Object, subject, verb. Self, friend, foe.

With 7,000 natural languages here on earth, it’s hard for an invented fantasy language to be unique - or even unusual.

AI-generated view of the Ruma, cradle of the Rumaran language and a comforting reminder that MidJourney has no idea how rivers work.

I started inventing the language Rumara pretty much blind: I was halfway through a book on inventing languages and dove right in. I had a few ideas and goals, outlined in my previous post, but I hadn’t considered their ramifications. What I discovered is that the arbitrary decisions I made reflected decisions, equally arbitrary, made by countless languages over thousands of years. Such decisions affect how we speak and, likely, how we think. When it comes to language construction and world-building, that’s an opportunity.

So what am I talking about?

Are “you” my friend or my enemy, and what does it mean for the grammar?

From top to bottom: bidiri, bidiru, bidirio, bidirun. I can already tell that these vowel forms would not survive one generation as written without simplification.

When it comes to verbs in Rumara, I thought of Latin where, like in most Indo-European languages, 1st-, 2nd, and 3rd- person verbs are conveyed by the endings. I sing, you sing, he/she/it sings; canto, cantas, cantat. Naturally, I generated my own endings. For bidiri, song, we get the following verb forms (I am often dropping the final vowel in verb forms based purely on how they sound. I like to imagine that Rumara has a set of rules for this that I don’t understand.)

• bidiri, I sing (1st person)

• bidiru, you sing (2nd person)

• bidirun, he or she sings (3rd person)

In the process, I was inspired to add another form - a new 2nd-person form used, not for friends, but for enemies.

• bidirio, you sing (2nd person, oppositional)

Why? I don’t think of the people of Ruma as being especially violent towards outsiders. Perhaps it’s the opposite: they are so welcoming that, when they find themselves facing an enemy, they can’t bear to use the same words. I imagine the River Kingdom as an ancient culture tied together by the river itself, surviving with a shared language as empires rise and fall around it. The language reflects this history.

Regardless, where in English we might simply say “You sing,” in Rumara you would say either “You, my friend, sing” or “you, my enemy, sing.” So:

Rakee xu bidiru
Please sing, my friend!

Rakee xio bidirio
Go ahead and sing, asshole (implied)!

• Rakee is a frame that means “I propose or demand that…”, with the context supplied by the rest of the phrase.

• xu and xio are the “you” pronouns for a friend and an enemy, respectively. With no suffixes, they are the subject of the sentence - the “you” in “you sing”. Clearly I decided they should resemble the verb endings, to the point where the sentence makes sense without the pronouns - you could say Rakee bidiru to your friend, and they would sing (if they felt like it.) I imagine including the pronouns is more formal.

• Fun fact, in Rumara most words are stressed in the second-to-last syllable. So bidiru would sound bi-di-ru, while bidirio puts the stress on bi-di-ri-o. Despite being different by only two vowels, the words would sound very different - which I like.

What are the ramifications of this? Well, it does mean it can be very clear when a speaker classifies you as their enemy. I’m creating this language for a D&D game. Just knowing it works this way means I can say something like:

The soldiers from the garrison keep their bows trained on you, and the leader asks, ‘Why do you come here?’

She is using the word xio for “you,” meaning she’s treating you as her enemy. What do you do?”

It’s a minor touch, but if later in the scene she uses xu for friend, it could be a dramatic moment. My players won’t even need to learn another language.

Do any languages actually do this, though? I am going to say - sort of. The closest I can find is that many languages have inclusive/exclusive versions of “we”: a “we” that means “me and you” and another that means “me and my friend, but not you.” This is called clusivity, and now that I’ve read a Wikipedia article on it it seems pretty interesting. Vietnamese apparently does this - chúng ta is the inculsive “we”, and chúng tôi is exclusive. (I say apparently because I don’t speak Vietnamese.) Some languages also do this with the verb ending. So I am not the first to use pronouns and verb endings to reinforce in-groups and out-groups.

I can’t find an example of a language with multiple 2nd person forms for enemy, but languages can distinguish between formal and informal - French has tu vs. vous, for example. So clearly you can say something about the person you are addressing. In real life I imagine that says a lot about the way a culture operates - and I suspect a language might change to lose these distinctions if the culture itself changes.

