A New Method For Manipulating Circuits, Application To Quantum Adders

Scott Wesley; William Schober

REVIEW

A New Method For Manipulating Circuits, Application To Quantum Adders

William Schober, Scott Wesley

Quantum Physics · Submitted 22 Sep 2026

Pith's verdict: A plausible and potentially important structural equivalence between QFT and ripple-carry adders, but the load-bearing rewrite identities are asserted from the authors' own prior work rather than proven here.

T0 · AI review

How this review was made: An AI referee (deepseek/deepseek-v4-flash) read the paper on September 24, 2026 and answered a fixed list of questions about it. T0 is the starting level every paper gets. The levels describe how deeply Pith has checked the paper, never the authors.

T0 AI review An AI read the paper and answered a fixed list of questions.
T1 Audit Automatic checks that the references and files the paper relies on exist and match.
T2 Compute The paper's code or calculations re-run with the same results.
T3 Formal statement The claims written out in a formal math language a computer can read.
T4 Proof Those claims proved, and the proof checked by computer.

How Pith checks papers →

· 4 major objections · 5 minor · Ask about this paper · The compiled record · Read the full review →

Challenge this review Comment (0) Read on arXiv ↗

Cite ▾

arXiv ID2609.26591 (version 2)

Pith IDTDI5RMYV, Pith's permanent record of this paper and its review

CitationScott Wesley, William Schober. "A New Method For Manipulating Circuits, Application To Quantum Adders." arXiv:2609.26591 (2026). https://pith.science/paper/TDI5RMYV

Cite this reviewBibTeX for Pith's review of this paper

Claim paper

Why Pith says this

Highlight any text to comment on it.

Editor's note

read the letter

Quick take: this is a genuine structural result if the rewrite engine is sound—an explicit route from Draper's QFT adder to Takahashi's ripple-carry adder with a new ancilla-free carry-lookahead circuit in the middle. The recursive definitions and the two inductions are coherent, and the target equivalence is independently plausible because both adders implement the same map. The paper doesn't oversell resources: it says the naive CL-Adder compilation is exponentially deep and frames the contribution as transpilation, not as a better adder.

The soft spot is the equational foundation. Everything in Sections 3 and 4 depends on identities (3)-(5), (11) and (13), taken from the authors' own hierarchical-circuit theory [9]. Those identities are asserted, not re-derived or machine-checked in this manuscript. Eq. (4) is the most delicate: it carries an orthogonality condition B⊥D, and the support argument in §3.1 is exactly the kind of step where a subtle mistake could hide. Eq. (13) is thinner still—it's a diagram plus one line about Pauli anticommutation, and it's what merges C^(3) into stages 3-5 of Takahashi's circuit. If any of these local rewrites is wrong, both inductions collapse. Relying on prior work is normal; the problem is not self-citation per se, it's that these identities are load-bearing and the reader cannot check them here because [9] is forthcoming. To be clear, I don't see an internal contradiction, and the gap is addressable with an appendix or machine-checked lemmas, but as submitted the paper's correctness is contingent on prior work being exactly right.

Second soft spot: the novelty claim. The paper says the CL-Adder is the first ancilla-free carry-lookahead adder, but it cites Remaud and Vandaele [5], whose title is exactly "Ancilla-Free Quantum Adder with Sublinear Depth," and Remaud [8] on the ripple-carry/carry-lookahead link. The paper needs a short comparison explaining why those results don't overlap. This is minor and fixable.

Third, there is no resource improvement and no code or formal verification artifact. Again, that's fine for a structural paper, but it sets the ceiling on significance: useful for people studying adder structure and circuit equivalence, not for practitioners who want faster addition.

My take: this deserves a serious referee. The skeleton is sound and the central claim is important. I'd send it out, with a clear request that the next version either prove the loaded identities or point to machine-checked proofs, and add the comparison with [5]/[8]. I wouldn't cite it as settled until that happens.

Referee report

Significance

The claimed equivalence is notable and aesthetically pleasing: it connects two structurally different adder families by a chain of local rewrites, and it introduces a concrete, ancilla-free Carry-Lookahead adder. The paper gives explicit recursive definitions and a legible induction structure, which are strengths. If the underlying equational identities are correct, the result provides a novel 'path' through circuit space and a useful intermediate circuit. However, the result rests entirely on a set of local identities, Eqs. (3)-(5), (11), and (13), that are neither proved nor precisely referenced in this manuscript; they are asserted from the authors' own forthcoming paper [9]. Without those identities the two inductions collapse, so the significance is conditional on the validity and correct application of those identities.

