>>40173
>By the 1964 coup, Brazil already went through 6 or 7, some was successful, some were foiled
Then two more in the following years: Costa e Silva pressuring Castelo Branco to accept him as his successor in 1965, and the moderate President's countercoup on the hardliner Minister of the Army in 1977. Both successful.
The latter was less of a countercoup against a mounting coup, and more of triggering a coup when the countercoup was already in place, ensuring its failure. There'd been for a while but the Minister wasn't really constructing a proper conspiracy. He was called by the President and fired; in the next 12 hours he tried to improvise a coup, launching a manifesto against the process of thaw, invoking the reds, and summoning commanders. As the moderates had planned this in advance and had all their pieces in place, it was immediate checkmate and by nighttime the Minister gave up.
>Do any of your sources - especially the Oral History - reflect anything about how they thought about coup as a tool?
Legitimacy - There was an interesting exchange between a legalist commander and a rebel emissary, the former insisted on his legal duties and the latter, with the example of previous coups and revolutions, arguing that such duties become void with an illegitimate government. Other times this is framed as between legal duty and duty to the fatherland. One interviewee mentions how a coup d'état is the military's way of expressing discontent. And there's a strong emphasis on popular support as necessary for a legitimate coup.
Luck - the balance of power is influenced by the random fluctuations of military life. Commanders might be away on vacation leaving subordinates from the other side in charge. In some units conscripts will be well trained, deep into their cycle, while in others they'll be recently incorporated. To some extent can be addressed. Commanders with recent troops can subject them to accelerated training. If they keep their sharp they'll have the advantage when the crisis comes.
Cohesion - if a general is undecided or on the other side, his subordinates can arrest him or take over, but this will leave scars in military discipline and unity in the year to come - this was the case in the 30s. Ideally subordinates should apply pressure to commanders and units should pick a side as a whole.
Distributing information - an untrustworthy regiment can be dispatched without information. Within one's own garrison, however, not taking a public stand will just create confusion and division. One can make the stand, let the opposition speak and arrest it, make the stand and not leave any room for the opposition to appear, or ask everyone from the lowest to the highest rank (the order can matter).