The core idea
An inscription is a material object carrying a situated statement. Script identifies the writing system; language supplies words and grammar; context connects the record to people, institutions and places. None of these automatically guarantees that every claim in the record is true.
1. Begin with the inscribed object
Epigraphy studies writing on materials such as stone, metal and terracotta. The support matters: a copper-plate charter could be kept and transported, while an inscription on a temple wall addressed people encountering that place. Ask where the object was found and whether it still stands where the writing was commissioned. A reused stone may have travelled into a later wall, separating the date of the text from the date of that wall. Surface weathering, broken edges and overwritten lines can limit reading. Documentation therefore records location, material, dimensions and the relation between text and object before proposing a narrative. A photograph of detached words is useful, but loses evidence that a whole-object photograph or find record may preserve. The inscription is more than its modern typed transcript.
Sources: ASI Bharat SHRI: Documenting and classifying inscriptions ↗ · ASI Bharat SHRI: Estampage documentation ↗
2. Script is not language
Write the same Hindi sentence once in Devanagari and once in Roman letters: the script changes while the language can remain Hindi. Ancient records similarly require separate identification of language and script. Many Ashokan records use varieties of Prakrit in Brahmi; northwestern examples also use Kharoshthi, written from right to left. ASI records Sanskrit inscriptions in Brahmi too, so seeing Brahmi does not by itself establish Prakrit. Letter forms change across time and place, helping palaeographers compare records. But a resemblance between one damaged mark and a modern letter is not a decipherment. A convincing reading must fit repeated signs, word forms, grammar and the surrounding statement. Nor can an undeciphered sign system be translated confidently merely because a proposed reading sounds historically attractive. Script identification is the beginning of interpretation, not its completion.
Sources: ASI Bharat SHRI: Epigraphy and scripts ↗ · ASI Bharat SHRI: Inscription gallery, Junagadh and Besnagar records ↗
3. Separate copying, transliteration and translation
A photograph records appearance; a transcript records a proposed reading; transliteration represents letters in another script; translation expresses meaning in another language. Each step introduces decisions. If a surface is damaged, a responsible edition distinguishes surviving letters from supplied restoration instead of silently completing the sentence. Editorial conventions differ, so consult its legend before interpreting brackets or dots. An estampage is a paper impression used by trained specialists to preserve fine surface details for study away from the object. It can be compared with photographs and earlier copies when a reading is disputed. These are professional conservation procedures, not an invitation to wet or rub a monument. For a classroom exercise, use an original drawing or paper facsimile: the reasoning can be practised without touching historical material.
Sources: ASI Bharat SHRI: Estampage documentation ↗ · ASI Bharat SHRI: Documenting and classifying inscriptions ↗
4. A year needs a reference system
The phrase “year twelve” is incomplete without knowing what is being counted. It could be the twelfth year of a ruler or a year in a named era. A regnal year begins from a convention connected to a reign; it is not automatically a CE year. ASI identifies Rudradaman’s Junagadh record with Saka year seventy-two and dates it to 150 CE. That documented conversion belongs to a specified era, rather than a general rule for every inscription containing seventy-two. In an invented chronology, if a ruler’s first regnal year begins in 1000 CE, the twelfth begins in 1011 CE under that counting convention. The eleven-year difference matters. An exact day would additionally require month, day, calendar conventions and the reign’s starting point. State the precision the evidence permits rather than manufacturing a full modern date.
Sources: ASI Bharat SHRI: Inscription gallery, Junagadh and Besnagar records ↗ · ASI Bharat SHRI: Epigraphy and scripts ↗
5. Worked reading: a model tank grant
Here is an invented record, not a translation of an ancient object: “In the third regnal year, the assembly assigned income from two fields to repair the village tank; Devika, daughter of a potter, gave one lamp.” Separate its claims. The assembly is presented as an allocating institution; two fields supply income; the purpose is tank repair; Devika is identified as a potter’s daughter; her own occupation is not stated. The text does not state field area, income amount, repair completion or whether all villagers participated in the assembly. The named woman is evidence of a represented donor in this model, not proof of equal rights throughout an entire society. If another hypothetical account lists payment to masons, it strengthens the claim that work was funded. It still does not establish how long the repaired tank functioned. Each inference needs its own supporting link.
