Category Archives: Grammar
Discussions about Dothraki grammar.
Rytsas! I’ve busied up nice and good in recent days. I’m not sure if I’m going to be able to keep up with this blog. To keep up momentum, I’d be happy to feature user-generated content. If you have any ideas, throw them at me! I’m down.
Today I’m going to briefly discuss the number system in High Valyrian. Valyrian numerals are a bit more complicated than Dothraki numerals, but there are some nice bits in the system that improve its usability. First, all numbers are adjectives. In effect, you could treat them like participles, for those who are familiar with Valyrian grammar (for those who aren’t, I’ll show you how that plays out in a second). Here are the numbers 1 through 10 in High Valyrian in the lunar class (showing both cardinal and ordinal numbers):
As a refresher, all three adjectival endings are utilized in the table above. The nominative endings for each adjective type in the various genders look like this:
Anyway, you’ll notice that with the exception of tȳne, “second”, all ordinals are Class III, which should be helpful. The rest of the numbers split their class membership with one important exception, which I’ll explain in a bit.
Essentially, numbers agree with the nouns they modify in case and number. This should be fairly simple for certain things, but not for others. Let’s start with a couple ordinary examples. First, here’s an example using lanta, “two”, and a noun of each gender (vala “man”; azantys “knight”; dōron “stone”; hāedar “younger sister”) in the nominative:
- Lunar: lanti vali “two men”
- Solar: lantyz azantyssy “two knights”
- Terrestrial: lanta dōra “two stones”
- Aquatic: lantra hāedri “two younger sisters”
As you can see, all these nouns are in the nominative plural, and so the number matches in case and number. As all numbers are adjectives, though, they do display the same agreement that other adjectives do outside of the singular and plural numbers. Here are a couple examples (lentun “community”; mentyr “army”):
- Paucal (Terrestrial): mēriar lentun “one community”
- Collective (Solar): mēre mentyr “one army”
So above, even though we’re only talking about a single community, the agreement on the adjective “one” is plural (i.e. mēriar as opposed to mērior), just as the agreement on “army” is singular. Things are complicated slightly when these terms become words in their own right (falling into Declension Class VI). Some words do indeed jump the shark, so to speak, and become words of a more usual class (I know this was a question that came up before). For example, lentor, originally the collective of lenton, “house”, is now just an aquatic noun of Declension Class III, rather than a collective of Declension Class VI. In that case, lentor (the word for “family line” or “house”, in the Westerosi sense) would behave in the usual manner. A word like tembyr, though (“book”, lunar), behaves differently. Here it is in its two numbers:
- Singular: mēre tembyr “one book”
- Plural: lanti tembyri “two books”
Here even though it’s built off a collective, the adjective “two” gets plural agreement in the plural. Similarly, even though a paucal would ordinarily get plural agreement, it will get singular agreement in the singular if the word is being treated as a separate, relexified word.
All of this, of course, is much simplified when dealing with ordinal numbers. A couple of examples appear below:
- Singular: ēlie vala “first man”
- Plural: ēlī vali “first men”
That latter might look familiar (or its meaning, at least). Anyway, ordinal numbers agree entirely in case and number with the nouns they modify, since the number of an ordinal doesn’t actually determine or interact with the number of a noun in any way.
Now for the slightly more complicated part (although its effect will be to simplify things). Though lanta, “two”, and ampa, “ten”, might look similar, they are different in that ampa is never inflected. Thus:
- Lunar: ampa vali “ten men”
- Solar: ampa azantyssy “ten knights”
- Terrestrial: ampa dōra “ten stones”
- Aquatic: ampa hāedri “ten younger sisters”
The number ampa never changes for any reason, though its ordinal, amplie, does (in the usual fashion). Ampa is not the only number to do so. To see more, here’s another table with the numbers up to twenty:
|11||mēre ampā||kūrie||16||bȳre ampā||byllie ampā|
|12||lanta ampā||ñallie||17||sīkuda ampā||sīglie ampā|
|13||hāre ampā||saelie ampā||18||jēnqa ampā||jēnqelie ampā|
|14||izula ampā||izunnie ampā||19||vōre ampā||vollie ampā|
|15||tōma ampā||tōmelie ampā||20||lantēpsa||lantīblie|
A couple of things to note about the above. First, note the special ordinal forms for “eleven” and “twelve” (holdovers from the old days). Also note that all other forms use a modified version of ampa that ends in a long consonant. This is the result of the standard juxtaposition process of coordination. In short, the final vowel is lengthened, and main stress shifts to the last syllable (as with commands). The result, ampā, is still never modified, and is used in both cardinal and ordinal constructions. The word for “twenty”, lantēpsa, is likewise indeclinable.
