structure of the urban watershed, but at the urban scale it is a floodable area due to the organization of its drainage network. For planners, this is a problem because they must resolve its "multifunctional dimension" and predict its evolution simply by being affected by what, in terms of complexity, defines error: exogenous variables. We can continue talking about problems in "nondeterministic" physics, such as the three-body problem and the probability of "determining" the position of a body when it is affected by two others. No urban process can be reduced to two variables!
Glad that you are going to read the book.
My main concern was the “vengeance” that permeated your comment. You could have phrased things in much better way. It was, after all, a very good article. I know something about villages and cities in history, and Pueyo was largely correct.
As for the book, the author believes that cities, biological organisms, and all other complex system can be explained by “ Quarter-power scaling laws” where positive consequences outweighs negative consequences as a thing grows in size.
It is actually a very deterministic argument.
Tomas, it might be interesting to study how these patterns apply to Amish farmers in the US. Their 30 minute radius is determined by the speed of a horse-drawn buggy and they have certainly been require to expand to new villages as their population has grown. From beginnings in eastern Pennsylvania they have expanded to new communities in Ohio, Maryland, Indiana, West Virginia and further west.
Interesting! Thanks for putting this on my radar!
Leonardo, you say you “don’t mean to be disrespectful,” but then call the article “misinformation,” “outrageous,” “disingenuous,” and “simplistic.” That feels more dismissive than constructive.
I agree complexity matters, and of course no single model explains everything. “Deterministic” is a strong word and does need nuance, but I didn’t read the article as trying to replace archaeology or demography..It is pointing out that clear regularities show up in urban data and making those patterns visible for a wider audience. That’s not misinformation, it’s sparking curiosity and debate.
Calling it “misinformation” just shuts down a useful conversation. The references you shared look great, and adding probability and more cases is a solid next step. But if we’re talking humility, maybe that reminder cuts both ways. We’ll get further by building on these ideas instead of writing them off.
You're right, Patricia. I didn't mean to be offensive, but unfortunately, those of us who deal with territorial planning face simplistic analyses on a daily basis (I swear I don't mean to be offensive) that reduce urban rivers to piped drains, draining surfaces to dry plazas, and turn any land surface into a business opportunity, even when there's evidence that the soil may be unstable or unhealthy. I can assure you that the history of a city's development doesn't obey deterministic laws, and Tomás exposes some of this in the explanation, but not in the data (surely due to the nature of his article). Of course, there are patterns that define trajectories, especially throughout history, but urban science is moving in a different direction and is laboriously working on approaches diametrically opposed to what Tomás proposes. I insist, a quick review of the scientific literature on the subject over the last 30 years reveals a different approach based more on function than on parts.
I’m curious what you’ll think about the upcoming articles!
I’m also curious what specific claims you disagree with in this article, rather than its overarching goal.
Yes, I would be as well, because I think (with a few minor exceptions) you got it exactly right.
I think the commenters reaction was more about keeping people from outside self-identified urban design experts quiet. As a professor, I saw this attitude all the time when outsiders suggested a new approach or theory. But the best ideas often come from outsiders.
Yeah, and after reading the one readable link from the bunch he sent, I remained unimpressed.
Thanks for explaining, Leonardo. I get where you’re coming from, and I see how frustrating it can be when complexity gets reduced to oversimplified plans. What I took from Tomás piece, though, was less about dismissing that complexity and more about showing the broad regularities that stand out.
And I think there’s room for both, the detailed work you mention and the broader patterns that make these ideas accessible to people like me. That is exactly what I enjoy about this newsletter, different angles coming together and making the conversation richer (at least that's how I learn!). Appreciate you taking the time to add to your initial comment.
The author of the book that I linked gives a very persuasive argument that cities do, in fact, obey deterministic laws: Quarter-power scaling.
