David Peterson: “The Art of Language Invention” | Talks at Google


JEREMY: Just say that we’re
delighted to welcome back to Google, David Peterson,
linguist, language creator. And his new book is “The
Art of Language Invention.” [APPLAUSE] DAVID PETERSON: Grab this. All right, all right. By the way, Jeremy, I just
want to mention beforehand, you have a great one of
those teacher stares, the “I can wait just
as long as you.” Wonderful. All right. So anyway, thank you. Thank you all for coming out. I’m delighted to be here again. This is my first
time in this room, and I absolutely love the
color of the chairs, which of course, nobody watching
this on YouTube can see. But they’re randomly colorful. I would definitely be
sitting in an orange chair. Anyway, so as we’re
getting started, let me get out of the way
my fancy little animation. Check this out. Ooh, sparkly. All right, so that’s
the title of the book. Anyway, so yeah, my
name is David Peterson, as he mentioned. This is a picture of me. That’s my mother buttoning
my top button there the morning of my wedding
because I couldn’t do it. Anyway, so I’ve created
a number of languages for television and film. Some of them, you’ll know. So for example, for
“Game of Thrones,” I created the Dothraki
and Valyrian languages. And this bottom one is Mag Nuk. That was a language I created
for the giant for last season. I kind of thought that
he was going to have more lines than just one. Probably didn’t merit an
entire language, but you know. Hey, now there’s a possibility
for a great big giant monologue sometime in the future. I’d look forward to it. Anyway, so I’ve also
created languages for “Defiance,” specifically
these languages. And I got the
opportunity to create these fancy scripts for them. I love that I was
able to do that. Creating scripts is one
of my favorite things to do as a language creator. And you can’t always
convince the production that it’s worthwhile,
but sometimes, you can. So “Defiance,” and then I
did that for “Star Crossed” as well, where I created a
language and a writing system. This was a show that
was on the CW last year and had a one season run. After that, I created a language
for Thor’s “The Dark World”– it was the sequel to
Thor– called Shivaisith. And I did create a
writing system for it. But this was actually–
it’s a bizarre– it is a fan-created writing system. So I created it for a
fan of the language. It wasn’t actually
used in the film. But there’s basically one
real big fan of this language, and I thought, well, all right. I also created a language
for “Dominion” on Syfy called Lishepus. It’s the one that the angels
and the Eight-Balls speak, if you watch that show. I created a language for the
second season on for “The 100.” The language that
the Grounders speak, it’s called Trigedasleng. And then for “Penny Dreadful,”
which is an amazing show, I created a language
for season two. It’s basically the
demonic language that you hear Vanessa and the
witches speaking periodically. And it’s pretty cool. And I’m also happy
to announce that I’m working on a new show
called “Emerald City” that’s going to be
on NBC next year. And all the episodes
are directed by Tarsem, which is just crazy. He did “The Fall.” He’s so good. I’m really excited about it. And so the two languages
that I’ve created are called Munja’kin and Inha. You can look forward to that. But I am not here
today to talk to you about any of those
languages at all, any of those fancy languages. I’m here to talk to you
about this language, Megdevi. How many people have
heard of this language? Are you serious? Wow. I really did not expect
to see any hands. All right, so this is the
first language I ever created. I created it as kind of
a freshman, sophomore, over-the-summer type of
thing at UC Berkeley. And if you’re wondering why
the language is called this, I had a girlfriend at
the time named Megan. My name is David. And then you’ve seen,
fantasy languages sometimes have names like this–
Cheysuli, Dothraki, Ferengi, Pakuni. And so that was how
you got the name of that embarrassing language. So this was my very
first take at it. I had never heard
of anybody creating a language for any
purpose other than international communication
prior to doing this. So I truly believed I was the
first person ever to do it. And in fact, at the
time, I thought, oh, wow. I could create a language
for myself and that maybe me and my girlfriend could speak. And maybe there’s actually
marketability in this since nobody’s ever thought
of this idea before. This is what I believed. This is 2000, bear in mind. It’s just that conlanging wasn’t
very much in the public eye. Anyway, as for how
this language worked, I need to take a sidetrack
and introduce you to a little bit of Arabic. I’m sure there are some
people here who know Arabic, but for those that don’t, I took
a year of Arabic at UC Berkeley and absolutely fell in
love with the language. It’s a wonderful language. I never imagined a language
that, grammatically, could work in that way. And one of the unique
features of Arabic is that it has something
called a root system, usually a triconsonantal
root system, where you have something like these
three consonants– K, T, and B. And by the way, Arabic is
written from right to left. So notice how each
consonant matches up with the orthographic
form there. And what happens is that
this root has something to do with books and writing. But the way that
it is transformed into words in the language
is that you actually arrange vowels inside and without
these consonants, and also suffixes and prefixes, to
form different types of words. And the only thing
that remains constant is the linear order of
these three consonants. So for example, these are all
verbs in different tenses. At the top, we have “naktub,”
which means “we write.” Then on the lower left, we have
“aktub,” which means “I write.” And then in the past, “katabat,”
which means “she wrote.” And notice that the only
thing that’s happening here is K, T, and B are always
in the exact same order. Sometimes, they’re right
next to each other. Sometimes, there are
vowels in between. And you can make
even more words. So for example, “kaatib,”
“writer.” “Maktab,” “office.” And “kitaab,” “book,”
just like that. So I thought that was a really,
really cool way for a language to work. And I never heard of it before. And so basically,
when I was creating my own language, Megdevi,
I just copied it wholesale. And I also created
this terrible font, which I included
just so that you could see how terrible it is. Actually, when you’re
standing that far back, it actually looks
better than it is. But there are just so
many little shaky lines and everything, just god awful. Anyway, so that you could
have a root like this– ja, su, fe– and have
jaesif– these are real words from the
language– “flame thrower.” Jisfe, “flame-colored.” And jasif, “flame,” like that. Hm, flame thrower, whatever. And you can also see that
with the writing system, I was heavily
inspired by Arabic. So if you were to
spell something that would be roughly
equivalent to the pronunciation of “writer” in Arabic, “kaatib,”
it would look like this. And so you can see the K,
it’s got that thing on it. And it just has a little
bubble on the top. And the B has one dot
except it’s above. And there’s another
little bubble. And the T looks more T-like. It was clearly just derivative. The worst part about it
though, was, in my mind, the way that I
understood language to work, and specifically,
