Fantastic Toponymy: Inventing Place Names

When we invent a place for a game, we need to name it. We can just mash together some syllables until they sound cool, or we can try to make a name that seems to have history, that sound natural and hints at the culture of the place by using real elements from real cultures. Or if we're really adventurous, we can invent sounds that have invented meaning, and use those to build our place names - like Tolkien famously did.

"Toponymy" is the naming of places, from the Greek "topos" place and "onoma", name. The places all around the Old World have names that have developed over centuries or millennia- names usually derived from a simple description of the place.

The New World, by contrast, is named far more recently, with names that are usually simple, current language terms that are easily understood, or native names perhaps spelled poorly by the settlers. Both follow similar conventions once you look into the language.

Reflecting culture

Where is this place you're naming? What are the people there like? Are this a place of familiar, homely folk, or do you want to invoke a feeling of "otherness" about your invented place?

The names of the British Isles are familiar sounding to we English speakers (even those outside the UK). They can seem homely, even rustic and Olde Worlde (especially to those outside the UK!).

Similarly, the names of other nations' places, made up as they are by different language elements, can seem foreign, but familiar - "Porto Nova" reflects that it's inhabitants speak some Mediterranean, Romance language, so we have a handle on the sort of culture we might find there, just from the name.

A place called "Klak'amtuu" or "Shissiissii" tells us that this is a culture we are not familiar with.

Be careful not to just borrow real world names directly for your fantasy locations, though. Those have cultural baggage that you probably don't want in your game.
Of course, if you deliberately want to bring in the implications of a real-world name to your game, that's up to you - just be sure you're aware of them!
Calling your pseudo-Britain "Albion" (as in the popular CRPG Fable) is fine if you include all the people of Great Britain - the Cornish, English, Scottish and Welsh), but not so good if you just treat them all as English. We have whole political parties here in the UK dedicated to independence for those nations.

Settlements and other places 

Since you're reading this in English, I'll start by looking at English place names.

Nottingham is a worn-down version of "Snotta-ingas-ham", meaning the settlement ("-ham") of the people ("-ingas") of Snotta (a Danish chieftan). That "-ingham" ending is all over England.
Canterbury is from "Cantware-burgh", meaning the Kentish ("Cantware-") Stronghold ("-burgh"). Again, that "-bury" ending, and the related "-burg", "-borough", and "-burgh" ending are all over the British Isles.

So armed with this knowledge, your invented place could be called Notesbury ("Snotta's Stronghold"), or Cantingham ("Settlement of the people of Kent") - some other combination from the name elements appropriate to the culture.

There are plenty of resources for such elements of British place names on the web, so I won't just reproduce a list here.

Similarly, you can find that sort of list for other cultures: Here's a list for Japanese place names, and a short list for Maghreb toponyms.

From these examples, you can see that place names tend to be derived from a few common parts:

• geographical feature (river, hill, valley, forest, etc)
• type of settlement (farm, fort, village, town, etc)
• direction or position (north, south, upper, middle, etc)
• person of note (founder or clan leader, saint, etc)
• description (dark, cold, red, windy, etc)

It can be helpful to make up a plain English name with such elements, and then translate it into whatever language is appropriate - for example: Hill Fort Town might be "Lawtonbury" in pseudo-Britian, or "Okajo Machi" in pseudo-Japan.

Country and regional names 

Nations and regions are often named for who lives (or lived) there (usually this is the name given to the place by its inhabitants), or a descriptor of the region (usually this is a name given by outsiders to the region).

Thus we get Scotland, the land of the Scots; France, named for the Franks; and Afghanistan the place of ("-istan") the Afghans, and so on.
While on the other hand we get Cameroon, the land of shrimp (Portuguese "Camaroes", via French "Cameroun"), from the abundance of shrimp found by European explorers in the Wouri River; and Wales, named by the Saxons for the foreigners ("Welisc") who lived there.

This can be useful - dual names for a place can help show historical enmity between the natives and the outsiders who imposed the other name.

Here's a list of country names and etymologies - and a list of regional names for places within countries that aren't on the former list.

You can see from those lists that national names - in the Old World - tend to be derived from a few common elements:

• people's or tribal name (Franks, Angles, Danes, Sicels, etc)
• description of people (foreigners, bearded ones, etc)
• geographical description (green, forest, mountainous, etc)
• place indicator (place, land, home, etc)

So again using your language of choice to reflect the culture you want to imply, you can name your region in a similar manner to settlements and places, but using these elements.

Geographical feature names

The great geographic features of the land were often named long long ago, in languages far removed from everyday. The new folk invading or colonising a place would just use the local name, not knowing or soon forgetting what it meant.
The River Avon in England means the "River River" - "avon" is derived from an ancient British word for river.

Alternately, the name may be clear and descriptive in the current common language - the Misty Mountains in Tolkien's Middle Earth, for example.

Features like mountains, rivers, lakes, marshes, and headlands will usually include some element that means just that - "mountain", "river", "lake", etc. Together with that element, there will usually be another part that simply describes the feature - "misty", "dark", "wide", or similar - or perhaps the name of a person or tribe.

Once again, we end up with the same sort of naming method, choosing from a this set of word elements.

Alien and bizarre places

To suggest a strange an exotic place, you can use language elements the players will not be familiar with, or invented languages like Klingon, or Quenya - or even your own invented gibberish sounds.

So in Middle Earth, we have Imladris, meaning "Cloven Valley" for Rivendell, and and Hithaeglir meaning literally "Towers of Mist" for Misty Mountains.

