Divergences, Counterterms, and Scheme Dependence

CV and CA are ultraviolet divergent, and their finite parts admit proposal-specific counterterms and scheme choices. A comparison is meaningful only after translating regulators to the same induced boundary data and declaring all finite local terms. Matched differences and some time derivatives can be scheme insensitive within a restricted class, but no universal absolute complexity follows from subtracting divergences.

Required background. Complexity Equals Volume Proposals and Complexity Equals Action Proposals supply the bare functionals.

Helpful background. Finite Counterterms, Schemes, and Multi-Trace Data supplies holographic-renormalization logic. Regulator Dependence and Continuum Complexity gives the boundary interpretation.

Leading cutoff structure

In Poincaré AdS$_{d+1}$,

$$\mathrm ds^2=\frac{L^2}{z^2} \left(\mathrm dt^2-\mathrm dz^2-\mathrm d\mathbf x_{d-1}^2\right).$$

The $t=0$ maximal slice regulated at $z=\epsilon$ has

$$V(\epsilon) =L^dV_{d-1}\int_\epsilon\frac{\mathrm dz}{z^d} =\frac{L^dV_{d-1}}{d-1}\epsilon^{-(d-1)}+\cdots.$$

Thus

$$\mathcal C_V^{\mathrm{div}} =\frac{L^dV_{d-1}}{(d-1)G_N\ell_V} \epsilon^{-(d-1)}+\cdots.$$

Curvature of the boundary slice generates subleading local divergences; even boundary dimensions can produce logarithms. Covariant counterterms on the cutoff surface cancel these terms, but allowed finite local invariants shift the renormalized answer.

CA has a related power expansion but different subleading coefficients. Null-boundary and joint terms are essential: including the reparametrization counterterm removes an otherwise stronger divergence and makes the leading term proportional to boundary spatial volume Reynolds and Ross 2017. Timelike regulator counterterms and the null scale $\ell_{\mathrm{ct}}$ remain part of the scheme.

Regulator translation

Suppose one computation uses $z=\epsilon$ and another a radial coordinate $r=r_c$. Fefferman–Graham matching gives a state-dependent expansion

$$r_c=\frac{L^2}{\epsilon}+a_1\epsilon+a_2\epsilon^3+\cdots.$$

Keeping only the leading relation can leave a spurious finite difference. One must equate the induced boundary metric and sources to the order that contributes to the finite term, then transform every joint and counterterm consistently.

First application

Renormalize the Poincaré-AdS CV result by subtracting the leading local volume counterterm. Repeat in $r=L^2/z$ coordinates and verify equality after exact cutoff translation. Then add an allowed finite term

$$I_{\mathrm{fin}}=c_{\mathrm{fin}} \int_{\Sigma_\epsilon}\!\mathrm d^{d-1}x\sqrt\sigma\,L^{d-1}\mathcal R[\sigma]^{(d-1)/2}$$

when dimension and locality permit it. The absolute renormalized complexity shifts with $c_{\mathrm{fin}}$. A formation difference between geometries with identical boundary data may cancel the same term; a comparison with different curvature need not.

Adversarial control

Add every symmetry-allowed finite counterterm and require a claimed universal number or bound to remain unchanged. Also vary the null counterterm scale in CA and the reference length $\ell_V$ in CV. If the claim moves, report it as scheme or normalization dependent. Do not choose a finite coefficient after seeing the desired answer.

Regime, evidence ceiling, and handoff

The asymptotic expansion assumes an asymptotically AdS solution and a regulator within the semiclassical region. Higher-curvature, $\alpha'$, loop, and KK terms generate their own boundary terms and can change both divergences and finite pieces. Large $N$ makes the leading result large; it does not make the scheme unique.

The evidence ceiling consists of invariant differences or derivatives demonstrated to be independent of an explicitly enumerated scheme class. Continue to Complexity of Formation and Time Growth for matched subtractions and to Operational Meaning, Nonuniqueness, and Evidence Status for proposal comparison.

The chapter overview contains the structure diagram and validity and failure diagram. They are embedded there once so that their shared chapter-level context is not repeated on every article.

For the chapter-wide comparison of assumptions, counterevidence, falsifiers, and claim ceilings, see the claim-domain table.

References

• Akhavan, A., and Omidi, F. (2019), “On the Role of Counterterms in Holographic Complexity,” Journal of High Energy Physics 2019(11), 054. arXiv:1906.09561.
• Reynolds, A., and Ross, S. F. (2017), “Divergences in Holographic Complexity,” Classical and Quantum Gravity 34, 105004. DOI; arXiv:1612.05439.