Blog Archives

Hash Yer Ast Fin…?

The finale has come and gone, so I thought it might be fun to take a look back at my last Dothraki adventure from season 2. In addition to elucidating a key piece of dialogue from episode 210, it’ll also give you a glimpse into what it’s like to work as a conlang translator for a show like Game of Thrones.

As was mentioned in the last post, the final Dothraki line of season 2 is Jorah’s, and it’s shown below (with its on screen translation):

  • Mas ovray movekkhi moskay.
  • “Take all the gold and jewels.”

As those who’ve studied a bit of Dothraki probably realized, there’s not much tie between those words and that translation. In fact, one might say that none of those Dothraki words corresponds with any of the English words of the translation—and such a one would be correct. This is what happened.

At around 1:45 a.m. Pacific Time on October 10th, 2011 I was feeling sleepy, and was thinking about going to bed. This was unusual for me, because generally when I’m doing work for Game of Thrones, I go to sleep between 4:00 and 5:00 a.m., as it’s not uncommon for me to get a translation request around 4:00. It’d been a few weeks since I’d done anything at all for the show, though, and I was, for all intents and purposes, done, so on this night in particular I thought to myself, “You know what? I’m going to switch to a more normal schedule: In bed at 2:00 a.m., up at 10:00 a.m.” Pleased with my decision, I shut my computer and everything down and went to sleep at 2:00 a.m.

As luck (or fate) would have it, I received an e-mail from Bryan Cogman at 4:03 a.m. entitled “EMERGENCY Dothraki line!!!” He said they needed the Dothraki for “Take all the gold and jewels”, and they needed it in a couple hours. I ended up reading this e-mail at around 1:00 p.m. on the 10th, because I way overslept (you have reasonable control over the time you go to sleep; not necessarily over the time you wake up). Even though it was late, I quickly translated the line and sent it off to Bryan at 1:09 p.m. The line translated into Dothraki was:

  • Fichas ei hoshor ma dan.
  • “Take all the gold and jewels.”

Unfortunately, it did not, in fact, make it in time. Bryan wasn’t on set that day, but he said he thought they did it in Common—which is unfortunate (the more Dothraki, the better!), but what could I do? So I chalked that one up to bad luck, and promptly forgot about it.

Until May 29th, 2012.

At 2:17 p.m. I got an e-mail from the Game of Thrones postproduction supervisor asking for the Dothraki version of “Take all the gold and jewels”. While I think it’s primarily for foreign language versions of the show, they use the actual Dothraki lines for something in post (you never see them in the standard English broadcast without subtitles), and every so often something gets in that they can’t find the Dothraki translation for, so they ask me. In this case, I was a bit baffled, as I could’ve sworn my Dothraki translation for that line didn’t make it in. I sent her the correct translation above and asked if it sounded right, and she said that it sounded a “little different”. She then e-mailed me a recording of the line:

…and I’m all like, Ki fin yeni?! Then I realized what happened: They wanted Dothraki, so they had Iain Glen ad-lib. Now, we’ve seen some ad-libbing before, but never a sentence this long—and never for a sentence for which an actual Dothraki translation was already available (translating the above required no new Dothraki words). So I asked if I could have a day (bearing in mind that this episode would be airing in five days), and set to work.

My narrow transcription of what Iain Glen says is something like this (treating it like one big word with several main stresses):


There’s no [b] in Dothraki, of course, but all that means is it makes no difference if you pronounce something with [v], [b] or even [β]—and there’s probably a fair amount of dialectal/idiolectal variation. So the fact that there is (to my ear) a clear [b] in his ad-lib is no big deal. No, the thing that tripped me up was the presence of two [aj] diphthongs. As both are stressed, my immediate reaction was: participle. Participles in Dothraki aren’t common, but they’re possible—and most end in -ay (or [aj]). With that in mind, then, I could break down the stream into at least the following:

[ˈbɾaj mo.ˈve.hi.mos.ˈkaj]

At this point, it was a matter of chopping up words even further. Both the participle syllables could not stand on their own (if they did, they would need to come from the words brat/bralat or kat/kalat, respectively, both of which would end up violating Dothraki minimal word constraints in one tense or another), which means that they’d need to borrow at least one mora from a previous syllable. Seeing as participles most comfortably modify nouns—and seeing as the word ma is already a word in Dothraki and isn’t a noun—I decided the first chunk absolutely had to be mas ovray. As for the second, there’s that other stress to contend with. Having non-initial stress on an open syllable is nearly impossible in Dothraki, and seeing as -i is a ready verb ending, I decided to make the last participle moskay, leaving the middle part [mo.ˈve.hi].

