Sanskrit and Lisp (2011) | Hacker News Reader

Sanskrit and Lisp (2011)

ifacethoughts.net

82 pointsyati61 comments

So there are several things in this article that are flat out wrong:

  (1) "No wonder so many languages are derived from it."

Not sure what he is referring to exactly, but Sanskrit itself is a descendent of proto Indo-European, from which many other Western languages were derived. So it is a near cousin to the original language of the Indo-European people, but not the direct ancestor of Western languages (this was a misconception of the 19th century linguists that "discovered" Sanskrit).

  (2) "It is more expressive"

I assume by this he means you can do more with less. To a certain extent this is true.

  (3) Very systematic

This is only true to a certain degree. Classical Sanskrit is itself not the original language, and probably the largest single pain point is the "sandhi" rules, in which words are combined together. Although someone eventually wrote down a large number of rules that shows how these changed together, they are more the product of slurring of speech over time, rather than "design" per se.

  (4) "Excellent grammar"

Is a lot of grammar "excellent grammar" ? In Sanskrit, it does allow certain structures that are not available elsewhere (like the existence of a dual case), which presumably facilitate density, and, consequently, presumably also facilitates efficiency. If so, one could argue that the grammar of Attic Greek is more "excellent" than that of English. However, to me "excellence" is something that depends more on the specific use case -- a language that someone can speak and use is better than one that one cannot.

> Not sure what he is referring to exactly, but Sanskrit itself is a descendent of proto Indo-European, from which many other Western languages were derived. So it is a near cousin to the original language of the Indo-European people, but not the direct ancestor of Western languages (this was a misconception of the 19th century linguists that "discovered" Sanskrit).

OP never says Western Languages descended from Sanskrit, you are making that part up. The fact is that many Indian languages descended from it (or rather from the natural form, Prakrit, as opposed to the classical form). This is similar to how Romance languages (Spanish, French, Italian and Portugese) evolved from Vulgar Latin.

>OP never says Western Languages descended from Sanskrit, you are making that part up. The fact is that many Indian languages descended from it (or rather from the natural form, Prakrit, as opposed to the classical form). This is similar to how Romance languages (Spanish, French, Italian and Portugese) evolved from Vulgar Latin.

He's not necessarily :making it up_. It's quite common for Indians, and even Indian linguists to spread ridiculous ideas about Sanskrit, including it being "perfect", "better than other languages", "holy", and quite commonly "all other languages descending from it" as well as the claim of Sanskrit being "the first language".

There's a lot of things in that blogpost that would make most linguists facepalm (especially the part about Sanskrit being "more expressive" and "more systematic"). The idea that Sanskrit is an ancestor language for other languages (far away from the Indo-Aryan group) goes quite hand in hand with these ridiculous notions.

To be honest, he sounds just like the other parents trying to justify why their children should spend years learning Latin. (This was a snooty private school, not a parochial school.)

I think you're being paranoid, and he's just comparing it to other Indian languages, and saying that a bunch of Indian languages descended from it (as they did, by way of Prakrit.)

While it may be common for some Indian people to "spread ridiculous ideas about Sanskrit", a lot of people who have studied Sanskrit do come away with the feeling that, compared to modern languages in general (English, and other European or Indian languages), Sanskrit manages to pack a wealth of meaning into a few words. I don't see why a layperson saying that should be labeled as "ridiculous" and such.

> There's a lot of things in that blogpost that would make most linguists facepalm (especially the part about Sanskrit being "more expressive" and "more systematic"). The idea that Sanskrit is an ancestor language for other languages (far away from the Indo-Aryan group) goes quite hand in hand with these ridiculous notions.

There is a reason I have not addressed the other points you raised.

> It's quite common for Indians, and even Indian linguists to spread ridiculous ideas about Sanskrit, including it being "perfect", "better than other languages", "holy", and quite commonly "all other languages descending from it" as well as the claim of Sanskrit being "the first language".

You'll have to back up those claims. Southern Indians (> 25% of Indian population) would have a HUGE problem with anyone saying that. Just because a few quacks are saying it does not mean many Indian believe it.

The reason why Sanskrit seems so interesting for programmers is that it was first language to have formal rules for grammar over 2400 years ago.

