We characterise precisely what the G2 framework can and cannot derive, and show that the boundary is not arbitrary but follows from the nature of a compact Lie algebra. The framework derives, accurately and parameter-free, every dimensionless quantity tested: mass ratios (the top/bottom ratio to 0.01%), the Weinberg angle (3/8), the generation count (three), the Koide relation (2/3), the unified coupling (1/42), the reactor angle (1/42), and the colour content of a generation (16 = 2(3)+2(3̄)+4(1)). It does not derive absolute scales: the breaking scales, the overall mass scale, and the value of the Cabibbo parameter. We show this split is structural. G2 is compact and therefore scale-free — it encodes only angles, ratios, representation dimensions, and counts. Every dimensionful prediction factorises as (pure G2 structure) × (one scale), exactly as the mass formula m = v(1/42)ⁿ(ratio) displays. We argue this one dimensionful input is irreducible in principle: no theory built on pure structure can derive an absolute scale, since scale is meaningful only relative to a reference. The framework is thus a complete theory of dimensionless physics requiring a single scale input — more economical than the Standard Model's nineteen parameters, while honest that it does not solve the hierarchy problem.
Vali Ilyas (Wed,) studied this question.