GM Resources

Talking About Talking: Languages in RPGs

Lucas Asher

17 Mar 2025 • 10 min read

I've been reading a lot of other RPG blogs about the way that they tackle the topic of languages in their game. Dungeonfruit began this iteration of the conversation about language, and Murkdice followed it up with a really great article as well. But something that Murkdice neglected to consider was the extreme levels of illiteracy in the Medieval-Renaissance-Early Modern setting that is core to D&D specifically and a lot of high fantasy worlds generally.

Were People in the Middle Ages Illiterate?

This is a surprisingly difficult question to answer simply. While there has been a lot of scholarship about the concept of literacy, differing definitions of "literacy" make it hard to truly pin down a number. There was no "sixth-grade reading level" to compare people to.

Famed Welsh historian Derek Brewer "estimates that in England 'probably more than half the population could read, though not necessarily also write, by 1500.'" (source). But even this idea of reading can be broken down and challenged, as medieval people defined basic reading skills as knowing how to pronounce text correctly, without necessarily comprehending anything that they were reading.

Meanwhile, the term "literate" comes from the Latin littera, and would have meant formally educated. So even those who could read AND understand but who had not been to formal schooling would not be considered literate in the medieval sense. With formal education being expensive and restricted--usually only available to priests and nobles--we get our skewed statistics of the population being wildly illiterate.

Considerations for D&D

Still, a 50% rate of "being able to read at a basic level" means that I don't want to use Murkdice's definition of "fluency" in a language for a TTRPG. Particularly given the social origins of adventurers (often not nobles or priests and often not formally educated), I think it should not be quite so easy to be fluent including reading and writing... especially when, as Murkdice points out, being unable to read a text can make for a fun puzzle at the table.

It also opens up room for class or background features to include literacy and be relevant. Wizards and clerics, I would think, should be able to read and write. Monks maybe? The Noble background, sure. But for an Urchin Rogue? At best they would have that basic "can pronounce text"--but likely not even that, as that was often taught by parents, and a street urchin might not have access to even that level of learning if they do not have a strong family structure in place.

3 Levels of Fluency

Breaking language into three levels of fluency seems to be a popular division among the blog posts that I read. Extensive quoting below:

I see absolutely no problem with this division, but I want to complicate it just a bit, in order to bring it in line with how medieval people would have actually thought about literacy, from what our sources tell us.

Writing Fluency, Medieval Redux

Our sources tell us that there were three sorts of "literacy" to a medieval person.

Legere, meaning basic ability to read, in the sense of pronouncing text correctly. (source)
Intelligere, meaning understanding the text based on its grammar. (source)
Littera, meaning formally educated. My understanding is that we'd call this "media literate" today: you'd know how words were being used for rhetorical reasons, and you could analyze the text for meaning. (source)

So, to combine this with the ideas from AngryGM, Murkmail, and Dungeonfruit, here's the writing component of "language" in RPGs that I'm thinking about:

When attempting to comprehend a document, letter, or other written text, your level of proficiency with the written form of the language matters. Your proficiency with the spoken language and the written language are separate traits. If you read a text aloud, listeners (including yourself) can understand the text according to their spoken proficiency with the language, but capped at the reader's written proficiency.

Non-Proficient: You cannot read or understand the text of the document.
Legere or Basic: You can sound out the text to communicate it. However, you cannot understand the text by reading it silently.
Intelligere or Proficient: You can understand the text's meaning without any problems.
Littera, or Advanced: You can understand the text's meaning without any problems. You can also pose one question about the text--such as about its tone, or about the level of education that the author has--and receive an answer from the GM.

Most characters, unless they have a reason to be, should be Non-Proficient or Legere. Few should be Intelligere, and gaining Littera should be a fairly rare feature, reserved for scholars and the formally educated. It gives free extra information without a dice roll, so it should not be given out to most PCs. Let your scholars feel like scholars!

Speaking Fluency, Medieval Redux

The level of spoken proficiency with a language is far more based on game design than on history because it is even harder to know how medieval people spoke than how they wrote, just based on what gets passed down to us. We do know that many people in the medieval world would have been multilingual, which gives us a great excuse to encourage players to have their characters know a variety of languages. More on this later.

That said, I personally didn't love Murkmail's idea of having certain languages be naturally "opposed" by other languages. While speaking French to a German might mark you as an enemy historically, it is not the language that is offensive, and a German and a Frenchman might meet in Italy and agree to speak French because it is their common tongue. Instead, I'm sticking with the less granular version of the system from Dungeonfruit and AngryGM.

Here's my proposed parallel structure for the spoken form of language proficiency:

When having a conversation or otherwise listening to spoken language, your comprehension varies based on your proficiency with the spoken version of the language.