how I created words for this language. So grammatically,
it’s a neat system. But I just want you to look
at some of these [INAUDIBLE]. In order to be able
to appreciate this, you have to know these
three patterns right here. So the 1, 2, and 3 are for the
consonants that will change root to root. And then pay attention
to the vowels on top. If it has an “ae” as the
first vowel and an “i” as the second vowel, that’s
more or less a thing. If it has an “ay” as the first
vowel and an “i” as the second, that’s a person. And then if it has an
“ee” as the first vowel and “eh” as the second
vowel, that is a place. All right, so our
first word, there wasn’t a place one for it,
but “jaelif” is a sailboat, “jelif” is “sailor.” All right, fine. That makes more or less sense. Next, W, V, D.
“Wived” is “marina. “Wevid” is “mariner.” And what the difference is
between “sailor” and “mariner,” hm, I’m not sure there. We know what the
difference is in English. Basically, they came from
two different languages, and that’s why we have them. But yeah, so now we have both
of these words in this language. So I’m like, all right,
all right, fine, whatever. Then we have “thaenis,” which
was vessel, specifically, an aquatic vessel. So now we have a word for
“sailboat” and “vessel.” And “thenis,”
which is “captain.” Next, “boat,” which is
different from “sailboat,” which is different from “vessel.” That was “chaesich.” And we also have “chisich,”
“boathouse,” which is different from “marina.” And “chesich,” “boater,” which
is different from “sailor,” which is different
from “captain,” which is different from “mariner.” Then it goes even further. We have a very special root
that’s just for clipper ships, “baelil.” Presumably, “belil”
is somebody who sails a clipper ship or something. But I think that the
reason that we even have this word for clipper
ship is just because I was a fan of the Clippers. And so why not? I wonder, actually, if that
B-L-L root has something to do with the Clippers. All I’m thinking
of is Bison Dele. Somebody remember him? Anybody remember
his original name? Because I literally don’t. All right, anyway, moving on. Anyway, so there are a
bunch of other crazy things about this language. For example, I had
taken Esperanto. So the case system for
this language was identical to Esperanto’s. The number suffixes. So looking at number systems
for natural languages is a fascinating thing. But this language literally
had separate suffixes for number words that were in
the tens, hundreds, thousands, hundred thousands, millions,
billions, trillions, up to the quadrillions. So that you could take
the root for 1 and make 1, 10, 100, 1,000, 10,000,
so on, all the way up to 1 quadrillion. I don’t think that there
are other languages that do things like that. There is a special
word pattern just for colors, which is weird. I kind of like it. The tenses were
identical to Esperanto’s. It had this correlative
system that was, again, identical to Esperanto’s. All of it was purely lifted. And in the beginning,
when I’d just started out, I actually started to translate
Shakespeare’s “The Tempest” into this language. And it just keeps going like
that, which sounds impressive until you know how palely
derivative the language is. In fact, I get
this question a lot about, oh, wow, how do you
create a full language that’s functional? That’s actually not hard. It’s not hard to create a
language that is completely and fully functional, can
do everything you want. It’s fairly easy if
you’re going to copy a lot of things wholesale
from other languages and just basically recode them. Then all it takes is the simple
matter of just coding it over. It’s like if you’re
talking about drawing a human, it’s like, wow, it
must be tough to draw a human. It’s not hard. There you go. It’s a human. We all recognize it’s
basically a human. I don’t know. He’s rubbing his belly while
he’s pointing, I guess. But of course, what
you mean when you say, oh, it’s hard to
draw a human, you mean it’s hard to draw a
human that looks realistic. It’s easy enough
to draw something that’s basically a human. Kids get it by the time
they’re three or four. Same is true of a language. Easy to create a fully
functional language. Difficult to create
a language that actually looks like
a real language that we speak here on Earth. And so for me, the
real watershed moment was when I found other
language creators online, the early conlang community,
the conlang listserv. This is an early photo. And I’m sorry this
can’t blow up more. It was taken with an old camera. From the first Language
Creation Conference in 2006, where a lot of language creators
met for the very first time in person. Before that, it was
all online interaction. But basically, I found
this listserv in late 2000. I joined up. I started reading some messages,
people talking about all these different languages. And I was like, what do you
mean Tolkien created languages? The “Hobbit” guy? I didn’t know that. Didn’t know that
Klingon was a language. I didn’t know any of that. And so at first,
of course, which is what naturally
happens when you think you’ve been doing
something really great and you see something
vastly different, you think, well, they’re just
doing it wrong, of course. I’m the one who’s
really the cool person. And I thought that
at the beginning. But as I started to really
look at other people’s work, and study their work, and
follow all these conversations, and see what was going
on, that was when I said, oh, I think I understand
what I’ve done here. And this literally
225-page dictionary and reference grammar that
I’ve created for this language, Megdevi, is total garbage. And so it was then
on that it was like that was my moment zero. I started over, and
I started learning. And that’s basically what
I started in late 2000, and what I’ve continued
to do even now. And I think one of the
most important things that happened to me when I
keyed in with the community is I started to ask myself
questions that I had never considered before, questions
that I’d never imagined before. So I’ll give you an example. We have this word
“corner” here in English. And so presuming–
there’s no tricks here– that this box is just an empty
box, how many corners have we got here? AUDIENCE: Eight. DAVID PETERSON: There we go. Eight. I know somebody said
four, but yeah, we have eight corners right there. It’s slightly different if you
ask this question in Spanish. “Cuantos esquinas hay?” Anybody who speaks
Spanish is probably going to know what
the answer to this is. “Esquinas” is essentially
a word for “corner.” And the answer to
this question is four. There are only
four esquinas alla. And they are those
on the outside. If you ask a different question,
“Cuantos rincones hay?” Then you get a different answer. There are four, and they’re
the interior corners. Because, in fact, there
are two different words for that concept in Spanish. And if you think about it,
it’s actually quite important. When somebody says
something like, oh, watch out for that
table, don’t hit your hip on the corner, you know
exactly what corner they mean. It’s hard, and
frankly, it would be really impressive to hit your
hip on the corner of a room. Like you’d have to
get it right in there. It’d be really impressive. It’s like watching the little
DVD bouncing logo, right? And you just keep watching it. It’s like, god, it’s
gonna hit that corner, it’s gonna hit that corner. Oh, it just barely missed it. It’s like, yeah,
that’s an achievement. So anyway, that’s just