Since it’s been brought up, here’s a quick list of the powers of ten up to one hundred (note: none of the cardinal variants decline):
A number like 121 would be (in the lunar) gār mēri lantepsā, so until you get to 200, that should take care of everything. There are numbers that go even higher (including the number naena, which does decline, which just means “too many to count”), but those will have to wait for another day.
Again, I’ve been absurdly busy of late, so I’m not at all sure if I’ll be able to hit even two posts a month, let alone four. I will do my best to keep up, though, I promise.
I’d also like to mention The Speculative Grammarian Essential Guide to Linguistics. This is a book I contributed to along with some of the other authors over at SpecGram, the internet’s premiere site dedicated to satirical linguistics. I don’t recall if there’s any Dothraki in there off-hand (there may be), but there are a few conlang-related pieces I wrote for SpecGram that I’m a big fan of (and, in case you’re wondering, yes, there are things I’ve written that I’m not a big fan of). If you’d like to purchase a copy of the book, you can do so here. It makes a good gift/bathroom book for anyone who has even the slightest connection to language. As we all speak one human language or another, I think that covers most humans… Anyway, if you’re curious about whether or not you might like it, head over to SpecGram and take a look at some of the articles there. That will give you a fair sampling of the content you’ll find in the book.
Until next time, geros ilas!
Okay, as I write that title, I’m now thinking I can’t promise I’ll say everything about adjectives, but I’ll say some things. Is that cool?
High Valyrian was primarily head-final, meaning that adjectives usually preceded the nouns they modified. It actually makes more sense to start a discussion of adjectival inflection by looking at adjectives that appear after the nouns they modify, though, as prepositive adjectival inflection can be seen as a reduction of postpositive adjectival inflection. (I’ll explain this in a second.)
Though nouns have a number of different declensions, adjectives comes in three major varieties which I’ll exemplify using these three adjectives below (for expository purposes, the endings are lunar nominative singular):
- Class I: kasta “blue, green”
- Class II: adere “sleek, smooth, slippery, fast, quick”
- Class III: ēlie “first”
Couple semantic notes on the above. Kasta is a word that can refer to anything that’s in the blue-green spectrum. Such words are common in older languages which tend not to have as many lexical color terms as a modern language eventually does. For a more in-depth treatment of this phenomenon, check out this post on Dothraki color terms from a while back. Second note is that adere probably first meant “slippery”, and then went on to develop the other senses.
The adjectives above are grouped the way they are because they inflect differently. Class I adjectives are the most informative, as they will decline differently for every case, gender and number combination—or almost. As with subject-verb agreement, adjectives only display partial number agreement (all adjectives, not just Class I adjectives). While a noun can appear in the singular, plural, paucal or collective numbers, adjectives only have singular and plural forms. In agreeing with a noun, an adjective will show singular agreement with singular and collective nouns, and plural agreement with plural and paucal nouns. The same is true of subject-verb agreement.
With that out of the way, this is what the inflection of kasta looks like in the singular:
And here it is in the plural:
Adjectives of Class II and Class III are distinguished by not having declensions that correspond to each gender. Instead, both classes group the solar and lunar genders together and then the terrestrial and aquatic genders together. Thus (and what is, by far, the most exciting part for me) each class can be represented with a single table. Behold!
And now for Class III:
Now Class II has a couple of subclasses which I won’t get into here, but these are the main three declension patterns you’ll need to know to correctly inflect postpositive adjectives.
Now for prepositive adjectives.
