0:00:00.000,0:00:03.000
I'd like to begin with a thought experiment.

0:00:04.000,0:00:07.000
Imagine that it's 4,000 years into the future.

0:00:07.000,0:00:09.000
Civilization as we know it

0:00:09.000,0:00:11.000
has ceased to exist --

0:00:11.000,0:00:13.000
no books,

0:00:13.000,0:00:16.000
no electronic devices,

0:00:16.000,0:00:19.000
no Facebook or Twitter.

0:00:19.000,0:00:22.000
All knowledge of the English language and the English alphabet

0:00:22.000,0:00:24.000
has been lost.

0:00:24.000,0:00:26.000
Now imagine archeologists

0:00:26.000,0:00:28.000
digging through the rubble of one of our cities.

0:00:28.000,0:00:30.000
What might they find?

0:00:30.000,0:00:33.000
Well perhaps some rectangular pieces of plastic

0:00:33.000,0:00:36.000
with strange symbols on them.

0:00:36.000,0:00:39.000
Perhaps some circular pieces of metal.

0:00:39.000,0:00:41.000
Maybe some cylindrical containers

0:00:41.000,0:00:43.000
with some symbols on them.

0:00:43.000,0:00:46.000
And perhaps one archeologist becomes an instant celebrity

0:00:46.000,0:00:48.000
when she discovers --

0:00:48.000,0:00:50.000
buried in the hills somewhere in North America --

0:00:50.000,0:00:53.000
massive versions of these same symbols.

0:00:55.000,0:00:57.000
Now let's ask ourselves,

0:00:57.000,0:01:00.000
what could such artifacts say about us

0:01:00.000,0:01:03.000
to people 4,000 years into the future?

0:01:03.000,0:01:05.000
This is no hypothetical question.

0:01:05.000,0:01:08.000
In fact, this is exactly the kind of question we're faced with

0:01:08.000,0:01:11.000
when we try to understand the Indus Valley civilization,

0:01:11.000,0:01:13.000
which existed 4,000 years ago.

0:01:13.000,0:01:16.000
The Indus civilization was roughly contemporaneous

0:01:16.000,0:01:19.000
with the much better known Egyptian and the Mesopotamian civilizations,

0:01:19.000,0:01:22.000
but it was actually much larger than either of these two civilizations.

0:01:22.000,0:01:24.000
It occupied the area

0:01:24.000,0:01:26.000
of approximately one million square kilometers,

0:01:26.000,0:01:28.000
covering what is now Pakistan,

0:01:28.000,0:01:30.000
Northwestern India

0:01:30.000,0:01:32.000
and parts of Afghanistan and Iran.

0:01:32.000,0:01:34.000
Given that it was such a vast civilization,

0:01:34.000,0:01:38.000
you might expect to find really powerful rulers, kings,

0:01:38.000,0:01:41.000
and huge monuments glorifying these powerful kings.

0:01:41.000,0:01:43.000
In fact,

0:01:43.000,0:01:45.000
what archeologists have found is none of that.

0:01:45.000,0:01:48.000
They've found small objects such as these.

0:01:48.000,0:01:51.000
Here's an example of one of these objects.

0:01:51.000,0:01:53.000
Well obviously this is a replica.

0:01:53.000,0:01:56.000
But who is this person?

0:01:56.000,0:01:58.000
A king? A god?

0:01:58.000,0:02:00.000
A priest?

0:02:00.000,0:02:02.000
Or perhaps an ordinary person

0:02:02.000,0:02:04.000
like you or me?

0:02:04.000,0:02:06.000
We don't know.

0:02:06.000,0:02:09.000
But the Indus people also left behind artifacts with writing on them.

0:02:09.000,0:02:11.000
Well no, not pieces of plastic,

0:02:11.000,0:02:14.000
but stone seals, copper tablets,

0:02:14.000,0:02:16.000
pottery and, surprisingly,

0:02:16.000,0:02:18.000
one large sign board,

0:02:18.000,0:02:20.000
which was found buried near the gate of a city.

0:02:20.000,0:02:22.000
Now we don't know if it says Hollywood,

0:02:22.000,0:02:24.000
or even Bollywood for that matter.

0:02:24.000,0:02:26.000
In fact, we don't even know

0:02:26.000,0:02:28.000
what any of these objects say,

0:02:28.000,0:02:31.000
and that's because the Indus script is undeciphered.

