Tamil writing proposal in closing, and, as a bonus, a Theory of History

It is now time to close this series of posts. I have had this proposal for a new way of writing Tamil kicking around my (many) laptops for a while but posted it to this blog only following the Pizza and Panini column for Indiatimes.com. Panini, in a sense, lives on!

Below are the links to all the posts on the Tamil writing proposal:

• Overview
• A modest proposal towards modernizing the Tamil language
• Response to some comments
• Harvard-Kyoto - an explanation
• More comments, examples
• A note of thanks

And in closing, something pretty. Below the fold is the brAhmI alphabet/alphasyllabary. It is, of course, studied by Buddhists for them to be able to understand ancient Buddhist texts, some of which were written in Pali. Just in and of itself, the brAhmI alphabet bears witness to the extraordinary influence that Panini still has around the world. I have a Theory of History which I have never posted about before. Want to know more? Click below to go to the rest of the post.

Tamil writing system proposal - Overview

This series of posts is intended to have the same precision of language and exposition as an academic paper but is written in an informal style so as to appeal to a broad audience. Here is an overview of the proposal that is intended to serve the same purpose as an Abstract in an academic paper.

-+-

In this series of posts, we introduce new diacritical marks for the Tamil language. The proposal has the following five elements: (a) allow all consonant clusters; (b) represent the aspirated consonants using the aaytam symbol (the 'therefore' symbol in English, Unicode U+0B83) prefixing the consonant to be 'aspirated', i.e., to go from 'ka' to 'kha'; (c) represent the soft consonants using a horizontal bar (Unicode U+0955) going over the consonant to be 'softened' (, i.e., to go from 'ka' to 'ga', simply draw a horizontal bar over the 'ka'; optionally, place a backslash-underscore after the character); (d) use a horizontal bar to represent the 'R' (Unicode U+090B) sound found in Sanskrit in Tamil (the syllabic alveolar trill) (or, optionally, place a backslash-underscore after the character); (e) represent the additional vowel sounds in Modern Hindi using the crescent symbol (Unicode U+0945) to produce the two additional sounds for U+090D and U+0911 (or, optionally, place a backslash-parenthesis after the character); (f) represent long vowels using the 'avagraha' symbol from Devanagari (U+093D); and (g) represent consonant-vowel conjuncts using the 'pull' symbol (the 'dot' symbol in English going above the character in question) for the consonant followed by the vowel in question.

-+-

Tamil writing system proposal - more comments and examples

I am writing this post nearly a year after the first post and so in this post, I would like to summarize the proposal for the Tamil writing system and, furthermore, add a symbol for the 'R' sound (as in kRshNa) to Tamil and capture two more sounds from modern Hindi. The 'R' character would be represented as a horizontal bar on top of the Tamil 'r' according to this proposal. Here is the proposal thus far:

1. Soft and aspirated sounds: To get (ka), (kha), (ga), (gha) in Tamil, just use one of the following:

a. In Tamil script: k, .: k, k\_, .: k \_

(here, the ".:" denotes the 'therefore' symbol and the '\_' denotes the horizontal bar).

b. ASCII: To represent this in ASCII, one could use the convention above where a period followed by a colon is used to denote the 'therefore' symbol and the backslash followed by an underscore is used to denote the horizontal bar as used in Sanskrit.

Note that a different way to represent this in ASCII is, of course, Harvard-Kyoto.

2. Conjuncts/consonant clusters: We allow all consonant clusters. We simply use the dot character to allow for consonant clusters of arbitrary length (e.g. lakshmya would be written as 'la' + 'k' with a dot on top + 'sh' with a dot on top + 'm' with a dot on top + 'ya'.

3. 'R': We represented the 'R' character by a Tamil 'r' character with a horizontal bar on top. Optionally, this may also be represented by the 'backslash-underscore' combination following the 'r' character in Tamil.

4. Chandra/Crescent symbol for a: We represent the 'a' sound (as in 'hat') using the arc symbol used in Modern Hindi. The arc symbol would go either on top of the preceding consonant OR if there was no preceding consonant, it would go on top of the character for 'a'. (this is the same as in Hindi). This can also be represented using a backslash followed by a parenthesis. e.g. 'hat' could be written as 'ha' with a horizontal bar on top + '\(' + 'T'.

5. Chandra/Crescent symbol for 'o': We represent the 'o' sound (as in 'got') using the arc symbol used in Modern Hindi. The arc symbol would go either on top of the preceding consonant OR if there was no preceding consonant, it would go on top of the character for 'A'. (this is the same as in Hindi). This can also be represented using a backslash followed by a parenthesis. e.g. 'got' could be written as 'kA' with a horizontal bar on top + '\(' + 'T'.

