I've been trying, but I can find no evidence to support the theory that voting or not voting was needed for a system to imply or deny consent. Non-compliance, however, is MUCH more effective and I can find all sorts of evidence to support that.
It's an extremely hard thing to prove, or disprove, using historical examples. Even if I were to cite specific cases of "low turnout -> secession/revolution", that does not prove causation, just correlation.
But it's easy to prove using logic.
Based on the below assumptions:
1) Voting sends a signal of consent for voting as a system
2) Signals of consent shape the perception of a government's legitimacy
3) Perception shapes reality
This leads to the conclusion that when you vote, your signal of consent is helping to shape reality. If we accept that your vote has a chance to flip an election, then your vote has a much, much greater chance of increasing a voter turnout statistic's significant digits. E.g., from 34% to 35%.
Let's say if it did hit 34% someone would have tweeted a tweet about the illegitimacy of the government, and at 35% they would have chosen not to.
Someone who sees that tweet, might be considering disobedience. But he didn't see the person's tweet about illegitimacy because you chose to vote and that tweet never happened. But if you didn't vote, he might have been encouraged into doing disobedience from that tweet.
When you vote, you are
suppressing other people's disobedience.
The likelihood of your vote being meaningful in a national election, is far
far less than your vote being meaningful in suppressing disobedience.
On an individual level, both chances are close to 0. But on the aggregate, it makes a huge difference, especially over long periods of time.
If disobedience is so important (and it is!) -- why take active steps to
discourage it?
(I would also add, that this signal of consent is sent when people
see you vote as well. Not just in voter turnout statistics.)