Anthony Shadid writes in the New York Times:
The semblance of a civil war has erupted in Homs, Syria’s third-largest city, where armed protesters now call themselves revolutionaries, gun battles erupt as often as every few hours, security forces and opponents carry out assassinations, and rifles costing as much as $2,000 apiece flood the city from abroad, residents say.
Shadid’s headline overstates the degree to which Syria’s unarmed revolution is “spiraling” into civil war. The armed “wing” of the uprising is largely comprised of military defectors in Homs, who took their guns with them as they sought haven in civilian homes. In my estimation, Syria will not “spiral” into violence, as most people on the ground have little opportunity to take up arms against Assad’s regime; neither are they interested in taking up arms against Assad’s regime. I base this impression on discussions with some Syrian activists on the ground, as well as reports that nonviolent mass mobilization continues in many cities with no hint of civilian-initiated violence on the horizon—all this despite continual massacres by the regime against unarmed civilians.
However, it is important to know how the armed wing may affect the strategic dynamics between the popular, civilian-led nonviolent movement and the regime. Kurt Schock and I have done some research on how violent radical flanks have influenced the outcomes of primarily nonviolent campaigns. Using a data set of 108 nonviolent campaigns from 1900 to 2006, we looked at how many of these campaigns were accompanied by contemporaneous violent movements, and whether the presence of these violent movements affected the success and failure rates of nonviolent campaigns.
The following figure is a bivariate cross-tabulation of the relationship between nonviolent campaign outcomes and the presence or absence of a violent radical flank.
What is evident here is that having an armed wing has a slight negative effect on the probability of success. However, this effect is not statistically significant—a finding that is confirmed in other statistical tests. There is definitely no evidence to support the notion that armed groups will help a nonviolent campaign. The bottom line: there is a slight tendency for armed wings to reduce the success rates of nonviolent campaigns, but this reduction is not common enough for there to be a real pattern from which to draw inferences.
Of course the most troubling possibility is that the armed wing will reduce the movement’s chances of success. Why might an armed wing reduce the probability of success for an unarmed movement? There are a few reasons. First, and most important, is that the emergence of an armed wing can reduce popular participation in a nonviolent campaign. See the following figure:
This figure shows that nonviolent movements without armed or “radical” flanks are much more likely to boast large numbers of members than campaigns with radical flanks. And as Maria Stephan and I show in our book Why Civil Resistance Works, participation is absolutely critical in the success of nonviolent campaigns.
Second, developing an armed wing can give the regime the pretext it needs to escalate widespread repression against all opponents—nonviolent and violent. Part of Assad’s propaganda has focused on how the uprising is comprised of armed gangs seeking to disrupt public order and destroy Syrian society. Such propaganda has heretofore seemed totally ridiculous, even among many security forces who have chosen to defect to the movement’s side. For a regime where loyalty within the security forces is crumbling, adopting armed struggle or an armed defense wing can actually reverse these trends in shifting loyalties. Security forces generally don’t surrender themselves to armed “traitors,” and Assad’s rhetoric may seem less crazy to the security forces when they suddenly find themselves under attack by their former comrades.
B. H. Liddel-Hart, who interviewed Nazi generals responsible for the German occupations throughout Eastern and Western Europe, observed the following:
[The Nazis] were experts in violence, and had been trained to deal with opponents who used that method. But other forms of resistance baffled them—and all the more in proportion as the methods were subtle and concealed. It was a relief to them when resistance became violent and when nonviolent forms were mixed with guerrilla action, thus making it easier to combine drastic repressive action against both at the same time. 
The armed wing in Syria will probably not be the deciding factor in whether the revolution succeeds or fails. If Assad falls, it will be because hundreds of thousands of unarmed protestors have withdrawn their cooperation from the regime; because security forces refuse to obey orders to crack down on unarmed protesters; and because business elites within the country will pressure Assad to abandon his post. It would help if a prominent member of the Alawite community would come forward and denounce the violence that Assad has perpetrated against peaceful demonstrators. It might also help if the international community figured out a way to give Assad a golden parachute out of the situation.
In sum, having an armed wing is risky, but not necessarily decisive. The armed wing won’t help the nonviolent movement in Syria. However, as long as the movement remains mainly nonviolent in nature, the campaign may succeed regardless.
 B. H. Liddel-Hart, “Lessons from Resistance Movements: Guerrilla and Nonviolent,” in Adam Roberts, ed., Civilian Resistance as a National Defence (Harrisburg: Stackpole Books, 1968), p. 205.