Major objections (4)

[Section 3.1, Eq. (4)] The orthogonality condition B⊥D is load-bearing for the entire merging stage, yet it is justified only by an informal support argument about Log B and Log D. The statement that 'Log B is unsupported on every subspace ... of the form |11>_{a_j b_j} ⊗ (-) while Log D is nonzero only on subspaces of this form' is not derived, and the decomposition of a ⊙-product into such subspaces is exactly the kind of step that needs a formal argument. Please provide a proof of Eq. (4), or a precise reference to a lemma in [9] with the statement reproduced.
[Section 4.3, Eq. (13)] The final step of Proposition 4.1 hinges on Eq. (13), which is presented only as a diagram with a one-line 'usual Pauli anticommutation identity' justification. This identity is used to merge C^(3) into stages 3, 4, and 5, and without it the transpilation to RC-Adder is incomplete. A formal statement of Eq. (13), including the exact gate ordering and the controlled-qubit structure, should be given, together with a derivation or a precise citation to a proved result in [9].
[Section 1.1 and Sections 3-4] The paper repeatedly applies identities (3), (4), (5), (11), and (13) as if they were established facts, but it only states that a sound-and-complete equational theory 'was discovered' in [9]. Since [9] is a single forthcoming reference and the identities are not reproduced or proved in this manuscript, the argument is not self-contained. Please provide explicit statements of all used identities, ideally in an appendix, or give a detailed mapping to the corresponding results in [9].
[Entire paper] No machine-checked proof, circuit simulation, or small-n numerical verification is reported. Given that the induction proofs involve intricate index bookkeeping and repeated application of the local identities, a verification for n=2, n=3, and n=4 would materially increase confidence and would also test the indexing of Eqs. (9), (10), and (12). This is not required for every formal proof, but it would be a valuable check for a paper whose main claims are of this type.

Minor objections (5)

[Section 1.1] There is a typo in 'The the empty wire id: 1→1'; also the notation 'we omit the J−K' is confusing, because the paper continues to use J−K in several places.
[Section 1, Introduction] The sentence 'the Quantum Fourier Transform adder is described used parameterizedZrotation gates' is grammatically broken and should be rewritten.
[Section 2, Eq. (1)] The recursive equation for QFTAprefix uses index conventions that are not fully aligned with the interleaved bit order defined in Section 1.2; please clarify the range of the product and the wire ordering in the diagrams.
[Section 4.1, Eq. (7) and Eq. (8)] The displayed identities in Eq. (7) and Eq. (8) appear as images and are not rendered in the text; the captions also do not specify the gate labels. Please ensure the figures are reproducible from the text alone.
[Section 3.2] The phrase 'expanding the recursion (and merging the X controls on the lowest wire)' is informal; the resulting closed form in Eq. (6) should be derived explicitly, since it is used in the induction proof.

How Pith checks papers → · Full audit of this paper →

The paper

The paper claims

This paper proves that the Quantum Fourier Transform adder and the ripple-carry adder are connected by an explicit sequence of local gate rewrites for every width n, with a new ancilla-free carry-lookahead adder along the way.

The reading

The paper's central claim is that two standard but seemingly unrelated quantum addition circuits—the Quantum Fourier Transform adder and the ripple-carry adder from classical reversible logic—are equivalent through an explicit gate-level transpilation for every adder width $n$. The proof builds a chain of local rewrites that passes through a third adder, introduced here as an ancilla-free carry-lookahead adder whose prefix distinguishes bit pairs that generate or propagate a carry. This closes the previously open structural gap between the Fourier-based and reversible-classical adder families, which use different gatesets and different recursive shapes. The authors acknowledge that a straightforward compilation of the new intermediate adder to the classical reversible gateset $\{X,\mathrm{CNOT},\mathrm{Toffoli}\}$ is exponentially deep (Section 3), so the contribution is the explicit structural bridge rather than an immediately practical adder. If the equivalence is right, it gives concrete reason to regard the two adder families as representations of one underlying circuit rather than separate species.

What carries the argument

The load-bearing object is the control product $\odot$, a connective from hierarchical quantum circuits: for circuits $C_1:n\to n$ and $C_2:m\to m$, the circuit $C_1\odot C_2$ acts as $C_2$ controlled on the eigenstates of $C_1$; together with circuit powers $C^\alpha$ it subsumes $Z\odot X = \mathrm{CNOT}$ and the small-angle rotations of the QFT adder. The proof runs on a small set of local rewrite identities—merging powers (Eq. (3)), merging under an orthogonality condition (Eq. (4)), Hermitian-prefix cancellation (Eq. (5)), and the Toffoli-pushing identities (Eqs. (11), (13))—together with recursive definitions of the Fourier transform, transform-addition, and seven-stage ripple-carry circuits. These identities let the authors merge every small-angle power $2^{-k}$ in the joint prefix, push the inverse Fourier prefix through the adder so it cancels, and then push the resulting carry-lookahead prefix through the first three ripple-carry stages. The output is an exact chain of local rewrites rather than a black-box unitary equivalence.

