REVIEW 2 major objections 5 minor 14 references
Vector current renormalisation in momentum subtraction schemes using the HISQ action
T0 review · 2 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash
Pith's one-line read This paper shows that the RI-SMOM momentum subtraction scheme yields $Z_V = 1$ for the conserved vector current on HISQ lattices, with no visible condensate contamination, whereas the standard RI$'$-MOM scheme carries a ~1%…
desk verdict A solid, practical HISQ demonstration that RI-SMOM gives Z_V=1 for the conserved current while RI'-MOM has ~1% condensate contamination; the explicit staggered-WTI caveat is real but the numerics carry the argument. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The load-bearing object is the lattice Ward–Takahashi identity (WTI), the exact relation between the finite-difference divergence of the conserved current matrix element and a difference of propagators. Inserting the amputated vertex into the SMOM definition and using the WTI gives Eq. (3.2), which reduces to $Z_V = 1$ for the conserved current; the same identity does not enter the RI$'$-MOM construction, which is why its $Z_V$ picks up condensate contamination. The SMOM kinematic setup ($p_1^2=p_2^2=q^2=\mu^2$) and the discretised momentum $\hat q$ are also essential to the derivation.
What would settle it
Take the RI-SMOM $Z_V$ for the conserved current on the same ensembles but with a different projector in the trace, for example $\gamma_\mu$ instead of $(\hat q_\mu/\hat q^2)\hat q$: the Ward–Takahashi identity forces both to give $Z_V = 1$, so a statistically significant difference between the two would reveal that the lattice form of the identity used in Eq. (3.2) is broken by staggered-quark effects. An explicit computation of the taste-non-singlet part of the amputated vertex would show whether the ignored subtleties contribute.
Extended reading notes
Core claim
The paper's central discovery is that the Ward–Takahashi identity protects the conserved vector current in RI-SMOM: after amputation and projection, the identity forces $Z_q/Z_V = Z_q$, hence $Z_V = 1$ for the conserved current, independent of the momentum scale $\mu$, quark mass, and lattice spacing. Numerically, this holds at the 0.05% level of statistical errors on the HISQ ensembles studied. In contrast, the RI$'$-MOM scheme does not use the WTI, so the conserved-current $Z_V$ there is not equal to one and receives ~1% condensate contributions at $\mu = 2$ GeV, making the continuum limit depend incorrectly on $\mu$. The paper also shows that the RI-SMOM $Z_V$ for the local current agrees with the previous form-factor determination, so the difference between them is purely discretisation effects, with no condensate contamination.
Load-bearing premise
The argument assumes the lattice Ward–Takahashi identity survives amputation and projection for staggered quarks; if staggered-quark discretisation effects break that identity, the prediction $Z_V=1$ and its protection from condensates would no longer follow.
Editorial extensions
If this is right
- RI-SMOM can determine $Z_V$ for the local vector current to high precision using only about 20 gauge configurations, instead of the O(1000) needed for the form-factor method.
- Because the WTI protects the conserved current, $Z_V=1$ in RI-SMOM holds independent of mass, momentum, and lattice spacing, within the 0.05% statistical errors of the calculation.
- RI$'$-MOM results for the conserved vector current carry a ~1% condensate systematic at $\mu = 2$ GeV and should not be used without explicit condensate corrections.
- The RI-SMOM local $Z_V$ differs from the form-factor result only by discretisation effects, confirming that both methods are free of condensate contamination and can be used interchangeably.
- Preliminary charm applications—the $J/\psi$ decay constant and the charm contribution to the muon anomalous moment—agree with the PDG and with earlier determinations, showing the practical value of the method.
Reading between the lines
- If $Z_V=1$ in RI-SMOM survives with higher statistics and finer lattices, then the ratio of local to conserved current renormalisation in that scheme is itself determined by the WTI, giving a nearly parameter-free way to calibrate local HISQ currents.
- The same WTI-protection logic should extend to other currents that sit in conserved or partially conserved Ward identities; testing whether an axial RI-SMOM $Z_A$ obeys the axial WTI and stays free of condensates would directly generalise this result.