One could with relative ease invent a range of place name word element and use them consistently for a new culture. This could work particularly well in contrast with a region of familiar sounding names - for example, Crick Hollow, Buckland, Chetwood, and Weathertop giving way to Khazad-dum, Lothlorien, Gondor and Cirith Gorgul as we progress further away from the homely Shire with Frodo and Company.

Similarly, this technique works even with names that while relatively familiar, don't have that homely feeling created by names like your local region's names - so English name elements giving way to Germanic and Scandinavian names, or Latin names will create an atmosphere that evokes those lands, and all the assumptions that go with them.

Conclusion

Because place names can be evocative, and can be laden with culture, we should be careful not to just use them at random. Names and style choices in game settings will create an atmosphere. With only a little research and work, you can bring life into a place just by speaking its name.

Morality & Reality - meta-ethics in your game

What is good? What is evil? Are they real physical forces, purely mortal moral judgements, or what?

What started me thinking about this again - after I've already laid out my opinions and tool box for both gods and morality in games - was a rather heated discussion about a particular real-world religion and its morality or lack of it. I'm naming no names. The discussion got rather entrenched, but it made me think of a few things that I found helpful.

Now there are plenty of discussion threads out there on alignment, and some very serious heavy debate about religion. I'm not going to try to replicate those here - I want to look at how the different points of view can be used when creating your game world.

Like my article on gods, I think that one's attitude toward Good and Evil in the game setting can have profound consequences which may be more wide reaching than you'd expect. As in that article, I'm going to present some stances that  you might consider for your game setting, and look at how these influence the setting.

I'm going to talk about Good and Evil with capitals, in that I want to distinguish these grand concepts from the less important good (meaning beneficial, competent, fitting, etc.) and evil (meaning nasty, ill, etc.).
I'm also going to talk about gods, angels and demons, and mortals as creatures that have different moral scopes within a fantasy game setting - but I'm using these terms generally: I don't mean to refer to D&D 3.5's demons, WoD's angels, or Zoroastrianism's gods, or any other specifics.

Debating ethics

Good & Evil are Actual Forces?

In this stance, Good and Evil are real, measurable forces, like temperature or mass.

As well as conscious beings having morality, Good and Evil can inhabit objects and places. A book may be inherently Good, its presence harmful to Evil creatures - or dreadfully Evil, driving its readers insane. An Evil ring may corrupt the wearer. A Good, sacred place may heal the sick.

If Good and Evil are tangible forces, then are sapient creatures left with a fixed morality? Can Good be driven out of a person? Can an Evil being become Good?

Of course, it's a common complaint about alignment systems that characters' mortality becomes fixed, and this stance seems to reinforce that idea - but it needn't be like that...

A flexible alignment system in action?

Only Actions are Good and Evil?

Only the actions of sapient creatures are Good or Evil in this stance.

Things cannot have morality, only people - and then only as the result of their acts. By doing Good, Bobby the Barbarian is Good. By doing Evil, Winifred the Wizard is Evil.The Book of Nasty Tales, however, is just a book, the Sword of Demonslaying is just a magic sword, and the Chapel of Holy Rest is just a building.

Perhaps the build up of moral actions can be detected in a person - their Evil becomes tangible through their repeated acts, and we can sense it, through divination magic, or as a creepy feeling.

This stance inherently allows the moral standing of a person to change as they choose to act differently. Over time, a Good person can be corrupted by Evil acts, and vice versa. We can have redemption and damnation as story arcs - a rich vein of role-playing to mine!

While this stance is flexible and arguably more realistic, it lacks the supernatural element of Good and Evil that we enjoy from  fantasy literature.

The Middle Way

The above stances can be somewhat restrictive - especially for a fantastic setting, which is of course where we most commonly want to play with Good and Evil ideas.

Perhaps there are beings who are inherently Good and Evil, who are exemplars of those morals - Angels and Demons - but the common folk, the mortals of the setting, are free to choose their moral acts, and in choosing, to determine whether they are Good or Evil.

Perhaps objects can be imbued with Good and Evil by exemplars like Angels and Demons, or particularly pious or profane mortal magic users.

In such a setting, mortals retain free will, and can change their moral codes over time, but the existence of creatures of absolute morality is permitted. We can have our inherently Evil monsters, our literally holy ground, AND our ambiguous and mutable protagonists, struggling to do the right thing.

Stealing settings for fun (and possibly profit)

Rather than trying to invent some wholly original game setting (and then trying to explain it to the players), often it's easier and more immediately fun to take an existing adventurous setting and modify it to your needs.

Love Star Wars and Spaghetti Westerns? Do both at once! Just watched Jason and the Argonauts and Pirates of the Carribean? Merge them!

How do you do that? Break it down into essential parts of the setting you want to emulate, and think about how to translate those into your planned setting.

So - what are the essential parts of the setting? Let's look at a couple of examples.

---

Cowboys in Space
(or wherever)

- making a Wild West themed setting in any genre

You will need:

• The Old World
• The planet (or planets) from which our Colonial settlers' ancestors hailed - representing Europe (and Asia and Africa) in our real world history.
• The Old World has its problems, which may still influence what happens in the New World, but usually only at arms' reach: a war in the Old World might send more emigrants to the New World
• Example: Earth-that-Was in Firefly. We assume this was our Earth, and Firefly is set in our future. Earth-that-Was is gone, and has no further influence over the setting.
• Example: Mundia in the Helios system is the Old World. It is home to the human race, and is thought to be the orignal planet on which humans evolved and developed all civilisation. Its continental superpowers fluctuate between cold and hot war.
• The New World 
• The region that has been settled by the colonials.
• The New World is largely virgin and unspoiled; rich in resources, but hard and wild.
• Including:
• The Established Colonies
• The early colonies which have had a hundred years or more to settle down and expand - the equivalent of the 13 colonies of the US.
  