6. Consonant-vowel conjuncts: We allow for consonants with a dot to precede vowels to create consonant-vowel conjuncts. (e.g. 'Union' could be written as 'y' with a dot on top + 'U' + 'n' with a dot on top + 'i' + 'ya' + 'n' with a dot on top). This is an optional feature that reduces the learning curve. Now, in order to able to write conjuncts in Tamil, you only need to know how to write the corresponding character and how to write the vowel.

7. Avagraha/long vowels: We borrow the avagraha symbol, again from Devanagari, for prolonging vowel sounds. It is particularly useful for representing the 'schwa'. (e.g. University would be written as 'y' with a dot on top + 'n' with a dot on top + i' + 'va' + avagraha symbol + 'r' with a dot on top + 'si' + 'T' with a dot on top + 'i'.

-+-
Below are some example words represented using the new diacritical marks.

Examples

1. rAma           = 'r' + the 'aa' symbol (the symbol that takes you from ra to raa) + 'm'

2. lakshmaNa  = 'l' (without the dot on top) + 'k' with the dot on top + 'z' (the first 'sh' character) with the dot on top + 'm' (without the dot on top) + 'N'
3. bharata          = '.:' + 'p' with a horizontal bar on top + 'r' (without the dot on top) + 't' (without a dot on top)
4. shatrughNa   = 'z' (the first 'sh' character) + 't' with a dot on top + 'ru' + .: + 'k' with a horizontal bar as well as a dot on top + 'N'
5. Tattoine         = 'T' + the arc symbol on top of the 'T' + 'T' with a dot on top + 'Too' + 'i' (small i) + 'n' with a dot on top.
6. Skywalker     = 's' with a dot on top + 'k' plus the prefix for the 'ai' (Harvard-Kyoto) sound + 'v' + the 'aa' symbol (the symbol that takes from ra to raa) with a crescent on top + 'k' with a dot on top + 'k' + 'r' with a dot on top.

Tamil writing system proposal - a note on the alphabet

A very short note on the Tamil alphabet. There are (around) 36 characters in the Tamil alphabet. Here is the Harvard-Kyoto table with the Tamil characters underlined and highlighted. Not all the characters in Tamil are part of this list and not all characters in the list, as can be seen, are part of the alphabet. Just so people can follow along.

 a A i I u U R RR lR lRR e ai o au M H
   k kh g gh G c ch j jh J
   T Th D Dh N t th d dh n
   p ph b bh m y r l v z S s h
                             

More on serendhomophones and serendreduplication

So I am coining two new sets of terms in the field of linguistics. The first set is 'serendhomophony'/'serendhomophones' and the second one is 'serendreduplicates'/'serendreduplication'. Perhaps, people have been implicitly aware that this stuff was out there. But bringing something to explicit awareness by codifying it in academic language counts. And I don't think anyone has actually described it in the mathematical language of chance and probability. So here is a very simple but very precise description of these two terms.

First, 'serendhomophone'/'pseudohomophone'. A 'serendhomophone'/'pseudohomophone' is defined as follows. Sometimes, you may have a set of words/phrases in language A that are homophones with words/phrases in language B. This homophony is usually quite apparent for speakers of language B who don't know language A. These word/phrase pairs are called 'serendhomophones'.

Why does this happen? Pure chance. Every sound in language B maps to some sound in language A. So when words in language A are read out in close succession as in speech or song, the words in language A map to some arbitrary words in language B. (This mapping is not fixed. Based on intonation et cetera, the mapping can even change.) So essentially, words and phrases X1, X2 and X3 in language A that happen to map to words and phrases Y1, Y2 and Y3 in language B when they are spoken in a particular way and in a particular sequence. Note that there is a lot of serendipity in terms of what the particular word Y1 will end up meaning in language B. (Sounds have a more or less arbitrary assignment to meaning, generally speaking.) Note also that languages often don't have the same set of sounds, and so the correspondence between pseudohomophones may not really be very exact and often is not.

Here is a 'misheard lyrics' version of the Tamil song "Kalluri vaanil kaayndha nilaavo?" (now often referred to as "Benny Lava") featuring the Tamil dancer Prabhu Deva.



Note that Buffalax does not use actual pseudohomophones, but a subset of the sounds in the 'misheard lyrics' version of the song are pseudohomophones.

The first few pseudohomophones for the words in the song would be the ones below.

