An Interview with David J. Peterson and Sai Emrys

An Interview with David J. Peterson and Sai Emrys
about Dothraki and the Language Creation Society

Author: Jim Henry

MS Date: 10-23-2010

FL Date: 10-01-2013

FL Number: FL-000019-00

Citation: Henry, Jim. 2010. «An Interview with David J.

Peterson and Sai Emrys about Dothraki and
the Language Creation Society.»
FL-000019-00, Fiat Lingua, . Web. 01 Oct. 2013.

Copyright: © 2013 Jim Henry. This work is licensed under

a Creative Commons Attribution-NonCommercial-
NoDerivs 3.0 Unported License.


Fiat Lingua is produced and maintained by the Language Creation Society (LCS). For more information
about the LCS, visit

Interview with David J. Peterson and Sai – Jim Henry – 1

An interview with David J. Peterson and Sai
Emrys about Dothraki and the Language
Creation Society

An abridged version of the first part of this interview appeared in Usona Esperantisto, 2010:5,
pp. 7-9 as «Dothraki and Esperanto: an interview with language creator David J. Peterson».

Since the interview was conducted, Sai has dropped the «Emrys» to become mononymic.
Historically, however, it seems suitable to refer to him here by his name at the time of the


David J. Peterson, a noted language creator, has been hired to flesh out the Dothraki language
from George R. R. Martin’s epic fantasy series A Song of Ice and Fire into a full-fledged,
speakable language. This constructed language is to be used for the upcoming HBO television
series based on the novels, titled A Game of Thrones after the first book in the series. The
producers of the show hired the Language Creation Society to find an expert language creator for
this job; a panel of nine LCS volunteers reviewed thirty-five proposed versions of Dothraki, in a
double-blinded process, and passed on four of the best to the producers, who selected Peterson’s
version of the language.

George R. R. Martin1 sold his first short story in 1970, and his first novel in 1977. Most of his
early work was science fiction; he has also written horror and superhero fiction, written
screenplays, and edited a number of anthologies. Among the best of his early works are the short
works «Sandkings» and «A Song for Lya» and the novel Fevre Dream. He started writing epic
fantasy with A Game of Thrones, published in 1996; the fourth book was published in 2005, and
he is currently working on the fifth.2 The language of the culture to which most of the viewpoint
characters belong is represented in the books by English, with most of the characters from this
culture being given European or Europeanesque names; the Dothraki, a nomadic horse-riding
people with some similarities to the Mongols of our world, are presented as foreigners. Martin is
no linguist or polyglot, however, and has said that he had not fleshed out the Dothraki language
beyond the few proper names, words and phrases used in the books.3 The producers of the
television series based on the books wish to have the Dothraki characters speaking their own
language on-screen, with English subtitles; this required the services of an expert language

David J. Peterson studied linguistics at UC Berkeley (B.A., English and linguistics) and UC San
Diego (M.A., linguistics). He started creating his own constructed languages in 2000, beginning
with a personal language based on Esperanto and Arabic;4 he has been active in the conlang
community since early 2001, including volunteer work for the Language Creation Society, and
he’s one of the editors of Speculative Grammarian.5 His website showcases thirteen of his
conlangs (some more developed than others); his main project, Kamakawi, is influenced

Interview with David J. Peterson and Sai – Jim Henry – 2

primarily by Hawai’ian (though more in its phonology than its grammar). Most of the linguistic
details of his instantiation of Dothraki remain confidential until nearer the time the TV series is
to be aired, but what’s been released so far suggests that it’s an a priori, fusional and synthetic
language with influences from Russian, Turkish, Estonian, Inuktitut and Swahili.6

A pilot episode with a certain amount of Dothraki dialogue has already been produced. Peterson
will be translating dialogue for individual episodes as their screenplays are written, and
supplying sound recordings for the actors to listen to as well as transcriptions in his romanized
Dothraki orthography. Peterson, Martin, and HBO are discussing possible use of the expanded
Dothraki for future books in the series.

I interviewed David J. Peterson, and Sai Emrys, president of the Language Creation Society, by
email in May 2010.