• Example: The Sapiens Cluster. The initial colonisation of the Cluster was acheived through the creation of an artificial wormhole, allowing vast insterstellar distances to be travelled in a single short hop. The Cluster's several adjacent star systems were swiftly colonised by sub-light expeditions, and are now teeming hubs of industry and commerce.
• The Wild Frontier
• The regions of expansion, where the established colonies are pushing into the hinterland - the Wild West of the mid-19th Century.
• Example: Since the colonisation of the Cluster, the invention of the FTL jump drive has allowed for interstellar expansion beyond the Old Colonies. The dozens of new frontier systems see explorers pushing back the edges of the known systems, prospectors following mineral rushes seeking fast riches, terraformers trying to create new farm lands, emigrants looking for a new chance - and the inevitable opportunists who support and prey on those waves of migrants: gamblers, grocers, doxies, doctors, engineers, bandits and marshalls.

Optional content: 
Some of the elements that go to make up the Wild West genre may be contentious and delicate, touching on real-world issues that are distasteful and perhaps out of place in an adventurous game setting - treat these with care.

• The Recent Civil War
• The Established Colonies and certain of the New Frontier regions have engaged in a bloody and devastating war over slavery or some similarly divisive ethical issue. The slump experienced by the losing faction causes resentment and division for decades to come. 
• Various Natives - oppressed, belligerent, friendly, etc
• Attempts to civilise the natives
• Annexation of native land and resources

---

Pirates of the Mediterranean
(other seas are available)

- a Greco-Roman themed take on swashbuckling piratical adventures

You will need:

• Civilised trading nations
• The sophisticated empire and other regions between which expensive and exclusive goods are traded - so that the pirates can steal them
• These will be your stand-ins for England, Spain and the "exotic" regions they traded spice with.
• Example: Greece and Carthage trade around the Mediterranean and beyond, clashing with each other, but more usually coexisting as rivals - up to a point.
• Naval authority
•  The civilised nations have their navies who protect their trade ships.
• Unexplored regions full of native riches
• Standing in for the Meso-American civilisaitons plundered by the European colonials
• Example: The African and European hinterlands are largely unknown lands to the Mediterraneans - so we can have hot and cold lands to explore and exploit, and all the alien cultures to go with them. With our mythic theme in mind, we can add legendary monsters and dangerous wizards into the mix.
• Lots of uncharted islands and coasts
•  Standing in for the islands of the Carribean, the Mexican Gulf, etc.
• Example: The Greek Islands themselves, as well as invented myriad islands for adventures among hidden coves, smugglers' stashes, pirate havens and isolated cultures. In this mythic setting, these islands can be plucked straight from the Odyssey, with cyclopes, harpies and all.

Table NSG1.0 & 1.1 - the Nemesis Speech Generator

Like many GMs, I like to make up tables for silly things. I found one of them today - my Nemesis Speech Generator.

This is for use when the Big Bad Evil Guy is monologuing.

NSG 1.0

NSG1.1

Fantasy settlements - part 2.5: Location, location, location!

People don't just throw a stick in the air and build a town where it lands - at least, those who do don't meet with enough long term success for your player party to encounter such a shambolic shanty.

What makes a good location?

A good location is one where the settlement can thrive. At its base level, the location provides something that other locations do not. That may be some physical resource, like water, or it might be a geographic resource, like a meeting of transport routes, or a social resource, like a holy site.

Here's a page where the topic is summarised nicely for high school level studies.

What can I add to that? Not a lot really - except some fantasy specific examples and possibilities.

Bridge point
The Twins - A Song of Ice and Fire
Osgiliath - The Lord of the Rings
 

Osgiliath by AbePapakhian (Deviantart)

Thandol Span / Dun Modr - Wetlands / Arathi Highlands, World of Warcraft

Nodal point
The Crossroads - Northern Barrens, World of Warcraft

Inter-dimensional and inter-planar portals - a town around the entrance to the Underworld


Dry point
Swamp Castle - Monty Python and the Holy Grail
Thousand Needles (post-Cataclysm) - mesas as islands in a now flooded valley, World of Warcraft

Atlantis (post sinking) - often depicted as an air bubble under the sea

Wet point
Ramkahen - Uldum, World of Warcraft

Ramkahen - Surrounded by desert

Defensive

Stormwind - fortified prominence, World of Warcraft
Minas Tirith - The Lord of the Rings

Minas Tirith guarding the narrow point between the mountains - Encyclopdeia of Arda

Aspect
Winterfell - hotsprings protect against the worst of winter, A Song of Ice and Fire

Shelter
Blackrock Mountain - a safe solid haven in a land of fire, Burning Steppes, World of Warcraft 
Light's Hope - a holy haven in a diseased land, Eastern Plaguelands, World of Warcraft

Resources
Nordrassil - magical resource - The World Tree of Mount Hyjal, World of Warcraft
Keep on the Borderlands, OD&D - a town supporting adventurers flocking to newly discovered dungeons or crypts

The growth point of a settlement will tend to set the form of that settlement: Aspect and Resource based settlements will often have dispersed buildings; Bridge and Nodal point settlements are frequently linear (clustered along one of the strips), and other settlement types tend toward the nuclear (closely gathered around  the central point).

