The Non-Quotient Majority

A hyperfield generalizes a field by allowing addition to be multivalued: the sum of two elements is a set, not a single element. Every field quotient produces a hyperfield (the cosets form one), so quotient hyperfields are the natural examples. Baker and Jin conjectured that these natural examples are vanishingly rare — that almost all finite hyperfields are non-quotient.

Ando et al. prove it. Using a probabilistic argument extending Hobby’s result (which covered even-order hyperfields), they show that for any order, the fraction of finite hyperfields that arise as field quotients approaches zero. The typical hyperfield is not a quotient of anything.

The proof reveals a structural reason. In almost every finite hyperfield, the sum of any four or more nonzero elements contains zero. This is a combinatorial property that quotient hyperfields cannot satisfy in general — field quotient structure imposes too much arithmetic regularity. The generic hyperfield is too combinatorially wild to have come from a field.

The authors also give precise asymptotics for the number of finite hyperfields on a given finite abelian group. The count grows fast — much faster than the count of quotient hyperfields, which is what makes the non-quotient majority inevitable.

Hyperfields were introduced to generalize valuation theory and tropical geometry. The result says the generalization goes far beyond its origins: the algebraic structures it creates mostly have no field-theoretic ancestry. The offspring outnumber the parents overwhelmingly.