David, how did you first get interested in languages and linguistics?

Kind of an odd story, really. One morning during my junior year of high school I woke up
horrified that there were millions upon millions of people on Earth who could speak
French and I wasn’t one of them. That day, I made it my rather naïve goal to learn every
language on the planet.

During my senior year of high school, I was determined to take advantage of the meager
language learning opportunities my school offered. I was already taking Spanish, so I
added to it German 1, and wanted to add French 2 (French 1 wouldn’t fit my schedule),
but the teacher wouldn’t let me in.

At Berkeley, where I went to college, I was determined to keep up with learning
languages, and took Arabic, Russian and Esperanto in my first year. In my second
semester, on my mother’s urging, I also took an introductory linguistics class, albeit
reluctantly (I didn’t see the point of linguistics, since you didn’t spend any time actually
learning languages). It’s a good thing I did, though, because I ended up liking it quite a
bit, and stuck with it for the next six or seven years.

When were you at Berkeley and UCSD?

I was at Berkeley from the fall of 1999 to the spring of 2003. I then moved to San Diego
and was at UCSD from the fall of 2003 until the fall of 2006.

What languages and areas of linguistics interested you most? What did you do your own
research on?

I’d always been interested in French, but studying one language, for me, invariably leads
to others. The languages that still interest me a good deal are Hawai’ian (probably my
favorite), Arabic and, well, all the rest. Those two I think are tops for me.

Coming from Berkeley, my two biggest areas of interest were phonetics and historical
linguistics (as well as pidgin and creole studies, which, in a way, is quite similar to

Interview with David J. Peterson and Sai – Jim Henry – 3

historical linguistics). I was extremely fortunate to take classes from the professors I did
— John Ohala’s class in phonetics is certainly one I won’t ever forget.

At UCSD, I discovered my real passion within linguistics was morphology. The program
there has graduate students doing two research projects in two different areas, and my
intention was to go with my strength first (phonetics), and then to spend the rest of my
time working in morphology. Things didn’t end up working out that way, though.

Taking so many language courses at different schools, and studying some other languages on
your own, I suppose you’ve encountered several different language teaching methodologies.
Which did you find most effective for you?

Different folks learn in different ways, of course, but I’ve found there seems to be a trend
in a lot of language courses to focus on culture. I don’t find that very useful — especially
when it’s done mostly in English. The best courses I took were all done exclusively in the
target language. There was a lot of fluent language input to hear and emulate. In addition,
though, the assignments tended to be what many would consider «boring»: learning a
particular paradigm, and then replicating it in writing or in speech (or sign). When it
comes to learning, for me, actually using a form helps imprint it on my memory the best.

Oh, and kind of a stray note. For many of the courses I took (Arabic, Russian and
Esperanto, for example), there were a series of instructional videos we watched in class.
They were usually silly, and most of the time we weren’t learning something new from
them, but what I liked about them was it gives one an opportunity to hear a different
fluent speaker using the language. One can quickly get used to the accent and manner of
speaking of one’s instructor, so mixing it gives one greater insight into how the language
can be spoken.

Who were the most interesting professors you had?

I don’t want to leave anyone out, certainly, but at Berkeley, John Ohala’s course in
phonetics was phenomenal. In addition I count myself as fortunate that I was able to take
a course in historical linguistics, and to take it from a truly outstanding professor, Andrew
Garrett. Another memorable class was John McWhorter’s class on pidgins and creoles —
a class that, unfortunately, may have been foreshortened a bit, due to its taking place in
the fall of 2001 (the story of Berkeley during 9/11 is a fascinating one, but not one for

At UCSD, I came to know a number of world class linguists, including the legendary
David Perlmutter. It’s hard to quantify any of this, really; I made so many friends there:
David, Sanford Schane, Grant Goodall (also an Esperantist), Masha Polinsky… Regarding
the way I look at language, though, the most influential professor I met was Farrell
Ackerman. It might not have seemed so at the time, but taking his course in morphology
was a life-changing event for me.

How did you first get interested in constructed languages?

