WEBVTT

00:00:01.102 --> 00:00:06.180
We have historical records that allow us
to know how the ancient Greeks dressed,

00:00:06.180 --> 00:00:07.458
how they lived,

00:00:07.458 --> 00:00:09.159
how they fought ...

00:00:09.159 --> 00:00:11.211
but how did they think?

00:00:11.535 --> 00:00:16.094
One [natural] idea is that the deepest
aspects of human thought --

00:00:16.094 --> 00:00:17.976
our ability to imagine,

00:00:17.976 --> 00:00:19.306
to be conscious,

00:00:19.306 --> 00:00:20.681
to dream --

00:00:20.681 --> 00:00:22.701
have always been the same.

00:00:23.092 --> 00:00:24.601
Another possibility

00:00:24.601 --> 00:00:28.351
is that the social transformations
that have shaped our culture

00:00:28.351 --> 00:00:32.428
may have also changed the structural
columns of human thought.

00:00:33.096 --> 00:00:35.627
We may all have different
opinions about this.

00:00:35.627 --> 00:00:38.344
Actually, it's a long-standing
philosophical debate.

00:00:38.724 --> 00:00:42.335
But is this question
even amenable to science?

00:00:43.116 --> 00:00:45.555
Here I'd like to propose

00:00:45.555 --> 00:00:50.160
that in the same way we can reconstruct
how the ancient Greek cities looked like

00:00:50.160 --> 00:00:52.777
just based on a few bricks,

00:00:52.777 --> 00:00:56.835
that the writings of a culture
are the archeological records --

00:00:56.835 --> 00:00:59.467
the fossils of human thought.

00:01:00.199 --> 00:01:01.390
And in fact,

00:01:01.390 --> 00:01:03.570
doing some form of psychological analysis

00:01:03.570 --> 00:01:07.184
of some of the most ancient
books of human culture,

00:01:07.184 --> 00:01:13.028
Julian Jaynes came in the '70s with
a very wild and radical hypothesis ...

00:01:13.028 --> 00:01:15.444
that only 3,000 years ago,

00:01:15.444 --> 00:01:20.857
humans were what today
we would call schizophrenics.

00:01:21.883 --> 00:01:23.575
And he made this claim

00:01:23.575 --> 00:01:26.752
based on the fact that the first
humans [inscribing] these books

00:01:26.752 --> 00:01:28.792
behaved consistently,


00:01:28.792 --> 00:01:31.708
in different traditions
and in different places of the world,

00:01:31.708 --> 00:01:35.444
as if they were hearing and obeying voices

00:01:35.444 --> 00:01:38.394
that they perceived
as coming from the Gods,

00:01:38.394 --> 00:01:40.413
or from the muses --

00:01:40.413 --> 00:01:43.893
what today we would call hallucinations.

00:01:44.102 --> 00:01:45.476
And only then,

00:01:45.476 --> 00:01:46.780
as time went on,

00:01:46.780 --> 00:01:50.361
they began to recognize
that they were the creators --

00:01:50.361 --> 00:01:53.078
the owners of these inner voices.

00:01:53.510 --> 00:01:56.273
And with this they gained introspection:

00:01:56.273 --> 00:01:59.221
the ability to think
about their own thoughts.

00:01:59.965 --> 00:02:03.378
So Jaynes' theory is that consciousness --

00:02:03.378 --> 00:02:06.538
at least in the way we perceive it today,

00:02:06.538 --> 00:02:10.225
where we feel that we are the pilots
of our own existence --

00:02:10.225 --> 00:02:13.164
is a quite recent cultural development.

00:02:13.545 --> 00:02:15.337
And this theory is quite spectacular,

00:02:15.337 --> 00:02:16.769
but it has an obvious problem

00:02:16.769 --> 00:02:20.761
which is that it's built on just a few
and very specific examples.

00:02:21.085 --> 00:02:22.872
So the question is whether the theory

00:02:22.872 --> 00:02:27.913
that introspection built up in human
history only about 3,000 years ago

00:02:27.913 --> 00:02:31.283
can be examined in a quantitative
and objective manner.

00:02:31.779 --> 00:02:35.338
And the problem on how
to go about this is quite obvious.