0:02:31.000,0:02:33.000
We don't know what any of these symbols mean.

0:02:33.000,0:02:36.000
The symbols are most commonly found on seals.

0:02:36.000,0:02:38.000
So you see up there one such object.

0:02:38.000,0:02:41.000
It's the square object with the unicorn-like animal on it.

0:02:41.000,0:02:43.000
Now that's a magnificent piece of art.

0:02:43.000,0:02:45.000
So how big do you think that is?

0:02:45.000,0:02:47.000
Perhaps that big?

0:02:47.000,0:02:49.000
Or maybe that big?

0:02:49.000,0:02:51.000
Well let me show you.

0:02:52.000,0:02:55.000
Here's a replica of one such seal.

0:02:55.000,0:02:57.000
It's only about one inch by one inch in size --

0:02:57.000,0:02:59.000
pretty tiny.

0:02:59.000,0:03:01.000
So what were these used for?

0:03:01.000,0:03:04.000
We know that these were used for stamping clay tags

0:03:04.000,0:03:07.000
that were attached to bundles of goods that were sent from one place to the other.

0:03:07.000,0:03:10.000
So you know those packing slips you get on your FedEx boxes?

0:03:10.000,0:03:13.000
These were used to make those kinds of packing slips.

0:03:13.000,0:03:16.000
You might wonder what these objects contain

0:03:16.000,0:03:18.000
in terms of their text.

0:03:18.000,0:03:20.000
Perhaps they're the name of the sender

0:03:20.000,0:03:22.000
or some information about the goods

0:03:22.000,0:03:25.000
that are being sent from one place to the other -- we don't know.

0:03:25.000,0:03:27.000
We need to decipher the script to answer that question.

0:03:27.000,0:03:29.000
Deciphering the script

0:03:29.000,0:03:31.000
is not just an intellectual puzzle;

0:03:31.000,0:03:33.000
it's actually become a question

0:03:33.000,0:03:35.000
that's become deeply intertwined

0:03:35.000,0:03:38.000
with the politics and the cultural history of South Asia.

0:03:38.000,0:03:41.000
In fact, the script has become a battleground of sorts

0:03:41.000,0:03:43.000
between three different groups of people.

0:03:43.000,0:03:45.000
First, there's a group of people

0:03:45.000,0:03:47.000
who are very passionate in their belief

0:03:47.000,0:03:49.000
that the Indus script

0:03:49.000,0:03:51.000
does not represent a language at all.

0:03:51.000,0:03:53.000
These people believe that the symbols

0:03:53.000,0:03:56.000
are very similar to the kind of symbols you find on traffic signs

0:03:56.000,0:03:59.000
or the emblems you find on shields.

0:03:59.000,0:04:01.000
There's a second group of people

0:04:01.000,0:04:04.000
who believe that the Indus script represents an Indo-European language.

0:04:04.000,0:04:06.000
If you look at a map of India today,

0:04:06.000,0:04:09.000
you'll see that most of the languages spoken in North India

0:04:09.000,0:04:12.000
belong to the Indo-European language family.

0:04:12.000,0:04:14.000
So some people believe that the Indus script

0:04:14.000,0:04:17.000
represents an ancient Indo-European language such as Sanskrit.

0:04:17.000,0:04:19.000
There's a last group of people

0:04:19.000,0:04:22.000
who believe that the Indus people

0:04:22.000,0:04:25.000
were the ancestors of people living in South India today.

0:04:25.000,0:04:27.000
These people believe that the Indus script

0:04:27.000,0:04:29.000
represents an ancient form

0:04:29.000,0:04:31.000
of the Dravidian language family,

0:04:31.000,0:04:34.000
which is the language family spoken in much of South India today.

0:04:34.000,0:04:36.000
And the proponents of this theory

0:04:36.000,0:04:39.000
point to that small pocket of Dravidian-speaking people in the North,

0:04:39.000,0:04:41.000
actually near Afghanistan,

0:04:41.000,0:04:44.000
and they say that perhaps, sometime in the past,

0:04:44.000,0:04:47.000
Dravidian languages were spoken all over India

0:04:47.000,0:04:49.000
and that this suggests

0:04:49.000,0:04:52.000
that the Indus civilization is perhaps also Dravidian.

0:04:52.000,0:04:55.000
Which of these hypotheses can be true?

