Does pwscf print out some phonon representations (eigenfunctions, eigendisplacements, whatever) at the gamma point? If so, then you can identify them by hand using the SMODES program in the ISOTROPY suite:
http://stokes.byu.edu/iso/isotropy.php
There is a similar utility at the Bilbao Crystallographic server:
http://www.cryst.ehu.es/
scroll down somewhere near the bottom, you look for the SAM utility among the Raman tools.
EIther of these will also give you information about the optical activity of the modes.
If you're using PwSCF, I assume you're using the PHonon package that comes as part of the Quantum Espresso distribution?
I haven't used this myself (I use VASP + Phonopy), but what you need to do is to analyse the symmetry of each of the vibrations in your spectrum, which will give you their irreducible representations and let you see which ones are A1g, B1g, A1u, E, etc.
Apparently, there is a symmetry analyser in PHonon: http://www.quantum-espresso.org/faq/phonons/#7.1
There is a list of the input parameters the program accepts here: http://www.quantum-espresso.org/wp-content/uploads/Doc/INPUT_PH.html; the one you want looks like it's called search_sym, and is enabled by default.
If by "initialise" you mean you want to generate "frozen phonon" structures, that is, models with the atoms displaced by the particular vibration, there doesn't seem to be an option to do this among the PHonon keywords, but if you can lift the eigenvectors from the output file, you could do "by hand", or, if you have a lot of modes to look at, script it in Python or some such language.