Rather than redoing the tables, I’ll just make some comments. For the most part, a prepositive adjectival form will lose its final syllable when the inflection is disyllabic. This means that you’ll lose the -ti in all forms that have it, as well as the -si in instrumentals and -mi in comitatives. Word-final -t is also lost unless the adjective modifies a vowel-initial word. Here’s an illustrative example using the dative:
- aderot ābrot “to the quick woman”
- adero Dovaogēdot “to the quick Unsullied”
This does mean that in the nominative and vocative plural you get, for example, kastyz rather than kastyzy (nominative) or kastyzys (vocative). That word-final -z usually devoices to -s unless the following word begins with a vowel or a voiced consonant. Another example:
- kastys hobresse “blue goats”
- kastyz dāryssy “blue kings”
Where a disyllabic inflectional form is simply VCV, only the final vowel is lost, not the final syllable. For example:
- ānogro ēlȳro “of the first blood”
- ēlȳr ānogro “of the first blood”
You’ll see this most often in singular instrumentals and comitatives, in addition to terrestrial/aquatic genitives of Classes II and III.
Finally, Class III needs some special attention. For forms that modify a solar or lunar word, where a shortening would leave the final syllable with ȳ, that vowel changes to io. The same is not true of the terrestrial/aquatic. Here are some illustrative examples:
- valosa ēlȳse “with the first man”
- ēlios valosa “with the first man”
- daomȳssi ēlȳssi “with the first rains”
- ēlȳs daomȳssi “with the first rains”
And a couple of final notes. First, as those who’ve been studying High Valyrian nominal declension will know, many paradigms often level the distinction between the instrumental and comitative (some using a comitative m form for both and some using an instrumental s form for both). When an adjective modifies a noun, it will agree with the split. All adjectives, as a result, have distinctive m and s forms, but for a particular paradigm, it may only inflect with one of the two.
Second, High Valyrian is in the process of eliminating word-final m (or, to put that more accurately, High Valyrian’s never liked word-final m), so contracted forms that end in m often only keep that m if the following word begins with a vowel or a labial consonant. Otherwise, that m becomes an n.
That should be enough to get things going with adjectives! To conclude, here are a couple notes on some things that came out in recent interviews. First, while I have provided translations to George R. R. Martin when he requested them (whether he used them or how can only be determined when the books the translations were requested for are published. I still haven’t gotten a chance to look at the maps book to see how those translations worked out), I never said I provided Valyrian translations. That was an assumption on the reporter’s part. Second, I recently did an interview for Entertainment Weekly’s radio program. Somehow my middle name came up, and at the end of the spot, one of the hosts guessed my middle name—or so I thought! When they repeated it at the interview’s close, I could have sworn they said “David Jasper Peterson”. If that is the case, then I’m afraid I misheard them the first time—i.e. they said “Jasper”, but I thought I heard my actual middle name. I hereby go on record to say that my middle name on my birth certificate is not Jasper, though I’d certainly like that name better than my actual middle name, which is terrible. My apologies to EW!
That concludes this initial look at adjectives in High Valyrian. I planned to include adjectives in Astapori Valyrian as well, but this post got too long… Another time.
OH! Almost forgot. The Valyrian section of the Dothraki Wiki is live, and it looks oustanding! Take a look at the High Valyrian vocabulary page, for example. There’s tons of interlinking examples throughout the wiki and a lot of good info. Excellent work!
A lot of hands went into putting the wiki together, but there are a few people who did the most work. Hrakkar did a lot of the behind-the-scenes work (with some help from our old friend Lajaki!) to make sure the wiki works the way it ought and all the links are correct. Then the bulk of the content was generated by Esploranto (a.k.a. Najahho) and Mad_Latinist, who’s rivaling me for the most frequent commenter on this blog. Kirimvose! It looks great!
Another season of Game of Thrones is in the books, which means that this blog will go back to discussing grammar—this time with Valyrian added to the usual Dothraki posts (though I will mention that the Dothraki posts have not disappeared. There’s more there yet!).
This week I wanted to talk a little bit more about verbs. I spent a lot of time on the verb conjugation paradigm, and am reasonably pleased with how it came out. We’ve already gotten a look at the present indicative tense, so let’s jump to the past. There are two main tenses that occur primarily in the past: the perfect and the imperfect. Each tense has a stem modification in addition to personal endings, but the stem modification for the imperfect is predictable. The perfect displays patterns of predictability, but is not 100% predictable based on the shape of the root.