As successful small settlements expand, their original purpose may be lost - they'll begin to attract people for the protection that comes from large numbers and the opportunities of towns and cities. These larger settlements become nuclear around the original centre.

EDIT - some additional examples:
From the various forums where I trawled for comments, a wealth of additional suggestions have popped up.

• To launch an offense - campaign headquarters may become permanent given a long enough conflict
• Military base - town supporting a military base, perhaps a training ground rather than a strategic point. This becomes significant with standing armies, rather than the predominant medieval muster of knights and militia.
• Religious - loads of examples. Probably fits the resource categor, under tourism. Pilgrims making devotional journeys to religiously significant sites were common in medieval times.
• Accidental / catastrophe survivors - refugee camps become permanent over time
• Victory / battle site - in the Excalibur movie, Camelot is founded on the site of a great victory
• Slave city for massive monument - the pyramids' construction was supported by a massive city of workers

• temples around the site of fissures that give off psychoactive fumes
• mage schools around areas where the boundaries between the planes grow thin
• mountaintop laboratories to catch the lighting
• village of vassals to a nearby dragon
• small settlements of people trying their luck at removing the sword from the stone
• summer/wintering grounds in the path of prey migratory paths
• boom towns/ghost towns - some enormously valuable resource has been discovered, and people are flocking to exploit it (boom), and the aftermath of the drying up of that resource (ghost town)
• outcast/leper colonies
• campsites for academic expeditions

Fantasy Settlements - part 2: Distribution

How often will you find settlements when travelling, either in civilised lands, or the wilderness?

Or, more to the point in fantasy gaming, if I randomly teleport the player characters to the middle of a country, how far are they from the nearest town?

Population density

How often you find settlements will depend on how densely populated the area is.

According to the 2007 Demographic Year Book of the United Nations, there were 49 persons per square km in the surface area of the world. Among the continents, Asia - with a density of 126 persons per square km - was the most densely populated continent, followed by Europe (32), Africa (32), Latin America (28), Northern America (16) and Oceania (4). [Source]

We can use these figures to give us some bench marks for the population density of the fantasy lands we're inventing.
Is our fantasy land crowded, like India? Or sparsely settled, like Australia?

Of course, these figures are from recent times - you might be planning a setting more like medieval Europe, or feudal Japan. We can use historical data to tweak our number fairly easily. For example, here's the Wikipedia entry for historical population - from which we see in the 14th Century, the global population of earth was about 300 million - about 5% of today's population - so for a simple formula, you could divide those figures above by 20...

Except that as we can see from these maps showing population through the ages, we've tended to cluster round some prime habitable area for most of known history. If we're going to assume that a populated area is going to be attractive to live in, then we shouldn't just arithmetically cut the density by 95%. Let's call it half instead, and add a "Wilderness" category, where practically no-one lives.

For a bench mark, I'll set the following density levels:

• Very high density population: more than 200 per square mile
• High density: 150 - 200 per square mile
• Medium density: 75 - 149 per square mile
• Low density: 30 - 74 per square mile
• Sparse density: up to 30 per square mile
• Wilderness: no random population

(I'm mixing my units in this entry, sorry, but population per square km is about one third of the population per square mile - and fantasy games prefer miles. So I've multiplied all those sq km figures by 3.)

Clumps of people

Where is that population, though? Populations are rarely distributed at random. People tend to settle in clumps - that's the technical term, honestly.
We're not just a blanket of people spread out over our respective countries, one per football pitch-sized bit of land. We gather together in settlements.

Urbanisation
As I mentioned before, the d20 fantasy rules seem to steer us to a relatively urbanised population.

In modern times, about 50% of the world's population live in urban areas. In the early stages of industrialisation, 10% of the population lived in towns and cities.
Before industrialisation, however, between 1 and 2% of the population lived in town and cities - the slower transport links of this era meant that 50 to 90 farmers were needed to support just one city-dweller. [sources]

For a typical fantasy setting, a 1% level of urbanisation would be appropriate - travel is by foot or horse, so getting fresh food into urban centres requires a close network of rural population: plenty of villages, hamlets and farmland.

Of course, in some high fantasy settings, magic can take the place of industrialisation, so that the early industrial or even modern distribution might be appropriate. Consider the level of magic, and the ability to rapidly travel. Food production might even be magically achieved. It might even be possible to exceed our real-world modern level of urbanisation.

I'm going to stick with a typical, pre-industrial feel for the fantasy setting. I tweaked the settlement categories so they would shake out to give a more rural population:

• Metropolis (25 000 +)
• City (5 001 - 25 000)
• Large town (2001 - 5000)
• Small town (501 - 2000)
• Large village (201 - 500)
• Small village (61 - 200)
• Hamlet (21 - 60)
• Thorp (up to 20)

I've crunched through the maths to find the percentage of the population that lives in each type of settlement, for an arbitrarily large population (large enough to fill all the categories).

• Metropolis 0.5%
• City 1%
• Large town 2.5%
• Small town 4%
• Large village 5%
• Small village 12%
• Hamlet 25%
• Thorp 50%

(Notice that over 90% of the population live in rural settlements - most live in communities of just a couple of dozen people... but we still have scope for some huge cities and metropolises.)

Now, from those percentages, for the benchmark densities above in an arbitrarily large area, I worked out how many settlements of each category there would be - from which I worked out how many miles there would be between each settlement...

