In the real world, humanity speaks hundreds if not thousands of languages, and has spoken thousands or tens of thousands in its history.
Most table top role-playing games (RPGs) make the following simplifying assumptions:
- PCs and most NPCs speak a “common language”.
- Some communities – nations, “races”1, religions, professions – speak less common or uncommon languages. There are at most a few dozen of these languages.
- PCs can understand languages in the communities they belong to, and possibly “foreign” languages depending on their class (game or social) and education.
- Languages are unrelated to each other, and must be learned separately if they can be learned at all after character creation.
To approximate how languages work in our world, these rules reverse the last assumption. But it requires a little work.
The GM groups languages in the game world according to their degree of similarity, using a graph or a table.
Rather than learning each langugage from scratch, player characters get a break in learning a language that’s similar to one they already know.
The first step is to place all the game world’s languages into groups based on their degree of similarity.
Degrees of Similarity
While linguists may debate the degree of similarity between real languages, for our purposes we’ll classify languages as one of the following:
- The languages differ only in accent, idioms, and occasional or obscure vocabulary changes. Speakers of one can understand speakers of the other with only a little difficulty. For example, some dialects in the North or West of England differ significantly from Standard British English.
- The languages diverged only within the past dozen generations or so, so they share enough grammar and vocabulary that speakers of one can generally understand the others. Examples include most dialects of English (particularly northern ones) and Scots, or Danish, Swedish, and Norwegian. In both cases, languages in the group descended from a common source, and speakers of each are in close physical proximity.
- One language descended from another, or both from a common source, and still share much vocabulary and grammar in common. The (Western) Romance Languages – Catalan, French, Italian, Portuguese, Spanish, etc. – are oft-cited examples; all descended from Latin nearly two thousand years ago, and share 70-90% of their vocabulary with Latin and each other as well as some common gramatical structures.
These rules will ignore what we’ll call “trivial dialects”, such as the varieties of North American English, which native speakers can understand with little difficulty and non-native speakers can decipher with only a little practice.
Likewise, we will ignore languages that may have a common ancestry but have few words or grammatical structures in common. English and German are both Germanic languages, but English has changed dramatically in the last dozen centuries. Knowing English really doesn’t help one to learn German. (Trust me on this.)
For clarity and convenience, the GM should create a table of languages. Suggested fields in the table:
- the name of the language.
- the name of one or more groups to which this language belongs, mainly for convenience.
- Native Speakers
- the ethnic group, region, species, or other community that uses the language.
- Scripts (optional)
- the writing systems for this language, if any. See a future article if the GM wants to define this in more detail.
- Written Only? (optional)
- a column, flag, or notation indicating that the language has only a written form.
- Extinct? (optional)
- a column, flag, or notation indicating the language is extinct. By definition an extinct language is Written Only.
Then the GM would create a Group Table with the following columns
- Language Group
- The name of the Group
- Whether all languages in the group are Dialects of each other, Close to each other, or simply Related.
- Languages in each group.
If the Group Table and the Language Table disagree about which groups a language belongs to, players should alert the GM but assume the Group Table is accurate.
It’s helpful to name dialects and dialect groups by the parent language. Instead of enumerating them all, the Languages list in the Group table can simply name the group(s).
Note that languages may belong to several groups, even groups with the same degree of similarity. For example, a widely distributed language can form a dialectic continuum, such that languages next to each other are (in our terminology) are Dialects, but those at the edges are only Close, Related, or worse. A table for such languages might look like this:
|Ologun, Mountain||Olokai, Ologun (Eastern)||…|
|Ologun, Plains||Olokai, Ologun (Western)||…|
|Ologun, Plateau||Olokai, Ologun (Central, Eastern)||…|
|Ologun, Valley||Olokai, Ologun (Central, Western)||…|
And the group table would look like this
|Ologun||Close||all Ologun dialects|
|Ologun, Central||Dialect||Plateau, Valley|
|Ologun, Eastern||Dialect||Mountain, Plateau|
|Ologun, Western||Dialect||Valley, Plains|
|Olokai||Related||Olokun (all dialects), Kaitan, …|
So, for example, Mountain Ologun and Plateau Olokun are Dialects, Mountain Ologun and Plains Ologun are only Close, and Mountain Ologun and Kaitan are only Related.
The game effects of Language Groups depends on how the rules and/or GM handle languages in their game.
For convenience, GMs should ensure all PCs have a language in common. If they speak different Dialects of that language, ignore those effects when they speak to other players. (Players can put on a brogue or funny accent if they wish and the other players are OK with it.)
In systems that assume characters either know a language or don’t, the degree of similarity affects the use of communications skills.
If the characters don’t have a language in common, assess a penalty to all communication skills based on the most similar languages they share. (Unless one party is deliberately trying to misunderstand.) The exact penalty depends on the dice used for skill checks; see the suggestions below:
|2d10 or 3d6||-1||-3||-6|
|d100||-20% or x2/3||-50% or x1/2||-75% or x1/5|
|Year Zero6||none||2 successes required||3 successes required|
The GM should let characters pick up a new Dialect of a language they already know after spending a few months of game time around speakers of that Dialect.