To start with, let’s look at the imperfect. The imperfect tense is used primarily to set up action in the past. It focuses on a specific action in the past that is viewed internally (i.e. is viewed as not yet having been completed). In a sentence like “He was talking to some lady when her dragon lit him on fire”, the verb “was talking” would be in the imperfect in High Valyrian. The imperfect tense is associated with the -il suffix (by the way, pay careful attention to my use of the word “suffix” there. I’ve seen “infix” thrown around, but such an analysis is inaccurate) plus the e set of personal endings. Here’s what the imperfect looks like with a consonant-final stem. Below I’ll use the verb pāsagon, which means “to trust” or “to believe”.
|Person/Type||Imperfect Active Tense|
The imperfect has no associated participle, and no stand-alone infinitive or imperative.
When a verb stem with a final vowel is put into the imperfect, the vowel of the suffix -il coalesces with the vowel of the stem to produce a long vowel. As our example, I’ll use the verb bardugon, which means “to write” (coined in honor of Leigh Bardugo, author of Siege and Storm, which just came out [plug!]. You may remember her from such Dothraki words as lei).
|Person/Type||Imperfect Active Tense|
As you can see, the tense isn’t that difficult to get a handle on. The only wrinkle is figuring out whether a stem is consonant- or vowel-final, and then what the result is if the stem is vowel-final. Here’s a summary (using the first person singular active indicative as an example):
- pās-agon “to trust” → pāsilen
- bardu-gon “to write” → bardīlen
- keli-gon “to stop” → kelīlen
- mije-gon “to lack” → mijīlen
- nekto-gon “to cut” → nektēlen
- penda-gon “to wonder” → pendēlen
The above should be fairly intuitive. Moving on to the next tense, the perfect probably enjoys much greater use than the imperfect. The perfect tense focuses on an act that has been completed. By definition this action will have occurred in the past, but it can often be used with present relevance (what is often called an anterior). In English you can actually use the simple past in just this way. For example, if someone offers you food but you’re full, you can say, “I’ve eaten”. This is the English perfect, and it’s fairly standard. You could also say, “I ate”—even better if you add “already”. Think of the High Valyrian perfect as both of those uses rolled into one, but without needing the word “already”. Using our example above, the verb “lit” would be in the perfect in High Valyrian.
In the perfect, it’s not enough to simply know whether the stem ends with a consonant or vowel to figure out what the perfect will look like. Most of the time it has a -t or -et suffix, but this isn’t always (or exclusively) the case. Here’s what our two example verbs look like in the perfect. First, pāsagon.
|Person/Type||Perfect Active Tense|
What a tasty verb… And now bardugon.
|Person/Type||Perfect Active Tense|
Again, the endings are fairly simple (the same as the present tense endings), it’s just figuring out the stem. Here are some examples of perfect stems (again using the first person singular) and their associated infinitives:
- gaom-agon “to do” → gōntan
- henuj-agon “to exit” → hembistan
- māzi-gon “to come” → mastan
- pikīb-agon “to read” → pikīptan
- pygh-agon “to jump” → pȳdan
- qanem-agon “to sharpen” → qanēdan
- rāpūlj-agon “to soften” → rāpūltan
- rij-agon “to praise” → riddan
- rȳb-agon “to hear” → ryptan
- sik-agon “to bear” → sittan
- tat-agon “to finish” → tetan
- urne-gon “to see” → ūndan
- verd-agon “to arrange” → vēttan
A lot of the major patterns are contained in that list along with a couple of the more bizarre ones.
At this point, I think it’s more than possible to put a few sentences together. I’ll see what else I have time to put out in the coming months. Until next time, geros ilas!
As many will have noticed, there’s no new episode of Game of Thrones this week. There’s also no new episode of Defiance, for fans of the Syfy show. In fact, there’s not much on TV this weekend except for sports. The reasons is evident, though it seems that networks are only catching on this year. This Monday is Memorial Day in America.
Now ordinarily, one would think that since it’s a long weekend, people would be gearing up to go home and watch TV—and that’s often true. But as a holiday, Memorial Day is all but guaranteed to have the best weather of any American holiday throughout the year. The weather may be nice on certain holidays in certain parts of the country on any given year, true, but Memorial Day is just about guaranteed to have great weather in every part of the country every single year. As a result, families use this time to get together and go outside. And while sporting events work great for such weather (you can drop in and drop out, catch a play while getting something to drink, etc.), sitting down for a serious drama seems to be at odds with the gorgeous weather outside. Consequently, American networks decided to bow to the weather and take a week off.