Settlement distribution by density category

• Very high density
• Metropolis - up to 200 miles
• City - up to 100 miles
• Large town - 30 miles 
• Small town - 15 miles
• Large village - 7 miles
• Small village - 3 miles
• Hamlet - 1 mile
• Thorp - 1/2 mile

• High density
• Metropolis - 250 miles
• City - 125 miles
• Large town - 33 miles 
• Small town - 15 miles
• Large village - 7 miles
• Small village - 3 miles
• Hamlet - 1 mile
• Thorp - 1/2 mile

• Medium density
• Metropolis - 300 miles
• City - 150 miles
• Large town - 50 miles 
• Small town - 25 miles
• Large village - 10 miles
• Small village - 5 miles
• Hamlet - 1 1/2 miles
• Thorp - 1/2 mile

• Low density
• Metropolis - 450 miles
• City - 250 miles
• Large town - 75 miles 
• Small town - 30 miles
• Large village - 15 miles
• Small village - 5 miles
• Hamlet - 2 miles
• Thorp - 1 mile

• Sparse density
• Metropolis - 750 miles
• City - 350 miles
• Large town - 100 miles 
• Small town - 50 miles
• Large village - 25 miles
• Small village - 10 miles
• Hamlet - 4 miles
• Thorp - 2 miles

Random settlement finding

The thorp is the most frequent settlement type, so the distances between those are smallest. But this is also the minimum distance to any settlement. If you randomly appear anywhere in a sparsely populated region (let's ignore true wilderness for the moment), then you are no more than 2 miles from some kind of settlement.
What kind of settlement? Usually, that'd be a thorp - a clump of families, some little ranch and its cowpokes, a remote chapel with its attendant monks, or a country house and its family and staff. But it could be a metropolis, a city, town or village.

We've got a set of percentages of settlement types, and that turns quite neatly into a table. First we check how far away the nearest settlement is, then we roll to see what type of settlement it is.

Once we've got that nearest settlement figured out, the other types of settlement are randomly placed around about. The nearest city will be 350 miles away, the nearest large town, 100 miles, and so on.

		Miles to next
d%		very high density	high density	medium density	low density	sparse density
1-50	Thorp	0.5	0.5	0.5	1	2
51-75	Hamlet	1	1	1.5	2	4
76-88	Small village	3	3	5	5	10
89-93	Large village	7	7	10	15	25
94-97	Small town	15	15	25	30	50
98-99	Large town	30	33	50	75	100
00	roll 1d4
1d4		very high density	high density	medium density	low density	sparse density
1-3	City	100	125	150	250	350
4	Metropolis	200	250	300	450	750

Caveats and tweaks

If you're rolling randomly for an utterly new region, this system will work nicely enough - but take a look at 2.5, next, where I discuss location selection by desirability: people tend to build settlements where there's some resource to make the settlement worthwhile.

But if you're designing a setting, you'll probably have some ideas about where the capital city is, and where some other key major places are. (These are almost certainly placed in the most desirable spots - be sure to check out 2.5 next time!)

Remove metropolises and cities from the table - they only count for 1.5% of the total, anyway. You'll be left with town, villages and smaller settlements - just the sort of size of place for you to quickly put together some notes and wing it.

Fantasy Settlements 1.5 - Excessively Urban?

Since last Fantasy Settlements blog post, I've not been idle - I've been digging around in population data and settlement models, and making spreadsheets.
Despite how dull that seems, I think that some of what I found may be interesting!

D&D demographics are post-industrial
D&D and derived d20 fantasy games give us a bunch of settlement types / names to work with, as I mentioned last time:

• Metropolis, large city, small city, large town, small town, village, hamlet, thorp

These settlements are assigned population sizes, purchasing limits, and so on.
Notice that we only have three types of rural settlement, and 5 types of urban.

When I studied urban development and settlement theories at school, one of the key theories was Christaller's central place theory. (I suspect I remember it so well through a combination of this topic being one that my father helped me understand by getting me to explain it to him, and the fact that it lays out settlements on a hex grid...)

I'll not bore you with all the details, but simply put, he proposed (and then to some extent proved in the real world) that settlements grow according to their place in a hierarchy - the settlements that provide the most important or unique services grow largest and influence a wider area, while those with common services are less influential and grow less. That is: there are many farms (small settlements), but only a few seats of government (probably in a metropolis).
You can look up more details, of course, but the practical issue I want to talk about here is that he said that each order of settlement would be served by on average six of the next lower order: a town would be surrounded by six villages, roughly equidistant, and those villages by six hamlets, each, and so on.

When I applied this idea to the categories from the d20 fantasy rules, I found that for 1 metropolis, I had:

• 6 large cities, 
• 36 small cities, 
• 216 large towns, 
• 1 296 small towns, 
• 7 776 villages, 
• 46 656 hamlets and ...
• 279 936 thorps

The d20 settlement rules give the following guide to population sizes:

• metropolis, 25000+ 
• large cities, 10001 - 25000
• small cities, 5001 - 10000
• large towns, 2001 - 5000
• small towns, 201 - 2000
• villages, 61 - 200
• hamlets, 21 - 60
• thorps, 20 or fewer

When I put those population data into my numbers of settlements, I got the following average:

• metropolis, 37500
• large cities, 105 003
• small cities, 270 018
• large towns, 756 108
• small towns, 1 426 248
• villages, 1 014 768
• hamlets, 1 889 568
• thorps, 5 598 720 or fewer

That gave me a total urban population of 2,594,877, total rural population of 8,503,056 (rural being anything smaller than a town), in a total population of 11,097,933. In other words, a 23.4% urban population.
(This is of course based on a relatively small "metropolis" - bigger metropolises will tip the balance even further.)