Interview with David J. Peterson and Sai – Jim Henry – 4

It was the convergence of three classes I was taking at the time: Esperanto, Arabic and
Introductory Linguistics. Due in large part to the influence of Esperanto and all the
natural language data I was getting in linguistics, I hit upon the idea that someone could
create a language not for international communication (Esperanto, I reasoned, fit the bill
for that), but just for personal use — say, between friends. As soon as the idea hit me (it
was in a linguistics class), I immediately started designing the phonology for my first
language: a poor project with the regularity of Esperanto and the structure of Arabic. I
would later discover that I wasn’t the first to have this idea.

Who taught that Esperanto course at Berkeley? What was it like?

That was actually the first time that particular Esperanto DeCAL course was taught at
Berkeley, and it was co-taught by Lana Shlafer and Mina (or Stelet) Kim. I thought it was
a unique experience since both Lana and Mina had been speaking Esperanto since birth
(they were raised bilingual in Esperanto and their native language [Russian and Korean,
respectively]). It’s one thing to take a course in a constructed language, and quite another
to take it from two native speakers.

I still remember everything I learned. All I would need to communicate comfortably in
Esperanto is more vocabulary (well, that and I’d probably avoid using any participles).
When it comes to a language instruction course, I can’t think of higher praise to offer —
though, of course, it was also a lot of fun. I still have my USEJ sticker on my guitar case.

What, if anything, did you already know about Esperanto before taking that course? What
motivated you to take Esperanto just then rather than one of the other language courses offered
that semester?

My mother told me about Esperanto when I was a senior in high school (though I still
can’t figure out how it ever came to her attention). That was the first time I’d ever
encountered the idea that a human could invent a language. It seemed pretty outlandish at
the time.

Several months later, we were at a library book sale and I came across Step by Step in
Esperanto by Montagu C. Butler (oh, and looking at it now, I see I wrote the course
control number and course entry code for that Esperanto class on the inside cover). I
looked that over, and took it with me to Berkeley.

As for the course, language classes at Berkeley are five units. That semester I was taking
Arabic and Russian, in addition to a class for my English major and introductory
linguistics. All told, that comes to 18 units — a hefty load for a freshman (average is 12;
heavy is 15). I wasn’t planning on taking any other classes that semester, but during the
first week of classes I returned to my dorm and saw a flier for the Esperanto DeCAL
posted on the front door. I had a good experience with the previous DeCAL I’d taken (on
the Doors and Jim Morrison’s poetry), and figured I already knew a bit about Esperanto,
so tacking on another class wouldn’t be too much trouble.

Interview with David J. Peterson and Sai – Jim Henry – 5

Which of your own conlangs is the best developed? Which is your favorite?

Same answer for both questions: Kamakawi.7 Kamakawi was the first language I created
after I realized that my previous languages suffered from serious flaws (beginners’
mistakes; everyone makes them). Even so, the language wasn’t perfect to begin with, but
I’ve stuck with it over the years, and have really gotten it to a good place.

You talk on your website about a couple of pidginization experiments you did while at Berkeley,
taking a simple conlang lexicon with no overt grammar and seeing what the subjects did to turn
it into a language.8 What influence did those experiments, or your experience with collaborative
projects like Kalusa,9 have on your subsequent conlanging?

First, I have to say that I’m extremely grateful to John McWhorter for letting me run a
crazy, ill-conceived experiment in lieu of an straight-laced research paper on a natural
pidgin or creole (which is what we were supposed to do for that course).

The experiment itself is simple: Take a list of words (a few pronouns, a lot of nouns and
verbs), give them to a group of people, and have those people speak using only that
wordlist. It’d be up to them to concoct a grammar and, hopefully, an actual language.

The project in John McWhorter’s class led to the Language Birth DeCAL course I
coordinated in the fall of 2002. The initial experiment, I realized, was poorly designed,
and there simply wasn’t enough time to execute it properly — in addition to the fact that
there was no motivation for the participants to actually create a language from the
wordlist I provided them (aside from the fact that they were my friends). I reasoned that
if students were there for a grade, that would be motivation enough.