Personally, I couldn’t be happier! This time of the year I often find myself out of town on the weekends (maybe not every weekend, but some weekends), which means that I have to miss a live airing of Game of Thrones, which is just not cool. This year I don’t have to worry! As with last year, I traveled up to the Bay Area for BayCon and also to visit with family (and with Shubert’s). And since there’s no Game of Thrones or Defiance, I can really enjoy the weekend!
While we take a breath as we prepare for the final two episodes of Game of Thrones, though, I thought I’d put up a couple of inflectional paradigms from High Valyrian. The hope is that these can be used as a general reference for the future. There’s been some excellent and fruitful discussion in the comments section of this blog, but as anyone who’s a regular commenter is well familiar with, it’s kind of hard to keep track of who said what when, and so I’m sure I’ve made some mistakes (misreading comments, saying comment x is incorrect when I really meant comment y, etc.). These paradigms I promise will be 100% correct (unless they need to be changed in the future [joking (kind of)]).
Starting with the verbs, those who’ve been following along will know that there are basically two types of verb stems in High Valyrian: those that end in a consonant and those that end in a vowel. In High Valyrian, a stem can end with any consonant or vowel, but those that end in vowels have paradigms which are quite similar to one another, and those that end in a consonant have paradigms that are quite similar to one another (in both instances, though, there will be variation in the perfect, which is the part of the paradigm most likely to be irregular). Here I want to give you the most regular versions of each paradigm so that you’ve got a base line to go off of. Let’s start with the easy one: consonant-final stems. As an example, I’ll use manaeragon, which means “to raise” or “to lift”.
|Person/Type||Present Active Tense|
A couple of comments on the table above. The (dark) grayed out part of the table are forms that don’t exist (there are no subjunctive participles or infinitives or imperatives). Where one form stretches across singular and plural, it means there’s no distinction. In the case of the participles, those are adjectives with regular adjective endings, and the first is used with a lunar or solar class and the latter with a terrestrial or aquatic (i.e. those specific adjective endings conflate lunar and solar into one class and terrestrial and aquatic into another). You’ll undoubtedly be able to glance at the table and pick out some patterns. Bear those in mind as we move to the next paradigm—this one for limagon, which means “to cry”.
|Person/Type||Present Active Tense|
Aside from the subjunctive, the tables should look quite similar (probably because the stem ends in -a), so it may prove instructive to do another vowel-final paradigm that should help to describe the rest of it. Here’s sōvegon which means “to fly”.
|Person/Type||Present Active Tense|
And with that, one should be able to figure out the rest. If you’re looking for something to hang your hat on, if you have a consonant-final stem, the first person plural present active indicative will always end in -i, and for a vowel-final stem, it will always end in -ī, regardless of the vowel in the stem. If you’re trying to fill out the rest of the vowel-final forms, yes, the first person plural and second person singular are identical with i-final stems, and in the subjunctive, the final o and u of o- and u-final stems both become v.
Since we’ve devoted a lot of space to verbs, I’d like to wrap up with a couple common noun paradigms. You’ll notice that a lot of names of Valyrian origin end in -ys. This is how nouns and names of that type decline. I’ll use the word loktys, “sailor” as an example (a solar noun of the second declension class. Most [but not all] words of this class are solar).
It might prove instructive to refer to the first declension lunar paradigm revealed last week and compare it to this one. Pay particularly close attention to the singular and plural numbers, and note where cases are conflated and where they aren’t. This is what defines declension classes in High Valyrian.
Oh, and since it doesn’t fit anywhere else but I feel like mentioning it, verb stems never end in a long vowel or diphthong, and you’ll run into the following diphthongs in High Valyrian: ae, āe, ao and āo. There are also some on-glide diphthongs which can serve as the nucleus of a single syllable: ia, iā, io, iō, ie, iē, ua, uā, ue and uē.
I hope you enjoy the week off from Game of Thrones! Come next week, things are going to start to get messy. Geros ilas!
It’s now February 20th, and this is the first Dothraki post of the month. Given that it’s a short month, this may very well be the last, as well. I feel obliged to offer up some sort of explanation, given that (most months) I’ve been pretty good about living up to my unwritten (until now) four posts per month goal.