Now, when I looked at real world data, I found that this level of urbanisation only started after the industrial revolution. Before that, there just weren't the transport links to make massive urban centres sustainable. We couldn't get fresh food to the city markets fast enough to support city dwellers on a large scale.
Prior to the industrial revolution, the split of urban to rural population was more like 1%, rising to 10% over the first hundred years of the agricultural revolution and early industrial era.
A 23% urban population is more like the level we saw in Europe around the mid to late Victorian era.
[source:http://isites.harvard.edu/fs/docs/icb.topic1050993.files/2-15%20-%20Kingsley%20Davis%20-%20The%20Origin%20and%20Growth%20of%20Urbanization%20in%20the%20World.pdf]

Urban sprawls need magic or technology
So it seems that the top heavy set of settlements that the standard rules give us means that the d20 fantasy setting is out of kilter with its usual pseudo-medieval idyll. It's a more modern balance, based on a time of technological advancement, railways and mechanised farming.

Of course, maybe a wizard did it.

In some high fantasy settings, magic can take the place of industrialisation, so that the early industrial or even modern distribution might be appropriate. Consider the level of magic, and the ability to rapidly travel. Food production might even be magically achieved. It might even be possible to exceed our real-world modern level of urbanisation.

But I think it's important that if you decide to have a wizard do it, you know what they have to do.

Next: Population density
Back to my intended schedule, in which I put some of this research into practice.

The Tube Map Solution: from railroad to network

I picked up some adventure modules to run as filler in my campaign / mine for ideas. A few were good, but I found that far too many were railroaded to death.

Railroading is what happens when the players' choices are forced or eliminated in the name of plot. The worst cases are ridiculous - I've played in a game where the GM actually told us that a hovering slab of concrete appeared overhead, ominously looming until we turned back and played the plot he had written.
It was kind of funny at the time, but not really in keeping with the tone of the game (this was in the supposedly serious and gritty World of Darkness game setting) - and it's certainly not the best way to deal with players moving away from your prepared scenario.

The Alexandrian made a great guide to railroading - a tongue-in-cheek list of points to embrace when deliberately writing a railroad plot.
He's also written some great blog posts on how to open up scenarios and to make the plot flow more freely, from the players decisions.

This is all great advice when you're writing from scratch. But I suspect that like me, many GMs have limited time, and want to draw on pre-published material to make life easier.
So what I thought would be useful would be to use a few examples from published adventures where heavy railroading happens, and see how we can expand the choices to allow the players to choose meaningfully.

There's more then one way to get from Paddington to Liverpool Street

Fatal deviation

The renowned Dragonlance adventures have a reputation for steering the party along a pre-planned path. That may or may not be deserved - plenty of pre-published scenarios presented a linear story path, from one dungeon to the next (not all of them, but plenty enough) - but when I picked one up recently, I found a few glaring instances of serious railing.

In one part of the scenario in question, the player party is asked to aid some elves escape the bad guys - if you don't know the story, suffice to say the bad guys are very, very bad, and numerous. If the party accept, all well and good: the story continues, and the game with it.
If the heroes refuse the elves (that being the title of the section dealing with that possibility), then the GM is instructed to first beset the player characters with nasty dreams showing their death, and if that isn't enough of a hint that they've done the wrong thing, then to attack them with a horde of bad guys.

What the text then says regarding these attacks is what stunned me when I read it - it's the most blatant piece of railroading I think I've ever read in a scenario. What it says is this:

"These skirmishes will continue, one every game hour, until all the PCs are dead."

That's right - if the players don't want to play the story as written, then their characters must all be killed. It's not very much different to the ominous floating concrete slab, is it?

In situations like this, I tend to think about what the consequences of the player characters' inaction or failure might be, and then allow those things to happen. Let the game-world be changed by the decisions of the players!

Let's consider a few examples:
Luke Skywalker misses his "one in a million" shot, and the Death Star destroys the Rebel Base.
Now the game is about a dark dystopia, with a furtive and desperate resistance, instead of the relatively strong Rebellion we see in the other two films. The Jedi don't reappear - who has any faith in Luke, even if he survives? He's just a failure, along with any other survivors.
Our story focus turns to underworld connections and lowlifes, and morality becomes far greyer than the Dark and Light Sides of the Force - who is interested in that mumbo jumbo now?
There are plenty of adventures to be had as rebels: smuggling guns, assassinating Imperials, and so on - but the game has shifted away from the heroic path that was expected.

Aragorn and Co try to follow Frodo instead of Merry and Pippin.
Merry and Pippin are brought straight to Isengard and - once Saruman figures out they don't have the Ring - used as bait for Gandalf (assuming they are PCs, we'll want to keep them involved and alive). Gollum is probably either killed or at least kept on a far more close watch - since there are now several of the Fellowship to watch him.
Frodo, Sam, Aragorn and Co make a much easier route to Mordor (he's a Ranger after all) - but vast tracts of Middle Earth are destroyed by the forces of Mordor (Aragorn and Gandalf do not save Rohan and bring the Rohirrim to the Battle of Pelennor).
Does the Ring corrupt the rest of the Fellowship? Can the larger Fellowship make it through Mordor unseen? What evil forces are left occupying the lands even if the Dark Lord is destroyed? The adventure continues, but not in the way that was planned.

Those of course are big scale examples, but I'm using them to make a point.
When you look at scenarios, you need to think about what the fallout will be if the player characters don't succeed, or if they don't follow what you think is the best path. Or if the players hit on a simple short cut...