something that, if you don’t know Spanish, you never
would have thought of before if your language was English. Corners are corners. Well, it turns out they don’t
necessarily need to be so. Another one, and this
is just phonologically, we have these two words,
“spit” and “pit” in English. They’re actually different. And you can tell if you
put your hand right up in front of your face. And focus on the P, and say
“spit,” and now say “pit.” And if you’ll notice
for the second one, there is quite more
of a burst of air because, in fact, the
second one is aspirated. In English though, we don’t
think of these Ps as different. Like the P in
“spit” is basically the same as the P in
“pit” and whatever. Sound engineers know
they’re different when you have a handheld mic
in front of your face because it’s really annoying. But otherwise, we think of
them as basically the same. Whereas in other languages,
they can be quite different. So these are two words of
Hindi– “tal” and then “thal.” And they’re completely
different words. And you can see that they’re
even spelled differently, “tal” and “thal.” The first one means “postpone.” And then the second
one, I guess, refers to some sort of a
place where you can buy wood. I’ve gotten “wood shop,” “lumber
yard,” “lumber shop,” something like that. I hope it’s actually
still a word of Hindi. If there isn’t, somebody
can leave a nasty note in the comments– that’s
not a word we use anymore. It means something
totally different. But anyway, it
illustrates the basic idea that this aspiration, which is
just a background note of how you properly speak
English, is functional in a different language
like Hindi and very, very important in order to
make yourself understood. In fact, even this
past Monday, I’m constantly learning new things. This was an example
that somebody from the conlang
community, William Annis, found and showed me. This is Amharic. And it’s an absolutely
fascinating little facet that I never would have
even imagined as possible. These are two sentences. “The
judge judged against Aster,” and then “The judge
judged in Aster’s favor.” They’re basically opposites. And what’s happening is the verb
is at the end of the sentence. And I want you to look
at these parts of it. Essentially, the
verb is agreeing with the little preposition that
means against somebody’s favor or in favor of somebody,
which is just wild. I’ve heard of applicatives,
but usually, that involves some sort of
switch up of the arguments and changing the
structures and everything. This looks just like agreement
with a really weird thing to agree with. And that’s just awesome. Oh! It just gets me
up in the morning. Anyway, so being a part
of the conlang community, this is what I gained. You gain, essentially, an
understanding of the ways that languages
can and do differ. And then once you have
that, you get to a point where you realize,
basically, all right, languages can do
pretty much anything. So then when you sit down
to create a language, you really are
saying, what is it that I want to do
with this language? I have all of the
options before me. And now what am I
going to do with it? And so for example, if
you ask me right now, just a quick example,
what the word for “corner” is in Dothraki, it’s “kemikhof.” And there actually
isn’t a distinction between an interior and exterior
corner, but not by accident. It’s because I wanted
it to be that way. And in fact, it has to do
with the unique etymology of the word. It ultimately comes from a word
“kem,” which means “conjoined.” And so how the
word came to be is that it’s talking about
the meeting of two walls or the meeting of two things. And so that’s what
it’s focusing on, not the interactional
properties of the human, but actually what it
is and what it’s doing. And so that’s why
there’s only one word for “corner” in Dothraki. It was something I did on
purpose, not by accident. So now that I’m here,
I wanted to give you an in-depth example of
the type of questioning I do when it comes to
translation and things that I work with. So this is a script page from
the first season of “Defiance,” where this entire exchange
is in the Irathient language. And we’re going to focus on
this first sentence, which is an elder Irathient is saying,
“It’s how the Castis killed us in the Great Diaspora. Caves filled with chlorine gas.” So I got that, and I
think, well, of course, I don’t have a word
for “chlorine gas.” So what am I going
to do with that? How am I going to
translate that? To catch you up to speed, for
those who are not familiar with “Defiance,” I’m going
to need to introduce you to some aliens. First, there are the Irathients. They look like this. They have orange hair like that. Their planet was previously,
before the show started, conquered by the Castithans. And basically, at some
time in the distant past, it was like the Norman conquest. They came. They took over. They became the
rulers for a while. Now, they’re on more
or less even footing, but there’s still some
tensions due to the history. And then a third set of
aliens are the Indogenes. The Indogenes are responsible
for pretty much all of the technology that
the rest of the aliens– called, collectively,
the Votans– enjoy. So they created all
of the spaceships, all of the nanotechnology,
everything. And they are, themselves,
an interesting creation. They modify themselves
genetically. They have implants
all over the place. Their eyes can act
like microscopes. Their hands can do really
fantastic little machinations. So for example,
with their language, their writing system is
actually a post-hoc creation. Every single glyph
in the language, whether it’s for a number
or a syllable or whatever, has these little
conjoined double hexagons. This is nothing that would
ever arise in a natural context because it’s way too complex
to draw with your hand and get it right. But Indogenes can do it exactly
right every single time because of their augmentations. And so they actually
created this writing system for themselves after
the point in time where they started
augmenting themselves. Anyway, so this is
their number system. Their number system is base 7. And you’ll see why it’s base
7 if you see how the number system works there. So you can see the
little tick mark going do do do as
we go from azu, kizu, fozu, dezu, nuzu, hazu. And then that’s
1, 2, 3, 4, 5, 6. And then ama, you switch
the suffix, and you get 7. So basically, their 7 is the
equivalent to our 10, like 1-0. So I started with
them when I need to come up with words
for the elements– the periodic table of elements. I reasoned that they would
be the ones to kind of codify and then really
make use of this, and they actually
invented a system for referring to the
periodic elements. They did it by protons. So chlorine has 17 protons. 17 in base 7 is 23. What that means is that the
digit 2 followed by the digit 3 equals 17 in base 7. Then the word for
17, 23 in base 7. You have kima which
is basically 20. Remember, that’s the number
2 with the ma suffix, and now you see the
2 lines and then also the straight line up
and down in the middle. That’s 20, and then 3
of course we know, fozu. So 23, or 17, is kimafozu. They also have a suffix, vun. This was an older suffix. Basically the word vun used
to mean just something like, stuff way back when. Now it’s a suffix
you can add to things to make a kind of
nominal compound, and specifically they
add it to numbers to create a word that
means “that element.” So when you add vun to
