those of us who live in comfort and
security which is probably everybody in
this room like to imagine that we
deserve all the good things that come
our way but let’s stop for a minute and
think about some of the advantages that
might have taken our powers towards
getting here in my own case I’m not
talking about money I don’t come from a
wealthy family and I wouldn’t have been
able to get my university degrees
without a full package of scholarships
and loans inside jobs but growing up in
a country with the ability and desire to
see its children through higher
education is one of those advantages
that I’m talking about for half of us
being male and it’s given us a big head
start our teachers assume that was
quicker in math and science we’re much
less likely to have to face sexual
harassment and on average we heard about
20% more than our female counterparts
where I grew up in America being white
was a huge advantage over my black and
Hispanic neighbors I was never subjected
to the police stop and for us nobody
ever thought I was going to their store
to shoplift my family was never a red
light for getting housing in
neighborhoods with good schools like
Donald Trump used to do as a wave over
New York and for the 91 percent of us
who are straight we never have to hide
where identities or people we love and
we’re not subjected to job
discrimination or in jail in some
countries these inequities are real and
if we think about race or sex or for
sexual orientation we can readily see
how they might affect the ways that
we’re treated in school the doors that
are open to us are slammed in our faces
the internships the jobs the references
the bank loans that we either get are
denied but these are also things that we
talk about that we have names for and
that people are working to try to
rectify so do you think that the bushes
up there are able to leave countries
because there
because they have better DNA more
capable genes than the kansan teams on
the right or is it because they were
born with a silver spoon in their mouth
with every advantage from prenatal care
right on up
and we stand on the shoulders of our
parents who also probably have a lot of
things advantages and have are able to
pass on to us all these numerous subtle
extras like extracurricular programs or
dental checkers cultural excursions to
museum shows or events like this travel
sufficient and healthy diet professional
connections when I was about 12 I got
this great gift it was a gift
certificate back to my local bookstore
and I spent hours just trying to pick
out the books that would add up to the
money money total on the gift card this
was a gift from my grandparents we’re
educators at who saw it as their duty to
posterity that I should loves learning
decades later my sister built on this
heritage of ours and sends my seven year
old books with positive the messages
that are going to overcome the sexism
that she’s going to encounter in her
life here in Switzerland and these are
more examples of advantages that were
born into but there are also examples of
the advantage that we don’t even have a
word to describe my family could sent me
to the bookstore the library my sister
can send books to her niece because
books are available are plentiful in our
native tongue because of the language
I’ve heard even starting in the wool I
had open to me a world of volcanoes and
Dragons and medieval conquests and
science experiments
my daughter has pilot said painters and
architects and explorers and now with
the internet it’s not just books we can
ask Google but anything we need to know
about the nutritional value of mushrooms
how to calculate the age of that big
tree outside without cutting it down a
good portion of the world knowledge is
available to us because of the language
that we as luck would have it happen to
speak English French speakers are
similarly advantaged and German speakers
and Russians and Japanese Italians
portuguese maybe the correlation doesn’t
prove causation and many additional
factors are at work income life
expectancy and access to knowledge in
one’s native language really do march
very very closely together but we had a
couple of more advantages that we all
enjoy and which we have games for health
care and electricity without electricity
you can’t run a computer or a smartphone
you do your homework by the light of an
oil lamp but this is something I’ve seen
in debt houses that have been in Africa
you can’t keep your food or your
medicines or your vaccines fresh and
safe without health care and it’s all
that’s right it’s so obvious that in
most places it’s illegal to do not deny
somebody health care on the basis of sex
or race or even sexual orientation
yet health care is denied all the time
on the basis of language an immigrant
who can’t speak to a doctor
can’t get proper care people with HIV
can’t follow the instructions on their
drug cocktails if they can’t understand
the languages that the instructions are
written and children can’t become
doctors and nurses and serve their own
people if they can’t understand the
language in which health sciences talk
to them but since linguistic exclusion
is an unnamed disadvantage it marches in
silence listen to a statement from the
Ford Foundation and you can read it
which highlights the advantages that we
gain from an
open Internet in the language that we
can understand the internet is not a
