A piece of dogma current in the physical and engineering sciences is the idea
that a theory is of no use if, given initial data, a unique outcome can NOT be
predicted. The concept of an envelope is precisely when uniqueness fails, yet
the useful concept of a set of Huygen wavelets building a wave front is
precisely the idea of an envelope. The concept of a discontinuity is also
evidence of non-uniqueness in the real non-ideal world.
From another point of view, the Froenius integrability theorem states the conditions required such that a Pfaffian expression (ordinary differential equation) has a unique direction field solution. The requirement in simple cases is precisely, the topological torsion A^dA, must vanish. Hence, somehow the concept that A^dA <> 0 is to be associated with non-uniqueness and existence of an envelope. This idea is exploited in the enclosed pdf file below.