The Alexander polynomial is a classical knot invariant, traditionally constructed via the Alexander-Conway skein relation — a local rule that relates the invariants of three knots differing only in a single crossing. The construction is algorithmic, well-understood, and canonical.
Harper derives the same polynomial from quantum sl3 through a mechanism that violates the skein relation the polynomial is supposed to obey.
The approach generalizes the Alexander polynomial to the non-abelian setting using the representation theory of quantum sl3. At specific parameter values, the new invariant recovers the classical Alexander polynomial exactly. But the underlying R-matrix — the algebraic engine that drives the computation — does not satisfy the Alexander-Conway skein relation. The computation arrives at the correct answer through what should be the wrong algebra.
This is not a fluke of parameter tuning. The sl3 framework provides a richer algebraic structure than the classical sl2 setting. When you project down to the abelian invariant, the additional structure collapses, but the computational path that produced it remains non-standard. The polynomial does not know which route generated it.
The result reveals something about invariants themselves. The Alexander polynomial is more robust than its standard derivation suggests — it can be reached through algebraic mechanisms that violate the rules typically assumed necessary for its construction. The invariant is deeper than any particular skein relation that generates it.
This is a general phenomenon in mathematics: objects defined by specific constructions often admit alternative constructions that violate the original's axioms. The object does not belong to its proof. The destination is more fundamental than any particular path.