Do virtual states in Raman spectroscopy have distinct and observable dipoles that can interact with and/or induce dipoles in other molecules (over extremely short femtosecond timescales)? This notion comes form the distinct absorbance profiles for virtual states observed in recent attosecond spectroscopy studies: e.g. see Figure 2 in (http://www.nature.com/srep/2013/130122/srep01105/full/srep01105.html) (an open-access paper by Chini, M. et. al. in a Nature magazine... something).
This is odd to me because I always thought that "virtual states" should appear to be identical to the ground eigenstate in terms of their real observables/properties. Yet, having a distinct absorbance profile means that these virtual states must have dipoles that interact with attosecond light pulses in a manner distinct from either the ground or first excited state of an absorbing material, right?