Quote by waterfall
he presence of a flat background metric n(ab) which one has in the linear approximation but not in the full theory, so the statement that, in general relativity, gravity is treated as a massless spin2 field is not one that can be given precise meaning outside the context of the linear approximation."

That's exactly what it means. In fact, that's the case in quantum chromodynamics as well, when we speak about quarks and gluons.
It's always the case in physics that some particular object only has ontological meaning in some well defined framework, which typically is only an approximation to the real thing.
For instance, the notion of a unique particle strictly speaking really only makes sense in flat space in the infinite past and future, where there is some sort of perfectly massive detector registering them. That doesn't prevent us from modeling reality by pretending like that condition is relaxed (and to a very high degree of accuracy, it is).