Something that helped me out tremendously was Iain Glen’s character being non-Dothraki. One thing that non-native speakers will do is mispronounce tough consonants. So if you have a verb that’s [mo.ˈve.hi], there almost certainly must be a geminate. And since a geminate velar fricative would likely be flubbed by a non-native speaker, I decided that this verb would somehow relate to vekhat: a verb so semantically empty I could make it mean just about anything.

Once I had the words chopped, I had an even greater challenge: To create something grammatical that would have the same intended meaning as “Take all the gold and jewels”. The presence of participles made this more difficult than it might have otherwise been, but I saw this as an opportunity to fill in some gaps in Dothraki’s vocabulary.

For the first chunk, I decided that mas should refer to valuables, and ovray should mean something like “remaining”. Since I already had a word for “remain”, though, I poked around to see what I didn’t have. While I had something meaning “slack” (as in a rope), I didn’t have a word for something that was not attached, or not fastened down. As a result, ovray came to mean “loose” or “moveable”—and (especially) “portable”. I was able to use the same stem to derive the word ovrakh, which means (depending on context) “opening”, “availability”, “opportunity”, “vulnerability”, “weakness” or “weak point”. From mas I created a word meaning “to decorate” (ammasat).

The word movekkhat is a strange one, I’ll admit. It derives from the phrase nemo vekkhat. The latter verb is no longer really used by itself and the product of an old derivational strategy that you see in verbs here and there (e.g. lorat “to wink” ~ lorrat “to blink”), and it’s used to mean “to be intended for” or “to be for”. Vekhat means “to exist”, so vekkhat kind of extends that meaning. Since it was conventionally used with a reflexive subject, the mo of nemo glommed on to the front of the verb, until it eventually became a new verb (not an unusual process).

As for the last word, moskay, it means “to load” (a moska is a pack or sack used for transporting goods). Here is where I, yet again, took advantage of Jorah’s non-native intuition. To say something like “Everything not nailed to the floor is intended for transporting”, you’d actually use the infinitive. So, properly, the translation ought to be:

  • Mas ovray movekkhi moskat.
  • “The loose valuables are for loading.”

You could also use athmoskar, I guess, but moskat makes more sense to me. Jorah, though, comes from a language (Common, a.k.a. English) that makes much greater use of its participles—and also has a form that doubles as a gerund (e.g. “running” is a participle and a gerund). So it’s understandable if instead of using the correct infinitive he uses the participle.

And that’s how the last line of the season went from 100% ad-lib to official Dothraki. I sent off the above translation the next day, and it became canon. If he ever happens across this post, let me give a big thank you to Iain Glen for coming up with some phonologically plausible Dothraki! That’s not as easy to do as one might think.

Yet again I’ve written too much, so I’ll close this up. Thanks for reading! There’ll be more Dothraki tidbits throughout the offseason.

To Be or Not to Be?

Always a tough question for a conlanger. Not existence, of course, but the copula, and how to deal with it.

First, by way of introduction, the copula in English is our friend “to be”. It performs a few different functions, as in the sentences below:

  1. A dog is an animal. (Category Membership)
  2. That is Maria. (Equivalence)
  3. That door is green. (Predication)

Of course, in English, the verb “to be” does a lot of other stuff (e.g. passives), but it’s this basic X = Y relationship that we’re talking about. In English, “to be” does a lot of hard work for us, but other languages do things in different ways—and we don’t have to go too far down the linguistic tree to find differences. In Spanish (and a couple other [but not all] Romance languages), there’s not just one copula, like in English, but two, as shown below:

  1. La casa es sucia. “The house is dirty.”
  2. La casa está sucia. “The house is dirty.”

The sentences above translate the same way, but mean slightly different things. In (5), está is used to indicate that the house is dirty at the moment and could use some cleaning. In (4), on the other hand, es is used to indicate that that house is just a dirty, dirty house. Reminds me of when I used to walk to school past these two houses that my friends and I would call the Clean House and the Dirty House. The contrast was just too stark—and it never changed: the Dirty House was always dirty; the Clean House was always clean. In fact, can we get a shot of those houses, Google Street View?

A picture of two houses I used to walk by.

Click to enlarge.

Ki fin yeni?! Dirty House be clean now! How about that… Must be under new ownership. Trust me, though: the contrast was quite apparent, like…fifteen years ago.

Anyway, the contrast in Spanish is between more or less permanent states and temporary ones (or status vs. state). Other languages draw a distinction between identity constructions and locative constructions (a lot of creole languages do this), or stative predicates (things usually translated by “be + adjective” in English) or other standard copular constructions. That’s not the topic of discussion today. Today I wanted to talk about the form of the copula construction.