How many is a lot? And how many is exactly one?

In the spirit of “decide first, think it over later,” I address plurals partly through verb endings, partly through adjectives. I haven’t really thought through the second part yet, so lets look at plural verbs.

First of all, verbs in Rumara don’t encode tense. That makes it easy to encode both person and number without getting too complicated. I took what is called an agglutinative approach, although I did not know that at the time. This means that the different pieces of the words - morphemes - are stacked together as discrete parts. For example (where azko means “in the future”):

Azko yi bidiri
I will sing (lit. In the future, I sing)

Azko yi bidiriot
I alone will sing

Azko yi bidiriok
Some of us will sing

Azko yi bidirion
We will all sing

I am adding a plural suffix on top of the suffix for 1st-person (which is +i):

• +ot solo, alone, only: bidiri + (i) + ot

• +ok plural, each: bidiri + (i) + on

• +on plural, all: bidiri + (i) + on

This is completely distinct from the person ending:

Azko dan bidirun’ok
Some of them will sing

• bidir + un + ‘ok: sings + he/she/they + plural (some)

Again, why? The short answer is that I thought it was neat - I am doing this for fun, after all. I didn’t want to just replicate English, and adding gradients of plurality to the verb endings seems fun. The more I think about it, though, it seems natural for a culture that prides consensus and community. (Yes, my ancient river culture is a quasi-socialist utopia, what of it?) It’s important to note if everybody is doing something together, or if there are a few people striking out independently.

I especially like the ending for “alone.” The verbs are singular by default, so you’d think an additional meaning for “solitary” would not add anything - but it does! It’s the difference between “I can do it!” and “I can do it by myself!” I imagine that most of the time the solo form of the verb is a bit of a warning:

Arako siraivo drakabo xu garrakenu’ot?
Do you propose, with a knife, the dragon to attack by yourself?

Once again, this is not an alien concept. In English, the word “alone” added to a verb does the job nicely - all we’re doing here is incorporating it into the verb suffix. But that’s just English. Other languages tackle plurals in ways I do not understand. The Wikipedia article on grammatical numbers is baffling enough; it might be easier to look at the programmers working on translating plurals. It’s clear that languages can have plural forms for some vs. many; for countable vs uncountable; and others. Plurals are usually marked on the nouns, but they can be marked in verbs. In Rumara, the verbs tell you the number of the subject but not the object, and frankly I still have work to do on that front. Still, I can say:

Raza zinga bidiriha dan bidirun’ot
He sings many songs (alone)

Raza vo’nga bidiriha dan bidirun’on
They sing one song (all together)

• Note zinga - “many” - vs. "vo’nga” - one (adj), while the object - bidiriha - remains unchanged.

• The subject dan is the same in both cases, he/she/it/they.

• Again, the verb ending - ‘ot vs. ‘on - is the only thing that shows the solo vs. plural subject

That has an economical feel to me, and not unnatural. It suggests subtleties to the language that would make it tricky to learn if I weren’t the one inventing it.

I’m going to talk about frames one more time.

Frames were my first innovation, born of laziness. I wanted verbs in Rumara to all be in the present tense - I like the way they sound translated, and maybe I am still traumatized by learning perfect, imperfect, and pluperfect verbs in Latin. So my concept was that every sentence - every phrase - begins with a frame. The frame establishes mood and tense for the rest of the phrase, which ends with the verb. Frame and verb are like bookends or parentheses, and arguably the frame could be considered part of the verb.

From a flavor perspective, it means most phrases can be translated in a poetic or rhetorical fashion, like somebody reading an epic poem. Once upon a time, it is written, so it shall be… all great devices. Much more interesting than adding -ed to a verb.

Here are some of the frames I started with. The first version is standard; the short version would be used casually or when the context is already clear.

• Raza, ‘Za It is, happening in this moment, as we speak (present)

• Azkan, ‘Kan It was, a thing that happened, something already fact (past)

• Azko, ‘Ko It is foretold, ahead, it will be (future)

• Vokij, ‘Vok It is impossible, negative, it is not the case that (negative)

• Rakee, ‘Kee The hero proposes, I suggest, let’s… (subjunctive…?)

