The Supreme Court acquitted a man convicted of rape, holding that courts cannot selectively accept parts of the prosecutrix's testimony while rejecting other parts. The bench found that the High Court had erred by accepting portions of the victim's statement that supported conviction while discarding portions that created doubt. The Court reaffirmed that a victim's sole testimony can sustain conviction, but only if it is wholly reliable — internal inconsistencies make selective reliance legally impermissible.

This ruling crystallises a well-established but frequently misapplied evidentiary principle in sexual assault cases: the 'sterling witness' standard. Indian courts have long held that the testimony of a prosecutrix, if credible and reliable, does not require corroboration — this principle flows from the general law of evidence and has been reinforced in cases like State of Punjab v. Gurmit Singh (1996). However, reliability is an all-or-nothing assessment of the core narrative, not a cafeteria selection exercise.

What the Supreme Court corrected here is a logical fallacy courts sometimes commit: using the victim's statement where it helps the prosecution and discarding it where it creates inconsistency. If a material portion of the testimony is unreliable, that unreliability infects the whole account — a court cannot quarantine the 'inconvenient' part while convicting on the 'convenient' part.

The distinction CLAT PG tests repeatedly is between corroboration (not required but permissible as supporting evidence) and reliability (mandatory — the sole yardstick when a prosecutrix testifies alone). The examiner often presents a passage stating one principle and then asks whether a given fact pattern — where a trial court accepted some parts and rejected others — constitutes a correct application. Understanding that selective acceptance is impermissible, and why, is the core testable concept here.