Omniscient NPCs

Years ago, when I was running a Cyberpunk 2020 game, I attacked the player team after they thought they had escaped from the street gang pursuing them. One of the players asked "What? How did they know where we went?"
I probably had some stock answer at the time ("Who are you asking?" or "You don't know"), but it made me think, and it made me improve. NPCs have to act only on the information they have available. Just as players must separate their own knowledge from their character's (just because Pete knows the abilities and weaknesses of dragons in the game doesn't mean that his character knows them too) - the GM must separate his or her knowledge from that of the NPCs.

I picked up a Living Greyhawk scenario for D&D 3.5 at the Free RPG Day one year. I understand these scenarios are quite quickly written - there's literally thousands of scenarios for the setting, which was published for just 8 years - and that administration of the many regions of the setting across the world would have been a mammoth task, so I'm prepared to cut plenty of slack for copy editing, spelling mistakes and so on. What I'm far less impressed by are the frankly bizarre NPC encounters and their behaviour.

In one encounter, the PCs are confronted by a gang of thugs who have been sent to "sound them out" (and attack them).
What puzzles me about this encounter, and no doubt would puzzle players too, is that there's no explanation given for how the thugs know about the PCs or their mission.
The PCs have literally just met with a new patron (who himself is absolutely ridiculous - he's described as having a completely empty house, if the text is to be taken literally*, and he gives the player characters no reason to trust him, but every reason to distrust him...) who has sent them off to do some job - and they are accosted by these thugs, who know who they are, who they've just been talking to, and that they are in conflict with the boss thug. (At this point, due to the somewhat scrappy writing of the scenario, it isn't necessarily particularly clear to the players that they are in conflict with this thug boss.)

The scenario says "as soon as the PCs walk outside they are accosted" [my emphasis] and that the thugs have been sent by their boss "to feel out the PCs." This implies strongly that the thugs aren't just watching the patron's house, and decide to take on the PCs as they look like a bunch of adventurers and thus mean trouble - no, they've been sent there specifically to encounter the PCs for some reason.

As if this first band of prescient NPCs wasn't enough - another one arrives in 4 rounds flat! That's less than 30 seconds later, with no explanation of why they're all suddenly converging on the players' characters' party. These new arrivals are allies, too - for some reason. They aid in the fight against the omniscient thugs, despite having never met the player characters. Neither has their boss any experience of the PCs at this point - but he clearly sends his minions to help the PCs before they actually need that help.

How do the NPC bosses know about the PCs? What if the player characters take steps to avoid being seen? What if the PCs scout out the area before leaving?

As crazy as all this seems, these are simple enough questions for a GM to think up answers to  (of course, if the scenario was properly written, you wouldn't need to). Here's how I might answer the issues:

How do the bosses know? The NPC bosses have been spying on each other, and the patron. The bosses are rivals, and this patron is clearly trying to manipulate things. When word gets to them that the patron has visitors - adventurer visitors - both bosses send their fellows round to see what's up. The allied boss's minions aid the players because they are fighting their rival's thugs.

The PCs are cautious. Instead of the patron's house being empty, there are a few unobtrusive servants. One of the servants is passing information to a spy of one or other of the bosses - and the other boss's spy is observing this leaked information. Thus once the PCs are safe inside the patron's house, the information can be smuggled out - and the various NPC groups can converge on the patron's house while the PCs are getting briefed.
If the PCs scout the area before leaving, they see the thugs scaring off the locals, ready to set on the PCs as they emerge. The PCs have the chance to try to avoid them, intervene, or whatever they wish - but the allied NPC minions arrive as scheduled, and the thugs start a fight with them. Our PC party is supposed to be a band of heroes - no evil player characters are permitted in this scenario - so hopefully they might intervene...

*Sure, I know the writer meant that there was nothing worth stealing, knowing that players tend to have their characters loot anything valuable, but that's not what the scenario said.

Dead-end maze

"Team Bravo: the first assignment" is a supposedly "mini" adventure scenario provided by Wizards of the Coast for the d20 Modern game, which immediately turned into a multi-session marathon, deviating from the original plot enormously.
It's not a bad scenario, but it's full of points where the players can easily and very rationally pursue other angles, or overlook something the scenario writers think is obvious. This isn't so much a true railroad, but a maze, full of dead ends, with only one path through it.

The plot is supposed to be essentially three encounters: one with a petty criminal who has witnessed some dinosaur killing his mate, one with the mad scientist who has cloned the dinosaurs and one of his specimens, and one with the remaining escaped dinosaurs.
However, the progression of the story hinges on a few set pieces:

• Discovery and correct interpretation of a name tag at the scene of a crime (not too difficult)
• Pursuit and live capture of the mad scientist after he's set a killer dinosaur on them (very difficult)
• Discovery and correct identification of some tracking devices to pursue the escaped dinosaurs (medium difficult)
• Facing the dinosaurs (utterly deadly)

Name tag, chase scenes and a few alternatives

Adventures need more than one path through them, or they run the risk of getting stuck. The Three Clue Rule is well established now, so I won't go over it again here.

If the players ignore or overlook the name tag in the first scene of the adventure, then there is no path to progress the plot.

To be fair, for the scientist chase, there is an alternate method provided to get the party to move on to the next part of the adventure: a set of scanners is present in the lab.

Deadly dinosaurs

The final encounter is a bloodbath, in which any of the player characters will be lucky to escape - let alone defeat the dinosaurs. Three dinos lurk in ambush in the sewer. The scenario is supposed to be written for 2nd level characters - very new adventurers. just starting out in their careers. I ran through the numbers for those dinos' attack capabilities, bearing in mind the heroes defensive stats.