Panini defined Sanskrit grammar using Bakus-Naur rules and meta rules.

http://mathomathis.blogspot.in/2010/08/backus-naur-form-pani...

This is very true, although one minor nitpick is that you should say "formal written rules" for grammar. The rules appear in fact appear to be more formal and consistent in the centuries before Panini, and the need to write them down (including the aforementioned sandhi rules) was partially due to inconsistencies derived from the evolution of the language.

To sum it up, the writer probably has the mentality of William Jones: sort of romantic, but essentially flawed.

The imaginary "proto indo-european" language is a very deep seated racist brother of the Aryan invasion theory. You can read the wikipedia article and the start of this mythical language construct with Sir William Jones who wanted to make his study of Sanskrit fit within his Christian worldview. He found the Sanskrit, Latin and Greek related and decided that there is some mythical imaginary language that preceded them. How could well educated Britishers and Europeans come to the conclusion that Sanskrit was indeed the ancestor of Greek and Latin.

You seem to agree with the linguistic consensus that Sanskrit, Latin and Greek are all related. You seem to depart from the consensus with the conclusion that because they are all related, they must share an ancestor.

If that's so, how do you account for the relation between Sanskrit, Latin and Greek (and Russian, and English...)?

Or do you deny the relation altogether? Do you have a competing theory that you could point me to?

You want me to compete with a theory that made up an imaginary language?

...Yeah? I mean, if the PIE theory is obviously wrong, then what's the right answer?

How come that "the theory that made up an imaginary language" correctly predicted the existence of laryngeal phonemes in the Hittite language before the Hittite language was even known to exist? Stroke of luck, that a group of unknown consonants from an "imaginary language" appeared in a real language in exactly the right place in the right words?

Also, the notion that the PIE reconstructions describe a "language" in the sense that we usually understand it (a vernacular tongue of a group of people in a single time and place) is a false one: even though the reconstructions with very high likelihood represent individual IE language features in an old form thereof, it would be a folly to try to slap them together and say "this is a language that someone spoke", for the same reason you can't reconstruct individual genes from the tree of mammals, slap them together and say "this was the first mammal". It doesn't work like that. What you get is a constrained probability function over possible languages, not a single solution with 100% probability.

It's not racist in the least. If Sanskrit is the ancestor of Latin and Greek, then there should be Sanskrit a superstratum in Latin and Greek. However, there isn't. But there are common elements in all three languages that could only have come from an ancestor language.

What you're doing is ad homimen: discounting the theory based on personal attributes of the person who came up with it.

Can the theory stand on its own regardless of the personal motivations of the person who came up with it? Then the theory is solid.

Language Complexity that you speak against here is a good thing! But only when the rules are simple, yes even when you have many similar rules. The language would still be more efficient than a language with much less rules.

The other thing you criticize is your notice about the evolution of the language. That makes only sense when you just wanted to be accurate, but then it shouldn't look like critics. If a language evolved, it means it was successful.

(1) "No wonder so many languages are derived from it." >Not sure what he is referring to exactly, but Sanskrit itself is a descendent of proto Indo-European, from which many other Western languages were derived. So it is a near cousin to the original language of the Indo-European people, but not the direct ancestor of Western languages (this was a misconception of the 19th century linguists that "discovered" Sanskrit).

This misconception is very much like that of creationists when propose that evolution means "people came from monkeys"; when in fact we simply share increasingly recent common ancestors as you move from primitive primates to the other great apes.

> Here is what I came up with during our discussion...

You are looking for reasons where there is none.

* Why Brazil speaks Portuguese? * Why do we speak English now? * Why do people in Ukraine speak Russian instead of Ukrainian?

Not because Russian is more expressive. But because they were killed for doing otherwise. It's not about language features, it's about politics, wars, colonization, economics.

When kids ask why, parents are at their creative best. The difference these days, is they blog about it.

People primarily speak Russian in Ukraine not because they were afraid of being killed but because it was the lingua franca of the region and it was taught in Soviet schools. There was a lot less literature available in Ukrainian because Ukrainian was a language spoken primarily by serfs (educated people spoke a mixture of French and Russian).

Maya people in America primarily speak Spanish not because they were afraid of being killed but because it was the lingua franca of the region. Oh, wait...!

Sources? I'm really interested to learn who has been killed for speaking Ukrainian.