That, unfortunately, turned out not to be true. The students in the course never actually
memorized the list of words (a prerequisite for doing the experiment correctly), and the
experiment was pretty much a failure. We did get to be on local Berkeley TV, though
(that was fun).

Despite the lack of overall success, the students did try to use the language as best they
could. As an instructor and conlanger, though, I noticed that they simply weren’t doing
what I would do. I had a ton of ideas about how to create a full-fledged language from the
wordlist I’d given them. After the course was over, then, I decided to go ahead and do it. I
revised the wordlist and created a language called Kelenala,10 which is my attempt to
create a language from a fixed wordlist. I’ve had a lot of fun with it over the years.

From all this (and from studying pidgins and creoles), I realized something about
language. First, I think language users tend to work with what they’ve got before creating
something new. If the meaning of word x can be extended to word y, why derive a new
word? And if word z can be derived via regular derivational process from word y, why
create a new phonological form?

Second, it became evident that most of what we see in language after language is not due
to any kind of biological restriction, but simply due to either historical accident or simple

Interview with David J. Peterson and Sai – Jim Henry – 6

common sense (or convenience). For example, no language distinguishes twelve numbers
(singular, dual, trial…duodecimal). This isn’t because our brains can’t handle it, but simply
because after a certain point, it simply becomes inconvenient (it’s more taxing to
remember those forms than it is useful).

On the other hand, if such a thing happens to be useful in a given language, it may very
well arise. For example, in ASL, a number can be incorporated into signs like WEEK and
YEAR. To say «two boys», you use the sign for TWO and the sign for BOY, but with
WEEK and YEAR, you modify the handshape to reflect the number desired and can say,
in one sign, «two weeks», «three weeks», «nine weeks» — any number that can be
represented by a single handshape.

Similarly, though one never expects to find a natural language with a singular, trial and
plural number (i.e. with no dual), learning such a language wouldn’t be impossible.
Coming across one in the wild, though, is beyond improbable not because of how human
brains work, but simply because it seems bizarre to human sensibilities to have a special
way to say one of something and three of something without having a special way to say
two of something.

All of this (if I can round up everything I said in the previous paragraphs) fundamentally
changed the way I conlang. Essentially, I changed my approach. Rather than trying to
emulate the end product I was familiar with in linguistics (a reference grammar), I began
trying to emulate the way humans actually evolve language. Since then, I’ve found the
results of my work much more satisfying on both an artistic and a technical level.

How did you get involved in the CONLANG mailing list11 and the conlanger community more

I had been working on my own languages for the better part of a semester and a summer
when I came back to Berkeley in the fall. Though I may have unconsciously avoided it
before, I decided to search the internet to see if I really was the only one doing this (to see
if I had a «corner on the market», as my grandfather put it). Pretty soon I hit upon David
Pablo Flores’ conlang page, and from there, found several others, and before long I found
the Conlang listserv. As Sally Caves noted at the First Language Creation Conference,
that moment of discovery when one discovers the larger language creation community is
both exhilarating and deflating. I think the deflation doesn’t last, though — especially
when one happens upon a really inviting, friendly and knowledgeable community like the
Conlang-L. It took me a little while, but I soon realized that I wasn’t among competitors,
but among friends.

You’ve mentioned before that you’re natively bilingual in English and Spanish. Do you reckon
that infuenced your interest in languages and linguistics, or in conlanging?

In the long run, perhaps, but initially, quite the opposite. It’s very difficult to raise a child
to be truly bilingual; usually one language predominates over the other. In my case, that
was English. I was doing pretty well with Spanish, but then I was moved away to Fresno,
away from my mother’s side of the family. There I had pretty much no regular contact

Interview with David J. Peterson and Sai – Jim Henry – 7

with Spanish speakers any longer, and English took over. By the time I moved back, I’d
lost too much time, and Spanish became a source of frustration for me. I tried to block it
out, ignore it — and the thought of learning languages other than Spanish was abhorrent
to me. That was a serious roadblock. I had to reclaim my fluency and tap into my native
knowledge (and for that, I have not only my Spanish teachers to thank, but my
grandmother. I’ve always understood her Spanish, and she never gave up speaking it to

Ellen Wright asked you a number of questions about the phonology and grammar of your
Dothraki in her interview for; I won’t go over the same ground again. What can you tell
me about Dothraki semantics and pragmatics? Are there pervasive structuring metaphors that
differ interestingly from those of western European languages, for instance, or different rules for
when things need to be explicitly mentioned and when they can be silently implied?