- A direct check of the staggered-quark subtleties set aside in Eq. (3.2) would be to compute the projected WTI with different taste structures: if taste-non-singlet pieces leak into the projector, deviations from $Z_V=1$ should grow with the taste-breaking scale, so pinning down that scale would bound the effect more tightly.
- Cheap momentum-source $Z_V$ calculations could become the default for HISQ charm and bottom physics, freeing the more expensive form-factor method for the few cases where only it applies.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper presents nonperturbative calculations of the vector current renormalisation constant on HISQ ensembles using the RI-SMOM and RI'-MOM momentum subtraction schemes. The central claims are that the conserved vector current has Z_V = 1 in RI-SMOM, verified numerically to about 0.05% statistical precision across several lattice spacings and momentum scales, and that the RI'-MOM scheme shows roughly 1% power-suppressed nonperturbative (condensate) contamination for the same conserved current. The paper also compares the RI-SMOM renormalisation of the local vector current with form-factor determinations, finding that the difference is consistent with being purely a discretisation effect, and shows preliminary applications to J/psi decay constants and hadronic vacuum polarisation moments in charm physics.
Significance. If the claims hold, the paper establishes a computationally cheap and robust alternative to the expensive form-factor method for vector current renormalisation on HISQ lattices, and it provides concrete evidence that the standard RI'-MOM scheme is unsuitable for vector currents unless condensate corrections are included. The numerical checks at three lattice spacings and multiple values of the momentum scale are a clear strength, as is the use of the exact lattice Ward-Takahashi identity to motivate the RI-SMOM result. The paper is a proceedings contribution, so the depth of the analysis is limited, but the central comparison between the two schemes is physically important for ongoing precision charm physics.
major comments (2)
- [Section 3, Eq. (3.2)] The derivation of Z_V = 1 for the conserved current in RI-SMOM rests on Eq. (3.2), which is obtained from the Ward-Takahashi identity 'ignoring subtleties related to our use of staggered quarks [4,5]'. This is load-bearing for the paper's central claim that RI-SMOM is protected from condensate contamination. The manuscript should either spell out how the staggered/HISQ case is handled, for example by stating that Refs. [4,5] establish the analogue of Eq. (3.2) in the presence of taste degrees of freedom and explaining how taste-breaking effects are controlled, or explicitly restrict the protection claim to the tested kinematics. As written, the conclusion in Sec. 6 overstates the theoretical basis: the 0.05% agreement is an empirical verification at the tested mu, a, and statistics, not a demonstrated theorem for staggered quarks.
- [Section 4.1, Eq. (4.1)] The conclusion that RI'-MOM has roughly 1% nonperturbative (condensate) contamination depends on the fit ansatz of Eq. (4.1), which combines discretisation terms, condensate terms, and an alpha_s^4 matching uncertainty. The reported chi^2/d.o.f. = 0.6 shows the fit is good, but it does not by itself establish the physical interpretation of the mu-dependent continuum limit. I request that the authors provide the fitted values of the condensate coefficients, show the fit without condensate terms (which they state gives a poor chi^2), and test the robustness of the mu-dependence of the continuum extrapolation against alternative discretisation ansaetze. This would strengthen the comparative claim that RI'-MOM is not suitable without condensate corrections.
minor comments (5)
- [Figure 1] The SMOM conserved-current Z_V data are not shown in the figure; please include them or provide a table to substantiate the stated 0.05% agreement of Z_V with 1.
- [Section 5] There are several typographical errors: 'imporves' should be 'improves', 'onnf' should be 'on n_f', and 'contributiuon' should be 'contribution'.
- [Eq. (4.1)] The summation ranges for the indices i and j in Eq. (4.1) are not defined; please specify them explicitly.
- [Section 3] The vertex function G_V in Eq. (3.1) is written for the local current, but the subsequent derivation refers to the lattice conserved current; please clarify which operator is used in each step.
