Fun with abugidas (Part 1)

(Posted by Sandeep)

Most major Indian languages can be separated into two major language families–with North Indian languages mainly classified in the geographically diverse Indo-European family (with distant cousins as far-flung as Persian and Irish Gaelic) and the South Indian languages in the Dravidian family, which is mostly limited to the southern part of the Subcontinent.

Although grammatically and structurally quite distinct, many Indian Indo-European languages and Dravidian languages have some critical elements in common.

First, the various scripts used to write Indian languages evolved from one script, Brahmi, which has been dated at least to the 3rd century BCE (on the Edicts of Ashoka) and perhaps earlier.

Despite their common derivation, Indian scripts can look very different from each other.

Consider the Sanskrit quote I posted a few days ago, written first in Devanagari (used to write Hindi, Nepali, Marathi, among others) and then in Kannada (used to write Kannada, Tulu, Konkani, among others). Sanskrit is now mostly written in Devanagari, but historically it was written in whatever was the script in vogue in various regions of India.

Pretty different, right?

The apparent visual differences between North and South Indian languages is often incorrectly conflated with the actual structural differences between Indo-European and Dravidian languages.

For one, South Indian scripts, such as Kannada, Telugu, Tamil, and Malayalam, are “curvier” than North Indian scripts, which utilize more straight lines. However, this is popularly explained by linguists in India by the different writing media historically in use: ancient South Indians wrote on large dried leaves; straight lines would have punctured the leaves and rendered them useless, so South Indian scripts evolved more curves.

Whether or not this explanation is true, I think recognizing the common ancestor of the scripts of India is a great (and missed) opportunity to build unity.

Wikipedia has possible derivations of some letters in some Indian scripts from Brahmi:

In a nation of 22 officially recognized languages and hundreds, if not thousands, more unofficial languages, linguistic differences are used to divide people. The apparent differences in scripts are a major part of this divisive arsenal–“Oh, look how different Tamil looks from Bengali; they must be so different from me.” Why not use it for the opposite purpose? “It’s remarkable that even though Tamil looks different from Bengali, we share a common ancestor script.”

Folk etymologies and false derivations are rampant in India–especially because fluid word borrowings, especially from Sanskrit, confuse true linguistic relationships–but this is an actual, demonstrated, linguistically and historically valid commonality.

A common ancestral script may be a minor thing to note, but Indians could use all the unity they can get, right?

Alphabets, or why Indians were awesome linguists

Indians were incredibly awesome linguists. More on this later, but a brief overview: the Aṣṭādhyāyī of Sanskrit grammarian Panini (c. 500 BCE) is the earliest known work of descriptive linguistics anywhere in the world. Still, even Panini refers to older Sanskrit works on grammar. Linguistic ideas are built into the oldest of old Sanskrit texts and Sanskrit morphology and syntactic rules are some of the most complex and most developed of any language in the world, past and present. Four of the six branches of Vedanga (the study of the ancient Hindu texts the Vedas) are linguistic: phonetics, etymology, meter, and grammar.

In short, Indians were badass at linguistics.

Part of this badass-ness (badassitude?) came in the form of the organization of many Indian alphabets. Unlike the Latin alphabet, which came to its present order (A, B, C…) through a series of historical serendipities, the standard organization of the Sanskrit alphabet is remarkably systematic.

Many Indian languages now, even some Dravidian languages (which aren’t structurally similar to Sanskrit), use the exact same organizational chart.

Consonants are organized in an implicit table. On one axis, consonants are distinguished by the type of closure required for their production:

kaṇṭhya (velar), tālavya (palatal), mūrdhanya (retroflex), dantya (dental), and oṣṭhya (labial)

On the other axis, consonants are distinguished by voicing and aspiration:

aghoṣa alpaprāṇa (unvoiced unaspirated), aghoṣa mahāprāṇa (unvoiced aspirated), ghoṣa alpaprāṇa (voiced unaspirated), ghoṣa mahāprāṇa (voiced aspirated), then anunāsika (nasal).

So in the first row of consonants, you have velar consonants, beginning with an unvoiced stop and ending with a nasal.

/k/ /kʰ/ /g/ /gʰ/ /ŋ/

The pattern continues. At the end of that collection, there are several antastha (approximant) consonants, three sibilants, and a voiced fricative.

Here is a lovely table, adapted from Charles Wikner’s A Practical Sanskrit Introductory (1996).

This table is misleading, though, because it’s not quite the exact order that the alphabet is recited in. The consonants ya, ra, la, va, sa, sa, sa, and ha are recited after ma. Here is a better representation of the order, here in Kannada, but without the linguistic tags (Omniglot):

As far as I’m aware, this order is used more or less in the following major languages: Hindi, Kannada, Marathi, Nepali, Bengali, Telugu, Malayalam, Konkani, and Gujarati, among others. Tamil uses a similar, but reduced, organization.

The Indian obsession with linguistics is built into the very structure of its languages. And it’s awesome.

[Competition Update: We decided to withdraw ourselves from’s Best Grammar Blog of 2011 competition because we felt that voting was proceeding in an unfair manner. We are no longer participating.]