What makes a city?

[The Simpsons Cinematic Universe]

An Australian friend and I were recently discussing administrative divisions. Unbeknownst to myself, the Australian federal government defines a "city" simply as an municipality with more than 100,000 residents, and a "town" as a municipality which does not meet that threshold. While the United States has no federal regulations regarding cities, this post is meant to highlight how strange some of the state regulations are.

For my home state of New York, the list of the largest cities is particularly misleading. The overwhelming majority of cities here are in upstate New York, the lesser densely populated part of the state. Before you continue, read the first 20 cities of that aforementioned list, from New York to Ithaca.

Now, look at the following chart which I've compiled, which shows what the 20 most populous cities in New York would be, if any municipality with a population greater than, say, 80,000 people, was considered a city.



Only two cities from upstate manage the crack the top ten. Only seven even make the top twenty. Surprisingly even to myself, Buffalo, the largest city in upstate New York, was displaced to sixth place, from second. Hempstead (word to Tweed), a town with almost a million residents, would be the second largest city in New York. Suffolk County, which lacks any cities whatsoever as is, would suddenly have the third and fourth largest cities in the state.

On top of that, if Hempstead was a city, it would be the 18th largest in the United States, below Charlotte, but above Detroit, Boston, Denver, Washington DC, Portland, Las Vegas and Sacramento.

It's an interesting scenario to consider.

[Hatman 🍁]

Ontario is strange too. Dryden is a city with 8000 people while Oakville is a Town with 200,000 people.

[muon2]

States typically recognize incorporated municipalities as something different from a county subdivision with electoral powers. State law will reserve certain functions to incorporated cities and villages while county subdivisions such as towns, townships and unincorporated villages must look to the county or state for those functions.

[Franknburger]

Before going into detailed discussion (which I am very interested in), I think it is useful to look at the international standards, namely the Nomenclature of Units for Territorial Statistics (NUTS), which is structured as followed:

NUTS 0 (not defined officially): Nation. The only 'real' city nation I can think of is Singapore.

NUTS 1: State (US/ Germany), Province (Canada), or equivalent (e.g. planning regions in France, Spain, Italy and Poland; Scotland, Wales, Northern Ireland and English planning regions in the UK). Ideally, NUTS 1 regions should have between 3 and 7 million inhabitants, though several US and German states fall outside these borders. NUTS 1 cities include Washington DC, Ottawa, Greater London, Berlin, Madrid, Hamburg, and the Brussels Capital Region. Borderline cases are Isle de France (Paris Metro), and Gibraltar and Monaco.

NUTS 2:  States within smaller federal nations such as Austria, and the largest administrative subdivision in non-federal countries, i.e. regions (F, I, DK), autonomous communities (E), provinces (NL, BE), voivodships (PL), and - as an anomaly, they should be NUTS 3 - UK counties. Not defined for the US. Ideally, NUTS 2 regions should have between 0.8 and 3 million inhabitants. NUTS 2 cities include several European capitals (Vienna, Prague) or their metros (Stockholm, Bukarest). Among the more exotic cases are Ceuta, Andorra, and San Marino.

NUTS 3: County or equivalent (French departments, Italian/ Spanish Provinces, etc.), ideally with a population between 150,000 and 800,000. NUTS 3 cities are quite numerous in Germany, the UK and Poland. They should also be widespread across the US, even though most of the larger US cities would be better placed in the NUTS 2 category (which does not exist in the US).

NUTS 4 (abolished in 2003): Largest county subdivision, typically an unitary community, or a group of smaller communities, but also subdivisions of larger (NUTS 1-3) cities. Ideally 20,000-150,000 inhabitants. In the US case, many rural counties would rather qualify as NUTS 4 than NUTS 3. Essentially, all smaller cities / towns are NUTS 4.

NUTS 5 (abolished in 2003): : Smallest territorial unit, such as the US Census Block, the UK Electoral Ward, individual German communities that form part of an unitary community or a community association (Samtgemeinde, Amt), etc.

Note that territories can fall within multiple NUTS groups. Estonia, e.g., is NUTS 0-2, Isle-de-France is NUTS 1&2, Berlin / Ottawa/ Washington D.C. are NUTS 1-3, etc.

[pbrower2a]

An Australian friend and I were recently discussing administrative divisions. Unbeknownst to myself, the Australian federal government defines a "city" simply as an municipality with more than 100,000 residents, and a "town" as a municipality which does not meet that threshold. While the United States has no federal regulations regarding cities, this post is meant to highlight how strange some of the state regulations are.