- [Figure 2 and Section 4.2] For the fit of Z_F(0)_V - Z_loc-SMOM_V, please report the chi^2/d.o.f. and the fitted continuum value, so that the claim that the difference is purely a discretisation effect can be quantitatively assessed.
Circularity Check
No significant circularity: the central RI-SMOM result follows from an exact lattice Ward-Takahashi identity and is checked numerically at the 0.05% level, while the RI'-MOM comparison is a fit-based observation rather than a forced prediction.
full rationale
The paper's central claim that Z_V=1 for the conserved current in RI-SMOM is derived, not fitted. Starting from the exact lattice Ward-Takahashi identity (Eq. 2.2), the paper amputates and projects to obtain Eq. (3.2), whose right-hand side equals Z_q in the continuum; the paper explicitly states the conditionality: 'If this remains unbroken by discretisation effects on the lattice then Z_V=1 for the conserved current in the SMOM scheme independent of mass, momentum and lattice spacing. Through explicit numerical calculation we see this to be true up to the 0.05% level of our statistical errors.' Thus the claim is supported by an analytical identity plus direct numerical verification, not by reusing fitted outputs. The RI'-MOM conclusion of ~1% nonperturbative effects is inferred from the data themselves: the fit in Eq. (4.1) requires condensate terms to obtain an acceptable chi-squared, and the disagreement between continuum extrapolations at different mu is shown in Figure 1; this is a data-analysis conclusion rather than a prediction forced by construction. The local-current comparison in Section 4.2 tests RI-SMOM against the independent form-factor method [2], so the agreement is an external consistency check rather than a self-referential input. The citations to [4,5] for staggered-quark subtleties are self-citations, but they are not load-bearing in a circular way: the paper openly labels the WTI step as '(ignoring subtleties related to our use of staggered quarks [4,5])' and then independently verifies the resulting equality numerically at the 0.05% level. That stated caveat is a correctness risk about taste-breaking effects, not evidence that the derivation reduces to its own inputs. No step in the paper equates a fitted parameter with a predicted quantity or imports a uniqueness conclusion from the authors' prior work.
Assumptions & free parameters
free parameters (4)
- Discretization coefficients c^(i), c^(i)_alpha in Eq. (4.1)
- Condensate coefficients c^(j)_cond in Eq. (4.1)
- alpha_s^4 matching uncertainty coefficient c_alpha
- Polynomial coefficients in the local-current difference fit (Fig. 2)
assumptions (4)
- domain assumption The lattice Ward-Takahashi identity survives amputation and projection for staggered quarks, 'ignoring subtleties related to our use of staggered quarks' (Section 3).
- ad hoc to paper The fit ansatz Eq. (4.1) correctly separates discretisation, condensate, and perturbative-matching contributions to RI'-MOM Z_V.
- domain assumption The form-factor method Z_V^{F(0)} from ref [2] is free of condensate contamination.
- standard math The MS matching for RI'-MOM through O(alpha_s^3) is correct and the residual alpha_s^4 term is captured by the fit.
Cite this review
Pith. "Pith review of Vector current renormalisation in momentum subtraction schemes using the HISQ action." pith.science (2026). https://pith.science/paper/52BEH4QJ
@misc{pith2026190810116,
author = {Pith},
title = {Pith review of: Vector current renormalisation in momentum subtraction schemes using the HISQ action},
year = {2026},
howpublished = {\url{https://pith.science/paper/52BEH4QJ}},
note = {Machine review of arXiv:1908.10116}
}
abstract
As the only lattice vector current that does not require renormalisation is the point-split conserved current it is convenient to have a robust, precise and computationally cheap methodology for the calculation of vector current renormalisation factors, $Z_V$. Momentum subtraction schemes, such as RI-SMOM, implemented nonperturbatively on the lattice provide such a method if it can be shown that the systematic errors, e.g. from condensates, are well controlled. We present $Z_V$ calculations for the conserved current in both the RI-SMOM and RI$'$-MOM momentum subtraction schemes as well as local current renormalisation in the RI-SMOM scheme. By performing these calculations at various values of the momentum scale $\mu$ and different lattice spacings we can investigate the presence of power suppressed nonperturbative contributions and compare the results to expectations arising from the Ward-Takahashi identity. Our results show that the RI-SMOM scheme provides a well controlled determination of $Z_V$ but the standard RI$'$-MOM scheme does not. We then present some preliminary uses of these $Z_V$ calculations in charm physics.