17 you get chlorine. Kimafozuvun. Kimafozuvun. Yeah, sorry, I can say it right. And of course you can
spell it like that with the numbers
in front, or you can spell it all the way out. It’s basically the equivalent
of this in English. You know spelling
twelfth like that. God I love it. Oh, it stopped. All right. Anyway, so twelfth,
12, all right. So we’ve created this
word now in Indogene, or the Indogene language
which is called Indojisnen. Next step is to borrow
that into Castithan. And the reason is that
Castithan is essentially the language of commerce,
politics, everything. It’s the English
of the Votan world. So they borrowed all of
their scientific terminology from Indogene, or Indojisnen. So that gets borrowed in, and
it gets borrowed in pretty well. So it’s kimafozuvun
in Indojisnen and then in Castithan it’s kimafozuvuno. You have to add the O suffix
on there to make it functional. Then you have this
word in Castithan. By the way there are some
quirks– little aside with this spelling because
Castithan spelling is kind of slightly more
difficult than English, but slightly less
difficult than Tibetan. Tibetan I think is
the most difficult that we’ve created
on the planet. So just some notes. Like for example, you could
write ki this way, or this way. The reason is that there
are 2 different types of K’s in the history. -So you have this older word,
which had a long vowel “li-ku,” which is now pronounced ligo,
and it refers to some sort of small animal. That’s using, essentially,
the traditional K. And you’ll see that the
traditional K is now pronounced G in certain circumstances. However, the old ejective K
which is here for the word “ni-k’u” which is now
pronounced “nico,” and it means, kind of, lot or parcel. That one is always
pronounced K no matter where it is in the word. So when they do borrowings they
tend to use that special K, just to make sure that everybody
knows it’s always pronounced K. Same thing kind of here. These are both
pronounced the same but the one with the
longer vowel– or that used to be the longer vowel–
is always going to be pronounced ma. So we use that, just to be safe. And same, the first
Z also changes in certain circumstances. The second Z does not. So you use the second Z
when spelling things out. That gets us kima and zu. This one has a slightly
different explanation. So these are the
non-sibilant fricatives that have a different story. F is pronounced F at the
beginning of the word. It tends to be pronounced to
V in the middle of the word. So for example we
have this word that used to be pronounced that way. It’s pronounced fozwo, ritual. It’s a nice F at the
beginning of the word. In between you had this
older word, hiifau, which is pronounced hivo. It’s that word for shirt. It’s pronounced like a V
in the middle of the word. But since there are
no other fantastic ways of representing the
A, it’s kind of known that you have to swap out
for the sibilant fricative since there’s two
versions of each one. There’s 2 Z’s and 2 S’s. There aren’t 2 F’s. So you just use the regular
F in a borrowed word and you know that rather
than being pronounced V it’s pronounced F
because it’s a borrowing. The same is true of the original
V, which is often pronounced H in between vowels. You know it’s pronounced V
because it’s a borrowing. So that’s how you get the
spelling of kimafozuvuno in Castithan. The reason we had
create it in Castithan is because a lot of the
borrowings in a Irathient that have to do with
technology come from Castithan. Usually via Indojisnen,
but Castithan is where they got a lot of
terminology because of the time when Castithans were essentially
in charge of everything. So that gets borrowed
into Irathient this way, and it’s now kimafozvun. Notice that we lost a U there. That’s because they’re used
to having to remove vowels in between sibilants and
secondary consonants, moving from Castithan to
Irathient because that just happens a lot with
Castithan words. So they saw this one and
said, well there probably isn’t a real U there. Even though there
was a real U there, they thought there
wasn’t a real U there. It’s the same kind of thing
that happened with– what’s a good example from English? Like, nuncle. I think it was
originally nuncle, but now we call it
uncle because we thought people were saying
“an uncle” as opposed to, “a nuncle”. Really weird. Anyway, so at it’s
that type of a thing. Now as for the
word itself– this is just another
kind of a fun aside. Irathient has a
series of noun classes that are indicated by
the initial consonant. And it just so happens that
since this word begins with a K it falls into the
dangerous animal class. So klaidi is a dangerous insect. Kombisi is a kind of a
tentacle beast comes from above and sucks you up. Then kangezi Is a word for like
the alpha of a pack of animals. So for that reason they treated
this like this was the root. That the root was actually
imafozvun and that the K was incidental, and
indicative of a class 4 membership. Which means that they created
this word, rimafozvun, which means chlorine the
element, or the substance. Then kimafozvun, which is
now the dangerous animal and has the concept
of danger in it, is used specifically
for chlorine gas, which is what the Castithan’s used to
kill them back in the old days. And that is basically the story
for translating just that word, in that sentence, in that
scene, in that episode, in the show “Defiance.” So that’s kind of the process I
go through when I create words, and basically it’s all
thanks to the education that I had coming up through
the conlang community. I know to ask questions that
I never knew to ask before. And so one of the
things that I realized when we had this great
shift in the internet and how we interacted with it. I don’t know if it
was 2009, 2010, 2011, but it was that time
where suddenly nobody was using hand-coded
HTML anymore for their own personal websites. People kind of were going away
from listservs and bulletin boards even, and
instead were interacting through micro-blogging
platforms, blogging platforms,
social media platforms. I started to see a lot of
conlangers coming to the art, especially with a lot of popular
media like “Avatar,” and “Game of Thrones,” who knew those
shows but didn’t know anything about the old
conlanging community, and who definitely
at this point in time were not going to
join an email based listserv where they communicate
with people via email. Like, it’s not even a part
of our vocabulary anymore. And so, the thing was though,
I felt like the education that I had, which I thought
was fantastic, was being lost, and is essentially
irrecoverable. Even though like you can
search all of the archives on the conlang listserv–
they’re publicly available, there’s great stuff there,
nobody’s going to look at it, nobody’s going to
pull it up, nobody’s going to go through
and read old emails. It’s not going to happen. And so that was
one of the things I hope to do with this
book, “The Art of Language Invention.” I’m kind of trying to rescue
some of that knowledge and put it somewhere. Oddly enough into old
media, hardcover, book form because I think people that are
conlangers that are primarily on Twitter and
Tumblr are actually more likely to get
this book than they are to actually pull up the
old archives of an email list, or join it. Which is a really weird
thing– weird reality we live in but it’s
kind of where we are. So that was kind of my
fondest hope and purpose for this book is to kind