For my home state of New York, the list of the largest cities is particularly misleading. The overwhelming majority of cities here are in upstate New York, the lesser densely populated part of the state. Before you continue, read the first 20 cities of that aforementioned list, from New York to Ithaca.

Now, look at the following chart which I've compiled, which shows what the 20 most populous cities in New York would be, if any municipality with a population greater than, say, 80,000 people, was considered a city.



Only two cities from upstate manage the crack the top ten. Only seven even make the top twenty. Surprisingly even to myself, Buffalo, the largest city in upstate New York, was displaced to sixth place, from second. Hempstead (word to Tweed), a town with almost a million residents, would be the second largest city in New York. Suffolk County, which lacks any cities whatsoever as is, would suddenly have the third and fourth largest cities in the state.

On top of that, if Hempstead was a city, it would be the 18th largest in the United States, below Charlotte, but above Detroit, Boston, Denver, Washington DC, Portland, Las Vegas and Sacramento.

It's an interesting scenario to consider.

Let's start with the distinction between a city and a suburb:

1. If you live in Hempstead, Brookhaven, Ramapo, Islip, or Yonkers  and you travel outside of the area and you are asked where you are from you will use New York City as a reference to where your 'community' is.  One has to be incredibly isolated or ignorant to not know where New York City is. 

2. If your 'city' doesn't pass spell-check, it is probably not a real city but instead a suburb. Hempstead, Brookhaven, Ramapo, and Islip failed that test. (Really, so fails Paramus, New Jersey). Of course, some place that is the name of something else (Garland, Mesquite, Grand Prairie, and White Settlement around Dallas/Fort Worth), common surnames (Richardson, Addison, Denton, Irving) would pass that test, as would places that exist elsewhere (Arlington,  Lancaster, and Saginaw) or have some commercial identity (Frisco -- OK as a reference to a railroad, but use it for San Francisco and you demonstrate gross ignorance). Bedford, Carrolton, Seagoville, Lewisville, and Euless fail that test. But for any of those places you would refer to the Dallas-Fort Worth Metroplex. 

3. Although there are cities with very low profiles (Paris due to catacombs which deny a good foundation, Washington DC because of a law that prohibits buildings taller than the Capitol), cities are where the skyscrapers are. In that sense a place like Battle Creek, Michigan is more a city than is Livonia, Michigan -- even if Livonia is larger.  One easily knows when one is approaching a city and one can rarely distinguish one suburb from another.

4. True cities have their areas of cultural congregation. They are the places after which airports are named.  Radio and TV stations identify themselves by the name of a city and not by the name of a suburb. As a contrast -- Pasadena, California is well-known, but TV stations in the Greater Los Angeles Area identify themselves as being from Los Angeles.

       

[Franknburger]

3. Although there are cities with very low profiles (Paris due to catacombs which deny a good foundation, Washington DC because of a law that prohibits buildings taller than the Capitol), cities are where the skyscrapers are. In that sense a place like Battle Creek, Michigan is more a city than is Livonia, Michigan -- even if Livonia is larger.  One easily knows when one is approaching a city and one can rarely distinguish one suburb from another.

Ooops - that may be true for the US, and some Asian cities (Tokyo, Shanghai, Jakarta), but not for most of Europe and the Arab World. Here, it is more of "You are leaving the city when you get to where the skyscrapers are".  It is not just catacombs, but the fact that the inner cities are full of historic buildings that has prevented Paris, Rome, Lisbon, Brussels, Vienna, Prague or Budapest (to just name a few examples) from building skyscrapers downtown. Laws prohibiting buildings taller than a historical landmark (usually a church) are quite common across Germany (e.g. Berlin, Hamburg, Cologne), and mandatory for any city applying for UNESCO World Heritage status (Lübeck, Dresden, Venice, Florence, Marrakesh etc.).

"Hochhausghetto" (Skyscraper-Ghetto) is a common term in German (896.000 Google search results). If the term isn't self explaining, check out the French equivalent - Banlieue

[Torie]

What makes a "city" to me has nothing to do with subdivisions. The best way to know, is to look at where the lights are, and are not, from aerial photographs from space, or an airplane high up.

