Ah, I gotcha. And yeah, I would say the mathematical form of the argument in the original post is not adequate for the application. It doesn't reveal anything about conditions for stability. The correct way is to examine the changes in energy as the cloud shrinks, and the key principles come from thermodynamics ([tex]dE = TdS - PdV[/tex] and the Ideal Gas Law), gravitational physics ([tex]U=-\frac{GM^2}{r}[/tex]), and some quantum mechanics (binding energy of the electron and H2 bonds).
So, once it starts collapsing, it heats up, but simple radiation goes like P=k∗T4P=k∗T4, so it loses a lot of energy (and cools) and then shrinks to heat up?
Basically, but think of it being more continuous than jerky, and the radiation isn't the most important cooling mechanism early on.
Initially the collapse may be virtually in free-fall, and proceed from the inside out (the inner regions become denser and collapse faster, which removes pressure support from the outer regions). The collapse is also isothermal during this phase, since hydrogen dissociation is important to preventing the increase in thermal energy (costs 4.5eV per H2 bond broken).
Then there is another phase of the collapse which is mediated by the gas being ionized (13.6eV per electron ripped from an H atom). When the gas is mostly ionized the temperature will rise more dramatically, and radiation becomes the dominant cooling mechanism. But by this time the dissociation and ionization have already allowed the collapse to result in a number of nearly protostar-sized, virialized regions. These continue to shrink as they radiate, and accretionary processes become important to shaping the protoplanetary system (then we have to account for transference of angular momentum and all sorts of fun things).
I think it's very cool to see what this whole process looks like in simulations.
[youtube]YbdwTwB8jtc[/youtube]