Hi Perado,
I believe what the authors are referring to is the fact that the effective inductive impedance is of opposite sign to that of the capacitive impedance. Hence, if one is considering what the total parasitic capacitance is, by adding some small amount of inductance, the net capacitive impedance appears reduced. This technique is employed to compensate for undesired lead inductance. Specifically, by adding in capacitive impedance 1/jwC = -j/wC to inductive impedance jwL, the net reactive impedance can be set to 0 at some frequency (-j/wzC + jwzL = 0).
Of course, as the frequency changes from wz, the impedance becomes non-zero for a fixed C and L. Often people use varactors for the C to allow for tuning an impedance.
Does this make sense to you?