Indian scripts

The variety of the world's scripts has long held a fascination for me, something which is independent of the variety of spoken languages they encode. Perhaps this has something to do with my musician's training, specifically training in Western notation. (In that sense all Western classical musicians could claim to have been using two scripts since an early age.) I think however that artists often have an equal penchant for scripts per se - even in the West, where they are not integrated into the mainstream of 'fine art' in the way they have been in Islamic and Chinese culture. At any rate they are not just, and even not primarily, a supplement to spoken language and thus a subject for linguists. One could begin to summarize Jacques Derrida's famous and intricate arguments on this point - but that would be rather unnecessary so early in the day. Better to observe the striking instances of non-verbal scripts: not only music notation, but dance, computer code, the notation of games (chess, go, bridge), certain kinds of assembly manual (decipherable or not)...

What I want to discuss in this post is however something that occurs particularly with language, and to some extent also with music. It seems to me that a script doesn't merely reproduce a language, which has its own identity already fixed through sound and meaning. It also constitutes part of the language's identity: the specific way in which written letters and the sound of words blend, or what we call spelling (although that word isn't normally used in the case of different scripts). This occurs despite the fact that many unwritten languages, and illiterate speakers of written languages, have existed for which none of this can hold true. What one can say is that, once a script is present to mind for a language and its speakers, it will begin to affect the way words in that language are imagined, and this involvement of the imagination will powerfully affect the identity of words, in the sense of how foreign or familiar they are felt to be.

To illustrate this very briefly first of all with a European example, I will appeal to any regular beer-drinker and ask them, if they are linguistically inclined, if they have ever idly peeled away the back label on a beer bottle, and in doing so noticed the spellings of the word "beer" and its ingredients in the various European languages (excluding those derived from the Latin "cervisia", such as Spanish). In German it becomes "Bier": the same sound more or less, although one somehow imagines it in a German accent, and the word as spelt would have gloomier connotations in English. Pronounce it with your best approximation to an Italian accent, and the Italian spelling "birra", with its extra syllable and rolled "r", immediately seems logical, although someone with no Italian might not recognize the word if it were printed, and out of context. For the most startling transformation however, you need a little bit of Greek. Although the pronunciation is roughly similar to Italian, not only is the script itself different, but it must cope with the odd fact that modern Greek has no single letter expressing the consonant "b" (since the second letter of the alphabet "beta" is actually pronounced "veta"). When forced to include the sound as part of foreign words, it has to improvise by shoving together "mu" and "pi". Beer is thus "mpira": a weird defamiliarization, splitting in two a consonant one had thought was utterly fundamental, atomic - a kind of linguistic equivalent of the first time one discovered that green could be produced by blending blue and yellow.

In the Indian languages this kind of thing goes on all the time with English words, and it is one of the first things you notice when learning their scripts since, of course, transliterations of English on adverts or shop-signs are both ubiquitous and among the first things one can decipher. The best way to present this is as a puzzle. I'll give the Indian spelling ("back-transliterated", of course, using the standard modern system for Brahmi and Devanagari minus the diacritics), and you guess the English word: kariyar; sarkas; kek; daktar; eyarplen; ilekshon; apel; ophis; meshin; pulis; injin. (Answers at the bottom of this post.)

Among the English consonant sounds that are trickiest to deal with are "t" and "d". The languages of north India all have about four letters that could reasonably correspond to each of these, distinguished by whether they are aspirated (pronounced with extra breath) or unaspirated, and whether the letter is taken to belong to the retroflex (tongue curled back against the palate) or the dental group (tongue touching the teeth). The problem is that English "t" and "d" sit in between the Indian categories: you pronounce them with a certain, non-emphatic amount of aspiration, and with the tongue loosely in front of the teeth. Naturally, both the script and the majority of Indian English speakers cannot deal with that, and so they have to make a choice. And this is always in favour of the "hardest" available sound: unaspirated and retroflex - the retroflex consonants paradoxically being the one group that native English speakers cannot produce unless they are physically shown how to pronounce them! The result is an (unnecessarily) inaccurate reflection of normal English pronunciation, and thus the most recognizable feature of an "Indian" accent. Get an Indian-born English speaker to say "dirty" or "today" and you'll hear what I mean.

A similar paradox occurs with "s", "z" and "j" in Bengali. Most native English speakers (with the exception of graffiti taggers) have probably ceased to notice that the plural ending "s" in many cases is actually pronounced as "z" - e.g. "things" is really "thingz". But Bengalis have registered this fact, and try to reproduce it when transcribing English words. Only the problem is, there isn't a letter "z" in the Bengali script. They can pronounce it, but the closest they can to writing it in their alphabet is a "j" - which isn't really the same thing as "z", bears no relation whatever to "s", and produces very humorous-looking spellings, such as "ledij" (ladies) or "chij" (cheese).