luxury broadband is an essential public
utility and must be equally accessible
to everyone because it enables people to
express themselves organize make a
living participate in policymaking get
rid of job earn an education and do all
of the critical things that we so often
take for granted but typing in the
Internet in the wrong language is like
laying water pipes with no clean water
with an opening water source to draw
from language is a thick glass wall to
an open Internet to political
participation to getting a job to
earning in to earning an education we
recognize that the internet isn’t
central utility like water like
electricity but we still not recognize
that it’s useless to the billions of
people who can’t read and interact them
in the lucky languages the typical
response in Europe in the US and even in
elite circles in Africa in India and
South America is what the learn English
let them learn French like and learn
Spanish we did it at work for us anyone
can learn a second language and who but
the unsuccessful people the little
people can can speak those local
languages well yes
you do speak the privileged language and
you’re successful but that’s either an
accident of birth or a lot of study that
your circumstance is made possible 95
percent of Africans don’t are not
literate the languages that their
government inherited from their European
colonizers and that’s going to be a
quarter of the world’s population by
2030 everybody that I speak to in India
tells me that everybody in India speaks
English and they’re right everyone does
but they’re cook doesn’t and their
housekeeper doesn’t
method and their children don’t and
their grandchildren won’t 95 percent of
Indians don’t and billions of people
whose mother tongues are on the wrong
side of military or economic power are
perpetually excluded for paths towards
knowledge of prosperity because people
in government’s and funding agencies
think the languages that they speak are
worthless if not actual impediments to
progress now despite the indifference or
hostility for many decision-makers we do
have solutions that we’re proposing to
eradicate language is a barrier in
reaching internationally declared goals
for Social Development 496 years ago
Martin Luther figured out that local
language was elemental to participation
and he translated the Bible into German
since then on the back of that a new
technology in that time called the
printing press missionaries have spread
their ideas using local languages in
hundreds and hundreds of languages but
we can’t pretend that moving from ink to
electrons from the printing press of the
tablet is going to magically make
billions of people conversing in one of
the languages that they already
privileged people happen to speak rather
what we’re saying is we can work with
much smaller populations from front from
these languages communities and collect
the information that we need to produce
linguistic equity for their languages
and then we can plug that information
directly into the technologies that have
already been developed for English for
transfer German and the universal
translation that the system enables can
open universal knowledge to everyone
including the cook and the housekeeper
and the driver and their children who
dream of becoming doctors and nurses now
I’ll close with a two-minute version of
what could be a two-hour dive into some
very complicated problems first the
question that many of you are dying to
ask doesn’t Google already do that and
the short answer is no they don’t and
the carrot
the basic issue is data for most
languages digitized data doesn’t exist
at all or in any meaningful sense and
where digitized data does exist it
doesn’t line up in any way the confused
across languages briefly and I’ll leave
on a lot of the complexity that explains
how Google has actually got pretty good
at translating between certain languages
in certain situations here’s why they’ll
always fail in 99% of translation
scenarios and how mu C has reengineered
linguistic data so that we can navigate
towards link with language equity now
there are only two million pixels on a
high-res screen and I needed 25 million
pixels to show this to you so I put
this is the universe of human languages
and I’m here are the roughly hundred
languages that are being worked on in
relation to English and here are all the
other languages that they worked on in
relation to each other
without respect to English things like
Chinese to Japanese so this is the void
in knowledge that we’re trying to fill
so what are we doing to try to do this
well most dictionaries start with the
word like light light is a lot of things
it’s not dark it’s not heavy it’s not
serious it’s not factor and in a
monolingual dictionary you can indicate
that by differentiating things with ones
and twos and A’s and B’s and C’s but if
we start adding another language this
French
Swahili and finish Thai trying to
connect them all on the basis of their
English spelling in the middle you
quickly find yourself in a hopeless
tangle that PSP went over so this tangle
something I call linguistic weenie what
are we doing differently we start with
something the concept we treat millions
of concepts not dark not
not serious not bad thing and we connect
them vertically through features that
they do share like spelling or
pluralization or pronunciation but then