0:04:55.000,0:04:57.000
We don't know, but perhaps if you deciphered the script,

0:04:57.000,0:04:59.000
you would be able to answer this question.

0:04:59.000,0:05:01.000
But deciphering the script is a very challenging task.

0:05:01.000,0:05:03.000
First, there's no Rosetta Stone.

0:05:03.000,0:05:05.000
I don't mean the software;

0:05:05.000,0:05:07.000
I mean an ancient artifact

0:05:07.000,0:05:09.000
that contains in the same text

0:05:09.000,0:05:12.000
both a known text and an unknown text.

0:05:12.000,0:05:15.000
We don't have such an artifact for the Indus script.

0:05:15.000,0:05:18.000
And furthermore, we don't even know what language they spoke.

0:05:18.000,0:05:20.000
And to make matters even worse,

0:05:20.000,0:05:22.000
most of the text that we have are extremely short.

0:05:22.000,0:05:24.000
So as I showed you, they're usually found on these seals

0:05:24.000,0:05:26.000
that are very, very tiny.

0:05:26.000,0:05:28.000
And so given these formidable obstacles,

0:05:28.000,0:05:30.000
one might wonder and worry

0:05:30.000,0:05:33.000
whether one will ever be able to decipher the Indus script.

0:05:33.000,0:05:35.000
In the rest of my talk,

0:05:35.000,0:05:37.000
I'd like to tell you about how I learned to stop worrying

0:05:37.000,0:05:39.000
and love the challenge posed by the Indus script.

0:05:39.000,0:05:42.000
I've always been fascinated by the Indus script

0:05:42.000,0:05:44.000
ever since I read about it in a middle school textbook.

0:05:44.000,0:05:46.000
And why was I fascinated?

0:05:46.000,0:05:50.000
Well it's the last major undeciphered script in the ancient world.

0:05:50.000,0:05:53.000
My career path led me to become a computational neuroscientist,

0:05:53.000,0:05:55.000
so in my day job,

0:05:55.000,0:05:57.000
I create computer models of the brain

0:05:57.000,0:06:00.000
to try to understand how the brain makes predictions,

0:06:00.000,0:06:02.000
how the brain makes decisions,

0:06:02.000,0:06:04.000
how the brain learns and so on.

0:06:04.000,0:06:07.000
But in 2007, my path crossed again with the Indus script.

0:06:07.000,0:06:09.000
That's when I was in India,

0:06:09.000,0:06:11.000
and I had the wonderful opportunity

0:06:11.000,0:06:13.000
to meet with some Indian scientists

0:06:13.000,0:06:16.000
who were using computer models to try to analyze the script.

0:06:16.000,0:06:18.000
And so it was then that I realized

0:06:18.000,0:06:21.000
there was an opportunity for me to collaborate with these scientists,

0:06:21.000,0:06:23.000
and so I jumped at that opportunity.

0:06:23.000,0:06:25.000
And I'd like to describe some of the results that we have found.

0:06:25.000,0:06:28.000
Or better yet, let's all collectively decipher.

0:06:28.000,0:06:30.000
Are you ready?

0:06:30.000,0:06:33.000
The first thing that you need to do when you have an undeciphered script

0:06:33.000,0:06:35.000
is try to figure out the direction of writing.

0:06:35.000,0:06:38.000
Here are two texts that contain some symbols on them.

0:06:38.000,0:06:40.000
Can you tell me

0:06:40.000,0:06:43.000
if the direction of writing is right to left or left to right?

0:06:43.000,0:06:46.000
I'll give you a couple of seconds.

0:06:46.000,0:06:49.000
Okay. Right to left, how many? Okay.

0:06:49.000,0:06:51.000
Okay. Left to right?

0:06:51.000,0:06:53.000
Oh, it's almost 50/50. Okay.

0:06:53.000,0:06:55.000
The answer is:

0:06:55.000,0:06:57.000
if you look at the left-hand side of the two texts,

0:06:57.000,0:07:00.000
you'll notice that there's a cramping of signs,

0:07:00.000,0:07:02.000
and it seems like 4,000 years ago,

0:07:02.000,0:07:04.000
when the scribe was writing from right to left,

0:07:04.000,0:07:06.000
they ran out of space.

0:07:06.000,0:07:08.000
And so they had to cram the sign.

0:07:08.000,0:07:10.000
One of the signs is also below the text on the top.

