The comments to my last
post were so involved that I could only answer them with another post.
Subbudu wrote of an imaginary dialogue between me and "LLover", someone who loves Language 'L'
Ambar: Trying to make me learn lang L is pointless. X people already know Lang E so I am going to forget L. Ls value is less as per Metcalfes law.
LLover: Well, as per Metcalfes law, Ls utility is lower. Therefore you should learn L to make its utility higher. As for the benefits accrued..and the benefits are this, that and the other and you can read fantastic literature and learn this and this and know more about this history etc.
I never said anything about forgetting a language, however impossible that might be.
I was arguing against making people learn languages which give disproportionately low benefits compared to the effort needed. What is happening now is that people like LLover want to make L popular because of their attachment to it.
LLover: What if I am sincerely convinced that IPX is a better protocol? Just because IP is more widely adopted it does not mean, IPX may not be/have been a better protocol. Certainly wide adoption cannot be the only way to say something is better. What if there is a new protocol PP that is a derivative of IPX that is good? Should I not advocate it?
Ambar: No point in that though..
LLover: If you compare Protocols just like you do languages - i.e. by the current state - number of users- then how do new protocols come about? Well I am convinced PP is better and I am going to try to advocate PP to whoever I can reach and sing the glories of PP.
Ambar: Metcalfe will eat you alive :-)
I was talking about Metcalfe, not Hannibal Lecter.
LLover: Maybe not.. Maybe I will deploy PP ..putting a gateway between PP and IP. My neighbor may start liking PP and it may spread in small pockets thus and grow, much like IP did.
Also, Currently my entire office uses PP and I stand much to gain in that my office can understand anyone who can speak PP..So I am just trying to do what I think is the best protocol.
The gateway analogy fails when it comes to natural languages. Do you mean that you'll have a bank of translators so that L speakers and E speakers can communicate? Even when the number of L speakers is miniscule compared to E. I can't think of any situation where this would work except the UN or inter-governmental talks.
Secondly, an Anon commenter had this to say:
As a computer scientist and mathematician, it always makes me cringe when people attempt to use numbers and laws sloppily to prove something. If you want to do that , you better be sure you can back your statements up.
Firstly, Metcalfe's law is a rule of thumb and not a law (somewhat like Moore's Law). All it really says is that in an n-node graph, there are O(n^2) possible edges.
Secondly, if Ambar had bothered to read his own wiki link, he would have mentioned Odlyzko & Tilly's March 2005 paper , which concludes that the value of a network is O(n log n) .
Anon, I agree that Metcalfe's "Law" is more an empirical rule than a Law. And as regards, Odlyzko and Tilly's paper, there's also Reed's Law which talks of an O(2^n) value. The very definition of 'value' is unclear here. The only consensus does seem to be that the value monotonically increases with N.
It doesn't matter whether the value is O(n), O(n log n ) or O(n^2), because languages aren't mere networks . The analogy is imperfect and oversimplified.Just why are languages not
"mere networks"? Is it because people have sentimental attachments to languages?
A few years ago I read the Tamil epic Ponniyin Selvan, which is magnificient and on par with anything I have ever read in any language. It blew me away. I would never have been able to read it if it hadn't been translated into English. I dont doubt that the original is way better than the translation.
THere are equally superb works in many regional languages, which i will never get to read unless I learn the languages, or someone who knows the languages translates it.
Good for you. What interests me is your underlying assumption that these works are in some manner indispensable and the ideas expressed within them will never occur anywhere else independent of them. This is what I mean when I refer to sentimentality being attached to language.
Of course my language has something invaluably unique in it!
Suppose there is a network connected to the internet with some amazing pile of info on it. The network is public, but uses a nonstandard protocol. Should nobody bother to learn the protocol, and thus let the information go to waste forever? I'm not saying everybody should learn the protocol, just that someone should. I would argue that the value of the network (equivalently, the obscure protocol used to access it) is either zero or positive. If it's positive, then we have nothing to lose and something to gain by not letting the protocol die out.
The trick is to extract the information from this nonstandard network and move it to a standard network. Come what may, the very fact that it is obscure and non-standard means that the information stored on it is much less valuable and accessible than that on the standard network. The effort required to extract the information makes it less valuable than it could have been.
Gaurav asks,
So when will we start learning mandarin ?Gaurav, the day that knowledge of Mandarin brings more benefit than learning English, all of us will learn it en masse. I guarantee you that.
P.S: Anon commenters, I would appreciate it if you used names, even pseudonyms so I can distinguish one Anon from another. :D