we match the concepts to equivalent
words for that idea in French and
Swahili and any other language and in
principle we can extend to a matrix that
connects all 7,000 human languages and
never get confused by the arbitrary
beyond words we can also stitch together
elements like verb conjugations and
dialect variations which was useful
concept here in Switzerland Swiss German
and we can work towards a repository
that combines extensive linguistic data
in complex ways within and between
languages on an equal basis so where
Google says this Arabic words light and
data are d’accord is like and we’re
going to guess which one you want based
on which one statistically appears the
most often in our purpose como si is
saying this arabic word is a device for
ninety fires which is make general in
spanish and shoot eaten and finished and
humidity in Swahili and that Arabic word
is electromagnetic radiation that can
produce a visual sensation which is
loose in Spanish and volatile finish and
longer in Swahili and we can make much
smarter translation and processing
building on these certainties we can
fill a plethora of technological and
social voids that stem from linguistic
inequity and in that detail these on our
webpage and they’re being worked on some
of my students second Lucy labs so for
example we can produce definitions in
languages like a common Sinhala
for the equivalents of electromagnetic
radiation and visual and sensation so
that students in Ghana and sri lanka and
anywhere else and others
the idea of light as energy and past
their business exams eventually we’ll be
able to untangle the like way to
automate pretty good translations
between any languages for which we have
the data and connect languages an
elector level that a racist languages as
a barrier at working towards a full
matrix of human expression across time
and space by recognizing our linguistic
privilege we can act over billions of
people who were slated for perpetual
exclusion from global engines of
knowledge and prosperity by accidents of
did the billions of people are excluded
by accident of being born into the wrong
language this is a wrong that we can
write we can choose legal injury in the
privilege of our linguistic Locke or we
can recognize the inequities assemble
the knowledge and build the tools that
we need to make every language the right
language right now we’ve got 1806
language pairs intimacy for 43 languages
and I’d like you to use those
dictionaries and I have a little card
that you can put them right on their
smartphones and to will be rolling our
tools for people to be able to help
build more and more resources for their
languages so I thank you for your attention
security which is probably everybody in
this room like to imagine that we
deserve all the good things that come
our way but let’s stop for a minute and
think about some of the advantages that
might have taken our powers towards
getting here in my own case I’m not
talking about money I don’t come from a
wealthy family and I wouldn’t have been
able to get my university degrees
without a full package of scholarships
and loans inside jobs but growing up in
a country with the ability and desire to
see its children through higher
education is one of those advantages
that I’m talking about for half of us
being male and it’s given us a big head
start our teachers assume that was
quicker in math and science we’re much
less likely to have to face sexual
harassment and on average we heard about
20% more than our female counterparts
where I grew up in America being white
was a huge advantage over my black and
Hispanic neighbors I was never subjected
to the police stop and for us nobody
ever thought I was going to their store
to shoplift my family was never a red
light for getting housing in
neighborhoods with good schools like
Donald Trump used to do as a wave over
New York and for the 91 percent of us
who are straight we never have to hide
where identities or people we love and
we’re not subjected to job
discrimination or in jail in some
countries these inequities are real and
if we think about race or sex or for
sexual orientation we can readily see
how they might affect the ways that
we’re treated in school the doors that
are open to us are slammed in our faces
the internships the jobs the references
the bank loans that we either get are
denied but these are also things that we
talk about that we have names for and
that people are working to try to
rectify so do you think that the bushes
up there are able to leave countries
because there
because they have better DNA more
capable genes than the kansan teams on
the right or is it because they were
born with a silver spoon in their mouth
with every advantage from prenatal care
right on up
and we stand on the shoulders of our
parents who also probably have a lot of
things advantages and have are able to
pass on to us all these numerous subtle
extras like extracurricular programs or
dental checkers cultural excursions to
museum shows or events like this travel
sufficient and healthy diet professional
connections when I was about 12 I got
this great gift it was a gift
certificate back to my local bookstore
and I spent hours just trying to pick
out the books that would add up to the
money money total on the gift card this
was a gift from my grandparents we’re
educators at who saw it as their duty to