[Franknburger]

What makes a "city" to me has nothing to do with subdivisions. The best way to know, is to look at where the lights are, and are not, from aerial photographs from space, or an airplane high up.

Agreed - sort of..  The lights won't tell us whether it is one or several cities. Check out the Bay Area on your aerial photograph, or Baltimore/ D.C.  Not to mention the Twin Cities..

[Filuwaúrdjan]

It's tempting to say 'a large urban area', but that's not actually right, is it? Hereford (pop. c. 59,000) is most certainly a city, while Dudley (which, if we take a pretty wide - and pretty absurd - definition of 'Dudley', contains more than 300,000 souls) is quite clearly not. And wouldn't really be one even if (God forbid) it somehow flukes city status at some point.

[Hatman 🍁]

Before going into detailed discussion (which I am very interested in), I think it is useful to look at the international standards, namely the Nomenclature of Units for Territorial Statistics (NUTS), which is structured as followed:

NUTS 0 (not defined officially): Nation. The only 'real' city nation I can think of is Singapore.

NUTS 1: State (US/ Germany), Province (Canada), or equivalent (e.g. planning regions in France, Spain, Italy and Poland; Scotland, Wales, Northern Ireland and English planning regions in the UK). Ideally, NUTS 1 regions should have between 3 and 7 million inhabitants, though several US and German states fall outside these borders. NUTS 1 cities include Washington DC, Ottawa, Greater London, Berlin, Madrid, Hamburg, and the Brussels Capital Region. Borderline cases are Isle de France (Paris Metro), and Gibraltar and Monaco.

NUTS 2:  States within smaller federal nations such as Austria, and the largest administrative subdivision in non-federal countries, i.e. regions (F, I, DK), autonomous communities (E), provinces (NL, BE), voivodships (PL), and - as an anomaly, they should be NUTS 3 - UK counties. Not defined for the US. Ideally, NUTS 2 regions should have between 0.8 and 3 million inhabitants. NUTS 2 cities include several European capitals (Vienna, Prague) or their metros (Stockholm, Bukarest). Among the more exotic cases are Ceuta, Andorra, and San Marino.

NUTS 3: County or equivalent (French departments, Italian/ Spanish Provinces, etc.), ideally with a population between 150,000 and 800,000. NUTS 3 cities are quite numerous in Germany, the UK and Poland. They should also be widespread across the US, even though most of the larger US cities would be better placed in the NUTS 2 category (which does not exist in the US).

NUTS 4 (abolished in 2003): Largest county subdivision, typically an unitary community, or a group of smaller communities, but also subdivisions of larger (NUTS 1-3) cities. Ideally 20,000-150,000 inhabitants. In the US case, many rural counties would rather qualify as NUTS 4 than NUTS 3. Essentially, all smaller cities / towns are NUTS 4.

NUTS 5 (abolished in 2003): : Smallest territorial unit, such as the US Census Block, the UK Electoral Ward, individual German communities that form part of an unitary community or a community association (Samtgemeinde, Amt), etc.

Note that territories can fall within multiple NUTS groups. Estonia, e.g., is NUTS 0-2, Isle-de-France is NUTS 1&2, Berlin / Ottawa/ Washington D.C. are NUTS 1-3, etc.

I don't think Ottawa counts as NUTS 1. Maybe NUTS 3. It's not its own province or territory in the way DC is.

[Torie]

What makes a "city" to me has nothing to do with subdivisions. The best way to know, is to look at where the lights are, and are not, from aerial photographs from space, or an airplane high up.

Agreed - sort of..  The lights won't tell us whether it is one or several cities. Check out the Bay Area on your aerial photograph, or Baltimore/ D.C.  Not to mention the Twin Cities..

As I said, legal boundaries have no real meaning to me, when it comes to what the polis means.

[Franknburger]

It's tempting to say 'a large urban area', but that's not actually right, is it? Hereford (pop. c. 59,000) is most certainly a city, while Dudley (which, if we take a pretty wide - and pretty absurd - definition of 'Dudley', contains more than 300,000 souls) is quite clearly not. And wouldn't really be one even if (God forbid) it somehow flukes city status at some point.

Check out Dudley for yourself on this aerial photograph . I tend to say it actually shows up as a separate light point, which - if I understand Torie's approach correctly - would make it a city (though it is clearly a borderline case).
The whole "urban area" or "agglomeration" point does not bring us further here - the Ruhr/ Rhine, the Midlands and the Dutch Randstad can all be identified as polycentric urban areas/ agglomerations, made up of several distinct city cores - just as you can tell from air that NYC and Philadelphia, while growing together, are still two separate cities.