0:07:10.000,0:07:12.000
This suggests the direction of writing

0:07:12.000,0:07:14.000
was probably from right to left,

0:07:14.000,0:07:16.000
and so that's one of the first things we know,

0:07:16.000,0:07:19.000
that directionality is a very key aspect of linguistic scripts.

0:07:19.000,0:07:21.000
And the Indus script now has

0:07:21.000,0:07:23.000
this particular property.

0:07:23.000,0:07:25.000
What other properties of language does the script show?

0:07:25.000,0:07:27.000
Languages contain patterns.

0:07:27.000,0:07:29.000
If I give you the letter Q

0:07:29.000,0:07:32.000
and ask you to predict the next letter, what do you think that would be?

0:07:32.000,0:07:34.000
Most of you said U, which is right.

0:07:34.000,0:07:36.000
Now if I asked you to predict one more letter,

0:07:36.000,0:07:38.000
what do you think that would be?

0:07:38.000,0:07:41.000
Now there's several thoughts. There's E. It could be I. It could be A,

0:07:41.000,0:07:44.000
but certainly not B, C or D, right?

0:07:44.000,0:07:47.000
The Indus script also exhibits similar kinds of patterns.

0:07:47.000,0:07:50.000
There's a lot of text that start with this diamond-shaped symbol.

0:07:50.000,0:07:52.000
And this in turn tends to be followed

0:07:52.000,0:07:54.000
by this quotation marks-like symbol.

0:07:54.000,0:07:56.000
And this is very similar to a Q and U example.

0:07:56.000,0:07:58.000
This symbol can in turn be followed

0:07:58.000,0:08:01.000
by these fish-like symbols and some other signs,

0:08:01.000,0:08:03.000
but never by these other signs at the bottom.

0:08:03.000,0:08:05.000
And furthermore, there's some signs

0:08:05.000,0:08:07.000
that really prefer the end of texts,

0:08:07.000,0:08:09.000
such as this jar-shaped sign,

0:08:09.000,0:08:11.000
and this sign, in fact, happens to be

0:08:11.000,0:08:13.000
the most frequently occurring sign in the script.

0:08:13.000,0:08:16.000
Given such patterns, here was our idea.

0:08:16.000,0:08:18.000
The idea was to use a computer

0:08:18.000,0:08:20.000
to learn these patterns,

0:08:20.000,0:08:23.000
and so we gave the computer the existing texts.

0:08:23.000,0:08:25.000
And the computer learned a statistical model

0:08:25.000,0:08:27.000
of which symbols tend to occur together

0:08:27.000,0:08:29.000
and which symbols tend to follow each other.

0:08:29.000,0:08:31.000
Given the computer model,

0:08:31.000,0:08:34.000
we can test the model by essentially quizzing it.

0:08:34.000,0:08:36.000
So we could deliberately erase some symbols,

0:08:36.000,0:08:39.000
and we can ask it to predict the missing symbols.

0:08:39.000,0:08:42.000
Here are some examples.

0:08:45.000,0:08:47.000
You may regard this

0:08:47.000,0:08:49.000
as perhaps the most ancient game

0:08:49.000,0:08:52.000
of Wheel of Fortune.

0:08:53.000,0:08:55.000
What we found

0:08:55.000,0:08:57.000
was that the computer was successful in 75 percent of the cases

0:08:57.000,0:08:59.000
in predicting the correct symbol.

0:08:59.000,0:09:01.000
In the rest of the cases,

0:09:01.000,0:09:04.000
typically the second best guess or third best guess was the right answer.

0:09:04.000,0:09:06.000
There's also practical use

0:09:06.000,0:09:08.000
for this particular procedure.

0:09:08.000,0:09:10.000
There's a lot of these texts that are damaged.

0:09:10.000,0:09:12.000
Here's an example of one such text.

0:09:12.000,0:09:15.000
And we can use the computer model now to try to complete this text

0:09:15.000,0:09:17.000
and make a best guess prediction.

0:09:17.000,0:09:20.000
Here's an example of a symbol that was predicted.

0:09:20.000,0:09:22.000
And this could be really useful as we try to decipher the script

0:09:22.000,0:09:25.000
by generating more data that we can analyze.

0:09:25.000,0:09:28.000
Now here's one other thing you can do with the computer model.

0:09:28.000,0:09:30.000
So imagine a monkey

0:09:30.000,0:09:32.000
sitting at a keyboard.

0:09:32.000,0:09:35.000
I think you might get a random jumble of letters that looks like this.