Just overall scope of history: [1] https://en.wikipedia.org/wiki/Chronology_of_Ukrainian_langua...

[2] https://www.youtube.com/watch?v=CDeujOxl3jc (they are talking about this in the video: [3] https://en.wikipedia.org/wiki/Bykivnia )

[4] https://en.wikipedia.org/wiki/Vinnytsia_massacre

As the policies of Ukrainisation were halted (1931) and replaced by massive Russification approximately four-fifths of the Ukrainian cultural elite, intellectuals, writers, artists and clergy, had been "eliminated", executed or imprisoned, in the following decade. The term was coined for this - Executed renaissance.

Many would argue that Holodomor (1932-33, 2-10 millions of Ukrainian speaking people died) was a part of Russification campaign.

Sanskrit is a festidious language that agglutinates huge numbers of tenses and forms in to a festival complexity that is a classical feature of the Indo-European language family. It might be great if you feel like being really damn specific, but it might also suck. It depends what you want to do with it.

I would posit that Sanskrit is more like assembly language (for its unique combination of specificity, table-thumping traditionalism, and verbosity). On the other hand, classical Chinese, which presents a roughly similar vintage literature, is a more fluid and combinatory system for type-indistinct component thoughts ... sort of like an extreme version of perl.

Sanskrit isn't verbose. I haven't seen any English translation that is very much lesser in size compared to original sanskrit text. I have to agree with author's comparison to lisp, very powerful ,very systematic (in fact the alphabets are arranged in the order in which tongue and palate interact to make that sound http://www.sanskritsounds.com/about-sanskrit/46/index.html) but lacks adoption of masses.

Sanskrit isn't verbose. I haven't seen any English translation...

Sure! Long multi-phoneme words with all the tense built in... the wordcount will be low. I meant semantically verbose, roughly: "average minimum unit of well-formed expressible information". In this facet, I found Sanskrit far worse than even Latin.

I recently noticed similar diffrence between Polish and English. Usually Polish text is shorter in written form (less words, and often less letters), but English text is shorter when spoken (less syllables). Polish has also a little more redundancy because word endings must agree.

It almost seems like Polish was optimized for reading/writing and correctness, and English for talking and speed.

It's late here so I'm not going to research this before responding. It is my impression from some previous reading that the closest European language to Sanskrit is German, with certain words such as luft (Lufthansa, Luftwaffe, etc.) being directly traceable to a common ancestor. With all of the Prussian-style border-shifting in the Germano-Polic region over time (the German History museum in Berlin is absolutely amazing illustrating it with maps!), I wonder if perhaps Polish shares some of that - for wont of a more informed phrase - 'old school Indo-European' quality? The answer is probably on Wikipedia, which also distinguishes betwen many types of Polish: https://upload.wikimedia.org/wikipedia/commons/1/15/Language...

Strange, I think Polish share much more with Sanskrit than German. Not that it makes it somehow better or worse, just my opinion.

For example we have flexion system working smilary to that in sanskrit (you glue up pre and postfixes to the word to change the tense, mark the gender of the actor (mandatory), or depending on the case of the noun). German has sth similar only in a few places (past tense), and in concatenating nouns together. Most of the time word order and "keywords" decides on the tense, like in English. Polish have 7 cases (1 rarely used), sanskrit has 8 (1 rarely used), German has 4 IIRC.

Polish only recently (a few centuries ago) lost dual plurality that exists in sanskrit, and we still have relicts of that in many places, like proverbs, names for body parts, the way nouns change with numerals, etc.

And these all things aren't specific to Polish - almost all Slavic languages share them. I've also heard Lithuanians say their language are even more similar to sanskrit.

Your view of Sanskrit is probably skewed by a few things. First of all, nearly the entire corpus of Sanskrit literature was written after (and in many cases long, long after) it was used as a day-to-day language. It was probably already losing ground at the time of Panini, roughly 500 BCE.

Second, what Sanskrit literature we have is highly artistic and edited literary text, replete with hard-to-get cultural references, not ordinary spoken language.

Finally, it's probably just plain different from whatever language you're used to, not intrinsically harder or more complicated.

IMHO Sanskrit is both harder and more complicated than Latin, which I spent years on before coming to a dalliance with Sanskrit and Pali. Then again, other than Latin (which I aced), I've never had much success with Indo-Europeans (except basic French, through drink).