I tried at all times to adhere to the ideals espoused by the Dothraki in George R.R.
Martin’s books. There’s a kind of Dothraki mindset that one gleans from the books.
They’re warriors, of course, but there’s also a series of dualities. For example, the
Dothraki are nomadic, yet they have one fixed city (Vaes Dothrak), and there’s a certain
code of behavior that applies to being in the city, and a different one for being elsewhere.
I tried to bring these different senses of place into the language where possible.

For example, in Dothraki something is in done in secret is done «torga essheyi» —
literally, «under a roof». The Dothraki believe that all honorable acts are undertaken
beneath the open sky, so to do something under a roof — out of the light of day — is to
do it secretly, and in a way which doesn’t merit approval.

Joshua Hartshorne recently wrote on the Scientific American weblog12 suggesting that you use
your position as Dothraki creator to test one or more apparent language universals — put in at
least one violation of a universal that’s common to existing natural human languages, but looks
as though it might be a historical accident rather than something inherent to the way our brains
work, and see if fans of the show are able to learn the language fluently anyway. Do you think
the experiment would be valid or meaningful, assuming it’s practical and ethical? Do you think
making such an experiment with Dothraki, rather than some other conlang, would be a conflict
of interest in business or artistic terms?

First, I think it’s difficult to create a language without violating at least one language
universal. With almost 3,000 of them on record, it would seem difficult not to. For me,
the question of whether or not language universals are valid is not as interesting as it is to
linguists. I hold the belief that the human mind is pretty flexible and can learn just about
anything you throw at it. We’ll have to wait and see if there’s something of interest in
Dothraki to be tested, but we are having an ongoing conversation with Joshua
Hartshorne. I imagine that things will become a bit clearer as we get nearer and nearer the
series premier.

As for using a conlang for such an experiment, I would have no problem with it, provided
doing so wouldn’t violate the language’s design principles. In my case, if I ever had to

Interview with David J. Peterson and Sai – Jim Henry – 8

decide between implementing something that might be useful in an experiment and
remaining faithful to George R.R. Martin’s vision and/or HBO’s interests, there’s really no
contest. I was hired to do a job, and that comes first.

Is the form of the Dothraki word «ido» (wooden, fake) an inside joke vis-a-vis the Esperanto/Ido

Hee, hee… You caught me! I would like to say, though, I don’t wish the Ido community
ill, by any means. Think of it as kind of good-natured ribbing — the way one might tease
someone from a rival school. But make no mistake: If I were choosing sides in an
Esperanto-Ido athletic competition, I’d proudly wear a green star sweater.

How would you say «ne al glavo sangon soifanta» (not to a sword thirsting for blood) in

Man, what a line to choose! You’re really making me work for this. I’m even going to
have to dust off the seldom-used present participle… Nice not to have to coin any words,

Vos arakhaan feveya qoyoon…

Dang, 11 syllables… Of course, if this were in verse, I’m sure the «vos» could be clipped
to «v'». That’d give us 10. Anyway, I would note here that this is the translation of a line
of verse, and that the phrasing, as a result, is unusual (if I had the full sentence with the
verb, I think I’d do it a different way, but without the verb, this is the best way).

Sai, you were the founder, or one of the founding members of the Language Creation Society,
right? How did that happen?

When I was at UC Berkeley (2004-2005), I created and taught a DE-Cal class on how to
create your own language.13 It was about half the scope of Linguistics 101; we went over
all the usual concepts but a bit more lightly and just in passing for the language creation.

The second year, I decided that I wanted to try to get reimbursed for what I was spending
out of pocket on MiniDV tapes (most of the video is online14). When I looked into the
paperwork of getting a special-purpose grant from the ASUC,15 our student association,
and just making a full-on student club, I noticed that the latter was not much more work
and would give me access to significantly more funding.