Non-Proficient: You cannot understand the spoken word in this language. It is complete gibberish, though you might still pick up basic tone and attitude (such as "this guy is angry"). You can make an insight roll to catch this basic attitude, but at disadvantage. (Or translate this–the skill and advantage–to your system of choice). Any rolls to communicate beyond catching these basic attitudes suffer a -10 penalty to the check.
Legere or Basic: You can catch a basic understanding of the spoken word. Any rolls associated with the conversation happen at disadvantage (ex: persuasion, insight, etc; if for understanding a spoken version of a document, this could also include history, for example).
Intelligere or Proficient: You can understand the meaning without any problems. You can make rolls without additional modifiers.
Littera or Advanced: You are able to use rhetorical skills to maximize the efficacy of your conversation. You can make rolls relating to the conversation with advantage.

I think this system--disadvantage, neutral, advantage--works nicely with D&D's framework and is simpler than adding in conditional, static modifiers. Having written text read aloud as it is structured, means that even the best speaker is sometimes hamstrung by a poor reader; while a scholar who excels at both the written and spoken language can truly thrive, getting advantage on many relevant checks and on bonus information. Plus, having the spoken and written word give different benefits (advantage vs bonus information) also gives people a reason to read things aloud to best understand them, something that is certainly true in an era when a lot of the things being written down were speeches and sermons.

What Is a Language?

Medieval and Renaissance people were polyglots. This was true of the upper classes when we think about languages.

I speak Spanish to God, Italian to women, French to men, and German to my horse. - Emperor Charles V of the Holy Roman Empire

But the lower classes would also speak a variety of what they'd consider languages, and which we would today consider dialects. The AngryGM goes into this concept and pitches how to use it for an RPG:

While the whole "fluency modifier" thing doesn't work directly with the setup that I have created, I like the "Language, Pidgin, Cant" framework. It covers a lot of the game-relevant bases, without getting extremely into things like regional accents as a factor--something that medieval people would have noticed and cared about, but makes games really overly complex (I've tried).

For languages and pidgins, I think the system can be translated as such:

True Languages operate under the system as written, with people being non-proficient, basic/legere, intelligere/proficient, or littera/advanced for both reading and speaking. No changes. This is for the classic languages: "elvish", "dwarvish" or their sort of country-level equivalents in your world.
Pidgins are simplified or blended languages meant to facilitate something and usually are not written (or rather, for written text, the more formal True Language is used). It is difficult to use formal rhetoric or grammar in these languages. So a blended example in the real world would be something like French Creole, while a "simplified" tongue would be Vulgar Latin--as opposed to Church or Imperial Latin. Pidgins follow mostly the same rules, with a few changes:
- Pidgins have associated True Languages--usually either one (for simplified pidgins) or two (for blended pidgins).

Pidgins do not have written proficiencies. It is just for speaking. If you want to write, you must use a written "true language" associated with the pidgin... which means you need to know some of the true language also.
Proficiency with pidgins is capped at Intelligere/Proficient. You cannot be Littera/Advanced in a Pidgin.

You automatically gain Legere/Basic proficiency in the spoken form of a pidgin's associated true languages if you are at least Intelligere/Proficient with the Pidgin.
You automatically gain Legere/Basic proficiency in any/all Pidgins that are associated with a True Language, if you are at least Intelligere/Proficient with the spoken form of that True Language.

Note: I do not have a comprehensive list of pidgins in my world. I let them spring up as the story requires, using that proficiency translation back and forth with associated true languages to facilitate.

Pidgins are also simplified when thinking about proficiency, which helps with this improv element--you either cannot speak the pidgin, can speak it sort of, or are proficient with it. No writing or advanced skill with it to worry about.

While it would be easy to consider cants just another tier of this, as AngryGM does, I think cants tend to fulfill a completely different role. They're not for daily conversation; they're for coded messages. Plus, I've never liked there being a universal "thief's cant"--surely thieves in one part of the world and thieves in another would have different cants, tied to their primary language.

As a result, I think cants are even simpler than pidgins:

Cants are essentially sets of codes (and indeed, you can use codes with this framework pretty easily):
- Cants have one associated language (but this can be a True Language or a Pidgin).

Cants are binary for both writing and speaking: you either are Non-Proficient or you are Intelligere/Proficient, for each of writing and speaking.
However, you are capped in both of these categories at the level of your ability with the associated language. So if you are non-proficient in elvish writing, you cannot become proficient with written elvish thief's cant.

Conclusion

And that's my pitch for a system for speaking/writing/fluency that draws from the three-level model of a lot of other blogs, but which also is a little more rooted in how medieval and Renaissance people would have understood proficiency with language themselves... or which at least draws from their terminology, and sets up a system where not everyone is going to be able to read and write.

Too complex? I get that! Let me direct you to Dave Clark's alternative to language entirely. It's not the idea behind this post, but I think that it is a great alternative way to think about language in RPGs if what you're looking for is to emphasize cultural identity in your world rather than focusing on your world's linguistic tree. It has the added benefit of being a neater tie-in for D&D's 2024 "species and background" system.

But, if you do want to keep language around and you want to make it more meaningful than is currently in the rules, I hope that between myself, Murkdice, Dungeonfruit, and the AngryGM you can find a system that works for you and your table.

You can help support the blog by sharing this post on social media to spread the reach, encouraging others to subscribe, or picking up my products on DMsGuild and DriveThruRPG. You can find all my products here!