If characters can learn new languages after character creation, assume that they can pick up a Close language in half the time, or with half the cost, of an unrelated language. At the GM’s discretion, acquiring a Related language can also receive a lesser, standard discount in time, experience points, etc.
In systems that represent languages as ordinary skills, characters automatically comprehend a language at a skill level based on the most similar language they know. Defaults for specific systems are below.
|d100 (e.g. BRP, OpenQuest)||skillx2/3||skillx1/2||skillx1/5|
|2d6 (e.g. Troika!)||skill-1||skill-2||skill-5|
Characters can also improve the dialect or language like any other skill.
For example, a character in BRP who speaks some standard form of English at 74% could speak non-standard Dialects at 49% but they’d only comprehend Scots (a Close Language) at 37%. They could, however, improve those skills with practice.
As a practical matter, players shouldn’t record the skill for a language on a character’s sheet until they improve it. Once they improve a Dialect, Close Language, or Related Language independently, improvements to the language they defaulted from don’t raise the new one.
GURPS 4th edition classifies language comprehension as Broken, Accented, and Native. Each level of improvement costs two character points (CP).
Under the Language Group rules, GURPS characters can automatically speak all Dialects of languages they know at one step lower. If they know the parent language like a Native they gain all Dialects at Accented, and if the parent language is Accented they comprehend the Dialect at Broken. Furthermore, if they speak a language at Native level they comprehend all Close languages at a Broken level. If a character spend points on a Dialect or Close Language to one they already know, they get those levels for free.
For example, a character who knows English (standard British or American) like a Native can speak and understand most Dialects from Britain, Ireland, and beyond at an Accented level. They also knows Scots (a Close language) at a Broken level; if they want to learn Scots better, they need to spend only 2 CP to raise that to Accented.
Characters can acquire a language for one less CP than normal if they speak a Related language at Accented or better. For example, someone who knows Spanish at Accented or Native level can speak Portuguese at a Broken level for one CP, Accented for 3 CP, and like a Native for 5 CP.
The GM can adapt this pattern for systems with similar comprehension levels.
Example: The Old Empire
In one game world I created, an empire had conquered nearly the whole known world a thousand years before, and then fallen. Instead of a single “Common” language, everyone spoke Close descendants of the ancient Imperial language, roughly separated into Northern, Western/Central, and Eastern Provincial, plus a few regional languages. Since I was using Magic World (a derivative of BRP), which treated every language as a skill, native speakers from one end of the old empire could only comprehend someone from the other end at half skill.
People outside the boundaries of the old Empire or from isolated enclaves spoke completely unrelated languages, which defaulted to 0%.
For the world of Telluria, I created the following language tables. At the time I hadn’t settled on a system.
Telluria has several languages, some common and some local. Player characters know at least two, and probably more, based on where they came from and where the campaign started.
|Aklo||sorcerers and worse|
|Chathic||Eastern||Great Desert, Wild Lands|
|Goblin||goblins of the Goblin Steppes|
|Keshite||Keshitic||Fertile Triangle, Great Desert|
|Keshitic, Old||Keshitic||Fertile Trianle (extinct)|
|Kymric, Bardic||Kymric||ancient bards of Kymry (extinct)|
|Kymric, Verdan||Kymric||Kymric settlers in Verda|
|Northlandish||Northern||Border Kingdoms, Kaerillus, North Lands|
|Russic||Ghost Lands (The Rus)|
|Sarkennian, Hieratic||Sarkennian||written only|
|Vethic, Common||Vethic||“common” elves of Verda|
|Vethic, Dark||Vethic (Classical)||Dark Elves of Verda|
|Vethic, Druidic||Vethic (Classical)||ancient druids of Verda (extinct)|
|Vethic, High||Vethic (Classical)||High Elves of Verda|
|Vethic, Trollish||Vethic (Classical)||“trolls” of Verda|
|Vethic Pidgin7||Vethic||human slaves in Verda|
Languages in bold are common languages, spoken as a second language in surrounding areas. PCs would know at least one of these, possibly two. Specific areas of influence:
- Keshite: Border Kingdoms, Great Desert, Kaerillus
- Kymric, Common: Border Kingdom, North Lands
- Northlandish: Kymry, Verda
- Sarkennian, Demotic: Wild Lands
- Theran: ports on the Theran Sea, Asterion/Qutub
Here’s a rundown of groups.
|Keshitic||Related||Keshite, Old Keshitic|
|Kymric||Close||Bardic, Common, Verdan|
|Sarkennian||Close||Hieratic, Demotic, various tribal languages|
|Vethic||Close||all Classical dialects, Common, Pidgin|
|Vethic, Classical||Dialect||Dark, Druidic, High, Trollish|
By which I mean elves, dwarves, goblins, etc., which should properly be called species. But that’s a whole other discussion. ↩︎
Here we use the term “dialect” informally. The difference between a “dialect” and a “language” is hotly contested, and more political and social than linguistic. Linguists sometimes say that a language is a dialect with an army. ↩︎
Assuming the speaker of the dialect doesn’t understand the more common or standard version of the language. ↩︎
Or any system using a dice pool where each die has a roughly 50% chance of success ↩︎
Or any system using a dice pool where each die has a 30%-40% chance of success. ↩︎
Or any system using a dice pool where each die has a roughly 1 in 6 chance of success. ↩︎