The cases that are really complicated are San Francisco / Oakland / San Jose, Athens/ Piraeus, even Berlin / Potsdam (aerial photograph on the latter two). Shall we really stop calling Piraeus a polis, just because it has from air become indistinguishable from Athens?

[Filuwaúrdjan]

Check out Dudley for yourself

I'd rather not, if that's alright with you.

[Filuwaúrdjan]

While we're on the subject of Dooodloooiiiiyyyy....



Via b3ta.

[muon2]

It's tempting to say 'a large urban area', but that's not actually right, is it? Hereford (pop. c. 59,000) is most certainly a city, while Dudley (which, if we take a pretty wide - and pretty absurd - definition of 'Dudley', contains more than 300,000 souls) is quite clearly not. And wouldn't really be one even if (God forbid) it somehow flukes city status at some point.

Check out Dudley for yourself on this aerial photograph . I tend to say it actually shows up as a separate light point, which - if I understand Torie's approach correctly - would make it a city (though it is clearly a borderline case).
The whole "urban area" or "agglomeration" point does not bring us further here - the Ruhr/ Rhine, the Midlands and the Dutch Randstad can all be identified as polycentric urban areas/ agglomerations, made up of several distinct city cores - just as you can tell from air that NYC and Philadelphia, while growing together, are still two separate cities.

The cases that are really complicated are San Francisco / Oakland / San Jose, Athens/ Piraeus, even Berlin / Potsdam (aerial photograph on the latter two). Shall we really stop calling Piraeus a polis, just because it has from air become indistinguishable from Athens?

I'd worry about leaning too much on light generation as a measure of cityness. Otherwise Williston, ND becomes a major US city.

[Franknburger]

While we're on the subject of Dooodloooiiiiyyyy....



Via b3ta.

Well, if they have a street market, they should be a city, shouldn't they?

In any case, the "cultural congregation" point raised by pbrower deserves further consideration. It's obviously not working too well on sports, particular soccer, as you may find more than one major team within a city/aggregation (I also see few reasons aside from football for not declaring Leverkusen a Cologne suburb). Radio/ TV stations may help in the US, but not in Germany, where their licensing is state (NUTS 1) responsibility, and stations also tend to refer to that state.
But there are probably other measurements that haven't yet been considered. Newspapers might actually work, at least as concerns the centre/ suburb question, and as long as we are in industrialised countries.

[patrick1]

As someone who currently lives in one of the villages in the Town of Hempstead, it is in no way a city.  The village of Hempstead has what resembles a small suburban city. I personally dislike the town organization of Nassau and Suffolk and would like them to do away with it.  There are just way too many overlapping layers of government.

[LastVoter]

Industry.

[RedSLC]

It's tempting to say 'a large urban area', but that's not actually right, is it? Hereford (pop. c. 59,000) is most certainly a city, while Dudley (which, if we take a pretty wide - and pretty absurd - definition of 'Dudley', contains more than 300,000 souls) is quite clearly not. And wouldn't really be one even if (God forbid) it somehow flukes city status at some point.

Check out Dudley for yourself on this aerial photograph . I tend to say it actually shows up as a separate light point, which - if I understand Torie's approach correctly - would make it a city (though it is clearly a borderline case).
The whole "urban area" or "agglomeration" point does not bring us further here - the Ruhr/ Rhine, the Midlands and the Dutch Randstad can all be identified as polycentric urban areas/ agglomerations, made up of several distinct city cores - just as you can tell from air that NYC and Philadelphia, while growing together, are still two separate cities.

The cases that are really complicated are San Francisco / Oakland / San Jose, Athens/ Piraeus, even Berlin / Potsdam (aerial photograph on the latter two). Shall we really stop calling Piraeus a polis, just because it has from air become indistinguishable from Athens?

I'd worry about leaning too much on light generation as a measure of cityness. Otherwise Williston, ND becomes a major US city.

Lol. The cluster of light around Williston is bigger than the tiny strip of light that I live in.

[RedSLC]

IMO, what distinguishes a city from a suburb or other similarly-sized unit is whether it has well-developed downtown area, where a lot of people from the surrounding areas work. i.e. it has to be more than just retail outlets, real estate offices, etc. that are usually found in the middle of suburbs.