'Kalluri' <=> 'Cull lure E'
'vaanil' <=> 'Vaughn nil'
'kaayndha nilaavo' <=> 'coin the nil ah woe''

Now, for 'serendreduplicates'/'pseudoreduplicates'. 'Pseudoreduplicates' are words/stems of different origins that are thrown together to form 'duplication' within a word. (These words are also called 'serendreduplicates' in my terminology). 'Pseudoreduplication' is the phenomenon wherein a word looks like a reduplicate in that there is some apparent duplication within the word but the word is, in fact, not a reduplicate, and the reduplication has, in fact, arisen by chance.

In a nutshell, 'serendreduplicate'/'pseudoreduplicate' words have some 'duplication' in them that has arisen by sheer luck. It is worth emphasizing this point. Why does reduplication arise? Well, by pure chance. Or, you could call it serendipity.

Serendreduplication and serendhomophony

First a note on the terms 'serendreduplication' and 'seredreduplicates'. A Google search for pseudoreduplication returns no less than 44 hits, out of which at least two ( 1, 2 ) point to academic papers in linguistics. So I have decided to avoid the term 'pseudoreduplication' since that term is already being used in linguistics.

Footnote 2 in the paper by Avram has this to say about pseudo-reduplication. 'Also called “quasi-reduplicated forms” (Bakker 2003: 40), “phonological reduplicated base form” (Miller 2003: 290), “fixed forms” or “fossilized forms” (Wellens 2003: 226). The comparative Austronesian dictionary says this about 'Quasi-reduplicates' : "Quasi-reduplicates" are those words that lack the corresponding unreduplicated forms e.g. didi 'pig, swine', nisnis 'beard'. It seems that the original unreduplicated form may have been lost within the language leaving just the reduplicated form as a sort of fossil. Note that this is not at all what serendreduplication is. It is, well, something completely different.

Now, serendreduplication may be used to motivate the idea that many things in languages may have arisen purely out of chance. I can think of so many things like that. For instance, the (Chinese) name Yao (as in Dennis Yao) is the slang word in Tamil for 'sir' or 'man' (and comes from the word 'ayya'). Pure chance that it is the case. (Think of what it would be like if a guy walked up to a Tamil speaker and told him his name is "Yao Yennaya"). And there is also the more unfortunate case involving the firm "Lund International'. The name of the firm has a somewhat funny meaning in Hindi.

But the point is that this sort of serendipity arises. And this sort of serendipity is, in fact, exactly what Buffalax of misheard lyrics fame exploits all the time. So what should we call the pairs of words and phrases that sound the same in the 'misheard lyrics'? I say let us call them 'serendhomophones'. They are words or phrases that similar to each other but it is pure serendipity that it is so. Note that the correspondence is not exact. It is just "close enough". And so, that leaves us with the minor matter of naming the phenomenon. And that is simple. This phenomenon we shall call 'serendhomophony'/'pseudohomophony'.

Update: Another example of 'pseudoreduplication' would be the word 'metametals', a word that I just coined to describe elements which occur next to metals in the Periodic Table. Of course, many metals are also metametals.

The term 'metametals' could also be used to describe artificial 'metals' engineered to have certain types of properties, that is, a metamaterial that is a metal. The word 'metametal' is a good example of 'serendreduplication', and is probably a better example of 'serendreduplication'/'pseudoreduplication' than 'metamathematics'.                          

Second Update: Googling for 'metametals' returns plenty of hits.

Final Update: Here is a 'misheard lyrics' version of someone doing "Gangnam style". It is one funny video. But if you are a speaker of an Indian language as well as English, I absolutely insist that you watch it. That will give you an idea of what is going on in the brains of non-Tamil speakers when they hear the "Benny Lava" song. There is a remarkable correspondence between the lyrics for this version of the song and what you hear in your head if you are an English speaker who doesn't speak Korean. It is exactly the same thing that is going on with the "Benny Lava" song as well. Words that may not seem like homophones to you may, in fact, be homophones inside the brains of English speakers who don't speak Tamil.

     

Serendreduplication, a new term in linguistics

I would like to coin a new term in linguistics called 'Serendreduplication'. (I also sometimes refer to it as 'pseudoreduplication' since that seems to be a more natural way to describe what is going on.) Now, reduplication, as the linguists among you will know, is a morphological process in which, as SIL International's site puts it, "a root or stem or part of it is repeated."

Well, 'serendreduplication'/'pseudoreduplication' is the phenomenon in which although it seems like the root or stem of a word (or part of it) is repeated exactly, it is actually two different words with entirely separate origins that are combined serendipitously (or there is some other type of "happy accident") such that it looks like there is reduplication going on.

Two examples of this are 'metamathematics' and 'abracadabra' in English. First, 'metamathematics'. The prefix 'meta' comes from the Greek preposition μετά = "after", "beyond", "adjacent", "self" whereas the word 'mathematics' comes from the Greek word μάθημα máthēm - "knowledge, study, learning".