00:02:35.338 --> 00:02:37.724
It's not like Plato woke up one day

00:02:37.724 --> 00:02:38.910
and then he wrote,

00:02:38.910 --> 00:02:40.596
"Hello, I'm Plato

00:02:40.596 --> 00:02:43.487
and as of today I have a fully
introspective consciousness."

00:02:43.487 --> 00:02:45.307
(Laughter)

00:02:45.637 --> 00:02:49.097
And this still is actually
what is the essence of the problem.

00:02:49.624 --> 00:02:54.071
We need to find the emergence
of a concept that's never said.

00:02:54.680 --> 00:02:58.942
The word introspection
does not appear a single time

00:02:58.942 --> 00:03:01.559
in the books we want to analyze.

00:03:01.971 --> 00:03:06.398
So our way to solve this
is to build the space of words.

00:03:06.794 --> 00:03:09.998
This is a huge space
that contains all words

00:03:09.998 --> 00:03:12.959
in such a way that they distance
between any two of them

00:03:12.959 --> 00:03:15.842
is indicative of how
closely related they are.

00:03:16.460 --> 00:03:17.456
So for instance,

00:03:17.456 --> 00:03:20.857
you want the words dog and cat
to be very close together,

00:03:20.857 --> 00:03:24.688
but the words grapefruit and logarithm
to be very far away.

00:03:25.008 --> 00:03:29.298
And this has to be true for any
two words within the space.

00:03:29.748 --> 00:03:33.109
And there are different ways that we
can construct the space of words.

00:03:33.109 --> 00:03:34.802
One is just asking the experts,

00:03:34.802 --> 00:03:37.199
a bit like we do with dictionaries.

00:03:37.199 --> 00:03:38.623
Another possibility

00:03:38.623 --> 00:03:40.688
is following the simple assumption

00:03:40.688 --> 00:03:44.640
that when two words are related
they tend to appear in the same sentences,

00:03:44.640 --> 00:03:46.214
in the same paragraphs,

00:03:46.214 --> 00:03:48.049
in the same documents,

00:03:48.049 --> 00:03:51.509
more often than would be expected
just by pure chance.

00:03:52.448 --> 00:03:54.305
And this simple hypothesis,

00:03:54.305 --> 00:03:55.792
this simple method,

00:03:55.792 --> 00:03:57.266
with some computational tricks

00:03:57.266 --> 00:03:58.679
that have to do with the fact

00:03:58.679 --> 00:04:01.995
that this is a very complex
and highly dimensional space,

00:04:01.995 --> 00:04:04.458
turns out to be quite effective.

00:04:04.458 --> 00:04:07.127
And just to give you a flavor
of how well this works,

00:04:07.127 --> 00:04:11.321
this is the result we get when
we analyze this for some familiar words.

00:04:11.607 --> 00:04:12.816
And you can see first

00:04:12.816 --> 00:04:16.278
that words automatically organize
into semantic neighborhoods.

00:04:16.278 --> 00:04:17.362
So you get the fruits,

00:04:17.362 --> 00:04:18.359
the body parts,

00:04:18.359 --> 00:04:19.360
the computer parts,

00:04:19.360 --> 00:04:20.359
the scientific terms

00:04:20.359 --> 00:04:21.357
and so on.

00:04:21.357 --> 00:04:25.699
The algorithm also identifies
the reorganized concepts in a hierarchy.

00:04:26.027 --> 00:04:27.027
So for instance,

00:04:27.027 --> 00:04:30.648
you can see that the scientific terms
break down into two subcategories

00:04:30.648 --> 00:04:33.608
of the astronomic and the physic terms.

00:04:33.608 --> 00:04:35.881
And then there are very fine things.

00:04:35.881 --> 00:04:36.878
For instance,

00:04:36.878 --> 00:04:38.054
the word astronomy,

00:04:38.056 --> 00:04:39.870
which seems a bit bizarre where it is,

00:04:39.870 --> 00:04:41.768
is actually exactly where it should be,

00:04:41.768 --> 00:04:43.219
between what it is --

00:04:43.219 --> 00:04:44.630
an actual science --

00:04:44.630 --> 00:04:46.165
and between what it describes --

00:04:46.165 --> 00:04:47.912
the astronomical terms.

