February 26, 2008

About Sanskrit

Today's Hindustan Times is only the latest publication to carry this assertion I've seen made, in different forms, for years:

"A report in Forbes magazine in 1987 said that Sanskrit is the most precise language and hence suitable language for computer software."

Does anyone have proof for this statement? That is, a link to the article in Forbes? Or if someone has a paper version of the said article, I'll pay for you to snailmail me a copy.

I want to know if there was indeed such an article. I'd also like to know what's meant by "most precise language" and "suitable language for computer software."

As a software dabbler myself, and because I was actively interested in Artificial Intelligence at the time, I remember Rick Briggs' paper that made waves in 1985, Knowledge Representation in Sanskrit and Artificial Intelligence. (The paper is available in all kinds of places, like here).

Briggs contended that Sanskrit grammar held profound lessons for AI researchers interested in natural language understanding. He saw parallels to the techniques of AI, remarking on the "surprising equivalence of the Indian analysis to the techniques used in applications of Artificial Intelligence". There's plenty more of interest in that paper -- I remember poring over it back in '85.

But if this is the source of the statement attributed to Forbes, there's something I'm missing. Briggs makes no claim about it being a "suitable language for computer software". Not least because such a claim, about any natural language, is meaningless.

Any thoughts welcome.

39 comments:

chica said...

I have zero knowledge about Sanskrit or AI for that matter. There's an article I once read on a blog. Its not about Sanskrit in particular but read the "First Verse" section of it. It gives a slight idea about the strength of Sanskrit grammar. http://hawkeyeview.blogspot.com/2007/12/ramayana-i.html
If there is such a defined preciseness to the language, it possibly could be used in AI.

Anonymous said...

"Most suitable language for computer software"
Diamond! :)

When I was learning Sanskrit in school I was told that Sanskrit is the mother of all languages. It isn't, but many Indians believe it to be. Sanskrit is getting a lot of coverage these days. I fear that the saffron brigade will soon demand that it be made compulsory for all students. And one government, if not the other, will relent.

km said...

Ah, that old chestnut keeps coming back every 5 years.

And even if Sanskrit could be used, how would that happen? Implementation of the technology is a far greater problem.

Anonymous said...

Also, please see this page:
http://mutiny.wordpress.com/2007/02/09/sanskrit-mother-of-european-languages-says-prof-dean-brown/

The comments are interesting.

Anonymous said...

Have you tried asking or writing to Forbes magazine?

Anonymous said...

Dilip,
Yes, that indeed appeared in Forbes. I have read the article. I'm not sure about the wordings though, but just because I didn't believe the claim initially I remember going and looking up for the article. If you can tell me the month I will check out my local library.

Dilip D'Souza said...

Chica, very interesting, thanks. That precision is fascinating.

junglibandar, that's a terrific comments section! I'm still wading through it.

Anon#1 above: answer is yes, on three counts. Two different postal queries in the last 3-4 years have drawn a blank (i.e. no response). email query awaits a response.

Anon #2 above: July 1987 is generally quoted. If you can find it, please make me a copy. If you leave an email address here, I'll send you my postal address. Thanks!

Anonymous said...

I don't know whether the exact quote appeared, but one of the reasons sanskrit is considered that way in computer circles is that Panini (who wrote sanskrit grammar)is now considered to be the inventor of Context free languages and sanskrit is a perfect context free language-Initially Noam chomsky was fully credited with it but now Panini's work is recognized (as of now) as the first authoritative expression of Context Free language .

Anonymous said...

In the online archive available in our library I could not find any article in the two July issues of 1987 related to Sanskrit. In a search from Jan 1987 till present the word Sanskrit is mentioned only once in 1999 in an article on Yoga. Computer and grammar are together mentioned once in 2000 on using computer in grammar schools.

Moreover, just the appearance of an article in a popular (and moreover nontech) magazine does not make it right. (we dont know what appeared there, if it actually appeared.) The nontechnical journalist can misunderstand a claim or worse might say the exact opposite in an attempt to give an analogy. (for eg. the quoted line in the post does not say programming language. what does "language for software" mean? how is it suitable?)
(Economist (and other news sources) recently published the development of a quantum computer that could solve hard problems in one-go. Both the assertions -- development of a quantum computer and power of q-comp to solve problems in one shot -- turned out to be wrong.)