If “ urban science is moving in a different direction and is laboriously working on approaches diametrically opposed to what Tomás proposes” it might well be driving itself into a dead end. This would not be the first time that urban design zealots have tried to reinvent cities only to make them worse.
Understanding why cities grow organically according to certain identifiable patterns and then working within them seems to be a much better approach.
Leonardo, my reaction to this article is quite different from yours. That any regularity at all arises in village placement and structure is wondrous. Rather than denying complexity, to me it brings to mind Phil Andersen's epithet "More is different". I gather, as an urban planner you see important differences between instances of urban growth, which is another aspect of complexity. To me that doesn't diminish the possibility that human-built things like villages arise with certain patterns because all humans share certain features. In a similar way I'm reminded that while all humans are unique, and human development is an extremely complicated process, yet most humans have 2 legs, 2 arms and 2 eyes. That doesn't in any way diminish the uniqueness of each human, but there is something in the complexity that produces some regularity and similarity between most humans. "More is different" hit the vernacular just as the study of complexity became mainstream in physics. To me this article explores that kind of complexity. It is a kind of determinism in the sense that "beings with these attributes in environments with these other attributes will develop villages with these kinds of observable features". Isn't that what all of us mean by complexity? (Or maybe you are saying that Tomas' data about villages is wrong, so his analysis and discussions are pointless. I can't speak to that.)
Dear Robert, I partially agree with what you say, because I don't deny that there are unique emerging patterns of urban relations throughout its history. But the deterministic tradition of physics and chemistry still doesn't allow emerging sciences of the 20th century, such as biology, ecology, and economics (and urban sciences are no exception), to change the approach to their study systems in order to understand certain interactions. Despite this, it was a biologist, Ludwig von Bertalanfy, who proposed General Systems Theory. And that's what I want to focus on: reducing reality (systems), even when studying urban systems, is necessary, but for that today we have more adequate statistical, mathematical, and geographical approaches. And even then, the delimitation of its variables is the most difficult task because defining the object of study in complex systems is not a given (I can assure you from experience). In order to be clearer about what I want to say, I propose this paragraph by Marc Barthelemy for better understanding:
"The beginning of statistical physics can be traced back to thermodynamics in the nineteenth century. The field is still very active today, with modern problems occurring in out-of-equilibrium systems. The first problems (up to c. 1850) were to describe the exchange of heat through work and to define concepts such as temperature and entropy. A little later many studies were devoted to understanding the link between a microscopic description of a system (in terms of atoms and molecules) and a macroscopic observation (e.g., the pressure or the volume of a system). The concepts of energy and entropy could then be made more precise, leading to an important formalization of the dynamics of systems and their equilibrium properties. More recently, during the twentieth century, statistical physicists invested much time in understanding phase transitions. The typical example is a liquid that undergoes a liquid-to-solid transition when the temperature is lowered. This very common phenomenon turned out, however, to be quite complex to understand and to describe theoretically. Indeed, this type of “emergent behavior” is not easily predictable from the properties of the elementary constituents and as Anderson (1972) put it: ”… the whole becomes not only more than but very different from the sum of its parts.” In these studies, physicists understood that interactions play a critical role: without interactions there is usually no emergent behavior, since the new properties that appear at large scales result from the interactions between constituents. Even if the interaction is “simple,” the emergent behavior might be hard to predict or describe. In addition, the emergent behavior depends, not on all the details describing the system, but rather on a small number of parameters that are actually relevant at large scales (see for example Goldenfeld 1992). Statistical physics thus primarily deals with the link between microscopic rules and macroscopic emergent behavior and many techniques and concepts have been developed in order to understand this translation – among them the notion of relevant parameters, but also the idea that at each level of description of a system there is a specifically adapted set of tools and concepts."