0:09:35.000,0:09:37.000
Such a random jumble of letters

0:09:37.000,0:09:39.000
is said to have a very high entropy.

0:09:39.000,0:09:41.000
This is a physics and information theory term.

0:09:41.000,0:09:44.000
But just imagine it's a really random jumble of letters.

0:09:44.000,0:09:48.000
How many of you have ever spilled coffee on a keyboard?

0:09:48.000,0:09:50.000
You might have encountered the stuck-key problem --

0:09:50.000,0:09:53.000
so basically the same symbol being repeated over and over again.

0:09:53.000,0:09:56.000
This kind of a sequence is said to have a very low entropy

0:09:56.000,0:09:58.000
because there's no variation at all.

0:09:58.000,0:10:01.000
Language, on the other hand, has an intermediate level of entropy;

0:10:01.000,0:10:03.000
it's neither too rigid,

0:10:03.000,0:10:05.000
nor is it too random.

0:10:05.000,0:10:07.000
What about the Indus script?

0:10:07.000,0:10:11.000
Here's a graph that plots the entropies of a whole bunch of sequences.

0:10:11.000,0:10:13.000
At the very top you find the uniformly random sequence,

0:10:13.000,0:10:15.000
which is a random jumble of letters --

0:10:15.000,0:10:17.000
and interestingly, we also find

0:10:17.000,0:10:20.000
the DNA sequence from the human genome and instrumental music.

0:10:20.000,0:10:22.000
And both of these are very, very flexible,

0:10:22.000,0:10:24.000
which is why you find them in the very high range.

0:10:24.000,0:10:26.000
At the lower end of the scale,

0:10:26.000,0:10:28.000
you find a rigid sequence, a sequence of all A's,

0:10:28.000,0:10:30.000
and you also find a computer program,

0:10:30.000,0:10:32.000
in this case in the language Fortran,

0:10:32.000,0:10:34.000
which obeys really strict rules.

0:10:34.000,0:10:36.000
Linguistic scripts

0:10:36.000,0:10:38.000
occupy the middle range.

0:10:38.000,0:10:40.000
Now what about the Indus script?

0:10:40.000,0:10:42.000
We found that the Indus script

0:10:42.000,0:10:44.000
actually falls within the range of the linguistic scripts.

0:10:44.000,0:10:46.000
When this result was first published,

0:10:46.000,0:10:49.000
it was highly controversial.

0:10:49.000,0:10:52.000
There were people who raised a hue and cry,

0:10:52.000,0:10:54.000
and these people were the ones who believed

0:10:54.000,0:10:57.000
that the Indus script does not represent language.

0:10:57.000,0:10:59.000
I even started to get some hate mail.

0:10:59.000,0:11:01.000
My students said

0:11:01.000,0:11:04.000
that I should really seriously consider getting some protection.

0:11:04.000,0:11:06.000
Who'd have thought

0:11:06.000,0:11:08.000
that deciphering could be a dangerous profession?

0:11:08.000,0:11:10.000
What does this result really show?

0:11:10.000,0:11:12.000
It shows that the Indus script

0:11:12.000,0:11:14.000
shares an important property of language.

0:11:14.000,0:11:16.000
So, as the old saying goes,

0:11:16.000,0:11:18.000
if it looks like a linguistic script

0:11:18.000,0:11:20.000
and it acts like a linguistic script,

0:11:20.000,0:11:23.000
then perhaps we may have a linguistic script on our hands.

0:11:23.000,0:11:25.000
What other evidence is there

0:11:25.000,0:11:27.000
that the script could actually encode language?

0:11:27.000,0:11:30.000
Well linguistic scripts can actually encode multiple languages.

0:11:30.000,0:11:33.000
So for example, here's the same sentence written in English

0:11:33.000,0:11:35.000
and the same sentence written in Dutch

0:11:35.000,0:11:37.000
using the same letters of the alphabet.

0:11:37.000,0:11:40.000
If you don't know Dutch and you only know English

0:11:40.000,0:11:42.000
and I give you some words in Dutch,

0:11:42.000,0:11:44.000
you'll tell me that these words contain

0:11:44.000,0:11:46.000
some very unusual patterns.

0:11:46.000,0:11:48.000
Some things are not right,

0:11:48.000,0:11:51.000
and you'll say these words are probably not English words.

0:11:51.000,0:11:53.000
The same thing happens in the case of the Indus script.