00:04:48.366 --> 00:04:50.097
And we could go on and on with this.

00:04:50.097 --> 00:04:52.181
Actually if you stare at this for a while

00:04:52.181 --> 00:04:54.034
and you just build random trajectories,

00:04:54.034 --> 00:04:55.725
you will see that is feels well --

00:04:55.725 --> 00:04:58.286
actually it feels a bit like doing poetry.

00:04:58.286 --> 00:04:59.286
And this is because,

00:04:59.286 --> 00:05:00.288
in a way,

00:05:00.288 --> 00:05:03.482
walking in this space
is like walking in the mind.

00:05:03.901 --> 00:05:05.778
And the last thing

00:05:05.778 --> 00:05:09.938
is that this algorithm identifies
what are our intuitions

00:05:09.938 --> 00:05:14.002
of which words should lead
in the neighborhood of introspection.

00:05:14.002 --> 00:05:15.072
So for instance,

00:05:15.072 --> 00:05:18.982
words such as Self, Guilt,
Reason, Emotion,

00:05:18.982 --> 00:05:21.102
are very close to introspection,

00:05:21.102 --> 00:05:22.102
but other words,

00:05:22.102 --> 00:05:24.432
such as Red, Football, Candle, Banana,

00:05:24.432 --> 00:05:26.072
are just very far away.

00:05:26.262 --> 00:05:28.882
And so once we've built the space,

00:05:28.882 --> 00:05:31.945
the question of the history
of introspection,

00:05:31.945 --> 00:05:34.277
or of the history of any concept

00:05:34.277 --> 00:05:39.055
which before could seem abstract
and somehow vague,

00:05:39.055 --> 00:05:40.754
becomes concrete --

00:05:40.754 --> 00:05:43.656
becomes amenable to quantitative science.

00:05:44.481 --> 00:05:47.137
All that we have to do is take the books,

00:05:47.137 --> 00:05:48.612
we digitize them

00:05:48.612 --> 00:05:51.420
and we take this stream
of words as a trajectory

00:05:51.420 --> 00:05:53.452
and project them into the space,

00:05:53.452 --> 00:05:57.087
and then we ask whether this trajectory
spends significant time

00:05:57.087 --> 00:06:00.361
circling closely to the concept
of introspection.

00:06:00.911 --> 00:06:02.262
And with this,

00:06:02.262 --> 00:06:04.263
we could analyze
the history of introspection

00:06:04.263 --> 00:06:06.183
in the ancient Greek tradition,

00:06:06.183 --> 00:06:09.044
for which we have the best
available written record.

00:06:09.761 --> 00:06:12.131
So what we did is we took all the books --

00:06:12.131 --> 00:06:14.517
we just ordered them by time --

00:06:14.517 --> 00:06:15.934
for each book we take the words

00:06:15.934 --> 00:06:18.194
and we project them to the space,

00:06:18.194 --> 00:06:20.765
and then we ask for each word
how close it is to introspection,

00:06:20.765 --> 00:06:22.728
and we just average that.

00:06:22.728 --> 00:06:25.986
And then we understand
that as time goes on and on,

00:06:25.986 --> 00:06:29.088
these books get closer,
and closer and closer

00:06:29.088 --> 00:06:31.062
to the concept of introspection.

00:06:31.062 --> 00:06:35.387
And this is exactly what happens
in the ancient Greek tradition.

99:59:59.999 --> 99:59:59.999
So you can see that for the oldest books
in the Homeric tradition,

99:59:59.999 --> 99:59:59.999
there is a small increase with books
getting closer to introspection,

99:59:59.999 --> 99:59:59.999
but about four centuries before Christ,

99:59:59.999 --> 99:59:59.999
this starts ramping up very rapidly

99:59:59.999 --> 99:59:59.999
to an almost five-fold increase

99:59:59.999 --> 99:59:59.999
of books getting closer,
and closer and closer

99:59:59.999 --> 99:59:59.999
to the concept of introspection.

99:59:59.999 --> 99:59:59.999
And one of the nice things about this

99:59:59.999 --> 99:59:59.999
is that now we can ask

99:59:59.999 --> 99:59:59.999
whether this is also true
in a different independent tradition.