of capture that put it somewhere so that other
people could benefit by it. And so it’s not
even like my stuff. It’s the stuff that I
learned and basically just put into this capsule. So anyway, that’s the
story behind this book. Thank you guys so much for
your time and listening, and if you get it,
I hope you enjoy it. [APPLAUSE] AUDIENCE: First of all,
thank you for coming. There’s this old chestnut
about cellar door being the most beautiful phrase from a
phonetic standpoint in English. Since the people who are
hearing your languages aren’t native speakers,
how does the sound of the languages you create
play into their creation? DAVID PETERSON: Of course
that’s always the thing. That was Tolkien that said that
most beautiful word was cellar door, and of course he didn’t
mean cellar door, he meant, cellar door. Which is how he would
have pronounced it. So that was the thing
that was supposed to be most beautiful to him. Because for us cellar door
just sounds, I don’t know. And even so of course, that
thing is entirely subjective, but you can tell
though, if somebody thought cellar door
was really beautiful. You can see how that really
informed Tolkien’s languages. When I’m creating languages,
especially for shows, I know that the
bulk of the audience is probably going to
be English speakers and because of that
I know that they have kind of the history
of cultural stereotyping that exists for
English speakers. That is, we have kind
of a predictable range of which languages we think
are going to be beautiful, and which languages we
think are going to be harsh. So you can play with those. So that if a producer
wants a language that sounds harsh or
guttural, you know what that means and
you can produce it, but it’s not actually
true that any language is beautiful or ugly. It’s totally just basically
cultural stereotyping, and just random resemblances
between a language you speak and the languages you
think are beautiful. So yeah, it’s an unfortunate
byproduct of our existence, but I can still use it to
my advantage if I have to. AUDIENCE: You mentioned that
you created your first language while you were studying
at UC Berkeley. I was wondering how much formal
linguistics education you have, and how that has helped or not
helped your language creating. DAVID PETERSON: I
actually started creating my first
language at the same time that I started studying
linguistics, or I think I did. I told a very– I told a
very strange story in a lot of interviews that when I’m–
I’m actually giving a talk later at UC Berkeley and I
went back to my old notes to photocopy some of them for–
this is for the linguistics club– to show at
that presentation. And going through my
old notes I actually discovered what I had
been saying in interviews was factually incorrect. So I thought that I created my
first language in the spring semester of 2000 when
I was a freshman, and when I was taking Arabic,
and Esperanto, and Russian, and Linguistics 5. It turns out I wasn’t taking
those classes at that time. I did take Arabic
and Esperanto, but I took Russian and linguistics
the next semester in the fall of 2000. So I took Russian
and linguistics. I had previously
taken a year of Arabic and a semester of
Esperanto, and I started creating my first
language at the exact same time that I was taking my very first
introduction to linguistics, which I had never
heard of before and I only took because it
was a breadth requirement. But I think it was
like 2 or 3 weeks in– I actually have the date
if I go back to my notes, but I forget. Anyway. So I kind of started growing
as a language creator as I was learning linguistics. Oh, but I eventually did finish
out with a linguistics major, and I also got my master’s
degree in linguistics at UC San Diego. Definitely the study did
inform a lot of what I did but at the same time
it was helpful to have the background of
a language creator so that I could know to, kind
of interpret what I was learning in linguistics with more of an
artistic and also critical eye. Especially a lot of
the theoretical stuff. Like theoretical
syntax has not really informed what I’m
doing at all, but a lot of what I studied in
morphology, especially like lexical morphology,
[? word and ?] paradigm morphology, and a
lot of what I studied in historical linguistics has
informed what I do enormously, enormously, and definitely
phonology to, and typology, typology is huge. There was a lot
of back and forth. But anyway, a lot of it came
in and definitely linguistics was an important part of my
education as a conlanger. But I would say equally
important, if not more important, was my education
from other language creators. AUDIENCE: I was interested
in birth of language. Have you investigated that? DAVID PETERSON: What,
like language at all? AUDIENCE: Yeah like,
6,000 years ago. What shreds we know, you know? DAVID PETERSON: I think
we do know some stuff about 6,000 years ago. But you know essentially– AUDIENCE: Hard consonants
and like how it directly translates to an object. You know phonemes. DAVID PETERSON: I
don’t really buy that. I don’t think that like the
earliest, earliest, earliest forms of languages–
they were necessarily like such great sound
symbolism like that. There is sound symbolism that
exists throughout the world’s languages so often. If you have vowel where
your mouth is closed, like E, it’s often
associated with small things, and if your mouth is
wide open, ah, it’s often associated with large things. But beyond that
there’s not a lot that isn’t language specific. So if you think about English
we have words like glean, glint, glisten, glitter, glow. You could probably
come up with others. There are others. And you’ll notice those
all kind of have something to do with shininess, right? But there’s nothing–
I mean that’s an English specific thing only. It has nothing to do with any
other language on the planet. There’s nothing about Gla–
Glee, it means Glee– no, there’s nothing about
Gla that tells you, oh this has to do
with shininess. It’s just a bizarre cultural,
historical artifact of English. Now having said
that, of course I study the history
of languages a lot. I don’t think that
we’re at a point, and I doubt we’ll ever be at
the point, where we can say this is where language was created. This is how it came to be. I don’t think we’re
ever going to get there. Honestly, we’re pressing it
when we go back to things like Proto-Indo-European. There’s a lot we know but
then there’s so much we don’t. There are definitely
people out there who have looked into doing this. Like Joseph Greenberg who’s
a famous, famous linguist and made some
brilliant typological discoveries was a big
believer in proto world. That you could actually find
the commonalities between all the languages on the planet. I would say most linguists
don’t accept that, wouldn’t even except some
of the less grandiose claims like there were 10
original language families, which I’ve also heard. I’d say most don’t accept that. Most know that the further
back in history you go the less and less certain
we can be about what we’re saying about these languages. Though, but hey, from
here on out everything’s going to be recorded. So that’s nice. AUDIENCE: Thanks for the talk. Really interesting. I’m curious, do you have
any particular features of your languages, or
of languages in general, that you feel are really
interesting, that you feel really enamored with? Like you mentioned
in the example that you gave you
have noun classes, you had spelling
variations, you had this very systematic writing
