The cities of the Indus civilization left behind seals, pottery, and tablets marked with repeated signs. Archaeologists can study the objects and their settings, but the symbols have not yielded an accepted reading of the words or language behind them.
The scale of the surviving writing is a serious obstacle: the known corpus is only a few thousand inscriptions, and most are very short. A 2025 BBC report cited an average of about five signs per inscription and noted that no bilingual text has been found to act as a translation key.
Researchers have used computational methods to identify recurring patterns, evidence that the signs are structured. That does not by itself tell us whether a sign represents a word, sound, or another unit—or which language to test. Proposed decipherments remain disputed; an intriguing match is not a verified translation.