defect
Why this theorem is linked from Text and Code Embeddings by Contrastive Pre-Training echoes
Pith linked this Lean declaration because the review connected a specific passage in the paper to this theorem. The relation tag says how strong that connection is; it is not a generic placeholder.
we show that contrastive pre-training on unsupervised data at scale leads to high quality vector representations
ECHOES: this paper passage has the same mathematical shape or conceptual pattern as the Recognition theorem, but is not a direct formal dependency.
plain-language theorem explainer
The defect functional is defined as the J-cost on the reals. It anchors the Law of Existence formalization by equating existence with zero defect. Researchers working on Recognition Science foundations cite it to link the cost functional to the existence predicate. The definition is a direct one-line alias to the J expression from the Cost module.
Claim. The defect functional is defined by $defect(x) := J(x)$ for $x : ℝ$, where $J(x) = ½(x + x^{-1}) - 1$.
background
In the Law of Existence module the defect functional formalizes the statement that x exists if and only if defect(x) = 0. The J-cost is the canonical cost functional J(x) = ½(x + x⁻¹) - 1, which satisfies the Recognition Composition Law, is nonnegative on the positive reals, and attains its minimum of zero at unity. This definition sits inside the Recognition Science framework that derives all physics from a single functional equation, with J appearing as the T5 uniqueness map (J(x) = cosh(log x) - 1). Upstream the toy Defect from CPM.LNALBridge returns zero when structural checks pass, while the main J is imported from the Cost module.
proof idea
This is a one-line definition that directly aliases the J functional. No lemmas are applied; the body simply unfolds to the expression (x + x⁻¹)/2 - 1.
why it matters
This definition anchors the Law of Existence theorems such as defect_zero_iff_one and law_of_existence, which state that x exists precisely when defect(x) = 0 and therefore that 1 is the unique existent. It feeds directly into the cost algebra certificate that verifies J satisfies RCL, normalization at 1, reciprocal symmetry, and non-negativity. In the framework it connects to the T5 J-uniqueness step of the forcing chain and to the eight-tick octave. It touches the open question of extending the defect to non-positive reals in empirical programs.
Switch to Lean above to see the machine-checked source, dependencies, and usage graph.
papers checked against this theorem (showing 1 of 1)
-
Contrastive pre-training yields strong text and code embeddings
"we show that contrastive pre-training on unsupervised data at scale leads to high quality vector representations"