• Rako, ‘Koz Fate proposes, you propose, an enemy proposes, doom arrives (um…)

• Rakai, Kai Another proposes, a third alternative, neutral statement of timeless fact (you get the idea)

• A- or Ak- Turns the frame into a yes/no question.

This has worked better than I expected. It requires me to chunk phrases in specific ways - in particular, it suggests that there are few or no subclauses in Rumara. Can that really work? I have not tested it fully, so I am betting there are some complex structures in store for the language. But the interplay between verb and frame is already interesting.

For example, Azko zirgai would mean “In the future, I leave” - essentially “I will leave.” (Zirga as a noun means “horizon”, and as a verb means to go out of sight.) Vokij zirgai presumably means “I am not leaving” or possibly “I cannot leave.” I intend that you can double these up: Azko ‘vok zirgai would mean “In the future I cannot leave.”

But what about Rako zirgai? I like the idea of a mood that suggests something is happening despite the speaker’s (or subject’s) will, but I have not defined how that works. It could mean “Fate decrees I must go.” It might mean “You say I must go.” Or perhaps it simply means “I go” with a sense of foreboding.

I don’t know what I had in mind for Rakai zirgai, but at this point I feel like it can speak to habitual action: “I leave pretty regularly”.

Again, the nuance would be a pain to learn in a natural language, but for a storyteller or game-master, where you can decree meaning on the spot, this is solid gold. I’ve enjoyed the exercise of trying to frame sentences in this way (hence the word). So, is this truly an invention of mine Did I come up with a unique syntax never seen before?

I doubt it. It’s not the kind of thing that’s easy to Google, but I did stumble across an intriguing parallel with the Quechuan languages. Quechua was the language of the Inca Empire, and this family makes up the most widely spoken pre-Colombian languages in the Americas, with over 8 million speakers in and around Peru. I had never heard of them before last week, and the fact that they are called Runasimi is a total coincidence.

I bring them up because verbs in the Quechan languages have evidential markers. In other words, the language tracks how likely a statement is to be true - whether the speaker is staking their reputation on it, repeating something they heard, or posing a conjecture. These markers are attached to other words, often the first word or topic of the sentence - but sometimes to the specific word that is in doubt or being attested to.

I am not going to describe the system in detail on the basis of a Wikipedia article, but Googling quechua evidential gets you some cool results. The article summarizes the cultural aspects of this linguistic feature, although I wouldn’t take it on faith, as it’s reported by a non-native speaker:

(Only) one's experience is reliable.

Avoid unnecessary risk by assuming responsibility for information of which one is not absolutely certain.

Do not be gullible. There are many folktales in which the villain is foiled by his gullibility.

Assume responsibility only if it is safe to do so. Successful assumption of responsibility builds stature in the community.

Needless to say, this is roughly one hundred times cooler than anything I came up with. I wish English had this level of nuance - it would make the internet a much friendlier place. We could stop writing AFAIK and FWIW all the time.

For Rumara, I need to think about what frames mean for the language. How does it work when every thought begins with the context? What happens when speakers start dropping the frames for convenience? Over time, would they start slapping the frame on the verb just to save time, turning the language into something a little simpler? I don’t know, but it’s fun to think about.

Everything I know about languages I have learned in the last month.

A compendium of knowledge. If I retain 1% of what I have read in this I will be a much smarter man.

If you find it interesting, please don’t rely on me! I’ve been reading An Introduction to the Languages of the World this month: a fat little tomb that surveys the language families of Earth. Every few pages present a fascinating example of how languages work, and the ways they can be very different from each other. It breaks up high-level overviews with slightly deeper dives into languages like Russian, Finnish, and Ayacucho Quechua (yes, this is where I first learned of the language).

My attempt to read it cover-to-cover may not succeed, but I’ve found it to be very valuable.

I have also been enjoying the Great Course Language Families of the World, by John McWhorter (professor at Columbia). It’s an eye-opening survey of the vast array of languages on Earth, in helpful 30-minute chunks.

I am not sure what happens next with Rumara.

Maybe it goes in my back pocket until I am called upon to run a D&D game. Maybe I’ll work on the writing system - right now it’s little more than sketch of potential characters. Or maybe I’ll dive into an entirely new language to support a writing project I am working on.

Or maybe I’ll add some adverbs. Those seem fun.