And that made me think: «Hey, nobody’s run a conference on conlanging since the late
1800s’ Volapük conferences. I could totally do this.» So I did.16

The first LCC [Language Creation Conference] was basically just me running everything
(though Millicent Chaney, the ASUC liaison, deserves kudos for her help). I managed to
pull it off in 4 months to what I think was a fairly strong success as an event. I even got
last-minute sponsorship by the CogSci department (thanks, Prof. Kihlstrom!) and
managed to turn a small profit. I wrote it up afterwards17 to help others learn from it.

Interview with David J. Peterson and Sai – Jim Henry – 9

That was my last year at Cal, though, and to continue the conference (and to be able to
use that profit) I realized that it’d have to go independent. I contacted several conlang
community members I knew to be well respected, and put out an open call to see who
might be interested that I didn’t know yet.

A few months later, I had a Board of Directors for the new non-profit corporation. We
filed for incorporation using some of the LCC1 profits, split out roles in the new
organization, filed for tax exemption, etc. Lots of boring paperwork, but that kind of
organization is something I have a knack for, so it wasn’t too hard.

Since then, I’ve divested even more of the day-to-day and philosophical control away
from myself. Board members pretty regularly disagree with me — something I encourage
— and we figure out how to do things in a well-balanced manner that addresses all
concerns but still gets it done at the end of the day.

LCC2 was co-run by Alex Fink (and the close working friendship we established through
that started a rather more intimate relationship); LCC3 by David Durand. LCC4 is still on
the way, but will likewise be a collaborative effort.

David Peterson, Sylvia Sotomayor, and Don Boozer have done a tremendous amount, and
as time goes on I expect this to continue to the point where someone else could easily
take over leadership. Frankly, I’m entering graduate school, so I’m soon going to have a
lot less time. And I could use a break. So other people will take over, the organization
overall will go on.

Practically speaking, the LCS — like any small non-profit — is a doöcracy more than any
kind of formal system. If you have an idea for what you’d like to see happen, and are
willing to actually get it done, then we’d all be thrilled to have you do it, take the credit,
and take primary control.

If I’m not mistaken, the LCS is the first organization that’s been devoted to constructed languages
in general

To my knowledge, this is correct. And it’s why I wanted to establish it.

Why do you think it took so long for something like the LCS to come into being, given that
numerous conlangers and relatively cheap communication technology have both been around for

I don’t know, honestly. I think most people simply don’t think to do it, like with many
other things.

And even if you do, you have to have some idea of what to do with it — like in my case,
with the class and LCC driving the creation of the LCS (both the UCB student club and
the non-profit corporation).

Interview with David J. Peterson and Sai – Jim Henry – 10

Just as a group, we already have multiple great collaboration fora (CONLANG-L, ZBB,
etc), and I think that fulfills most of what you want to get from an organization. All the
LCS does is add a few more things — more formal fora (conference, podcast,
publishing), a way to do things that require financial collaboration (membership fees),
and a single, helpful point of contact for the outside world.

How did this job of finding a conlanger for the A Game of Thrones television series come about?

Dan Weiss18 contacted Arika Okrent because he’d heard of her book In the Land of
Invented Languages.19. She sent him to me, as she knew me from her research on the
book, having attended LCC2, and some conversations we’d had since then.

By the way, I heartily recommend Okrent’s book. Even for someone who already knows a
decent amount about the history of conlanging, it’s extremely entertaining and
informative. It focuses primarily (by dint of historical preponderance) on conlanging
before the modern artlang-focused20 community of language creators, but has a chapter
on that too.

How did you publish the call for conlangers and filter the applicants, given HBO’s desire to limit
the publicity about the project until it was greenlit?

I think it was clear right off that we didn’t want to be like everyone else who’d done this,
and just hand the job to someone. We wanted to make sure that everyone in the
community had a fair chance at it, and that HBO got top-quality material. Likewise, we
didn’t want HBO to have to deal with trying to figure out what was linguistically well-
done, as it’s simply not their area of expertise; it’s what they hired us for.