Extended reading notes

Core claim

The headline result is Corollary 4.2: $\mathrm{RC\text{-}Adder}_{0:n-1} = \mathrm{QFT\text{-}Adder}_{n-1}$ for all $n$. The proof reaches it in two inductions. Proposition 3.1 proves $\mathrm{CL\text{-}Adder}_{0:n-1} = \mathrm{QFT\text{-}Adder}_{0:n-1}$ by merging the powers inside the joint prefix of the Fourier transform and transform-addition circuits, pushing the inverse Fourier prefix through the adder, and cancelling; Proposition 4.1 proves $\mathrm{RC\text{-}Adder}_{0:n-1} = \mathrm{CL\text{-}Adder}_{0:n-1}$ by pushing the carry-lookahead prefix through the first three stages of the seven-stage ripple-carry adder and recovering the three missing gates. The intermediate object, the ancilla-free carry-lookahead adder of Eq. (6), computes the most significant sum bit by using the $\pm1$ eigenspaces of $\mathrm{CLprefix}$ to see whether a carry would be generated or propagated, without first computing the lower carries. The whole transpilation is an equality between circuits, not merely between their unitary matrices.

Load-bearing premise

The load-bearing premise is that the equational theory for hierarchical circuits used here is sound and complete, so every local rewrite in the chain exactly preserves the unitary map; if any of the rewrite identities it supplies is invalid, both induction proofs collapse.

Editorial extensions

If this is right

For every width $n$, the QFT adder and the seven-stage ripple-carry adder coincide under an explicit sequence of local rewrites (Corollary 4.2).
There exists an ancilla-free carry-lookahead adder, displayed in Eq. (6) and Fig. 5, which the authors report as, to their knowledge, the first such adder with no ancillas.
The transpilation produces a connected path through the space of quantum adders whose intermediate circuits range from the QFT adder to the ripple-carry adder, with each step a local circuit equivalence.
The authors note that the naive compilation of the CL-Adder to the classical reversible gateset $\{X,\mathrm{CNOT},\mathrm{Toffoli}\}$ is exponentially deep, so the new adder is established as a structural intermediate, not as an efficient implementation.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The same control-product rewriting machinery plausibly applies to other controlled arithmetic circuits—comparators, incrementers, modular addition—where recursive prefix/suffix mismatches make direct transpilation awkward; the proof pattern of this paper is a natural template.
Because the CL-Adder's prefix has a highly degenerate spectrum, an efficient standard-gate implementation may exist even though the naive one is exponential; constructing it would turn the structural bridge into a practical adder.
The equality between a Fourier-based circuit and a reversible-classical circuit suggests that resource differences among adder families—native gateset, depth, ancilla count—are properties of the chosen representation of addition, not of addition itself; one representation can be deformed into another without changing the computation.

Original abstract

We use a new technique for manipulating controlled quantum circuits to convert between two distinct types of quantum adders, one based on the Quantum Fourier Transform and the other based on the Ripple-Carry technique from classical reversible logic. This conversion takes the form of an explicit gate-level transpilation. We also present a new quantum adder with a natural interpretation as a kind of Carry-Lookahead adder that uses no ancillas.

The compiled record

The paper's own sentences, arranged, kept beside the review and separate from it. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.

Compiled by classifier-1.13.0 via TypeSafe API, question bank abstract-p1-20260923, 2026-09-24: 1 request, 2,918 input tokens.

What the abstract itself states, as probabilities: claims first 0.10 · states limitation 0.08 · human participants 0.02 · controlled comparison 0.03 · randomized assignment 0.01 · states quantitative result 0.05 · compares with prior results 0.43 · mentions code or data availability 0.03.

The abstract, sentence by sentence

arXiv listing abstract v1. Results and the main claim are underlined; show every role.