posterity that I should loves learning
decades later my sister built on this
heritage of ours and sends my seven year
old books with positive the messages
that are going to overcome the sexism
that she’s going to encounter in her
life here in Switzerland and these are
more examples of advantages that were
born into but there are also examples of
the advantage that we don’t even have a
word to describe my family could sent me
to the bookstore the library my sister
can send books to her niece because
books are available are plentiful in our
native tongue because of the language
I’ve heard even starting in the wool I
had open to me a world of volcanoes and
Dragons and medieval conquests and
science experiments
my daughter has pilot said painters and
architects and explorers and now with
the internet it’s not just books we can
ask Google but anything we need to know
about the nutritional value of mushrooms
how to calculate the age of that big
tree outside without cutting it down a
good portion of the world knowledge is
available to us because of the language
that we as luck would have it happen to
speak English French speakers are
similarly advantaged and German speakers
and Russians and Japanese Italians
portuguese maybe the correlation doesn’t
prove causation and many additional
factors are at work income life
expectancy and access to knowledge in
one’s native language really do march
very very closely together but we had a
couple of more advantages that we all
enjoy and which we have games for health
care and electricity without electricity
you can’t run a computer or a smartphone
you do your homework by the light of an
oil lamp but this is something I’ve seen
in debt houses that have been in Africa
you can’t keep your food or your
medicines or your vaccines fresh and
safe without health care and it’s all
that’s right it’s so obvious that in
most places it’s illegal to do not deny
somebody health care on the basis of sex
or race or even sexual orientation
yet health care is denied all the time
on the basis of language an immigrant
who can’t speak to a doctor
can’t get proper care people with HIV
can’t follow the instructions on their
drug cocktails if they can’t understand
the languages that the instructions are
written and children can’t become
doctors and nurses and serve their own
people if they can’t understand the
language in which health sciences talk
to them but since linguistic exclusion
is an unnamed disadvantage it marches in
silence listen to a statement from the
Ford Foundation and you can read it
which highlights the advantages that we
gain from an
open Internet in the language that we
can understand the internet is not a
luxury broadband is an essential public
utility and must be equally accessible
to everyone because it enables people to
express themselves organize make a
living participate in policymaking get
rid of job earn an education and do all
of the critical things that we so often
take for granted but typing in the
Internet in the wrong language is like
laying water pipes with no clean water
with an opening water source to draw
from language is a thick glass wall to
an open Internet to political
participation to getting a job to
earning in to earning an education we
recognize that the internet isn’t
central utility like water like
electricity but we still not recognize
that it’s useless to the billions of
people who can’t read and interact them
in the lucky languages the typical
response in Europe in the US and even in
elite circles in Africa in India and
South America is what the learn English
let them learn French like and learn
Spanish we did it at work for us anyone
can learn a second language and who but
the unsuccessful people the little
people can can speak those local
languages well yes
you do speak the privileged language and
you’re successful but that’s either an
accident of birth or a lot of study that
your circumstance is made possible 95
percent of Africans don’t are not
literate the languages that their
government inherited from their European
colonizers and that’s going to be a
quarter of the world’s population by
2030 everybody that I speak to in India
tells me that everybody in India speaks
English and they’re right everyone does
but they’re cook doesn’t and their
housekeeper doesn’t
method and their children don’t and
their grandchildren won’t 95 percent of
Indians don’t and billions of people
whose mother tongues are on the wrong
side of military or economic power are
perpetually excluded for paths towards
knowledge of prosperity because people
in government’s and funding agencies
think the languages that they speak are
worthless if not actual impediments to
progress now despite the indifference or
hostility for many decision-makers we do
have solutions that we’re proposing to
eradicate language is a barrier in
reaching internationally declared goals
for Social Development 496 years ago
Martin Luther figured out that local
language was elemental to participation
and he translated the Bible into German
since then on the back of that a new
technology in that time called the
printing press missionaries have spread
their ideas using local languages in
hundreds and hundreds of languages but
we can’t pretend that moving from ink to
electrons from the printing press of the
tablet is going to magically make
billions of people conversing in one of
the languages that they already