I've never had much success with Indo-Europeans

Did you compose this text with the help of Google Translate, then?

Not really, there is no evidence of Sanskrit being used for conversation other than in learned circles. Most of the Upanishads, Aranyakas and Puranas and the last two vedas were written after 500 BC. Kalidasa - arguably the greatest litterateur in Sanskrit lived after 500 BC. Bhasa - the playwright lived after 200BC. Most of the Sanskrit oeuvre has been produced after 1st AD and prolific works continued till 16th century.

More of one's working knowledge is loaded into the specifics of vocabulary in classical Chinese, whereas more of one's working knowledge is loaded into grammar in classical Sanskrit. At least it seems so at first. Wrestling with the complexities of compounds in Sanskrit and syntax in Chinese blurs this distinction though. The animating morphologies transcend the vocabulary/grammar distinction in highly modular languages.

At bottom both Chinese and Sanskrit, like many other classical languages (Latin, Greek, Arabic), are quite modular. While Chinese's phanopoetic menagerie may not at first glance appear as RISC-y as Sanskrit's lexis (comparing radicals to roots may not be fair anyway), basically it's rare in both languages to stumble across a word that is just a black box, whose etymology you can't pull apart so as to guess what it means. Rarer, at least, than modern languages, which seem more atomized and dictionary dependent, as opposed to the classical languages, which by their systematic modularity were able to function as universalist languages of empire.

I know both Classical Chinese and Sanskrit and have no idea what you are talking about.

"It's rare in both languages to stumble across a word that is just a black box, whose etymology you can't pull apart so as to guess what it means."

Err, yeah I guess you can guess from the morpheme of a Chinese character that a particular character is a type of tree or has something to do with trees, but that's about it. Similarly with Sanskrit, because of sandhi rules often times you can't even figure out what the word is in advance, and I'm not sure how you are suppose to "guess" the meaning absent looking it up in Monier Williams (unless you are guessing backwards from other Indo-European languages, in which case your potential guessable vocabulary will be quite small).

Glad my post survived the glance of a pro! (PS. Saw your evr.gr: please send me an email, I am working on ~LETS stuff tangentially and would like to share info.)

Jd, how do the sandhi rules keep you from finding out what the word is? They make it easier to find out. The rules are just that -- rules -- that bring order out of the chaos.

As to how you guess, yes, you infer from the etymology.

The fact that it's guessing doesn't mean it's easy, or that guessing gives you as much information as actually looking it up, parsing it, etc.

What makes words easy to figure out is by putting spaces in between the words.

Iamafrog <- you can figure that out what that means, but it certainly isn't easier than modern orthorgraphy.

Aamafrog <- you could figure that out if you know that "i" + "a" sometimes combines to form "aa" -- but why would you want to do that?

In what way is that easier?

> phanopoetic

It's not every day that we meet a legitimate word for which Google can only find a paltry 119 results ;-)

(This result did the most for me: Do we privilege sight or sound? - Foetry http://foetry.com/forum/index.php?topic=330.0;wap2‎ Let us call these tendencies the Phanopoetic and the Melopoetic; [...])

One could take the analogy further - for the same reason people choose C versus LISP. :-) Sometimes structure isn't the best (or quickest) solution for the problem at hand.

Here's the Google Cache http://webcache.googleusercontent.com/search?q=cache:XaBQEvN...

It's quite likely that natural Sanskrit did not have those properties, and all that has survived is the well documented formalizations of its grammar.

EDIT: By natural, I meant Vedic. Speaking as a linguist, I doubt that all of Sanskrit was accounted for when it was formalized.

There is nothing called natural Sanskrit(refined) - Sanskrit is a specialized language built and used for only scholarly purposes - ordinary people always used to communicate in Prakrit(unrefined)

I doubt that is the case. Panini wrote Sanskrit grammer in around 4th century B.C. and it is still the base of Sanskrit education.

Sanskrit is considered a "refined" language, implying serious effort to systematize phonemes, grammar, declension and even verb classification distinguishing recipient of actions (1-10 ganas, PP/AP padas, etc.). Refinement can give subtlety, accuracy, concision etc. but does not automatically and objectively mean "universal", "perfect" or "excellent", even though proponents conflate those aspects.