Thanks Leonardo. That's an interesting paragraph. I've sort of lost track of where this discussion is going, but in response to: "deterministic tradition of physics and chemistry still doesn't allow emerging sciences of the 20th century, such as biology, ecology, and economics ... to change the approach to their study systems in order to understand certain interactions." I recommend you check out Jim Sethna's Stat Mech book and his class notes. https://sethna.lassp.cornell.edu/Teaching/562/
Tomas, I appreciate your clearly reasoned approach to complex subjects. I was intrigued with your dialogue with Leonardo in the comments about theories of complexity. To me it illuminated how what you’re trying to do here differs from the practices of academia. It seems to me that in these articles you take a fresh approach to complex human phenomena by asking basic practical questions which any intelligent, college-level educated person might engage in, and you show how asking the right questions can lead to greater clarity about our amazing multilayered world. You don’t make outsized claims, you don’t labor under ideological agendas, and you’re open to new information and evaluating alternative analyses. Academic institutions, on the other hand, are often invested in hierarchies of status and performance which do little to support the advancement of insight.
Highly recommend checking out Luis Bettencourt's work here (textbook on the "science of cities": https://mitpress.mit.edu/9780262046008/introduction-to-urban-science/
Marc Barthelemy (https://www.quanturb.com/), Michael Batty (https://www.amazon.com/New-Science-Cities-MIT-Press/dp/0262534568) and Geoffrey West (https://www.amazon.com/Scale-Universal-Growth-Organisms-Companies/dp/014311090X) all have a bunch of great research in this area, too!
Fascinating subject that I wish more people new existed. Thanks for the great breakdown here, and looking forward to reading more of the rest of the series!
What are the main insights you draw from these? I don't think I'll have time to read them actually
Thanks! Added to my reading list!!
I highly recommend The Dawn of Everything by the anthropologist David Graeber and the archaeologist David Wengrow. It is a total revision of what we thought we knew about how cities and communities emerged.
I read it! I disagree a lot with it! I think he had an agenda and overfit it to the evidence quite dramatically, even though the evidence itself is fascinating!
Are the first three photos AI generated? They look like they're from a video game.
They are! WDYT?
Great article.
One extra point. The density of population is directly tied to land productivity (which is not the same as labor productivity).
Unlike most other crops, irrigated rice has an amazing ability to stay productive on smaller and smaller plots of land. This is because the nutrients are in the water, not in the soil. Productivity scales with human labor in rice production, so it is possible to have a tiny plot of land growing enough rice to support a farm household, as long as they live close enough to their field to tend it regularly. So the result is that villages in rice irrigated regions can be much bigger than other agricultural villages.
As far as I know other crops are more tied to land productivity and added more human labor does little to improve yields (beyond a certain point).
I cannot prove it, but I think this accounts for East/Southeast/South Asian settlement patterns of very dense populations in rural area combined with low urbanization rates. And also the enormous populations of India and China. This was very different from Europe and other regions.
Very exciting!
I think Smil has some of the numbers you may be looking for in Energy and Civilization.
I am curious how this fits in to arguments you’ve listed in Why Some Cities Thrive particularly about being situated of Fall Lines.
Also, Nick Szabo has a cool post on the trade-off between security and productivity of farms.
https://unenumerated.blogspot.com/2005/12/security-and-productivity-of-farms.html?m=1
He argues that unified Japan and England got productivity dividends since their security costs were lower than non-island settings.
I’ve been obsessed with Szabo’s work for years. One day I’ll get to the roaming bandits theory! Fascinating.
Smil is very productive. Hard to catch up. I need to try!
I thought villages and later-on cities grow in crossroads or waterways where they did commerce. Farmers don't need to live in villages or cities.
Yes, cities tend to grow on crossroads or waterways because their main purpose is long-distance transportation (I.e. between cities and the local villages). Villages must be near farms as the farmer must grow his own food. So most cannot be on rivers or crossroads (at least not large ones).
You are correct that not all farmers cluster in villages. Some live in isolated farmhouses or hamlets (smaller than a village). My guess is that crop type, geography and culture determine how dense a settlement they prefer.
Yes; that’s in the next article!