Every time someone talks of "preciseness" of sanskrit I wish to hit them on the head with the poems that smart poets have written that can be interpreted in two (or more) different ways. (the proponents of the vedic math have the clever ways of interpreting some couplets to their convenience). (btw, I have not heard the alternate meaning of the first verse in chica's link before). How would a poor compiler know which meaning to use? (maybe, you might say, we will limit the scope of the features allowed. with the limitations imposed it is no longer a natural language. Welcome to the world of prog languages.)

Finally, I dont see why people need to overstate the coolness of Sanskrit.

Vivek Kumar said...

I think what you are looking for is this:
http://www.gosai.com/science/sanskrit-nasa.html

Dilip D'Souza said...

Vivek, that's the paper by Briggs that I mentioned and linked to -- in fact at the end of the page I linked to there is this: Reprinted by permission from http://www.gosai.com/science/.

That paper is not in question. I'm searching for the July 1987 Forbes article.

Vivek Kumar said...

Sorry, my mistake. I read your post yesterday.. and forgot that it has already been linked.

I guess this must be the source of the whole Forbes thing. That game of Chinese Whispers.. it works even better on the internet.

Someone makes a half-baked mention of an article in a journal.. one person working at NASA is made to represent NASA.. somewhere down the line 1985 becomes 1987 and AI journal becomes Forbes.. by that time enough Indians are on Internet to give this one a life of its own.

I don't think anyone needs to waste time finding the Forbes article that doesn't exist :)

Anonymous said...

Dilip: You may wish to check this or write to the author.
http://www.infinityfoundation.com/mandala/t_es/t_es_rao-t_syntax.htm

Anonymous said...

Anonymous 9:34
How dare you link to something positive about Indian heritage?
Dont you know the rules of our typist here? Anything good about Hindu heritage etc has to be trashed.
Do not link to Infinity foundation. Rajiv Malhotra is a saffronite and the said piece is a bunch of lies. The ultimate truth is that put out by FOSA and others like Vijay Parshad and others.

Jungli Bandar
Sure - I agree the saffron brigade will make things like sanskrit compulsory - why stop at that even yoga, Vedic Maths etc will be made compulsory.
Maybe it is the saffron brigade that is responsible for the popularity of yoga worldwide. Let us trash our own knowledge and past. All the better for westerners to usurp all that knowledge.
These are the same people who will believe all the fiction manufactured about Chinese and Arabic heritage.

Anonymous said...

> something positive about Indian heritage?...
> Anything good about Hindu heritage...

Just I'm wondering, are "Indian heritage" and "Hindu heritage" the same thing??

> Let us trash our own knowledge and past.

I am getting you're sarcastic tone so please tell, where is the trashing?? DD is asking for copy of the Forbs article. junglibandar I think is not agreeing that Sanskrit is "most suitable language for software." If you have the article, please share. Or please tell how it is most suitable for software. If not, plese dont immediately saying ppl are trashing. How is this "trashing"?

If you

Anonymous said...

Jayanthi
Sure - everybody who disagrees has to provide all the proof necessary, but of course for DD links to ramdom websites and words of sufi saints or his own conjectures are enough evidence. It is precisely this attitude I am questioning.
Please read my response to junglibandar. is he only disagreeing? I am responding to this comment of his - in case you did not get it -which seems apparent.
'I fear that the saffron brigade will soon demand that it be made compulsory for all students. And one government, if not the other, will relent.'

Do you detect sarcasem? I did not intend to be sarcastic but just offered a piece of advice to the earnest human being who linked to infinity home page.

Anonymous said...

Hindu heritage and Indian heritage are not the same thing in typist lingo.
All other influences such as Islam and Wwestern influences are to be accepted without question and if they are lauded - even better - but it is when Hindu heritage is lauded that there is saffron agenda at work

Dilip D'Souza said...

for DD links to ramdom websites and words of sufi saints or his own conjectures are enough evidence. It is precisely this attitude I am questioning.

What attitude is that, Nikhil? This post has two links: one to the Rick Briggs paper, which I read when it first came out, and which I found fascinating. The other is to the AAAI magazine ToC, only because I could not locate the paper itself in the AAAI archives. (Which is why I provided the other link).

What are the links to sufi saints, please let me know? All I would like to know is, was there such a Forbes article? Can you provide me a copy, or a link? If someone say such an article has appeared, is it not permitted for me to ask to see a copy without remarks about my "attitude"?

Anonymous said...

Nikhil: Earnest human being(anon of 9.34) is suitably warned, chastised and gone forever! I have/had no idea of any backgrounds, but even ehb are simple but not stupid. Thanx for your comment/note.

Anonymous said...

