The Brahmi alphabet is one of the three major scripts of the Kushan period. The other two are Bactrian/Greek, and Kharoshti. Brahmi is first known from the edicts of the third century BC king Asoka, where it was used to write the language Prakrit. It came to be widely used, with regional variations across all of India, and eventually the regional variations developed into the medieval and modern scripts of the sub-continent. In the Kushan period Brahmi is used mostly in the south-east of the Empire, around the city of Mathura. What makes it important, is that hundreds of religious inscriptions written in Brahmi have been recovered from that site.
Paleography is the study of letter (known as aksaras) forms. It is an important part of interpreting and understanding inscriptions. Firstly, because it allows epigraphers to read the symbols themselves, despite their tendency to change over time. Secondly, because it provides a method of dating the inscriptions. The aksaras used in an inscription of the Gupta period are not the same as the aksaras used in an inscription of the Indo-Parthian/Indo-Greek period. And since about three quarters of all inscriptions have no date, or political information (such as the name of a king), they can be dated only on the basis of paleography. No-one has any doubts about the effectiveness of such methods in dating inscriptions over a broad time frame (hundreds of years), however the technique has been made controversial in Kushan studies because it was employed by Lohuizen to try and separate the inscriptions of the second Kushan century. This is a dating of an extremely fine nature (distinguishing letter forms no more than a hundred years apart). It does however provide a good opportunity to test paleographic dating because Lohuizen was thorough enough to leave us details of her analysis.
Language and Translation
Before historians can begin to interpret inscriptions, they must first be read. Though the inscriptions are written in one script they are not written in a single language. The Kushan period is in the middle of a process known as sanskritization. The languages used in the early inscriptions were prakrits. This was gradually replaced by Sanskrit in Indian inscriptions. However, this process was not a simply matter of one language sharply replacing another, or of an older language (the prakrit) gradually developing into Sanskrit. No, there was a lengthy period when inscriptions were written in a curious mixture of Prakrit and Sanskrit. What is interesting is that some pure Sanskrit inscriptions are found alongside these mixed (or Epigraphical Hybrid Sanskrit, EHS) inscriptions, so it is clear that people knew Sanskrit and could have written it if they chose too. EHS also exhibits a wide range of hybridization, some inscriptions looking basically like Sanskrit with spelling mistakes, others looking like Prakrit with a few Sanskrit words. For the sake of simplicity all of these variations will be labeled EHS, and only pure Sanskrit epigraphs picked out for comment.
The act of reading an inscription is often quite tricky, and it is rare to find an inscription of any importance on which there is not some level of disagreement between epigraphers as to its content. These disagreements are usually small in nature, they often revolve around the translation of a single word or even in some cases the reading of a single letter. Despite this it is surprising how often small differences can have huge impacts.
Probably the most important part of the process for the historian is interpretation. This has been a task neglected in the past hundred and fifty years, because it first requires the editing and translation of the substantial number of inscriptions recovered from all over India. So neglected is it that there does not even exist a comprehensive list of inscriptions relating to the Kushan period. There has been a list compiled by Luders, more than 50 years ago, but it includes only inscriptions before 300AD and only those in Brahmi.An index of the inscriptions on this site is available, and though it runs to more than 500 entries it is still incomplete.