That's what Wikipedia says about Elementary Particle:
In particle physics, an elementary particle or fundamental particle is a particle whose substructure is unknown, thus it is not known to be composed of other particles.
Assumed the above sentence: Can we ever know the structure of an elementary particle "Z" if it is an agglomeration of other n sub-particles "T" that, toghether, cancel their actractive forces creating our singular Z particle?
An answer could be to collide 2 "Z" particles until they break and show us the "T" particles, but again, the same question can be formulated for the "T" particle. We can demonstrate that a particle is not elementary, but we can't say that it is NOT composed of other particles.
This leads to another concept: suppose we make and demonstrate an M-Theory called "X" which affirm that "n" different particles can stick making all other known particles with the same known properties , we're not able to demonstrate that there's no other M-Theory "Y" that generalize the "X" one.
Furthermore can even make sense to formulate tens of different "Y" M-Theories which logically generalize the "X" one and that, simulated on a utopian machine, generate the "X" environment. The core question is, will we ever know when we are done?