system– reminds me of Hangul, for instance. Is there anything that you
feel that you notice yourself saying, like I can’t put
this in another language again, because I
do it all the time. DAVID PETERSON: Yeah, well
definitely noun classes are one of my most
favorite things. It’s so easy to create
words, and to create really interesting
words that you never would have thought of
if you had noun classes. I love them to death. But I’ve only ever– the one
I did for Irathient, that was like my second big one and
I haven’t done that precisely again, but I could
just do that all day. Love noun classes. Noun case is another
one I’m fond of, but that you can get away what. I mean, most languages have
noun case in some form. But yeah, case systems
are loads of fun. I think where I get into
trouble is not something that I love but something
that’s a weak point of me. I just hate verbs. It’s so easy, it’s so easy
to do a verb system that’s basically perfect imperfect. Lots of languages
either are that way, or started out that way. Where there’s just basically
a binary distinction of perfect and imperfect,
so done and not done, and it’s so easy
to just do that. Say, all right we’re done,
but I kind of do it too much. I know I shouldn’t, but
verbs are just so complex. They change by the decade
precisely how they’re used. Just watch a football
game and the guy says during the replay–
it’s like he’ll say, if he doesn’t pick up
that block right there, he doesn’t run to the end zone. That’s all present tense
but that’s not actually the present tense. I mean it’s the
present tense, but it’s some sort of weird, bizarre,
hypothetical, counter-factual tense. We just use the present tense
for it in a sports context only. Right? That happens, that’s
real verbs right there, and it just pisses
me straight off. That’s what it’s the worst part
of learning any new language. I don’t care how simple a
language will tell you– it’s like, oh you have
these agreement patterns. We don’t have that in Chinese. Yeah, I don’t care. I’m not learning
that verb system. It’s incredibly complex. Just because the words
themselves don’t change, doesn’t mean the
system isn’t complex. It probably means
it’s more complex. Verbs! AUDIENCE: I’m curious
what’s your favorite word? What language is it in? What does it mean, and
what was the context that you like it used in? Do you have a favorite word? DAVID PETERSON: A favorite
word that I created? AUDIENCE: Yeah, more likely. You could tell us both. The ones you like
that others created, but I’m kind of
curious about yours. DAVID PETERSON: I don’t know. I actually– it’s funny,
I get this question a lot and I never have a very good
answer, and I’ll tell you why. The words that I
remember the most are the ones that
I use the most, and they tend to be
very prosaic words. Like the word for “need
to” in Dothraki, zigerelat. I know that one but it’s there’s
nothing very great about it. I just use it all the time. And actually I was just
talking about this. There’s a particular
reason I think that I don’t remember
a lot of the really fancy, interesting words. Like I’ll give you
an example of one that I know because
I used it recently. [? “Hej” ?] is the word
for “sticky” in Dothraki, and [? “hejov” ?] is a word
for a situation that is more trouble than it’s worth and so
you’ll likely want to avoid it. I thought that was a pretty good
word but I’ve never used it, and the reason is
that– especially for these shows, the words–
I’m acting as a translator. I get English
scripts and then have to translate what’s given
to me, and so the words that I end up
using are the words that the writers
come up with and they don’t think of these really
fantastic and interesting words that exist in these languages. Where perhaps if there
were native speakers, they would find uses for them. They would have different
patterns of use, but instead we’re working
with English patterns of usage and words that the
English speakers come up with. So consequently since my
most common interaction with my own languages
is when I translate, those words are the
ones I know the most. So there are scores of
words that I have completely forgotten about, and I
usually only run across them when I’m looking for some
word in my dictionary and it happens to be right
next to another word that’s really cool, I’m
like, oh my god, I forgot I created that word! It’s so cool! Now, here’s the word for
belt. I go on from there. I do have a least
favorite word in English. Crafts. That’s the most hideous
word I’ve ever heard. Just awful. AUDIENCE: Do you have a
word aversion for that? DAVID PETERSON:
First of all, crafts. It’s f-t-s right in a row. That’s barbaric. And it just– makes me think
of the term arts and crafts. It’s just weirds me out. There was a– who has the mic? Over there. AUDIENCE: You’d
mentioned complexity in one of the
previous questions. Are there– in the
conlang community are there any measures
of complexity, or precision, or granularity
for different languages? DAVID PETERSON: Yes,
but they’re not applied to artistic languages usually. As languages created
for the purpose of art. Usually where you see
things like that discussed are creation of
auxiliary languages. So, languages like Esperanto
that are built either for international
communication, or communication within a group. They try to measure things
much more objectively so they can say which one is better. We tend not to do that
in the artistic language community for the same reason
that you don’t typically go to paintings and say
which one is better, and then come up with
objective reasons for why one painting
is better than another. There are certain
measures that exist within linguistics, such as
the minimum word dependency. I’m going to get this wrong. Minimal length
dependency, I think it is. There’s a paper that just
came out of this– a kid just came out of MIT on this. Where if you subscribe to
the theory of morphemes, which I don’t, but if you
do, you can chop up a word and see how many of them there
are and come up with number that tells you
something or other. But I haven’t used
it myself, and I don’t think it
enjoys a lot of use within the artistic
language community, but definitely within the
oxlang and perhaps englang community, englang being
engineering languages that don’t necessarily
need to be natural. AUDIENCE: As somebody
who had studied a lot of languages and
studies a lot of linguists, I’m sure you’re aware that
there’s some languages that are– although they have
their own vocabulary, but structurally in
terms of lexemes, they’re very close to English. You basically can do a
word for word translation and maybe change the
word order a little bit. There are other languages
where in that culture, in that language, you just
express an idea structurally completely differently. You can’t really say in
translating one sentence to another that this
word corresponds to this, and that to that. When you do these construction
languages for TV and movies that you know you’re going to
be translating English scripts, does that tend to make
you create languages that are structurally
closer to English so your translation
job will be easier? DAVID PETERSON: No, not a bit. No. Not at all. But of course, just a
comment, the languages that are going to be closer to
English structurally are going to be those that are
related English genetically, like in the Germanic
languages, but even so, there’s a general tenant
that we pretty much believe to be true that you
