-| RBindingGroup : ∀ Γ Δ Σ₁ Σ₂ τ₁ τ₂ , Rule ([Γ > Δ > Σ₁ |- τ₁ ],,[Γ > Δ > Σ₂ |- τ₂ ]) [Γ>Δ> Σ₁,,Σ₂ |- τ₁,,τ₂ ]
-| RApp : ∀ Γ Δ Σ₁ Σ₂ tx te l, Rule ([Γ>Δ> Σ₁ |- [tx--->te @@l]],,[Γ>Δ> Σ₂ |- [tx@@l]]) [Γ>Δ> Σ₁,,Σ₂ |- [te @@l]]
-| RLet : ∀ Γ Δ Σ₁ Σ₂ σ₁ σ₂ l, Rule ([Γ>Δ> Σ₂ |- [σ₂@@l]],,[Γ>Δ> Σ₁,,[σ₂@@l] |- [σ₁@@l] ]) [Γ>Δ> Σ₁,,Σ₂ |- [σ₁ @@l]]
-| REmptyGroup : ∀ Γ Δ , Rule [] [Γ > Δ > [] |- [] ]
+
+| RJoin : ∀ Γ Δ Σ₁ Σ₂ τ₁ τ₂ , Rule ([Γ > Δ > Σ₁ |- τ₁ ],,[Γ > Δ > Σ₂ |- τ₂ ]) [Γ>Δ> Σ₁,,Σ₂ |- τ₁,,τ₂ ]
+
+(* order is important here; we want to be able to skolemize without introducing new RExch'es *)
+| RApp : ∀ Γ Δ Σ₁ Σ₂ tx te l, Rule ([Γ>Δ> Σ₁ |- [tx--->te @@l]],,[Γ>Δ> Σ₂ |- [tx@@l]]) [Γ>Δ> Σ₁,,Σ₂ |- [te @@l]]
+
+| RLet : ∀ Γ Δ Σ₁ Σ₂ σ₁ σ₂ l, Rule ([Γ>Δ> Σ₁ |- [σ₁@@l]],,[Γ>Δ> [σ₁@@l],,Σ₂ |- [σ₂@@l] ]) [Γ>Δ> Σ₁,,Σ₂ |- [σ₂ @@l]]
+| RWhere : ∀ Γ Δ Σ₁ Σ₂ Σ₃ σ₁ σ₂ l, Rule ([Γ>Δ> Σ₁,,([σ₁@@l],,Σ₃) |- [σ₂@@l] ],,[Γ>Δ> Σ₂ |- [σ₁@@l]]) [Γ>Δ> Σ₁,,(Σ₂,,Σ₃) |- [σ₂ @@l]]
+
+| RVoid : ∀ Γ Δ , Rule [] [Γ > Δ > [] |- [] ]
+