As it has turned out, this month has been pretty busy. In addition to the SWTX PCA/ACA Conference from last week, I’m giving a TED University talk at TED this month (a whole 6 minutes on the 28th!), and have been busy doing a lot of prep work for that and for TEDActive, where I’m giving a workshop. If you want to talk any Dothraki, the best place to catch me these days is on Twitter or at our weekly Dothraki chat on IRC.
I didn’t want this post to be completely devoid of Dothraki, though, so I thought I’d address an issue that came up on Twitter. Our latest (and quite prolific!) Dothraki speaker Tyene Sand was trying to translate a sentence using the Night’s Watch (that is, the name “the Night’s Watch”). That can be translated in a number of ways (I offered Vitihiraki Ajjalani), but the translation called for the phrase to be declined in some way. This is where one runs into a dilemma.
In Turkish, if you take a foreign noun and try to decline it, the word behaves a little differently from native (or assimilated) Turkish nouns. Turkish names take a number of case suffixes (similar to Dothraki), but these suffixes participate in vowel harmony. Here’s a small example:
|mağaza||store||mağazada||at the store|
|göl||lake||gölde||at the lake|
As you can see, in the Turkish forms in the third column, there’s a suffix that’s either -da or -de. Which suffix you get depends on the character of the previous vowel (for more, see this article on Turkish vowel harmony), but they both mean the same thing.
That’s fine and good. What happens, though, when you add these suffixes to a foreign word? Turkish, as it turns out, does a couple of things differently. First, the suffix is always attached with an apostrophe (kind of like how sometimes in English, acronyms are pluralized with an ‘s as opposed to just s [e.g. DVD's rather than DVDs]). Second, unless the quality of the vowels is quite apparent, Turkish just uses one of those two suffixes—specifically, the -da suffix. Here’s an example:
So, now that we know what Turkish does, what does Dothraki do?
First, Dothraki noun phrases are often declined on the head noun. This is the rough equivalent of “passerby” vs. “passersby” in English (the latter being the formal plural of the former). Take, for example, the phrase asavva evomen, which has various meanings depending on context (for now, let’s say “afterlife”). If one wanted to pluralize this phrase, the appropriate plural would be asavvasi evomeni (the latter adjective taking an -i on account of concord). That is, asavva is the head noun, so it takes the plural; one doesn’t treat the whole thing as a single noun and attempt to add some sort of inflection to the end of evomen.
That said, one may want to write in Dothraki and talk about modern people, companies, products, places, etc. For something like “Google”, one option would be to try to translate the concept (good luck) or to render it in Dothraki (Gogol?). This might end up making things more confusing than necessary, though. As a result, the kind of catch-all repair strategy used in Dothraki is the preposition haji. Haji means something like “because of” or “on account of” or sometimes “with respect to”. In Dothraki proper, its meanings are a bit more specific. When used in conjunction with foreign names or terms, though, it stands in for any preposition and/or the genitive, allative or ablative cases. Thus, one might say something like:
- Anha tih mae haji Reddit.
- “I saw it on Reddit.”
Technically haji there could be standing in for she, ma, irge, hatif, vi, ha, ki—or the ablative, genitive or allative cases. Really, though, given the context, it seems likely that it’s standing for she (a general locative. Not sure if anything more specific would be used to refer to something one sees on a webpage. Mra, maybe?). One might be able to supply a context that would force another reading, but the most obvious reading suggests that whatever was seen was seen on Reddit.
Though the solution is pretty simple, the drawbacks are that there could be confusion or ambiguity, so it behooves one to supply the proper context so that only the correct interpretation is plausible. If more specificity is absolutely required, one can always use the proper preposition. If a case is needed, it’s probably best to attempt to render the noun in Dothraki, as below:
- Anha dothrak Disneylandaan!
- “I’m going to Disneyland!”
To make it clear, one may (in the Turkish style) separate the case ending from the root with an apostrophe, but personally I prefer it without.
I hope your February’s going well and that it’s not too cold where you are! It rained today, so California will get a bit chillier for the next couple of days, but otherwise I can’t complain. For those of you who speak or are familiar with other case languages, what do those languages do with foreign proper terms? How would “Google” come out in the instrumental in Russian? Or the translative in Finnish?