could translate anything into anything. When it comes to– AUDIENCE: I’m not
saying that’s not true. DAVID PETERSON: But yeah,
like the word for word. AUDIENCE: You know what
you’re going to be doing and to make it
easier for yourself. DAVID PETERSON:
Yeah, no, not at all. No. Because that would be boring. The reason I do this is I
love creating languages. That’s my fun. I mean the closest thing
I’ve gotten to English is actually Old
Trigedasleng in “The 100” but that’s because it’s
actually supposed to be an evolved form of English. English 100 years later. So it’s like it’s
really– but there it’s supposed to be like English
because it’s kind of supposed to be English, right. But other than that, no. Honestly, the languages
couldn’t be more different. They range from marginally,
somewhat, kind of close, to totally,
unrecognizably, out there. That’s how I have my fun. AUDIENCE: So every year we
lose dozens, maybe hundreds, of human languages
to extinction, and we lose the cultural
payload with them. How do you feel about the ethics
of conlangs going into– people running around geeking
out on Klingon when we’re losing real human
languages that they could be learning. And why do you not consider
using obscure human languages for some of these shows? Bring them back a little bit. Give them a new life. Make people aware
of them, rather than artificial languages. DAVID PETERSON: Let
me ask you this– how do you feel about
people getting really, really excited about novels
like “The Hunger Games,” which are stories about
fake people, when we’re losing the stories
of real people every day that are dying? AUDIENCE: Honestly, I think
that’s sort of false analogy. We have– and I admit I come
from a very specific basis on this. But I guess I would
say, how do you feel about ISIS
destroying antiquities when we could make new stuff? DAVID PETERSON: I’m
sorry say the last part? AUDIENCE: How do you feel about
ISIS destroying antiquities when you could make new stuff? That’s equivalent to yours. Is it– and I’m not saying
what you’re doing is wrong– DAVID PETERSON:
You’re saying that I’m destroying natural
languages in order produce– AUDIENCE: I’m not saying you’re
destroying natural languages. I’m saying, could we not
use some of the energies that people go into learning
conlanguages, geeking out about con languages, going
to Klingon conventions. I’ll tell you I got very annoyed
when I see– I saw even here at Google recently,
people talking about looking for
Klingon resources to do something in Klingon,
and it’s like great, but let’s put that effort
to saving a real language because when we lose
those languages, when the last speaker dies,
it’s never coming back as a living language. MALE SPEAKER: Yeah
so, of course– and in fact if
they weren’t going to be doing Klingon they would
be doing the other thing, right? AUDIENCE: That’s how it ends. MALE SPEAKER: No, of course not. So let’s stop– AUDIENCE: You could use
real human languages for some of these– DAVID PETERSON: You
certainly could. Yeah. We could have found
a– for example, we could have found some
very rarely spoken language and given it to the
barbaric, Dothraki who are ripping people’s tongues out. How would that have been? AUDIENCE: I wasn’t
necessarily saying Dothraki. DAVID PETERSON: OK. OK, but that’s actually
the issue that comes up. For example, there
are a lot of languages that are dying out right now. Not the majority of them but
certainly a minority of them, that are dying
because the speakers don’t want other people
to use their language. They don’t want people
to write it down. They don’t want
foreigners speaking them. Do we respect that? Or do we just kind of sigh
and let the language die out and respect the wishes of these
people who don’t want contact with the outside world? It’s a different kind
of ethical question. But personally there are
two ways of looking at it. One, artistically if you’re
talking about a totally fictional reality it
breaks the reality to use a language that
actually is spoken by other people on Earth. It’s kind of an odd thing. I always find it
odd when people are speaking English
when they oughtn’t to be– like “Amadeus.” I mean it was a fantastic movie. It was a fantastic
movie, but honestly why were they speaking English? Especially like, I don’t know. Well, it’s actually kind of
a bizarre thing of movies that we think that British
English would have been better than American English. I mean it wouldn’t, it’s the
same thing, but I don’t know. So artistically it
simply breaks it, and honestly there’s
no alternative here. The alternative to using
creative language in “Defiance” was English. They were never, ever
going to consider using any other
natural language, or the expense involved
in finding somebody to translate into
those languages. And then furthermore,
if you think about, especially some very–
some minority languages that aren’t spoken very
well– by very many people, and you start to translate some
of the dialogue in “Defiance.” For example, this
sentence that I showed you about the chlorine gas. Odds are that Piraha doesn’t
have a word for chlorine. So then it’s a question
of what you do. It seems doubtful that you
could get a speaker up there who would have to learn and
understand English well enough and perform on a deadline to
be able to translate into it, so you need somebody else. And undoubtedly they would
have to create words. They would have to
create words either by working with
native roots, or have a whole load of English
borrowings in there, or worse, create words they kind of
look like the real language and use them in those scripts. Which to me seems
really offensive. Essentially you’re
creating words to plunk down into somebody’s
actually existing language. You’re not a part of the
culture, and you’re saying, well we’re going to represent
this as a word in your language for the purposes of
this television show. I just can’t see that
being a good idea. On a separate note, if you just
to kind of remove the created language aspect entirely. The fact, of course, that
there are dying languages is terrible. It’s a question of what
one can do as an outsider in order to either prevent it,
or to preserve those languages. I’m not sure how a lot of
speakers of minority languages would feel if there were,
say, a bunch of teenagers from Southern
California who started to learn their language
just because they thought it was cool, and
funky, and it was used for these quirky
people in this show. That, at least for
me, that would make me feel a little uncomfortable. So I think that the
sentiment is well founded. I don’t think that the
solution you presented is a good one, or feasible,
or necessarily respectful. I also think that for standard
preservation of languages there are people that
are trained to do that. It’s not an easy thing. And of course I know
them, and in fact have studied under them coming
through linguistics. It’s not just something
that anybody can do. It takes a lot of
work with the culture, and having worked with Hollywood
a bunch, it’s a lot of work that I know that nobody
would be interested in doing. They wouldn’t be interested
in paying for it, they wouldn’t be interested
in taking the time with it, because they have
deadlines and they’re doing what they’re going to do. I’m not sure if that
answered every single aspect of your question but if not,
I’m happy to talk about it further because this
is an issue that comes up every
single time created languages are discussed. So thanks. [APPLAUSE]