0. Re. backgrounds

There is an almost vanishingly small grain of truth at the core of Nikhil's complaint IMHO that may be discernible to regular readers of dcubed.

Eg. On a post on say how Urdu or Arabic is being touted as the perfect language for computers, somebody chimes in with a comment on how Islamists are using this to claim supremacy for their faith and compulsory etc.

My reading of Dilip, formed over the past 4-5 years, is that he will not let that go unchallenged. It would be:

"What does this have to do with the topic at hand, which is Urdu as a computer language, or more precisely XYZ article abt Urdu?"

1. The text above is pure conjecture on my part, not an actual response from Dilip. I place it here to test how true it may ring for other readers.

2. Nikhil is way overstating it with sufi saints etc. i dont have a clue what he is referring to.

3. Dilip's is a slight selectivity or sensitivity. He has every right to have his biases. We all have them and are mostly not even conscious of them.

4. The references to the perfection of Sanskrit in those links are scary, and inspire in me some of the same fear that I find in junglibandar's comment.

5. I am not even suggesting this is a test not just because I know Dilip doesnt take tests, but also because I dont like giving them. Its just an observation.

6. On the topic:
From by brief reading, Sanskrit a natural language is not suitable for computer programming, the most pluses it has is a strict grammar, there is a Panini Normalised Format reference somewhere. Its also been pointed out that actual Sanskrit where used deviates a little from PNF. Rules to straitjacket it into PNF can well be extended to other languages to make them computer-like. So maybe a slight edge over these, mostly because its not a very active living language. Other languages they looked at: Barambic, Swiss German.

From google groups discussion. Just for anybody interested. My knowledge of Sanskrit is more than vanishingly small.

Thanks,
Jai

Anonymous said...

Meant(more than vanishingly) small not, more than (vanishingly small). Man, the ambiguity :-)

Jai

Anonymous said...

....
anyone found the article yet?does it even exist?
or is it just the ai '85 one?

Anonymous said...

Dilip,
I don't know what your intentions were when you wrote this. I wish you would have done some research.
Nowhere does the article actually say that Sanskrit is the most suitable language for computers. It is a conclusion derived from a lot of other facts.

Panini (according to Western historians circa 400 BC based on the arbitrary date of The Creation 4004 BC) in his Ashtadhyayi (lit. Eight Chapters) made an attempt to formalize the Sanskrit Language. As he himself admits he built on the works of many past grammarians. In 4000 sutras (closest English approximation: aphorisms) he sets down concise rules for formalizing the structure of the language. After a few millenia when the theory of computer science is developed Backus and Naur came up with the concept of a formal language for computers which needs extremely precise rules. The structure that is now called Backus Naur Format (BNF) has a large number of similarities with the rules from Ashtadhyayi written hundreds of years back. Hence the structure of Paninian/Classical Sanskrit is very similar to that of a formalized Computer Language and hence the claim. No other language has this kind of structure as defined by Ashtadhyayi. More detailed information can be easily found on the internet.

"Not least because such a claim, about any natural language, is meaningless."

In a sense the classical Sanskrit or the Paninian Sanskrit is not a natural language. Sanskrit itself means refined which is the opposite of Prakrit (I am referring to Vedic Sanskrit, and not the Prakrit languages: Pali, Ardhamagadhi and so on) which means natural or unrefinded or of the common folk. So what Panini did was formalize Vedic Sanskrit to the minutest details and gave us a very formal Classical Sanskrit.

Anonymous said...

I have checked Forbes Magazine 13h and 27th July 1987. There is nothing about Sanskrit, etc. Nothing about India as far as I could see. The latter issue has a cover story on Japanese Bilionaires. India was not in the business news in 1987.
Stig Toft Madsen
Copenhagen Denmark

Unknown said...

Yes Dileep D' Souza,
Why are you so concerned about Sanskrit all of a sudden. Agreed, perhaps the Forbes article never existed. May be it was 1987. But then again, what bothers you?
The quote or claim in itself, the year, the magazine name or simply Sanskrit.
Your botheration about this sounds like me worrying about why Ann Coulter and you broke up?

Anonymous said...

agreed with nikhil

and 2nd thing to my frd. who was asking about hindu and india relations?
hindu culture is 1000% same as indian culture as india/hindustha is the reason why vedic culture following ppl r called as "hindu".

and its such a shame that the question was asked by a hindu of india.

Anonymous said...

and the blog owner should feel proud that a hindu/vedic language was praised.

just tell me one reason to make me not to say that
"you r communal"

Dilip D'Souza said...

