Some forms of analytic reconstructivism take natural language (and common s
ense at large) to be ontologically opaque: ordinary sentences must be suita
bly rewritten or paraphrased before questions of ontological commitment may
be raised. Other forms of reconstructivism take the commitment of ordinary
language at face value, but regard it as metaphysically misleading: common
-sense objects exist, but they are not what we normally think they are. Thi
s paper is an attempt to clarify and critically assess some common limits o
f these two reconstructivist strategies.