42 Replies to “David Peterson: “The Art of Language Invention” | Talks at Google

  1. Great talk David.  It's quite sad that sometimes people with good intentions, like the guy at the end concerned about recording dying languages, take their frustrations out on completely benign artistic endeavors like conlanging.  It's sad because that attitude is basically implying that people who have a genuine love for language and want to engage artistically with it are doing something wrong because they aren't devoting their energy to getting out in the field.  It's a really bizarre argument.  David's analogy about the fiction writer and historian is a good one, and it threw the guy off.  He didn't explain why David's analogy was fallacious despite claiming it was, he just rebutted with a false analogy himself. 

    Choosing to devote your interest and skills in language to using language artistically is in no way equivalent to ISIS destroying cultural artifacts.  In the ISIS analogy, people are actively destroying cultural artifacts, not merely neglecting them because they're choosing to devote their energy to do art instead of artifact protection lol; what a strange argument.  And in David's case, he shouldn't be considered doing a disservice to dying languages because he's not himself out there recording languages.  Conlanging isn't done at the expense of language preservation.  To argue otherwise, you're basically saying that art of any form is a disservice to humanity because in a hypothetical world artists with any practical skills could be devoting their energy to other things.  I don't imagine many people would defend that perspective when it comes to traditional art forms, yet for some reason when it comes to people playing artistically with language in this manner (an art form just really starting to catch on) some people get all riled up.  The position is very illogical.  Furthermore, conlangs in art media seem to be getting a lot of people interested in language and linguistics who otherwise wouldn't have become interested in these things.  People getting excited about language, no matter how that initial excitement begins (through art or otherwise), seems to only be a good thing to me.

  2. People don't understand that a mass of people can't be controlled. Why isn't more energy put on learning real dying languages? Because people think it's more fun to make their own languages, and the community isn't a bunch of robots that can just be told to start focusing on other things.

  3. When that guy at the end used the term "geeking out" again and again it reminded me that being known as a "geek" used to be universally a bad thing, rather than now it it simply means being passionate about something out of the ordinary

  4. HAHA that ethics guy… wtf! "I think the sentiment is well founded, I don't think the solution you present is a good one, or feasible, or necessarily respectful." haha CRUSHED! he is like the linguistics equivalent of a SJW, i.e. self righteous and ignorant.

  5. That guy at the end couldn't have been serious. Use a real language for a made-up race of people? Or even aliens?

  6. In the subtitles:
    7:42 [INAUDIBLE]: Lexeme
    http://www.youtube.com/watch?v=Z50T-tslrgs&t=35m10s [? word and ?]: Word and paradigm

  7. Correction ¿Cuántas esquinas hay? with an "a" … or is it? Google throws up lots of results from spanish native speakers which don´t seem to care about conjugation for the question word, "cuánto(s)?" Anyone know?

  8. If you are talking about creating endangered language media for the purposes of preserving that language, then that media should really be made by and for that language's community of speakers, or if appropriate, people wishing to join, or rejoin, that existing community, because those are the people you need to engage in order to keep the language alive. Conlanging is really a completely different audience and market, from minority language resources, especially conlanging in the context of mass media sci fi and fantasy works like Avatar or Game of Thrones.

  9. Using existing natlangs would be appropriate if you're setting the story in the culture in question. If you're setting it in a fictional culture, go ahead and use a conlang.

  10. I think the man talking about old languages had a completely valid question, to be honest. I think it'd be a nice idea

  11. At about 7:44, the subtitle claims that David is inaudible. It is wrong. He clearly says "lexemes" and I'm assuming the subtitler didn't know that was a word.

  12. I don't like the word 'sixth' because I can't pronounce it, and the sound of it is just unpleasant. I have nothing against the number six, I just don't like 'sixth'.

  13. I'm sure the dying languages don't have a way of saying, "engage the cloaking device," or "beam the dilithium to cargo bay three." Or that word that commands the dragons to kill its enemies.

  14. base 7 barely makes any sense, because it’s probably one of the most impractical bases. it seems really odd that such a smart culture would willingly pick such a bad base system tbh. well at least that should add to this mystery of them?

    nonetheless, outstanding talk, this was quite inspirational and got me quite inspired. so thanks for taking your time to do it!

  15. I agree with the practical reasons David gave about why real dying languages aren't used in this way, but I don't know why the real native speakers would be offended if they were. If I was a speaker of a dying language, I can't imagine being offended if I heard it spoken by a fictional race on TV, and I would probably be happy if it got people interested in the language in the real world. At least the language would be getting exposure, and maybe it would be saved in this 'disrespectful' way.

    But all the reasons he gave about the difficulty in finding a native speaker who would be able to translate everything, I completely agree.

  16. my dear David, you are so cool, i wish can meet you up
    https://youtu.be/NdmBpPA6eYk
    nedimkufi.com

    I m in your direction, originally from Arab

  17. Just visiting to say that's this is a terrifying thumbnail.
    Will be back to watch though… David's a peach 🙂

Leave a Reply

Your email address will not be published. Required fields are marked *