Devanagari Ligatures

From The Sannyas Wiki
Jump to: navigation, search

It will be noticed from time to time that certain Devanagari character combinations in text form in the wiki are different from the forms shown in images, mostly of Contents pages of Hindi books published in the 90s and earlier. Likely this reflects not only a change of fashion but also the fact that Unicode, the current near-universal system of rendering characters in all writing systems as text electronically, is not capable (yet) of representing all of these specialized combined forms, called ligatures or conjuncts. And the most commonly used Hindi fonts are incapable of using all the potential that Unicode has developed, so the result is what we see.

Ligatures exist in many writing systems, including a wide variety of forms in different languages using the Latin alphabet. Even in English, a few survive in current-though-uncommon usage, relics of more ornate conventions from olden days. These are the a-e and o-e run together as æ and œ respectively, having still enough status to be in the ASCII 256 character set. The upper-case versions of those two, Æ and Œ, and the German ligature ß, representing two s's run together, are also in that set.

See also wikipedia:Typographic ligature for a general overview of how ligatures are used in various writing systems around the world and wikipedia:Devanagari#Conjuncts for a more detailed look at their use in Devanagari. The latter page has two pertinent tables, the first illustrating exactly what happens when two consonants run together without a vowel separating them for each possible combination -- that's 37x37=1369 possibilities -- the second a table of the 360 most common ligatures. That's a lot!

Wikipedia also outlines some of the "rules" that specify how these interactions might occur, many of them common-sense and explaining how to easily recognise most of those 360 most common ones. But a lot of combinations are not covered by these rules and have to be learned or recognised on a case-by-case basis. And this does not even start to deal with combinations of three or more consonants, sometimes as many as five.

The last general point to be observed is this: When a font or Unicode itself does not supply any "fancy" special conjunct character, but it is known that previous printing processes used ligatures, the convention is to attach a "virama", a ୍ diacritic indicating its inherent vowel is dropped. The next consonant then appears in its full normal way.

Here in the wiki we will settle for something more modest, a table of how current fonts treat ligatures differently from how typesetting happened for Osho's books twenty-five or more years ago. These are just a few known examples, more will be added. So far we have:

past image current text "joining" remarks
DW-lig.jpg द्व (dva) द and व (da and va) the most common in the wiki, appearing in
"dvar/dwar" ("gate") and "dvait" ("divided")
HM-lig.jpg ह्म (hma) ह and म (ha and ma)
NN-lig.jpg न्न (nna) न and न (double na) visually looking very much like त्र (tra),
easy to mistake, fortunately not common
TT-lig.jpg त्त (tta) त and त (double ta)
SHT-lig.jpg ष्ट (shta) ष and त (sha and ta) well, not exactly: ष is between
a स (sa) and श (a different sha)
KT-lig.jpg क्ट (kta) क and ट (ka and ta) still a common ligature but done in a simpler way
than previously. Note also a different ta than above