So we ran it as a competitive, reviewed job application.

As for NDA: I ran everything by their lawyer. Basically we were constrained from saying
anything publicly that would identify our client or the project. So the call for applications
was a bit vague about the details, just giving the main points that someone would want to
know to see if they’d be interested or not. If someone was interested (and could
demonstrate conlanging/linguistics ability), we had them sign an NDA, and then gave ‘em
all the information they needed.

How did you select the reviewers who were to read the proposals and select some of the best to
send to the show’s producers? Did you look for a certain range of expertise among them, e.g.
people knowledgeable about different areas of linguistics or creators of different types of

That was a combination of knowing people who’d make good reviewers, getting
recommendations, a couple self-nominations, and a couple erstwhile applicants who
decided not to go through with it themselves but who wanted to help out reviewing.

Interview with David J. Peterson and Sai – Jim Henry – 11

We definitely wanted to have a diversity of opinion. Reviewers covered the spectrum of
artlang tendencies, plus a couple who were more auxlang- or engelang-leaning, as well as
a fair diversity of formal linguistics background.

In assigning reviewers for applications, I made sure that each application got reviewed by
a good cross-section of different kinds of reviewer.

The conlang community is pretty small; a lot of the reviewers and applicants know each other
pretty well online, and some know each other in person, through the Language Creation
Conference or elsewhere. What kind of safeguards did you use to avoid bias on the part of the
reviewers vis-a-vis applicants they knew, or with respect to David being a member of the LCS
board of directors?

All applications were pseudonymized; all the reviewers saw was a set of documents
called «Hyacinth» or the like. Likewise, in giving reviewers’ feedback to applicants, they
were given pseudonyms (which still haven’t been unblinded, to ensure reviewer
sincerity). Reviewers didn’t see the résumés or cover letters, just the linguistic documents.

I also asked all reviewers to tell me which applicants they knew IRL, whose voice they
could recognize, whom they had any relationship with (positive or negative). Anyone
listed was completely excluded from being a reviewership pairing.

I was the only person who knew which person corresponded with which pseudonym, and
I tried to make sure that documents didn’t have lingering traces of authorship in metadata
or the like. I did not participate in reviewing at all other than as an organizer; both
because I was not blinded and because, frankly, the other reviewers generally had more
artlanging expertise than me. (I’m primarily an engelanger.)

David was, like any applicant, locked out of all discussion between reviewers and
anything else that might compromise double-blinding.

When I asked reviewers after they made the final decision and unblinded it, a lot were
surprised by some of the results. The only leak point was that some people have
distinctive artlanging «signatures» — like one might recognize a Matisse by the
brushwork alone — but only a small number of reviewers recognized a small number of

I think we did about as good a job making it double-blinded and fair as could possibly be

Incidentally, I asked reviewers to rate the applicants by name as to how well-known or
good they thought the applicant was a priori (before they saw any samples). The result of
that was only weakly correlated (0.26) with the average scores applicants actually got. So
I think it definitely wasn’t nepotistic.

About how many reviewers studied each proposal?

Interview with David J. Peterson and Sai – Jim Henry – 12

Overall, at least 3; on average 3.5. Each of the 9 reviewers did on average 10 reviews.
(Lots of work!)

Plus, everything that was in the top tranche — i.e. likely finalists — got a second pass by
every reviewer when we had a live chat about how to decide on finalists.

What criteria did the reviewers use to judge the proposed instantiations of Dothraki?

Copying straight from the reviewer rubric:



fidelity to canon (or, goodness of reasons for deviating from it)
aesthetics, appropriateness to and expressiveness of the Dothraki culture, etc.,
especially for phonaesthetics and culturally-specific phrases / sayings / etc.
linguistic cohesiveness, ingenuity, depth, naturalism, verisimilitude, etc.
thoroughness, clarity, and overall quality of work submitted
fudge factor

Each was rated 1-10, with 5 being unmarkedly neutral. Reviewers had flexibility for how
to interpret things; the way I averaged the scores (using each reviewer’s z-scores)
corrected for tendencies of one reviewer to be overall harsher or more generous than

Did you see broad similarities or trends among the applications, over and above the minimum
similarity necessitated by consistency with the words and phrases in George R.R. Martin’s

Not really. They all had the needed words and were naturalistic artlangs (because that was
part of the spec), but other than that, they were all over the map — inspiration languages,
documentation style, linguistic decisions, grammar, phonetics, everything.