99:59:59.999 --> 99:59:59.999
So we just ran this same analysis
on the Judeo Christian tradition,

99:59:59.999 --> 99:59:59.999
and we got virtually the same pattern.

99:59:59.999 --> 99:59:59.999
Again you see a small increase
for the oldest books in the old testament,

99:59:59.999 --> 99:59:59.999
and then it increases much more rapidly

99:59:59.999 --> 99:59:59.999
in the new books of the new testament,

99:59:59.999 --> 99:59:59.999
and then we get the peak of introspection

99:59:59.999 --> 99:59:59.999
in the work Confessions
of Saint Augustine,

99:59:59.999 --> 99:59:59.999
about four centuries after Christ.

99:59:59.999 --> 99:59:59.999
And this was very important,

99:59:59.999 --> 99:59:59.999
because Saint Augustine
had been recognized

99:59:59.999 --> 99:59:59.999
by scholars, philologists, historians,

99:59:59.999 --> 99:59:59.999
as one of the founders of introspection.

99:59:59.999 --> 99:59:59.999
Actually, some believe him to be
the father of modern psychology.

99:59:59.999 --> 99:59:59.999
So our algorithm --

99:59:59.999 --> 99:59:59.999
which has the virtue
of being quantitative,

99:59:59.999 --> 99:59:59.999
of being objective,

99:59:59.999 --> 99:59:59.999
and of course of being extremely fast,

99:59:59.999 --> 99:59:59.999
it just runs in a fraction of a second --

99:59:59.999 --> 99:59:59.999
can capture some of the most
important conclusions

99:59:59.999 --> 99:59:59.999
of this long tradition of investigation.

99:59:59.999 --> 99:59:59.999
And this is in a way,
one of the beauties of science,

99:59:59.999 --> 99:59:59.999
which is that now this idea can translated

99:59:59.999 --> 99:59:59.999
and generalized to a whole
lot of different domains.

99:59:59.999 --> 99:59:59.999
So in the same way that we asked
about the past of human conciousness,

99:59:59.999 --> 99:59:59.999
maybe the most challenging question
we can pose to ourselves,

99:59:59.999 --> 99:59:59.999
is whether this can tell us something
about the future of our unconciousness.

99:59:59.999 --> 99:59:59.999
To put it more precisely,

99:59:59.999 --> 99:59:59.999
whether the words we say today

99:59:59.999 --> 99:59:59.999
can tell us something of where
our minds will be in a few days,

99:59:59.999 --> 99:59:59.999
in a few months,

99:59:59.999 --> 99:59:59.999
or a few years from now.

99:59:59.999 --> 99:59:59.999
And in the way many of us
are now wearing censors

99:59:59.999 --> 99:59:59.999
that detect our heart rate,

99:59:59.999 --> 99:59:59.999
our respiration,

99:59:59.999 --> 99:59:59.999
our genes,

99:59:59.999 --> 99:59:59.999
on the hopes that this may
help us prevent diseases,

99:59:59.999 --> 99:59:59.999
we can ask whether monitoring
and analyzing the words we speak --

99:59:59.999 --> 99:59:59.999
we tweet,

99:59:59.999 --> 99:59:59.999
we email,

99:59:59.999 --> 99:59:59.999
we write --

99:59:59.999 --> 99:59:59.999
can tell us ahead of time whether
something will go wrong with our minds.

99:59:59.999 --> 99:59:59.999
And with Guillermo Cecci,

99:59:59.999 --> 99:59:59.999
who has been my brother
in this adventure,

99:59:59.999 --> 99:59:59.999
we took on this task.

99:59:59.999 --> 99:59:59.999
And we did so by analyzing the recorded
speech for 44 young people

99:59:59.999 --> 99:59:59.999
who were at a high risk
of developing schizophenia.

99:59:59.999 --> 99:59:59.999
And so what we did is we
measured speech at day one

99:59:59.999 --> 99:59:59.999
and then we asked

99:59:59.999 --> 99:59:59.999
whether the properties
of the speech could predict --

99:59:59.999 --> 99:59:59.999
within a window of almost three years --

99:59:59.999 --> 99:59:59.999
the future development of psychosis.