privileged people happen to speak rather
what we’re saying is we can work with
much smaller populations from front from
these languages communities and collect
the information that we need to produce
linguistic equity for their languages
and then we can plug that information
directly into the technologies that have
already been developed for English for
transfer German and the universal
translation that the system enables can
open universal knowledge to everyone
including the cook and the housekeeper
and the driver and their children who
dream of becoming doctors and nurses now
I’ll close with a two-minute version of
what could be a two-hour dive into some
very complicated problems first the
question that many of you are dying to
ask doesn’t Google already do that and
the short answer is no they don’t and
the carrot
the basic issue is data for most
languages digitized data doesn’t exist
at all or in any meaningful sense and
where digitized data does exist it
doesn’t line up in any way the confused
across languages briefly and I’ll leave
on a lot of the complexity that explains
how Google has actually got pretty good
at translating between certain languages
in certain situations here’s why they’ll
always fail in 99% of translation
scenarios and how mu C has reengineered
linguistic data so that we can navigate
towards link with language equity now
there are only two million pixels on a
high-res screen and I needed 25 million
pixels to show this to you so I put
this is the universe of human languages
and I’m here are the roughly hundred
languages that are being worked on in
relation to English and here are all the
other languages that they worked on in
relation to each other
without respect to English things like
Chinese to Japanese so this is the void
in knowledge that we’re trying to fill
so what are we doing to try to do this
well most dictionaries start with the
word like light light is a lot of things
it’s not dark it’s not heavy it’s not
serious it’s not factor and in a
monolingual dictionary you can indicate
that by differentiating things with ones
and twos and A’s and B’s and C’s but if
we start adding another language this
French
Swahili and finish Thai trying to
connect them all on the basis of their
English spelling in the middle you
quickly find yourself in a hopeless
tangle that PSP went over so this tangle
something I call linguistic weenie what
are we doing differently we start with
something the concept we treat millions
of concepts not dark not
not serious not bad thing and we connect
them vertically through features that
they do share like spelling or
pluralization or pronunciation but then
we match the concepts to equivalent
words for that idea in French and
Swahili and any other language and in
principle we can extend to a matrix that
connects all 7,000 human languages and
never get confused by the arbitrary
beyond words we can also stitch together
elements like verb conjugations and
dialect variations which was useful
concept here in Switzerland Swiss German
and we can work towards a repository
that combines extensive linguistic data
in complex ways within and between
languages on an equal basis so where
Google says this Arabic words light and
data are d’accord is like and we’re
going to guess which one you want based
on which one statistically appears the
most often in our purpose como si is
saying this arabic word is a device for
ninety fires which is make general in
spanish and shoot eaten and finished and
humidity in Swahili and that Arabic word
is electromagnetic radiation that can
produce a visual sensation which is
loose in Spanish and volatile finish and
longer in Swahili and we can make much
smarter translation and processing
building on these certainties we can
fill a plethora of technological and
social voids that stem from linguistic
inequity and in that detail these on our
webpage and they’re being worked on some
of my students second Lucy labs so for
example we can produce definitions in
languages like a common Sinhala
for the equivalents of electromagnetic
radiation and visual and sensation so
that students in Ghana and sri lanka and
anywhere else and others
the idea of light as energy and past
their business exams eventually we’ll be
able to untangle the like way to
automate pretty good translations
between any languages for which we have
the data and connect languages an
elector level that a racist languages as
a barrier at working towards a full
matrix of human expression across time
and space by recognizing our linguistic
privilege we can act over billions of
people who were slated for perpetual
exclusion from global engines of
knowledge and prosperity by accidents of
did the billions of people are excluded
by accident of being born into the wrong
language this is a wrong that we can
write we can choose legal injury in the
privilege of our linguistic Locke or we
can recognize the inequities assemble
the knowledge and build the tools that
we need to make every language the right
language right now we’ve got 1806
language pairs intimacy for 43 languages
and I’d like you to use those
dictionaries and I have a little card
that you can put them right on their
smartphones and to will be rolling our
tools for people to be able to help
build more and more resources for their
languages so I thank you for your attention
Please follow and like us: