84% Got Bail. The 16% Who Didn't Tell You What the State Actually Had to Prove

You Are the Sessions Judge Deciding Bail for the

43rd Worker That Week Imagine you're the sessions judge in Noida, working through your forty-third bail application connected to the same FIR. The prosecution's case reads almost identically each time: a crowd of workers entered company premises, carried lathis and stones, raised slogans, caused damage.

This particular accused was in that crowd. Nothing in the file says what he specifically did. Do you deny bail because a genuinely violent riot did occur and someone has to answer for it or do you grant it because "he was present" is not the same as "he did something"?

This is not a hypothetical for dramatic effect - it is, almost verbatim, the dilemma 222 bail orders in this case actually worked through and the state's own numbers show how hard the question actually is: 84% granted, 16% denied and the line between the two groups wasn't randomness.

It was evidence.

The Framework: The State's Duty to Prosecute Real Violence Versus the Presumption Against

Guilt by Association Both sides of this dilemma carry real weight and it's worth naming what each one actually costs if you get it wrong. If courts grant bail too readily, genuine organisers of coordinated violence - arson, assault, property destruction that real companies and real employees suffered - walk free while investigation continues and future crowds may read leniency as permission.

If courts deny bail too readily, ordinary workers who showed up to protest lawfully for higher wages spend weeks or months in jail for a crime they may have had nothing to do with, simply because they were standing in a crowd where a smaller number of people did something violent.

The courts in this case resolved the tension not by picking a side in the abstract, but by insisting on a specific evidentiary standard: individualised proof of a specific act or role, not membership in a crowd. Where the prosecution could show WhatsApp group administration, recovered banners tied to a specific person or CCTV footage of a specific individual's conduct, bail was denied.

Where the prosecution offered only "he was there," bail was granted - repeatedly and explicitly on that reasoning.

The Resolution This is the

right standard and it's worth being honest about what it costs rather than treating it as a costless, obviously correct answer. It means some people who genuinely supported or even quietly approved of the violence, but whose individual conduct the state couldn't specifically evidence, go free pending trial.

That is a real cost - to the companies whose property was damaged and to the principle that violence should have consequences. But the alternative - using presence at a protest itself as sufficient grounds for prolonged detention - effectively criminalises protest itself, treating everyone in a crowd as equally liable for whatever the crowd's most violent members did.

Between two imperfect outcomes, requiring the state to prove an individual's specific conduct before locking them up for weeks is the standard that protects the right to protest without abandoning accountability for those the evidence actually implicates - which is precisely why the NSA cases against Verma and Chaudhary, built substantially on reading material and prior protest history rather than specific violent acts, sit on much shakier ground than the FIRs where CCTV and WhatsApp records did the work.