When you rip, tear or mechanically deform a polymer (for example a piece of plastic) you are putting energy into the material. The energy from this deformation causes the polymer chains in the vicinity of the deformation to attempt to align. To some degree this partial alignment makes continued deformation easier. To continue tearing the polymer apart you just need to overcome the van der Waals forces holding the chains together.
In addition to the polymer chains being pulled apart, some chemical bonds will be broken. I don't know how many bonds are broken per chain pulled apart, but I suspect it is low. Low or not, some free radicals (or ions) will be generated when the chemical bonds break. If the experiment is performed under ambient conditions, the radicals or ions will quickly react with oxygen or ions in the air. If the experiment is performed in an inert atmosphere, then the radicals will persist (dangling bonds) and can be observed in an esr spectrometer. It has been demonstrated that cross-linked polymers, which are less prone to flow deformation and therefore more likely to undergo bond breaking, do produce larger esr signals.
So most of what is going on is molecular deformation and pulling apart of the polymer chains - but some bond breaking does occur.
It would be an interesting experiment to cut two pieces of plastic and then cut one of them into finer shreds. Place the two samples into separate tubes and then record their esr signals. How much more intense would the signal be in the finely cut sample compared to the bulk sample? Polycarbonate-A, straight out of the reagent bottle, produces an esr signal.