I should "feel proud that a hindu/vedic language was praised" by using what is apparently a false claim? (i.e. the citing of a nonexistent Forbes report).

All I want to know in this post is, was there such an article. As Madsen says above, there was not such an article.

You want to feel proud of a false claim, please go ahead.

I have no interest in making you not say or otherwise whatever you want. Please go ahead.

Anonymous said...

I have also gone through the topic when searching for this great news - I found something more interesting here - http://www.hinduwisdom.info/Sanskrit.htm

Jay ho 'Dev Bhasha'!

Anonymous said...

Call me "Anonymous 101".
I am an Indian, Hindu graduate student (Engineering) in USA. I am not an expert in AI or Lingustics. I have studied Sanskrit for 5 years in India. (But, that does not make me an expert in Sanskrit either!)

The only aspect of Sanskrit language, that I find advantageous in any software application, is the one-to-one correspondence between pronunciation and script. e.g. "u" in "put" and "but" has different pronounciations. This NEVER happens in Sanskrit. Thus, if one were to write a code that allows a computer to accept "spoken" commands, Sanskrit is a very suitable language. In a lose mathematical sense, Sanskrit defines a one-one and onto (hence, invertible) function between set of letters and set of pronounciations. This feature may as well be used in Artificial Intelligence. Layman example : Its easier to create Robots that speak and understand Sanskrit that any other language.

But, I do not have an article that supports what I have said above.

Thanks.

H1B Express :) said...

I studied Sanskrit for about 7 years. Now I can't use my tongue properly :-). I think it just like people who don't know nothing about their own religion boasts about it, people who never learned Sanskrit is making these claims. People have a fantasy towards things that they don't completely know, like the neighbors wife :-)

Anonymous said...

I am trying to understand why a question about the abilities of a language gets mixed with religion and culture? Question started about if any such article exists or not as it was claimed. Almost all of the guestions/answers or questions and answered focused on almost any thing other than the topic.

Qasrani said...

Above is my post plus this:
Being a computer programmer, I know it that Sanskrit is not suitable for computer programs if used as natural language for computers. I read that it is a language where meaning of each or any word depends on context or words coming before it. So just imagine how many nested if-else conditions are needed to complete even a single command.

Ketan said...

It is sad that such a genuine question asked by you was given the anti-Hindu spin by few readers.

Even if you wrote this article only to show that people are more apt to buy silly claims when they show 'their own' culture in glowing terms, then it was an important and precious message. Because it is this tendency that makes people go in denial of existence of things that are wrong in their culture/religion/nation.

Thanks for writing this!

Anonymous said...

you all are fuckers, sanskrit is worlds greatest language and it is the most precise language compared to any other eg- english,french etc.sanskrit is the mother of all languages some or the other words of a language have been taken out from sanskrit.though i agree that it is not suitable for computers.

BLAZZER12 said...

Hello,

I was skimming the web just about this claim. Since you have asked about the Forbes magazine reference, its from the "issue of July 1987". Or that is what this page claims to be :

http://www.aurovillelanguagelab.org/Research/The%20Greatness%20and%20Glory%20of%20Sanskrit.htm

And yeah, to hell with all those who want to bias this topic towards their own political and religious agenda...

Atanu said...

Sanskrit is a good language at least most of the Indian languages stem from it, so if it is made compulsory, I don't find any harm in it ! In fact it will help us learn a lot about ancient India and will also help us stop those Hindu fundamentalists too. And answer to someone who asked about Indian = Hindu equation ! Ok let me tall you one thing, Hindu is basically not a religion but an Indian way of life, its probably the life style and the collection of philosophies constructed in ancient India, so people sometimes use Indian and Hindu interchangeably, there is no harm ! But when fundamentalists and politicians color it they mean Hindu the religion and so there is harm in it. It depends on the view.

articlehack said...

http://www.aaai.org/ojs/index.php/aimagazine/article/view/466

this is all about AI .. their research paper is also attached:

http://www.aaai.org/ojs/index.php/aimagazine/article/view/466/402

They conclude that Sanskrit is best for AI.

Anonymous said...

Thanks for the article. While I have respect for world's ancient languages such as Tamil, Kannada and Sanskrit ,
I strongly object hoax and nonsense spread by people stating that it came in forbes etc.
There is no need to magnify our glory through false propaganda/rumour/hoax.PERIOD.
This article is a hoax buster.
Let us debate anything with rational approach.
-Raj