What other projects would you like to see the LCS involved in in the future?

Well, the journal is still nascent; I’d like to see it really get off the ground. I’d like to see
the LCC — hopefully the next one — be held outside the US for the first time.

I think we could do more to provide a good introduction to newbies and outsiders. There’s
the Language Creation Kit21 (and now the Language Creation Kit book!22), Don’s
Conlangers’ Library,23 etc., but they’re really more aimed at people who already have
some idea of what conlanging is all about and have some connection with our

And I’d like to see people just come to us and say «I’ve got this idea for something we
could do that’d be great, and I’m willing to make it happen». We’re happy to give people
any resources we have (hosting, naming, advice, social connections, sometimes money) if
they can take point on a project.

Interview with David J. Peterson and Sai – Jim Henry – 13

Ultimately, like I said, that’s what it always comes down to. There’s only so much any
individual can do. And this is why we have an organization. 😉

The last couple of Language Creation Conferences had around 30-40 attendees, if I recall
correctly. That’s not much smaller than the typical attendance for the national convention of
Esperanto USA, and about the same as some of the annual or every-few-years regional
Esperanto conventions. Do you think it would make sense to combine the LCC and one of these
conventions in some year, to share facilities, reduce overhead costs, and exchange ideas between
the communities?

Potentially, sure. I’d be open to it at least. And it certainly would have financial and
organizational benefits.

My only concern with this would be the distinction that the LCS makes between
conlanging and conlangs. We are non-partisan; we cannot advocate one language over
another. As you mentioned earlier, we’re more about the craft itself than particular
speaker communities (though to the extent that we can support them neutrally, we’re
happy to do so).

So for instance, I think talks about Esperanto (as about any other conlang, auxiliary or
otherwise) would be perfectly welcome at an LCC — but proselytization would definitely
not. We want to give our audience new perspectives, new ideas, new tools for
approaching whatever they want to do in their own conlanging, not to try to convince
them to get behind someone else’s project.

I would also wonder how much interest there is. People who make languages, and people
who learn languages for use, have in my experience tended to have a fairly different
perspective and reasons for interest. That could result in a neat cross-community
collaboration, or it could leave both sides bored with the other half; I just don’t know
which would be the result.

I think it’s a pragmatic question that’s worth exploring.

Indeed; but I suspect the overlap between the groups (people who make conlangs, people who
enjoy learning others’ conlangs, people who are interested in learning about others’ conlangs if
not in learning them) is larger than generally realized. A significant subset of the Esperanto
speakers I know are either conlangers themselves, or interested in other conlangs besides
Esperanto; that could be sample bias on my part, though.

Like I said, I think it’s an empirical question that we don’t really know the answer to yet.
We can always try it and find out, or poll our respective audiences.



Interview with David J. Peterson and Sai – Jim Henry – 14


3. «Klaatu Barada Nicto…» Weblog post on David Peterson’s Dothraki.



6. «Creating Dothraki: An Interview with David J. Peterson and Sai Emrys» by Ellen B. Wright.






12. «Fantasy TV in the service of science: An open letter to HBO about ‘Dothraki'» by Joshua
See also: «The Dothraki response to a call for science in a created language» by Sai Emrys and
David J. Peterson.






18. One of the executive producers of «A Game of Thrones», and the writer of the pilot episode.


Interview with David J. Peterson and Sai – Jim Henry – 15

20. artlang: artistic language, part of the commonly accepted typology of conlangs. Constrast
«auxlang», international auxiliary language (e.g. Esperanto) and «engelang», engineered language
(e.g. Lojban).

21. By Mark Rosenfelder.


23. http://library.conlang.orgAn Interview with David J. Peterson and Sai Emrys image

Descargar PDF

(Visitado 1 veces, 1 visitas hoy)