No Stupid Questions

In a nutshell, revolution turns on information and power asymmetry. Sometimes just one of those is sufficient, having both makes the results highly likely, but not guaranteed.

Stating with information asymmetry, we start with the identifying the usual groups of people. I will be using terminology more akin to the Westminster style of governance, which does not so clearly distinguish the roles. In a governmental revolution, there are those in power (eg a president, prime minister, members of parliament, monarch), there is the state and its institutions (eg military, judiciary, civil service workers in those departments, treasure, welfare, foreign representatives), and then there's the citizenry (aka the people).

Of these, the citizenry are the absolute largest group but the least organized. In a monarchy or autocracy where power is concentrated in the very few, the citizenry are often denied the means of communication or it is strictly censored or controlled. In a republic, the state is created as the organization which is meant to serve the people, and I'm not aware of any republic that has ever created two duplicate organizations to guard against usurpation. To that end, the citizenry are the most dependent on the state and the government for information. Even when it's now technically possible to exchange information using mesh networks, online forums, ham radio, and even plain ol letters, the fact is that convenience means that the majority just aren't dialed into the situation, or that the official mouthpieces have enough sway to quell the public.

But it need not only be the citizenry that are kept in the dark. The government itself can end up being split apart by those who know versus those who don't. As an example, look to the former South Korean president that attempted to impose martial law. In the chaos that ensued, members of the legislature needed to understand what was going on first, in order to combat the situation. It eventually emerged that the legislatute was being blockaded and that a vote would be held to nullify the imposition of martial law. Photos of some legislators scaling the outside wall of the assembly made international headlines. That was only possible because enough representatives got word that a vote was going to happen, and that it wasn't a trap.

The South Korean example also shows what happens when the state is not on the president's side. The military was doubtful that the president could lawfully declare the legislature as acting against the interests of the country, and so they did not substantially mobilize. Likewise, the citizenry were not having it either and protested in public. Perhaps it would have been different if the president was able to sever communications lines, an often-used tactic in the hours prior to a coup.

As for power asymmetry, that's much easier to explain. The same groups as before each wield separate powers, some of which are more effective at times and some less. For example, the military has all sorts of hardware that could be used against the citizenry or against the state institutions. Shopping mall, tax offices, and city halls aren't exactly built to repel RPGs and mortar fire. The government also benefits from having authoritative power, meaning they can claim a mandate (eg from heaven, from the monarch, or from the people) that legitimizes their attacks on the state institutions or the people. See the Stalinist era of the USSR.

Meanwhile, the people have the power of populism, where the influence of social mores can and does have tangible impact. Look to the UK where MPs and cabinet members have been forced to resign "due to scandal", where their position "becomes untenable". From an American perspective, this would seem unusual since a corrupt politician would still end up serving their term. Yet in the UK, they recognize that they cannot continue in their job if nobody will ever look at them with a straight face. No committee would keep them on, they could never hold a cabinet portfolio, they can't effectively represent their constituency, and can't represent the country in good terms overseas. They could just sit there and collect the paycheque, but ultimately, they know their days are numbered or the government will have the police service investigate them. So they resign, simply because of the crushing weight of public opinion. That is power.

Finally, there's the institutions themselves that have power. With the presumption of regularity, institutions hold tremendous soft power. That is, without firing a gun, an IRS tax agent or DMV worker can make someone's day, or make it their worst day. A judge can grant search warrants that authorizes someone's house to be turned upside down. Or a department of transport can start eminent domain proceedings to acquire someone's home. Meanwhile, the central bank can change the value of money, even the banknotes in your wallet, overnight. So powerful are institutions that in at least two places in California law, one of which is the open government act, the law opens with a declaration that "The people of this state do not yield their sovereignty to the agencies which serve them". This is a warning against the institutions to not abuse the power they are entrusted with.

So, what does this mean for revolution? For both information and power, it's not about how much is possessed but how it is used. Sometimes information can coerce power to be used. The Zimmerman telegram was a large part of how the USA joined WWI, because the British intercepted it and realized it would spur Americans to support the war against Germany. Domestically, small power can be used to test a larger power, basically to try calling a bluff. If the police declare a curfew due to false allegations of rioting, protesting is a response to the dare: will the police actually try to pepper spray and arrest thousands of people that show up anyway? If they don't, they've folded. If they do, there is now information (eg video, photos, TV) that can be leveraged to encourage more power (eg more protests, or state intervention against local police). In the most extreme case, the police could respond with overwhelming force (see Kent State Massacre). But in that situation, it was so uncalled for that other powers responded: the USA's involvement in Vietnam and Nixon's presidency became more unpopular than ever, causing mandatory conscription to end in 1973. It has not come back since, because people will still remember that event. Even as the shooters in question escaped legal culpability, it has cost the nation the effective power to call the citizenry into military service. Such power would be tough to regain, because the citizenry would fight it. Hence why all such attempts since in the USA have failed to reintroduce conscription.

TL;DR: the balance of information and power ebbs and flows over time, sometimes yielding unique opportunities or colossal failure.