Interesting paper by Mark Katz on why some democratic revolutions succeed and others fail:

Certain theorists, including Crane Brinton and Timothy Wickham-Crowley, have argued that the role of the armed forces is the key factor in deciding whether a nondemocratic revolution succeeds or fails. If the armed forces protect the ancien regime, then the revolutionary opposition is unable to seize power. If, however, the armed forces do not protect the ancien regime, then the revolutionaries usually do come to power. I will argue that just as in attempts at nondemocratic revolution, the role played by the military is also a key factor in determining the outcome of democratic revolution. When the military is willing to use force to protect the ancien regime, democratic revolutionaries cannot prevail. It is only the refusal of the armed forces to use force that allows democratic revolutionaries to succeed.

What, then, determines whether the armed forces of an authoritarian regime will use force to suppress a democratic revolutionary movement? Using a comparison of the cases mentioned, I will argue that the decision by the armed forces not to protect an authoritarian regime is not the result of a democratic conversion on the part of the military as a whole, but that it results instead from an overwhelming desire to prevent conflict within the military. Thus, if even a small number of key commanders defect to the democratic opposition, this can neutralize the armed forces as a whole, even though most military leaders may be wary of, or even hostile toward, democratization. But if these key defections to the democratic opposition do not occur and the military remains unified, it is able to crush easily the democratic revolutionaries.

Since the obvious case of concern here is Iran, let's talk Iran. The military really has three branches here. There's the Armed Forces proper, for one. By all accounts, it's pretty loyal—though some analysts think that a downturn in the economy could provoke dissent. Next we have Iran's Revolutionary Guard (the Pasdaran), which was created in 1979 to protect the Revolution and enforce Islamic codes. These folks are perhaps the weakest link. In 1994 the Pasdaran showed a wee bit of disloyalty, when anti-clerical riots broke out in Ghazvin, and a number of Pasdaran and Army Commanders refused to suppress the crowds. Furthermore, I believe a number of Revolutionary Guardsmen voted for Khatami in 1997. Anyways, after Ghazvin, Tehran began relying mostly on the basiji--volunteer groups of vigilantes to maintain domestic order. This is the third group, funded mostly by religious charities. I'm not sure, but I don't think Ayatollah Khameini has direct control over the basiji. I believe they are mostly controlled by the Combatant Clergy Association, a conservative party which forms the largest bloc in parliament.

Anyways, the key here, I think, is that there are so many overlapping armed forces that someone is bound to crack down on a revolutionary movement. In Tiananmen back in 1989, as Nick Kristof reported, the 38th Army refused to fight the protestors, but it also refused to defend the protestors. So Beijing just brought in the provincial 27th Army to open fire on the square. Had the 38th Army been more aggressive in its defense, the revolution might have worked. But they weren't.

Something similar seems the likely fate for Iran, (albeit for different reasons) at least in the near future: Revolution breaks out, perhaps one branch of the military (the IRGC) averts its eyes, but the other branches—particularly the basiji brigades—come in and crush the rebellion. This isn't a technical analysis by any means; I'm just pointing out that with fractured and overlapping military forces, Iran doesn't resemble, say, Philippines circa 1986 all that much.