The curious relationship between place names and population density

Political map with toponyms

Giving a name to a place is an important act. It says a place has meaning, that it should be remembered. For thousands of years, the way we kept track of place names—or toponyms—was by using our memory. Today, we’re not nearly so limited, and the number of toponyms seems to have exploded. Yet oddly enough, the number of places we name in a given area follows a trend uncannily similar to one seen in hunter-gatherer societies.

Eugene Hunn, now a professor emeritus of anthropology at the University of Washington, stumbled upon what appears to be a fundamental relationship between toponyms and population density when he published a paper on the subject in 1994. His discovery stemmed from a literature survey of twelve hunter-gatherer societies from around the globe. Hunn tabulated each society’s toponym repertoire and the size of their home territory to calculate the number of toponyms per square mile, or toponymic density. From this data, he distilled two trends.

First, the average number of toponyms converged on what he called the “magic number 500”. Hunn found that trend in a few other papers on topics like folk taxonomies of plants and animals, and he posited that the number was an inherent limitation of the human mind—that when relying on memory alone, individuals tend to retain names to 500 items per category. A hunter-gatherer, for example, may be able to name 500 different types of plants. Unfortunately, Hunn’s “magic number 500” wasn’t all that magical given the variability about it—individuals in the hunter-gatherer groups he studied actually recalled between 200 to 1000 toponyms. The concept doesn’t appear to have caught on in the academic world.

Hunn’s second finding, though, is more compelling. When he arranged the toponymic and population densities of the twelve hunter-gatherer groups on a graph, a clear relationship stood out. Where people lived closer together, the number of place names per square mile skyrocketed. Where they lived farther apart, they named fewer places per square mile. The figure, which I’ve reproduced below, appears to have a linear relationship. That’s an artifact of the logarithmic scale of the axes, which compresses the data as you move away from the origin. The scale is hiding a subtle curve, one that bends down as though the x-axis has roped the line and is pulling it closer.

Toponymic and Population Density of Twelve Hunter Gatherer Groups

The general trend in Hunn’s figure—that we name more places when living at higher densities—makes such good sense that I knew there had to be a modern corollary. Despite all our sophisticated maps and petabytes of computer storage, I suspected that we still hew to the same basic pattern as our hunter-gatherer forebears. So I dove into a simple yet relatively modern set of toponyms—the U.S. Postal Service’s ZIP code system.

First proposed in the 1940s, ZIP codes were meant to speed the processing of mail at sorting facilities. Most major cities at the time were already divided into postal zones, like “Milwaukee 4”, but small towns and rural areas had no such system. Mail volume swelled after World War II, so the postal service introduced the Zone Improvement Plan in 1963. From what I can tell, there don’t appear to be any hard and fast rules about the size of ZIP codes. Exactly how they are delineated seems to be a postal service secret and one that likely depends on their logistical needs. They can even overlap. But none of that really matters, because ZIP codes give names to places. They’re toponyms. I suspected that the more densely populated states had a higher density of ZIP codes, just like in hunter-gatherer societies. And sure enough, they do.

ZIP code and population density by state

The wrinkle lies in the trend line’s curve, which is masked by logarithmic axes the same way the curve in Hunn’s figure is hidden. The best way to read both graphs is backwards, from right to left, from high population density to low population density, paying special attention to the scale of the axes. Before we start, we should assume one thing, that people name places at the same rate per square mile regardless of population density. In other words, people will name seven things per square mile regardless of whether they live at ten or 100 people per square mile. Returning to the graphs, if we start at high population densities on the right and move left to lower population densities, the curve drops below our straight line assumption. Not only do people name fewer things at lower population densities, they name fewer things per square mile than our fixed rate assumption would have predicted. In other words, a hypothetical group living at ten people per square mile will name only four things per square mile, compared with the seven named if the population density were 100 people per square mile.

That’s key. There are plenty of gullies and hillocks of grass in the Great Plains, for example, but few people. As such, we name fewer things per square mile. It makes navigation easier—fewer waypoints to remember when traveling—and keeps us focused on the resources that matter. After all, population density is often driven by resource availability, whether that be food, water, shelter, or some other necessity. It’s as though our minds can’t cope with vastness, and so we name fewer things to compress the interstitial space.

The intriguing part is that ZIP codes and Hunn’s hunter-gatherer toponyms are described by one particular mathematical relationship (a power law, for the interested math-types). Not only that, they’re following the trend in a strikingly similar way.¹ As humans, we seem to have settled on a comfortable way of describing the world regardless of whether we remember it with neurons or silicon.

  1. Toponymic density = 0.3675(population density)0.8388
    ZIP code density = 0.0005(population density)0.6944


Hunn, E. (1994). Place-Names, Population Density, and the Magic Number 500 Current Anthropology, 35 (1) DOI: 10.1086/204245

Photo by Tim De Chant.

Related posts:

Density in the pre-Columbian United States: A look at Cahokia

Thinking about how we think about landscapes

Scientific American Guest Blog: Why we live in dangerous places

Thanks to the support of readers like you, Per Square Mile remains independent and ad-free.