0:11:53.000,0:11:55.000
The computer found several texts --

0:11:55.000,0:11:57.000
two of them are shown here --

0:11:57.000,0:11:59.000
that have very unusual patterns.

0:11:59.000,0:12:01.000
So for example the first text:

0:12:01.000,0:12:04.000
there's a doubling of this jar-shaped sign.

0:12:04.000,0:12:06.000
This sign is the most frequently-occurring sign

0:12:06.000,0:12:08.000
in the Indus script,

0:12:08.000,0:12:10.000
and it's only in this text

0:12:10.000,0:12:12.000
that it occurs as a doubling pair.

0:12:12.000,0:12:14.000
Why is that the case?

0:12:14.000,0:12:17.000
We went back and looked at where these particular texts were found,

0:12:17.000,0:12:19.000
and it turns out that they were found

0:12:19.000,0:12:21.000
very, very far away from the Indus Valley.

0:12:21.000,0:12:24.000
They were found in present day Iraq and Iran.

0:12:24.000,0:12:26.000
And why were they found there?

0:12:26.000,0:12:28.000
What I haven't told you is that

0:12:28.000,0:12:30.000
the Indus people were very, very enterprising.

0:12:30.000,0:12:33.000
They used to trade with people pretty far away from where they lived,

0:12:33.000,0:12:36.000
and so in this case, they were traveling by sea

0:12:36.000,0:12:39.000
all the way to Mesopotamia, present-day Iraq.

0:12:39.000,0:12:41.000
And what seems to have happened here

0:12:41.000,0:12:44.000
is that the Indus traders, the merchants,

0:12:44.000,0:12:47.000
were using this script to write a foreign language.

0:12:47.000,0:12:49.000
It's just like our English and Dutch example.

0:12:49.000,0:12:51.000
And that would explain why we have these strange patterns

0:12:51.000,0:12:54.000
that are very different from the kinds of patterns you see in the text

0:12:54.000,0:12:57.000
that are found within the Indus Valley.

0:12:57.000,0:12:59.000
This suggests that the same script, the Indus script,

0:12:59.000,0:13:02.000
could be used to write different languages.

0:13:02.000,0:13:05.000
The results we have so far seem to point to the conclusion

0:13:05.000,0:13:08.000
that the Indus script probably does represent language.

0:13:08.000,0:13:10.000
If it does represent language,

0:13:10.000,0:13:12.000
then how do we read the symbols?

0:13:12.000,0:13:14.000
That's our next big challenge.

0:13:14.000,0:13:16.000
So you'll notice that many of the symbols

0:13:16.000,0:13:18.000
look like pictures of humans, of insects,

0:13:18.000,0:13:21.000
of fishes, of birds.

0:13:21.000,0:13:23.000
Most ancient scripts

0:13:23.000,0:13:25.000
use the rebus principle,

0:13:25.000,0:13:28.000
which is, using pictures to represent words.

0:13:28.000,0:13:31.000
So as an example, here's a word.

0:13:31.000,0:13:33.000
Can you write it using pictures?

0:13:33.000,0:13:35.000
I'll give you a couple seconds.

0:13:35.000,0:13:37.000
Got it?

0:13:37.000,0:13:39.000
Okay. Great.

0:13:39.000,0:13:41.000
Here's my solution.

0:13:41.000,0:13:43.000
You could use the picture of a bee followed by a picture of a leaf --

0:13:43.000,0:13:45.000
and that's "belief," right.

0:13:45.000,0:13:47.000
There could be other solutions.

0:13:47.000,0:13:49.000
In the case of the Indus script,

0:13:49.000,0:13:51.000
the problem is the reverse.

0:13:51.000,0:13:54.000
You have to figure out the sounds of each of these pictures

0:13:54.000,0:13:56.000
such that the entire sequence makes sense.

0:13:56.000,0:13:59.000
So this is just like a crossword puzzle,

0:13:59.000,0:14:02.000
except that this is the mother of all crossword puzzles

0:14:02.000,0:14:06.000
because the stakes are so high if you solve it.

0:14:06.000,0:14:09.000
My colleagues, Iravatham Mahadevan and Asko Parpola,

0:14:09.000,0:14:11.000
have been making some headway on this particular problem.

0:14:11.000,0:14:13.000
And I'd like to give you a quick example of Parpola's work.

0:14:13.000,0:14:15.000
Here's a really short text.