99:59:59.999 --> 99:59:59.999
But despite our hopes,

99:59:59.999 --> 99:59:59.999
we got failure after failure.

99:59:59.999 --> 99:59:59.999
There was just not enough
information in semantics

99:59:59.999 --> 99:59:59.999
to predict the future
organization of the mind.

99:59:59.999 --> 99:59:59.999
It was good enough

99:59:59.999 --> 99:59:59.999
to distinguish between a group
of schizophrenics and a control group,

99:59:59.999 --> 99:59:59.999
a bit like we had done
for the ancient texts,

99:59:59.999 --> 99:59:59.999
but not to predict the future
onto the psychosis.

99:59:59.999 --> 99:59:59.999
But then we realized

99:59:59.999 --> 99:59:59.999
that maybe the most important thing
was not so much what they were saying

99:59:59.999 --> 99:59:59.999
but how they were saying it.

99:59:59.999 --> 99:59:59.999
More specifically,

99:59:59.999 --> 99:59:59.999
it was not in which semantic
neighborhoods the words were,

99:59:59.999 --> 99:59:59.999
but how far and fast they jumped
from one semantic neighborhood to another.

99:59:59.999 --> 99:59:59.999
And so we came up with this measure,

99:59:59.999 --> 99:59:59.999
which we termed Semantic Coherence,

99:59:59.999 --> 99:59:59.999
which essentially measures

99:59:59.999 --> 99:59:59.999
the persistence of speech
within one semantic topic,

99:59:59.999 --> 99:59:59.999
within one semantic category.

99:59:59.999 --> 99:59:59.999
And it turned out to be

99:59:59.999 --> 99:59:59.999
that for this group of 44 people,

99:59:59.999 --> 99:59:59.999
the algorithm based on semantic cohernece

99:59:59.999 --> 99:59:59.999
could predict with 100 percent accuracy

99:59:59.999 --> 99:59:59.999
who developed psychosis and who will not.

99:59:59.999 --> 99:59:59.999
And this was something
that could not be achieved --

99:59:59.999 --> 99:59:59.999
not even close --

99:59:59.999 --> 99:59:59.999
with all the other
existing clinical measures.

99:59:59.999 --> 99:59:59.999
And I remember vividly
while I was working on this,

99:59:59.999 --> 99:59:59.999
I was sitting on my computer

99:59:59.999 --> 99:59:59.999
and I saw a bunch of tweets by Polo.

99:59:59.999 --> 99:59:59.999
Polo has been my first student
back in Buenos Aires

99:59:59.999 --> 99:59:59.999
and at the time he was living in New York.

99:59:59.999 --> 99:59:59.999
And there was something in this tweet --

99:59:59.999 --> 99:59:59.999
I could not tell exactly what
because nothing was said explicitly --

99:59:59.999 --> 99:59:59.999
but I got this strong hunch,

99:59:59.999 --> 99:59:59.999
this strong intuition
that something was going wrong.

99:59:59.999 --> 99:59:59.999
So I picked up the phone
and I called Polo,

99:59:59.999 --> 99:59:59.999
and he was not feeling well.

99:59:59.999 --> 99:59:59.999
And this simple fact that reading
in between the lines

99:59:59.999 --> 99:59:59.999
I could sense through words
his feelings,

99:59:59.999 --> 99:59:59.999
was a simple but very
effective way to help.

99:59:59.999 --> 99:59:59.999
What I tell you today

99:59:59.999 --> 99:59:59.999
is that we're getting
close to understanding

99:59:59.999 --> 99:59:59.999
how we can convert this intuition
that we all have,

99:59:59.999 --> 99:59:59.999
that we all share,

99:59:59.999 --> 99:59:59.999
into an algorithm.

99:59:59.999 --> 99:59:59.999
And in doing so,

99:59:59.999 --> 99:59:59.999
we may be seeing in the future
a very different form of mental health,

99:59:59.999 --> 99:59:59.999
based on objective, quantitative
and automated analysis

99:59:59.999 --> 99:59:59.999
of the words we write,

99:59:59.999 --> 99:59:59.999
of the words we say.

99:59:59.999 --> 99:59:59.999
Gracias.

99:59:59.999 --> 99:59:59.999
(Applause)