15 November, 2010

Artistry at Girnar


Chandrakant Doshi

Abstract: This note examines certain artistic rendering of Brahmi script at Girnar and postulates its influence on modern Gujarati script.

KEY WORDS: Brahmi, ancient, script, inscriptions, Asoka, Girnar, Gujarati

Asoka, who ruled in India in the third century BCE, left a series of inscriptions all over his empire. These inscriptions comprise his famous edicts, found on rocks, pillars and in caves. A set of these edicts can be seen on a boulder near Girnar, in Saurashtra, Gujarat.

These edicts could be read in modern times following their decipherment by James Prinsep. Alexander Cunningham, who published a collection of these inscriptions1 in 1877, noted variations in the rendering of specific letters in the Girnar script2 when comparing with the Delhi-Topra and Dhauli edicts. He made no mention of variations in medial or diacritic marks attached to the glyph of a syllable. This note examines some of these variations observed in Girnar edicts.

The inscriptions are in the ancient Brahmi script. List of Brahmi script, along with Romanized Devanagari transliteration are available on various web sites, some of which are given in the bibliography3. I have prepared a table of Brahmi, Devanagari and Romanized Devanagari using the Kyoto-Harvard transliteration scheme and a copy is reproduced in the Appendix (Image 12). The Brahmi characters are hand written and generally follow the examples found at Girnar. I prefer Kyoto-Harvard transliteration scheme for its compactness and use it throughout this note, printing such occurrences in red.

Eugene Hultzsch4 published a fresh compilation of these edicts in 1926. He worked with fresh mechanical copies furnished by Archaeological Survey of India.

A comparison of Cunningham's copy of the rock edicts at Girnar with the estampages of Hultzsch shows that in Cunningham's reproduction the Brahmi characters are schematics, closely following the calligraphic quality of Delhi-Topra5 edict. A section of the North Face of this pillar edict is reproduced for reference, Image 13, in the Appendix. The first edict is delimited by green arrows. A copy of the Devanagari transliteration of the same is given below it (Image 14). Both these examples are from Hultzsch's corpus.

Asoka's first edict from Girnar is widely used as an illustration on many web sites. It is based on Cunningham’s copy and shown on a pink background but true colour of the rock is grey. This can be seen in the photographs of the boulder (Images 7 and 8), reproduced further down.

The Girnar edicts on the rock face are arranged in two columns, separated by a vertical line. Each edict is also separated by a horizontal line. The left hand column carries the first five edicts followed by the 13th at the bottom. The right hand column carries the edicts 6 to 12 and the 14th at the bottom. The characters in the edicts are 1.2 inches6 high.

The image of the first five edicts from Cunningham's corpus7 can be seen in the Appendix, Image 15. The first two edicts from a plate in Hultzsch’s corpus8 are also reproduced to illustrate the variations in script, Image 16.

For the close comparison of Brahmi characters in the two images, specially enlarged and annotated images of the first edict from each source are reproduced below, Images 1 and 2. A Devanagari transliteration9 of this edict from Hultzsch is also included, Image 3.

The diacritic or medial mark for the short i is a vertical line above the glyph of a syllable. In most cases it is attached to the glyph via a horizontal line, the two lines being normally at right angles. The long i is marked by two vertical lines. An example of both these can be seen in the first line of Girnar edict. In the reproduction from Cunningham, Image 1, these two occurrences are marked by blue arrows.

Examples of immaculate rendering of these diacritic marks can be seen in the first line of the Delhi-Topra edict, Image 13.

An examination of the same first line in Hultzsch's estampage, Image 2, shows a completely different situation. The short and long i in the first line, pointed by the blue arrows, are free flowing, more cursive rather than angular. Looking at the pointed glyph on the fifth line, which is the syllable mhi, the short i is rendered as an arc atop the glyph of the syllable ma. This is done in a manner completely different from the first two examples. Script of Girnar edicts represents an artistic flair not seen in other places.

This individuality of expression is clearly seen in five examples of the syllable pri, pointed by the red arrows in both the above images.

Instances of the immaculately rendered short or long i in the first edict are more an exception than a rule. Two such examples occur in the tenth line. One can be seen in the word dhaMmalipI, though both medial marks on the glyph of the syllable pa are not clearly visible on the estampage. Even Cunningham shows only one medial mark but Hultzsch’s Devanagari transliteration (Image 3) has the long i. Damage to the rock makes it difficult to make out both the diacritic marks. The other instance can be seen towards the end, where tI has two vertical marks for the long i clearly visible.

In the second edict, Image 4, both the perfectly vertical and cursive form of the diacritic mark for i are observed. However, slackness has entered on the very last word, mainly in regard to the glyph of the syllable sa.

The word is pasumanusAnaM and is enclosed in a red rectangle in the image of the estampage below.

On the left, the blue arrows point to sa (top) and pa (bottom). They both have an upright limb ending in a semicircle at the bottom. What distinguishes them is the little wing to the left of the limb attached to sa but not pa. This is clearly seen for the marked characters and can also be observed in the table of scripts, Image 12.

An examination of pasumanusAnaM shows that while pa has its limb upright, that for sa is slanted. This is observed twice in this last word but nowhere else in this edict. The improved engraving in the second edict indicates the presence of some supervision which seems to have disappeared at the last moment. The experimentation with the glyph of sa in that last word is not repeated elsewhere.

