Reconsidering the origin of the X-ray emission lines in GRB 011211

We reanalyze the XMM--Newton data of GRB 011211 showing that the spectral features, interpreted by Reeves et al. (2002, 2003) as due thermal emission from a collisionally ionized plasma, can be also reproduced by a reflection model (with ionization parameter $\xi\sim 10^2$). We discuss the implications of this interpretation, estimating the total mass required in the simplified case of a funnel geometry. We conclude that a moderate clumping of the reprocessing material (corresponding to a filling factor of the order of $f\sim 10^{-3}$) is required. Finally we show that, if this interpretation is correct, a bright quasi--thermal component is expected in the optical--UV band (containing about 90% of the luminosity of the illuminating continuum), whose presence can be used to test the reflection model.