IMO, Sanskrit persisted as the language of metaphysical and scientific treatises for a long time because of practical advantages. A large library of concepts, allowed easy reuse and concision [1]. Being trained in Vedanga (phonetics, metre, grammar, etymology, etc.) allowed practitioners to compose new sutras easily, much like PhD students using mathematical notations and proofs in papers.

It was used conversationally, at least in courts and debates. Its decline into a quasi-venerated read-only language over the last 1000 years would be an interesting (and contentious) topic for dissertation.

[1] Patanjali Yoga Sutras http://www.arlingtoncenter.org/Sanskrit-English.pdf

Much like Latin, a good deal of the regularity is because Sanskrit is only a written language.

Not sure what you mean but I believe Sanskrit was a spoken language, written down in each region's local script before being formalized in Devanagari.

Sanskrit was not a natural spoken language. I was created as a language for worship. Although Sanskrit was spoken among the educated circles in India, but not in a natural setting. The example that our language teacher gave us, was of couple of students speaking French (say) who are learning it as a foreign language.

Sanskrit was only used for scholarly conversation between learned people - Ordinary folks always conversed in Prakrits (unrefined tongues)

And there is no formalization of Sanskrit in Devanagari - We (Keralites) still use Grantha script to write Sanskrit - And a lot of late Sanskrit works have their origin in Kerala. Infact Malayalm script was invented to write Sanskrit with ease on palm leaves.

I think what he means is that what people actually spoke every day was as different from our bookish notions of what Sanskrit is as the language of a poor farmer or a slave in Ancient Rome was from the writings of Cicero. You really can't equate the two things.

Yeah, this is what I was going at. As well, when a language like sanskrit is spoken as vernacular, then you end up with exactly what we have now: a bunch of related languages that have all developed their own changes and quirks over the last few thousand years.

I completely agree here. The moment common people start speaking a language - any language - there will be irregularities and native elements introduced into the spoken variant. This will especially be true if the language is as tough to learn as Sanskrit or even Classical Chinese and Latin.

Very apt comparison considering I have the same problem with Lisp (and Erlang) as Sanskrit

https://news.ycombinator.com/item?id=5452038

Expressive, systematic, powerful - it looks atractive but none of that counts. What counts is ease of being copied and easy assimilation of unprecise copies. That's why PHP rules the web and DNA rules the life and english rules the world despite their messines.

English is not popular because it is "easy to copy" or any nonsense like that. As already mentioned, people who didn't speak English were killed. Likewise in the PHP boom, people who did not write PHP were underpaid(not all, but you get my point) - and BTW, the Web is not all a programmer has to/can hack on. Not sure why you put DNA in there, but I think the DNA is pretty expressive, powerful(packs a lot in a lot less) and Darwin has helped us understand evolution as a pretty systematic process. At least do some basic homework before making bold claims.

English is not popular because it is "easy to copy" or any nonsense like that.

English may not be easy to copy but it's comparatively "bland". "Bland" in this case means "has a smaller number of marked features". For example, it has no tones. It's always easier for someone speaking a tonal language to learn a toneless language than the other way round. Sanskrit has a complex system of synthetic inflection, which is a rather marked feature, and all speakers of analytic languages would hate your guts for forcing it onto them if you tried to make them learn it. In many ways, Sanskrit is like Ancient Greek. You spend a year learning something that other languages can do without (and do so without muddling the meaning of the text), and even if you learn it, you'll still be doing mistakes quite a lot of the time. Such a language would never have a chance of spreading as an international medium of communication.

Non PHP web devlopers underpaid? PHP devlopers are the bottom feeders. I know. I'm one. They are paid lowest because there's a lot of them.

There are a lot of them because they are most abundant. They are most abundant not because PHP is well desinged language but because ability to use PHP is easy to copy.

As for DNA read about non-coding sequences, telomers, pleiotropy and such. I'll state again. DNA is the pretty dumb chaotic stuff, built out of the most common intersting chemical element, carbon. It's everywhere not because it's well designed but because it copies well. It has 4bln years of incremental chaotic patches and cruft all around. I don't think any system that suffered through such history would be admired by any software architect.

Sanskrit is an artificial language. No wonder why it resembles programming languages.

At the level you are talking about, All spoken/written languages are artificial.

haha, that's true (1:0 for @ankitml)

Reply on news.ycombinator.com