Note that the prefix 'meta' is always what is used to describe a concept that is an abstraction from another concept. It is just sheer chance that the prefix finds itself added in front of a word that sounds like it. Not my favorite example (and trust me, there is a far better one coming), but it gets us started on the right track.

Okay, onto the word 'abracadabra'. The word 'abracadabra' is a word that is not only fun to say but interesting in its own right. It is known to have origins in Aramaic. As Wikipedia puts it, 'Although at first glance "Abracadabra" appears to be an English rhyming reduplication it in fact is not; instead, it is derived from the Aramaic formula "Abəra kaDavəra" meaning "I would create as I spoke")" Princeton's Allison Chaney has a helpful introduction to the word on her site.

The first known mention of the word ABRACADABRA was in the 2nd century CE in a book called Liber Medicinalis ^[1] (sometimes known asDe Medicina Praecepta Saluberrima) by Quintus Serenus Sammonicus,physician to the Roman emperor Caracalla, who prescribed thatmalaria^[2] sufferers wear an amulet containing the word written in the form of a triangle:^[3]

A - B - R - A - C - A - D - A - B - R - A
A - B - R - A - C - A - D - A - B - R
A - B - R - A - C - A - D - A - B
A - B - R - A - C - A - D - A
A - B - R - A - C - A - D
A - B - R - A - C - A
A - B - R - A - C
A - B - R - A
A - B - R
A - B
A

This, he explained, diminishes the hold over the patient of the spirit of the disease. Other Roman emperors, including Geta and Alexander Severus, were followers of the medical teachings of Serenus Sammonicus and are likely to have used the incantation as well.

I came to think about this idea due to a quiz question by my friend Govind Krishamurthi which went as follows :

QUESTION: In Tamil examples of this linguistic construction are:
Mada-Mada (faster)
<stuff deleted>

In Hindi, 
<stuff deleted>
Examples would be

<stuff deleted>
Khaana-Vaana
Shaadi-Vaadi
<stuff deleted>

There are also several examples in Telugu.. but I didn't post all of them.. Some of you can give examples in Telugu, Kannada and other languages too.

So the question is what is this linguistic construction called?

ANSWER : REDUPLICATION.

That gives you a good idea of what reduplication is. It also gives you enough examples to play around with so that you can infer what serendreduplication/pseudoreduplication is as well.

The Tamil writing system as an app - a response to some comments

At least a couple of people raised the concern regarding whether the Tamil writing system proposal would require people to learn new sounds. It seems that the concern is about the audience for the writing system - - or as I would like to put it, the concern is regarding the market and any potential requirements placed by this product on the market's population. It would be useful to clarify at this point that the target market for this is people in the United States and Singapore. Furthermore, it is only for specialized users. It is only for people who would want to represent these sounds. It is not intended for elementary and middle school children, for example. Again, I would like to emphasize that this system is only intended for use by specialized users.

The reason for my choosing this population is not that this system would only work for specialized users from the United States and Singapore. The reason is that, first, this subject area is highly political and so I want to basically avoid all the politics around this by simply saying that this system is being made available for those who would like to use it. If you don't like it, don't use it. As simple as that. This also allows me to not worry too much about politics involving people in the Third World, academics in various departments in various universities, et cetera. These people may not have had the best educational opportunities and may therefore have trouble learning entirely new sounds (because they may be from extremely disadvantaged backgrounds). I am referring, of course, to the people in the Third World, the first population referenced. The question of whether this former population of people should be made to learn new sounds and letters arises (given that the educational system in the Third World is uneven in terms of quality) and this is certainly a valid concern. Second, there are many sociological concerns about whether new sounds should be introduced into a language in the first place. However, as I clarify below, I am not proposing to introduce any new sounds into Tamil. This system is intended simply to represent sounds that journalists and writers in Tamil already are required today to represent as part of their jobs.

Indeed, I have kept my system to what  I would like to call an optimal minimum - neither so little is proposed that no significant value is attained nor so much that it creates political problems. When I think back to some results I have seen in the field (based on some work I was doing while at Harvard Business School), it would appear that it would be better for languages to have fewer symbols and sounds, not more, especially since having a very large number of symbols and sounds may impose a heavier burden on poorer children (and note that I am talking about poor children in the Third World) who may already know to speak the language (as native speakers) but do not have family members who can teach them to read. This is indeed a good reason why policy makers would not want to allow new sounds in a language. In fact, I think that this may be the only one good reason why issues around language would acquire a political hue. It is for this reason that I am not proposing to introduce any new sounds into Tamil.