If you enjoy what you just read, please consider supporting the site with a monthly recurring donation as a Sustaining Member.

You can also support Per Square Mile with a one-time donation.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>

  1. A curiosa relação entre nomes de lugares e densidade populacional. Postagem interessante, que traz pesquisa indicando que, quanto maior a densidade populacional, maior o número de nomes de lugares por quilômetro quadrado.

  2. Sounds to me like this is based on communication: one person or a family would give names to the local features they used — the place where they draw water, the place where bears are frequently seen — as a means of talking about them or giving directions. The more people, the more things get identified and need naming.

    Sounds similar to Dunbar’s Number [’s_number] which some scientists think might be one of the reasons for human cognitive development. Keeping track of 150 faces and relationships takes some processing power.

    Both of these seem related to human social characteristics and communication.

  3. Pienso que la relacion que sustenta mayor cantidad de nombres por unidad de superficie versus mayor densidad poblacional tiene que ver con el hecho que a mayor densidad poblacional mayor necesidad de precisar territorios en escalas menores (calles, cuadras, manzanas, minibarrios, piso x en un edificio, bloque x en un conjunto habitacional, etc.) mientras que a menor densidad, mayor territorio por unidad familiar o clan (tribus o comunidades campesinas o aborigenes que dominan amplias y extensas superficies que no requieren mayor precision identificatoria – la gran meseta, el valle bajo, etc.) Creo que tambien es respuesta al grado de comunicacion social como dice Paul.

    I think this curious relationship among place names and population density has to do with the fact that higher-density areas inhabitants need to identify smaller-scale territories with more detailed precision (such as lanes, driveways, plots, small neighbourhoods, floor x in a building, block x in a residential compund and alike), whereas in less dense territories there is a trend of larger tenements by few families or clans (i.e. tribes or aboriginal communities owning extense tracts of wilderness) that not necesarilly requiered precise identification other than just the “great plain” or the “lower valley” as enough toponimical naming. I also agree with Paul´s remark as an answer to certain social comunicational intensities.

  4. The more people per square mile, the more ZIP Codes per square mile… makes sense to me, as different ZIP Codes signify different sorting centers and post offices, and the higher the volume of mail, the more of those you need. Could it be any other way, given what ZIP Codes signify?

  5. I wonder about cause and effect here. The ZIP code system was deliberately arranged to make it easier for mail services, so it makes perfect sense that they would be arranged according to population. If a state only has a small population, there is less need to divide it into hundreds of different codes – it is perfectly reasonable to expect Wyoming, say, to have fewer separate codes than a city the size of Chicago. As such, the idea that there is some form of correlation between population density and ZIP code area size is simply explained since the system was designed to have exactly that type of relationship in order to ease modern mass communication.

    A far better test would be whether topographical names match the same sort of relationship, which is also a far harder test to get clear results for – there are names for a good many places in Antarctic for instance, possibly as many as there are for places within a reasonably-sized city, so any correlation is going to be much more shaky.

  6. James and Benjamin, it’s true that there should be fewer ZIP codes in lower density areas, the odd thing I found are that there are fewer than you would expect if the relationship were strictly linear. The fact that the pattern is the same as in hunter-gatherer societies makes the trend all the more intriguing.

  7. The density/toponym relationship can be demonstrated much more concretely with informal neighborhood names. I have noticed that the more densely populated the area, the more fine-grained the neighborhood names. In Manhattan neighborhood names might only describe a few blocks square. In the core of my town Seattle, every 10 blocks or so a new neighborhood name is used (3 block in downtown). In the outer fringes of the city, neighborhood names are not well known and degrade in generic terms like “Northwest Seattle” for at least 50 blocks. In the suburbs, it even worse. In suburban muncipalities that spreads across several square miles often the only well-known neighborhood concepts are “South xyz” or “North xyz.”

  8. I would think that the names would be a function of sociological rather than linear relationships. Related to both Paul and Chad’s comments, the denser the population, the tighter the community. A large population area would have multiple communities that people would relate to, thus requiring exponentially greater naming (sometimes different names for the same places depending on the community using the name). When I lived in Costa Rica, addresses in San Jose were based on directions using a community based landmark as a pivot (i.e. I live 100 meters south, 50 west of the Casa Italia, #24). In some cases, the landmark was no longer in existence, but those in the neighborhood had a collective memory (through the generations) of where the landmark used to be. Those outside of the neighborhood had difficulty finding addresses.

    In Budapest in the 1980’s, there were both Hungarian historical names to streets and locations and “Russian”. The use depended on who was asking for and who was giving directions. In my own town, there is a building still referred to as the D&H building which has not had that name for 40 years. However, those who grew up in the area, know when someone refers to that building which one it is. Look at old maps and names are community based. Bilingual or multi-cultural locations have multiple names, based on the communities. In rural or less densely populated areas, the community tends to be distributed over a larger area so there would be greater homogeneity in naming.