We use a new technique for manipulating controlled quantum circuits to convert between two distinct types of quantum adders, one based on the Quantum Fourier Transform and the other based on the Ripple-Carry technique from classical reversible logic. This conversion takes the form of an explicit gate-level transpilation. We also present a new quantum adder with a natural interpretation as a kind of Carry-Lookahead adder that uses no ancillas.

background contribution method result interpretation main claim

What the classifier reports for each sentence. The number is the probability it gave the label.
#	sentence	role	the authors' own work	states the main conclusion	strength
s0	We use a new technique for manipulating controlled quantum circuits to convert between two distinct types of quantum adders, one based on the Quantum Fourier Transform and the other based on the Ripple-Carry technique from classical reversible logic.	contribution 0.84	0.93	0.40	asserted 1.00
s1	This conversion takes the form of an explicit gate-level transpilation.	method 0.89	0.87	0.10	asserted 1.00
s2	We also present a new quantum adder with a natural interpretation as a kind of Carry-Lookahead adder that uses no ancillas.	contribution 1.00	0.95	0.27	asserted 1.00

What kind of paper the classifier reports this is: method or tool 1.00 · evidence from proof or derivation 0.91.

What was examined, and what was not

sources read

arXiv listing abstract v1

not examined

full text
figures and tables
references and citing sentences
relations to other papers

levels run

P abstract

not run

S sections
T body sentences
C citations
R cross-paper relations

Every quoted sentence on this record is the authors'. Every label is from Pith's closed vocabulary, with the probability shown beside it. Every number in the measurements table is the authors', checked by code. Record as JSON

Figures

**Figure 1.** The graphical language for hierarchical quantum circuit diagrams. Note that

**Figure 2.** The Quantum Fourier Transform Adder as introduced in [

**Figure 4.** The Ripple-Carry Adder first introduced in [

**Figure 5.** The Carry-Lookahead Adder described in Section

**Figure 6.** Recursive definition of the Quantum Fourier Transform circuit.

**Figure 7.** Recursive definition of the Transform Addition circuit.

**Figure 8.** Recursive QFT-sandwich form of the Quantum Fourier Transform Adder.

**Figure 9.** The Lookahead and CLprefix circuits on 5 qubit pairs.

**Figure 10.** The two steps comprising the first merging stage for QFTAprefix with

**Figure 11.** QFTAprefix after all n steps of both merging stages, with n = 5. |(ab)0:n−2⟩ CLAdder |an−1⟩ |bn−1⟩ = CLprefix CLAdder |(as)0:n−2⟩ |an−1⟩ |sn−1⟩

**Figure 12.** Recursive definition of the Carry-Lookahead Adder.

**Figure 13.** Recursive definition of the Ripple-Carry Adder.

**Figure 14.** Pushing C (0) 0:5 through 10:5 results in C (1) 0:5 . In other words, C (j) 0:n is what remains of C (0) 0:n after being pushed through the first j stages of 10:n ◦ 20:n ◦ 30:n−1. To prepare for the transpilation, we first compute C (1) 0:n , C(2) 0:n , and C (3) 0:n . 4.1 Pushing through stage 1 First we will compute C (1) 0:n using the two following identities. Eq. (7) is a corollary of Eq. (5) when A = Z and α = 1, though it is better known as one of the standard ‘Pauli pushing’ identities. Eq. (8) describes how CZ = Z ⊙ Z changes when pushed through a CNOT. =

**Figure 15.** Pushing C (1) 0:5 through 20:5 results in C (2) 0:5 . Pushing C (0) 0:n through 10:n therefore gives the following expressions for C (1) 0:n , where in the last step we have only regrouped terms. An example of this stage is shown in

**Figure 16.** Pushing the rightmost controlled box contained in

**Figure 17.** Pushing C (2) 0:5 through 30:4 results in C (3) 0:5 . CLprefix CLAdder = CLprefix CLAdder

**Figure 18.** Preparing CNOTs to merge into stages 1,2,6, and 7 of the CL-Adder.

**Figure 19.** Merging CNOTs into stages 1,2,6, and 7 of CL-Adder. Stages 3,4, and 5 are still missing

**Figure 20.** Pushing C (3) 0:5 through 40:4 results in the three missing gates to recover stages 3, 4, and 5 of Takahashi ripple-carry adder [4]. Pauli anticommutation identity = . = = (13) Note that the three gates Toffoli(an−1, bn−1, an), CNOT(an, bn), and Zan−1 ⊙ (−Z)an−1 ⊙ Xan in Eq. (13) are exactly the three missing gates in 30:n, 40:n, and 50:n respectively. Hence we can merge them to recover 30:n, 40:n, and 50:n. This is shown in

Showing 19 of the paper's 20 figures. See all figures in the PDF ↗ Figures reproduced from arXiv:2609.26591 by the authors.

References

Discussion

Disagree with the review, or know something it missed? Leave a comment, highlight any passage above to comment on it, or challenge this review formally.