As I mentioned, both English and Spanish use verbs. Some languages do it without verbs, though. In Arabic, for example (and also Russian), standard copular expressions are done without any verb at all where one would otherwise expect a verb. Here are some examples from Arabic:

  1. Hiiya taktub. “She’s writing.” (Non-Copular Expression)
  2. Hiiya mutarjim. “She’s a translator.” (Copular Expression)

In (6), the verb taktub is fully conjugated in the third person feminine present tense; in (7), mutarjim simply translates as “translator”. All you need is the two nouns (or pronoun and noun) and that does the job. In the past tense, though…

  1. Hiiya kaanat mutarjiman. “She was a translator.”

…the copula (of the wazan k-w-n) reasserts itself to indicate that the expression is past tense (the object is also marked with the accusative). This also happens in Russian (another well-known zero copula language). Some languages, though, never have a copula in any tense, and simply use the same expression in all contexts.

Dothraki, as has been noted, is a zero copula language, as shown below:

  1. Hazi eshina. “That’s a fish.”

In fact, you can get away with using adjectives predicatively in this way, with a subtle difference in meaning:

  1. Hazi eyelie. “That’s spotted.”
  2. Hazi eyeli. “That’s a spotted one.”

This is kind of a status vs. state distinction, as in Spanish, except that the standard construction in (10) (using the stative verb) can be used for both a state and a status (i.e. for saying something is, at the moment, spotted [say it got splattered with mud], or for saying that something [say, a toad] is spotted all the time), and the latter can only be used with status statements.

Adjectives aside, the main place you see a zero copula expression is in equivalence statements (“He’s a warrior”, “That’s my horse”, “This is the arakh I’m going to cut your tongue out with”, etc.). Outside of the present tense, though, where one would see the reintroduction of a copula in, say, Russian, one sees a change in case in Dothraki. In fact, we can break it down rather simply as follows:

Tense Case Example Translation
Future Allative Me khalaan. He will be khal.
Present Nominative Me khal. He is khal.
Past Ablative Me khaloon. He was/used to be khal.

Seems like a nice neat system, but it didn’t come out of nowhere. In fact, this tripartite system derives from an older innovation from when Dothraki was a true zero copula language (and, well, it still is, I suppose, but it started to develop some copula-like constructions).

In the oldest form of the language, as I mentioned before, the word order of Dothraki was VSO. In order to augment the tenseless zero copula, the following expressions were developed:

  1. Ee me khalaan. “He will be khal.” (Lit. “Goes he to khal.”)
  2. Jada me khaloon. “He was/used to be khal.” (Lit. “Comes he from khal.”)

In the modern form of the language when the word order changed, the verbs were simply dropped, since they weren’t necessary to express the content. Even so, these verbs can be reintroduced (in their original verb-initial position) in the modern language if further clarity is demanded.

Today, the two separate systems overlap a little bit. The zero copula expression is still used in tenses other than the present when simple equivalence is desired. That is, if, for example, one was telling a narrative and the context is understood as past, the zero copula expression serves, as shown below (I apologize for the long block of text):

Ma anha dothra Qarthoon heshjim, ma anha arthas lajakasaraan. Mori lajish k’athvezhvenari, vosma anha drozh mora nakhaan. Irge vilajeri, fansa anni laz vos irvoso k’athnithmenari, majin anha fono chiories jimmoon, zhey Fenni. Me koalakeesi.

And all of that was just for the very last sentence. Here’s the translation:

And I rode southwest from Qarth, and I encountered a group of warriors. They fought bravely, but I slew them all. After the battle, my dapple couldn’t trot without pain, so I sought out a woman from the west named Fenni. She was a healer.

Even in English, in fact, you could translate that in the present tense, given the appropriate context (so not in a novel, but if your listener was immediately on hand, “She’s a healer” is fine [or if it’s in the near past and the woman in question is still alive]). In fact, if one were, instead, to say Me koalakeesisoon, it would mean, in that context, she used to be (but no longer is) a healer.

Apart from a context like this, using the ablative can mean “X was Y” without any comment about whether they still are or not, or “X used to be Y”. So a statement like this wouldn’t be a contradiction:

  1. Me koalakeesisoon, vosma me vos koalakeesi ajjin. “She was a healer, but she’s not a healer now.”

That’s about all there is to copulae in Dothraki. If you haven’t got anything else to do, feel free to enjoy this fine song about Hamlet.