Not to overwhelm this post with maths, the short version is that the average damage deal by these 6 hits is enough to immediately drop any 2nd level character in this game system, and more than enough to utterly kill most characters - and there is a third dino also in this ambush...

Also, the sewer itself is a death trap. An earlier point in the sewer has a severe undertow current, which requires a swim check to avoid submerging. The difficulty of that swim check equates to something like a 5% chance of success for an average character, or about 50% for a strong swimmer - but that's assuming the PCs are unencumbered by armour or gear. Wearing armour hugely impedes swimming chances in this game.
So, effectively, the scenario has an encounter practically designed to strip the armour off the characters, immediately before the dinosaur ambush...

Now, it's not necessarily a problem to have an overwhelmingly deadly encounter in a game - that depends on the tone. Maybe your game is supposed to end with one single survivor barely making it out alive (a horror action story, like Alien, or the Predator movie) - or maybe you expect the player characters to recognise just how deadly dangerous one of the dinos is from the earlier fight, and tool up ready for the hunt.
The scenario writers didn't plan that, though. The writing implies that defeat of the dinos is a foregone conclusion - "After the heroes defeat the deinonychuses in the sewer, they can go back and investigate ..." it says. There's no acknowledgement of the deadly nature of that encounter - nothing like "Assuming the heroes defeat..." or "...the surviving heroes can ..."
Nor is there enough time to prepare for the hunt: they're expected to rush in before they even fully investigate the lab. Okay, one can easily give the players time to prepare, but the wealth system of this game means that there's not a lot of extra equipment they can gather to help them out.

Fixing things
In running this game, I had to make several changes.

• I added more clues - we need at least three, remember, and the writers had only given us two each time. (Two is better than most scenarios, to be fair.)
• I gave the party time to prepare for the hunt, and more gear, and I re-arranged the encounter to remove the flanking ambush.
• (It didn't help much, though - I still had to use a deus ex machina of some rival secret organisation to extract the nearly dead unconscious heroes, in the end. Thankfully, I'd been foreshadowing the existence of this other organisation throughout the adventure, and it also allowed me to add extra plot to the ongoing story - my players' characters were now indebted to their rivals...)

They worked at the time, but since then I've thought about it more, and I think I can do better. It still smacks of a railroad adventure: the players are pulled through a plot, each event happening on cue when they show up. Of course we need the players to be engaged by the story, but if they wander off to do something else, are the NPCs really going to sit around doing nothing?When I write adventures of my own, I like to write plans, not plots. The NPCs all have their motivations and agendas, and will carry on with them despite the PCs' actions. In the dinosaur adventure above, I'd include the dinosaurs as NPCs, too.

As written, those dinosaurs are just waiting in the sewer. Surely, they'd be more likely to get out there and attack more prey? And with more attacks - not just people, but animals too - the party might be able to find more clues.
Those clues need not lead them to the same events that the written adventure planned - can the dinos be tracked to their lair? Can the players lay a trap for the dinos?

The scientist in the scenario as written waits until the PCs come calling before he goes hunting the missing dinos. Shouldn't he be more proactive? Maybe some witnesses say they've already been interviewed, and give a description of the scientist. Maybe the PCs are tasked with investigating a missing person, when he gets eaten by his escaped creations...

Lastly, who hired the scientist? In the scenario, it's written that some secret military organisation commissioned the dinosaur project - what are they doing about all this?
In my version of the adventure, I used this complication - and it allowed me to pull my player characters out of the fire at the end of the scenario when the dinos proved to be as deadly as I've suggested above.

Mapping the tube

Ultimately, all these solutions boil down to one thing: preparation.
Read the scenario (thoroughly - don't skim it!) and make notes where it seems to be lacking depth, or is forcing the players into one railroad path.
For every railroaded scene or encounter, you should consider (and note down) the possible fall out, what the NPCs are up to while the PCs are footling around somewhere else, why the NPCs are acting the way they are written, and so on.

Plans not plots
NPCs carry on with their plans regardless of the PCs.
Consider what those plans are, and how they progress while the PCs are busy. Don't just leave scenes primed and waiting for the PCs to find them.
(Of course, you can provide a set piece scene now and then - and it'll work better because it isn't the norm.)

Fallout and consequences
Rather than the story grinding to a halt, or all the PCs being killed when they stray from the prepared plot, it pays to have an idea of how the possible outcomes will feed into the NPCs' plans.
Who will lose out, and who will benefit? Think of a few ways in which the event can play out, and note down what the consequences are. This will make sure the players feel that their actions are really important in the game, rather than just steered toward your planned plot.

Realistic and limited NPCs
Knowing how your antagonists know what they know means that you can think about what they don't know, as well. You're not trying to defeat the players, just challenge them - and that needs to be a fair challenge.
Of course, in fantasy and some sci-fi settings, some NPCs might really be omniscient. But when you  decide to legitimately used omniscient antagonists, the players will be all the more worried and alarmed because this hasn't been the norm.
And even such omniscient NPCs should have some source to their knowledge - so you can consider whether it too can be thwarted.

When I first thought of writing this post, my first instinct was to bitch about how awful those scenarios were, but it's far more karmic and constructive to show how to turn those glitches on their heads.
Hopefully, then, I've provided a few ideas on how you might use an hour or so of thought and a few notes, to take a railroaded scenario and turn it into some thing more like a choice-filled tube map.