Figures
Figures from the paper (1 more)
Reference graph
Works this paper leans on
-
[1]
E. Follana et al. [HPQCD and UKQCD Collaborations], Phys. Rev. D 75 (2007) 054502 doi:10.1103/PhysRevD.75.054502 [hep-lat/0610092]
arXiv 2007
-
[2]
B. Chakraborty et al. [HPQCD Collaboration], Phys. Rev. D 96, no. 7, 074502 (2017) doi:10.1103/PhysRevD.96.074502 [arXiv:1703.05552 [hep-lat]]
arXiv 2017
- [3]
-
[4]
A. T. Lytle and S. R. Sharpe, Phys. Rev. D 88, no. 5, 054506 (2013) doi:10.1103/PhysRevD.88.054506 [arXiv:1306.3881 [hep-lat]]
arXiv 2013
-
[5]
A. T. Lytle et al. [HPQCD Collaboration], Phys. Rev. D 98, no. 1, 014513 (2018) doi:10.1103/PhysRevD.98.014513 [arXiv:1805.06225 [hep-lat]]
arXiv 2018
-
[6]
K. G. Chetyrkin and A. Retey, Nucl. Phys. B 583, 3 (2000) doi:10.1016/S0550-3213(00)00331-X [hep-ph/9910332]
arXiv 2000
-
[7]
A. Bazavov et al. [MILC Collaboration], Phys. Rev. D 87, no. 5, 054505 (2013) doi:10.1103/PhysRevD.87.054505 [arXiv:1212.4768 [hep-lat]]
arXiv 2013
-
[8]
J. A. Gracey, Nucl. Phys. B 662, 247 (2003) doi:10.1016/S0550-3213(03)00335-3 [hep-ph/0304113]
arXiv 2003
Show all 14 references
-
[9]
B. A. Galloway, P. Knecht, J. Koponen, C. T. H. Davies and G. P. Lepage, PoS LATTICE2014, 092 (2014) doi:10.22323/1.214.0092 [arXiv:1411.1318 [hep-lat]]
2014 arXiv
-
[10]
Tanabashi et al
M. Tanabashi et al. [Particle Data Group], Phys. Rev. D 98, no. 3, 030001 (2018). doi:10.1103/PhysRevD.98.030001
2018 doi
-
[11]
G. C. Donald, C. T. H. Davies, R. J. Dowdall, E. Follana, K. Hornbostel, J. Koponen, G. P. Lepage and C. McNeile, Phys. Rev. D 86, 094501 (2012) doi:10.1103/PhysRevD.86.094501 [arXiv:1208.2855 [hep-lat]]
2012 arXiv
-
[12]
K. G. Chetyrkin, J. H. Kuhn, A. Maier, P. Maierhofer, P. Marquard, M. Steinhauser and C. Sturm, Phys. Rev. D 80, 074010 (2009) doi:10.1103/PhysRevD.80.074010 [arXiv:0907.2110 [hep-ph]]
2009 arXiv
-
[13]
Chakraborty et al
B. Chakraborty et al. [HPQCD Collaboration], Phys. Rev. D 89, no. 11, 114501 (2014) doi:10.1103/PhysRevD.89.114501 [arXiv:1403.1778 [hep-lat]]
2014 arXiv
-
[14]
Borsanyi et al
S. Borsanyi et al. [Budapest-Marseille-Wuppertal Collaboration], Phys. Rev. Lett. 121, no. 2, 022002 (2018) doi:10.1103/PhysRevLett.121.022002 [arXiv:1711.04980 [hep-lat]]. 6
2018 arXiv
Reviewed August 14, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.