But there are other cases when English loan words have apparently been changed on the phonetic level in a way that can't be explained by the differences between the languages' phonemic systems. Why should "biscuit" become "biskut" in Bengali, for instance, and not "biskit"? Why "haspatal" instead of "hospital"? A few words are distorted almost out of recognition: why "ketli" instead of "ketil" (i.e. "kettle")? The word "English" itself is "ingriji", as if the Bengalis were like the Japanese and couldn't tell the difference between "r" and "l". "Geometry" is "jemiti", and yet the "tr" combination is ubiquitous in Bengali. But one can't pin blame on Indians alone for this, for the most numerous cases of such distortion occur the other way, in the adaptation of Indian words into English. Some spellings, such as the nuances of the different types of Indian "t", cannot be represented at all in Roman script without immediate recourse to pre-defined systems of diacritic marks, so the issue of distortion does not even arise. (This is one symptom of Indian scripts' superior phonetic power.) But in other cases some quite interesting linguistic mangling has gone on. Yule and Burnell, the editors of the famous Anglo-Indian dictionary "Hobson-Jobson" chose a radical and strangely symbolic example for the title of their work. "Hobson-Jobson" is a heavily Anglicized representation of the Shia cry during the procession of Mohurram, "Hussain, Hossan". The anguished, guttural invocation of the names of the early Muslim martyrs, screamed out in the midst of extraordinary scenes of collective self-flagellation, is "cleaned up" so far that the grandsons of the Prophet sound like two elderly Edwardian butlers being summoned to fetch more tea.

In the end however, rather than simply ridiculing such examples I think they deserve some degree of celebration too. They contain within them the sensible acknowledgement that the world's languages vary from each other so much on every level that we will never be able to reproduce even the sounds of one language accurately within the framework of another. (In fact, as I stressed earlier, even a language's "own" script doesn't "reproduce" that language in an exact or scientific fashion - something that so bothered George Bernard Shaw that he agitated for English to be written in a more phonetic script, the Shavian alphabet, resembling Pitman shorthand.) In the present academic orthodoxy, when including words from a foreign script in an essay, one tries to be as consistent and complete as possible in transliteration, establishing tables and adding diacritics or special characters until every letter in the original script has its one-to-one counterpart in an adapted Roman alphabet. This exactitude is desirable if you're a professional linguist. But the consequence of such an adaptation is that one has to learn how to read a good proportion of the altered alphabet anew. In the logically extreme case, the International Phonetic Alphabet, designed to be able to represent every language on earth, the result is something often resembling computer code.

The "Hobson-Jobson" solution is more creative: to use the idiomatic features of one's own language as effectively as possible, resulting for the English speaker in lots of doubled letters, endings in "-y" or "-ee" instead of "-i", or "-oo" instead of "u" ("Hindoo"), use of "c" in preference to "k" and so forth. If the result looks more like English than like Hindi, well it was always going to - because the only way to really make Hindi look like Hindi is to write it in Hindi script. Why not, then, have the quaintness of spellings like "chupatty" - which in this case happens to remind one of an English word close in meaning ("patty") - instead of the false conscientiousness of chapātī? (Here the "u" represents the short "a" of Devanagari, just as in another culinary loan, "chutney", which one would "properly" have to write chatnī.) Why not the familiar Calcutta place-name Chowringhee - even though chaurangī would be more precise? As in other cases (such as Indo-Saracenic architecture), the Victorians seem to have already anticipated postmodernism, in this case via the principle that "there is no metalanguage"; that we will never create a perfect symbolic representation of reality (whether human or natural) but will instead forever be embroiled in the imperfect, adaptive and improvisatory task of translation.

Just recently there has been a belated confirmation of the failure of any attempt to establish the Roman alphabet as standard for the globe: the decision of the Internet Corporation for Assigned Names and Numbers (ICANN) to allow Web addresses in non-Latin scripts, billed as an important move to establish more equal international access to the internet. And quite right too. Indeed there would undoubtedly be some more enthusiastic Hindus here who would say that, if any script ought to be universally imposed, it should be Devanagari, the writing system of classical Sanskrit. With more than double the number of letters in the Roman alphabet, organized systematically according to natural phonetic principles (the consonants are grouped according to how far back in the mouth their point of articulation is), Devanagari's logic and power are a good reminder of how intellectually far advanced ancient Indian civilization was - at a time when the "English", if they could be said to exist, did not even have a writing system of their own.

(Answers: career; circus; cake; doctor; airplane; election; apple; office; machine; police; engine.)