Girnar is located in Gujarat, where the regional language is Gujarati. Its script is based on Devanagari, with two essential differences. This is best illustrated by an example.

The horizontal line on top of each character is absent in Gujarati. In Devanagari, the vertical limb in a character or a diacritic is a rigidly straight line. In Gujarati such lines end in neat little loops.

The Brahmi script of the Delhi-Topra pillar edicts with its perfect calligraphy is also observed in some of the other pillar edicts as well as many rock edicts. The artistic flair observed in the Brahmi script of Girnar edicts is not found elsewhere.

The Devanagari script reflects the calligraphic quality of the Delhi-Topra edict while Gujarati reflects the cursive nature of characters found in the Girnar edicts. The rendering of the Brahmi in the Asokan edicts at Girnar, with their artistic flair, makes its appearance in the adaptation of Gujarati script from Devanagari.

Asokan edicts are over 2200 years old while modern Gujarati script has been in use for only a few hundred years. Examples of the loops in which the vertical limbs end, as noted above, can be seen in the centuries old copper plate inscriptions in modern Gujarati. One such example10 is reproduced in the Appendix, Image 17. It is dated Vikram Samvat (VS) 1711 (=1655 CE) and confirms a land grant originally made in VS 1611.

Photos of the Girnar rock.

It is rare to see many photos of the rock or pillar with the inscription readable. As luck would have it, while searching the web for material about Girnar, I chanced upon two photographs of the boulder with the Asokan edicts. With the kind permission of the photographer, Manish Khamesra, these photos are reproduced here.

The first photo, Image 7, shows the full height of the boulder with the right column in view. Edicts 6 to 12 and 14 are engraved on this face, though it is difficult to read the edicts at the top of the nearly conical rock. However parts of other edicts are possible to read. Horizontal lines separating each edict are also clearly visible.

Second photo, Image 8, is a close up of the lower part of the rock and shows the three lower edicts. Two horizontal lines separate the three edicts. The top edict, which has only a few lines showing in the photo, is the eleventh edict. Below it, between the horizontal lines, is the twelfth and at the bottom is the fourteenth edict. A part of the left hand side of the inscription is not visible in this photograph but can be read from the first, Image 7.

What is most striking about the photo is the crystal clear quality of the inscription. The estampage does not do justice to this beautiful work of art. A close examination of the photo shows areas where the rock has flaked. The area is lighter in shade and the incision of the character is shallow. These photos demonstrate the need to have all these edicts photographed.

A copy of the Devanagari transliteration of the fourteenth edict along with Hultzsch’s estampage of the same is reproduced here (Images 9 and 10). Blue arrows on the image of the Devanagari transliteration mark the start position of each line of the inscription visible in the photograph above. Hultzsch’s translatation12 of the fourteenth edict is also included (Image 11).

The cursive diacritic for i, already observed, is still to be seen in the final edict, Image 10. Green arrows point to these in the first line of the 14th edict. Yet errors do creep in and can be seen at the very beginning of the first line, pointed by red arrows. The horizontal arm of the diacritic should be attached at the very top of the vertical limb of each glyph. After the first errors at the very beginning, the supervision is ever present, as pointed by the blue arrow; a missed glyph is being squeezed in. This is clearly visible in the photo above.

The technology of digital cameras has placed in the hands of researchers a tool that can provide copies of inscriptions quicker and of a better quality than those produced by earlier methods. They can augment the copies and estampages first produced nearly two centuries ago. A gallery of photos in public domain, such as Images 7 & 8 above, would surely inspire a few minds to study these ancient script and inscriptions.

This brief examination has revealed not only the artistry in the engraving of some characters but also occasional absence of supervision during engraving. The artistry in the glyphs points to a local tradition of the third century BCE that has shown itself in the modern Gujarati script.

A detailed examination of all the edicts at Girnar will help to identify the artistry mentioned in brief in this note but also enable us to understand the processes involved in engraving the edicts in the third century BCE.


I wish to express my thanks to Norman J. Street for providing access to the books not easily available. Many thanks also to Manish Khamesra for the kind permission to use the photos of the boulder at Girnar. Thanks also to my son Samir and his wife Susan for valuable comments and helpful discussions.



1/. Corpus Inscriptionum Indicarum Vol. I
Inscriptions of Asoka by Alexander Cunningham.
Office of the Superintendent of Government Printing, Calcutta, 1877

2/. Cunningham, op cit, page 14 - 15

3/. The following web sites have tables of Brahmi characters.

4/. Eugene Hultzsch
Corpus Inscriptionum Indicarum, Vol 1,
Inscriptions of Asoka, Government of India, 1925

5/. Hultzsch, op cit,
Image, p 122, inscription, p 119

6/. Cunningham, op cit, p 14

7/. Cunningham, op cit, Plate V

8/. Hultzsch, op cit, p 4

9/. Hultzsch, op cit, p 1

10/. The Cave-Temples of Western India by James Burgess and
Bhagwanlal Indraji Pandit, 1881, p112

11/. Manish Khamesra is a contributor on the website http://www.Ghumakkar.com
where the photos originally appeared.

They are part of a travel story "Junagadh through the ages", which can be accessed here: http://www.ghumakkar.com/2010/01/05/junagadh- %E2%80%93-a-journey-through-ages/

12/. Hultzsch, op cit, p26

© 2010 Chandrakant Doshi