0:14:15.000,0:14:18.000
It contains seven vertical strokes followed by this fish-like sign.

0:14:18.000,0:14:20.000
And I want to mention that these seals were used

0:14:20.000,0:14:22.000
for stamping clay tags

0:14:22.000,0:14:24.000
that were attached to bundles of goods,

0:14:24.000,0:14:27.000
so it's quite likely that these tags, at least some of them,

0:14:27.000,0:14:29.000
contain names of merchants.

0:14:29.000,0:14:31.000
And it turns out that in India

0:14:31.000,0:14:33.000
there's a long tradition

0:14:33.000,0:14:35.000
of names being based on horoscopes

0:14:35.000,0:14:38.000
and star constellations present at the time of birth.

0:14:38.000,0:14:40.000
In Dravidian languages,

0:14:40.000,0:14:42.000
the word for fish is "meen"

0:14:42.000,0:14:45.000
which happens to sound just like the word for star.

0:14:45.000,0:14:47.000
And so seven stars

0:14:47.000,0:14:49.000
would stand for "elu meen,"

0:14:49.000,0:14:51.000
which is the Dravidian word

0:14:51.000,0:14:53.000
for the Big Dipper star constellation.

0:14:53.000,0:14:56.000
Similarly, there's another sequence of six stars,

0:14:56.000,0:14:58.000
and that translates to "aru meen,"

0:14:58.000,0:15:00.000
which is the old Dravidian name

0:15:00.000,0:15:02.000
for the star constellation Pleiades.

0:15:02.000,0:15:05.000
And finally, there's other combinations,

0:15:05.000,0:15:08.000
such as this fish sign with something that looks like a roof on top of it.

0:15:08.000,0:15:11.000
And that could be translated into "mey meen,"

0:15:11.000,0:15:14.000
which is the old Dravidian name for the planet Saturn.

0:15:14.000,0:15:16.000
So that was pretty exciting.

0:15:16.000,0:15:18.000
It looks like we're getting somewhere.

0:15:18.000,0:15:20.000
But does this prove

0:15:20.000,0:15:22.000
that these seals contain Dravidian names

0:15:22.000,0:15:24.000
based on planets and star constellations?

0:15:24.000,0:15:26.000
Well not yet.

0:15:26.000,0:15:28.000
So we have no way of validating

0:15:28.000,0:15:30.000
these particular readings,

0:15:30.000,0:15:33.000
but if more and more of these readings start making sense,

0:15:33.000,0:15:35.000
and if longer and longer sequences

0:15:35.000,0:15:37.000
appear to be correct,

0:15:37.000,0:15:39.000
then we know that we are on the right track.

0:15:39.000,0:15:41.000
Today,

0:15:41.000,0:15:44.000
we can write a word such as TED

0:15:44.000,0:15:47.000
in Egyptian hieroglyphics and in cuneiform script,

0:15:47.000,0:15:49.000
because both of these were deciphered

0:15:49.000,0:15:51.000
in the 19th century.

0:15:51.000,0:15:53.000
The decipherment of these two scripts

0:15:53.000,0:15:56.000
enabled these civilizations to speak to us again directly.

0:15:56.000,0:15:58.000
The Mayans

0:15:58.000,0:16:00.000
started speaking to us in the 20th century,

0:16:00.000,0:16:03.000
but the Indus civilization remains silent.

0:16:03.000,0:16:05.000
Why should we care?

0:16:05.000,0:16:07.000
The Indus civilization does not belong

0:16:07.000,0:16:09.000
to just the South Indians or the North Indians

0:16:09.000,0:16:11.000
or the Pakistanis;

0:16:11.000,0:16:13.000
it belongs to all of us.

0:16:13.000,0:16:15.000
These are our ancestors --

0:16:15.000,0:16:17.000
yours and mine.

0:16:17.000,0:16:19.000
They were silenced

0:16:19.000,0:16:21.000
by an unfortunate accident of history.

0:16:21.000,0:16:23.000
If we decipher the script,

0:16:23.000,0:16:25.000
we would enable them to speak to us again.

0:16:25.000,0:16:28.000
What would they tell us?

0:16:28.000,0:16:31.000
What would we find out about them? About us?

0:16:31.000,0:16:34.000
I can't wait to find out.

0:16:34.000,0:16:36.000
Thank you.

0:16:36.000,0:16:40.000
(Applause)