Any consistent first order theory $T$ has a model.
For a second order theory this is false.
Even for geometric theories this is false.
But what is the state of art in this subject, can we somehow characterize logics for which there is a model?
This question is somehow located in a framework of questions I am asking in these days.