What does the model inscription establish?
| Recorded statement | Supported reading | Still unknown |
|---|---|---|
| Third regnal year | A date within a reign | Modern date without the reign and convention |
| Income from two fields assigned | An allocation for the stated purpose | Field area and amount of income |
| To repair the village tank | Purpose of the allocation | Whether repair was completed |
| Devika, a potter's daughter, gave a lamp | A named donor and a stated relationship | Her own occupation; rights across society |
Sources: ASI Bharat SHRI: Documenting and classifying inscriptions ↗ · ASI Bharat SHRI: Epigraphy and scripts ↗
6. Worked reading: a king and a reservoir
ASI’s account of the Junagadh inscription reports Rudradaman’s restoration of the Sudarshana reservoir after flood damage. The record also praises royal action and presents a claim that the repair avoided oppressive taxation or forced labour. Read the two functions together: information about water management and a public representation of good kingship. The existence of the carved claim is directly observable; the claim about how every cost was borne needs independent assessment. Do not either accept every boast literally or discard the entire record as propaganda. The text’s references to earlier work under Mauryan rulers connect memory of infrastructure with a later political present. A careful historical question is therefore double: what happened to this reservoir, and why did the ruler choose to narrate its past in this way? Material and textual comparison can answer different parts.
Sources: ASI Bharat SHRI: Inscription gallery, Junagadh and Besnagar records ↗ · ASI Bharat SHRI: Epigraphy and scripts ↗
7. Build a connected, limited account
An inscription becomes more informative when connected to other evidence without forcing agreement. A royal order prescribing a payment establishes an instruction; a receipt may indicate collection; a petition may show resistance or a dispute. Different genres can reveal different stages of the same process. The absence of a name in one inscription does not prove that the person or group never existed: stone texts preserve selected acts and voices. Conversely, a spectacular new reading should not overturn a chronology unless the object’s provenance, letters and dating are adequately established. Your final account can be concise: identify the object, state the reading, explain the inference and mark the limit. This lets inscriptions illuminate administration, labour, devotion and exchange while keeping the difference between a recorded voice and a complete society visible.
Sources: ASI Bharat SHRI: Documenting and classifying inscriptions ↗ · ASI Bharat SHRI: Epigraphy and scripts ↗
PUT IT INTO PRACTICE
Practice: test four claims
- A label reads “Sanskrit, Brahmi”. Explain it. Answer: Sanskrit identifies the language; Brahmi identifies the script. The two labels answer different questions.
- In a model reign starting its first year in 1200 CE, find the start of year five. Answer: 1204 CE under the stated convention, not 1205; four completed years precede the fifth.
- A damaged word may be “gift” or “tax”. What should the edition do? Answer: record the uncertainty and compare surviving letters and context; the difference changes the action described.
- The tank model states an allocation. Can you report completed repairs? Answer: no; seek expenditure, construction or subsequent-use evidence before strengthening the claim.
Check your understanding
Can one script write more than one language?
Yes. ASI documents Prakrit and Sanskrit records in Brahmi; script and language must be identified separately.
Why record a stone’s findspot?
It helps reconstruct context and possible movement; the present location may differ from the original setting.
Is transliteration a translation?
No. Transliteration changes the script representing the text; translation expresses its meaning in another language.
Does royal praise make an inscription useless?
No. It is evidence of political self-presentation and may contain testable factual claims. Assess each claim separately.
What does one named donor establish?
That the record represents a particular donation and identity; it does not establish the distribution of rights across society.
Why not silently restore a missing word?
Readers would mistake an editorial inference for surviving evidence and could build further conclusions on an unmarked uncertainty.