Uniqueness of the solution of nonlinear singular first order partial differential equations

Hidetoshi Tahara

Uniqueness of the solution of nonlinear singular first order partial differential equations

T0 review · 0 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The paper proves that a weak double-limit growth condition—after time tends to zero, the solution size on a disk divided by $R^2$ tends to zero as the disk shrinks—forces any solution of the singular first-order equation to coincide with…

desk verdict A technically sound, honest uniqueness paper for a narrow class of singular first-order PDEs; the double-limit condition is a real weakening and the sectorial case is new, but Case 2 is sketched and the sectorial condition lacks motivating examples. read the letter →

arxiv 1908.08182 v1 pith:HTER257B submitted 2019-08-22 math.AP

Hidetoshi Tahara This is my paper · ORCID

classification math.AP MSC 35A0235F2035B60

keywords uniquenessofsolutionsnonlinearsingularpartialdifferentialequationsfirst-orderPDEBriot-Bouquettotallycharacteristicirregularsingularitymethoddouble-limitgrowthcondition

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper proves a uniqueness theorem for nonlinear singular first-order partial differential equations of the form $t\,\partial u/\partial t=F(t,x,u,\partial u/\partial x)$, where $F$ is continuous in $t$ and holomorphic in the other variables. The result is that a very weak two-step smallness condition selects exactly one solution: after measuring the size of a solution on the time strip $(0,\sigma)\times D_R$, letting $\sigma\to 0$, and then letting the disk radius $R\to 0$, the quotient $\frac{1}{R^2}\sup_{(0,\sigma)\times D_R}|u|$ must tend to zero. Under the sign assumption $\operatorname{Re}\lambda(0,0)<0$ for $\lambda=\partial F/\partial u$ at the origin, any solution satisfying this condition agrees with the unique distinguished solution $u_0$ produced by earlier existence theory. The same conclusion is obtained for totally characteristic equations and for irregular singular equations on sectorial domains, with extra sign conditions on the coefficient $c$. A consequence is that local holomorphic solutions of Briot–Bouquet and totally characteristic type equations that satisfy this condition extend holomorphically to a full neighborhood of the origin.

What carries the argument

The load-bearing object is the double-limit growth condition (2.3), and the load-bearing mechanism is integration along characteristic curves of the linearized equation. Subtracting the distinguished solution reduces the equation to a linear first-order system for $w=u-u_0$ and $q=\partial w/\partial x$, of the form $t w_t - b(t,x)w_x = (\lambda(t,x)+a(t,x))w$, with a companion equation for $q$. Along the characteristic curves $t\,dx/dt=-b(t,x)$ the coefficient $\lambda+a$ keeps real part below $-a$, so solutions decay by the factor $(t_1/\tau)^a$ as the curve is followed backward in time. The smallness condition, through standard estimates on holomorphic functions, bounds $w$, $q$, and the characteristic coefficients so that every characteristic starting from a small disk or sector reaches $t=0$; letting $t_1\to0$ then kills the terminal value. In the irregular singular case an additional explicit representation $x(t_1)=\xi/\varphi(t_1)\bigl(1-p\xi^p\int_{t_1}^{t_0} c(\tau,x(\tau))/\varphi(\tau)^p\,d\tau/\tau\bigr)^{1/p}$, together with the sign $c(0,0)<0$ and a sectorial derivative estimate, keeps the characteristics inside the sector until $t=0$.

What would settle it

Compute the double limit for the explicit equation $t u_t=-u+(u_x)^2$: the nontrivial solution $u=x^2/4$ has limit $1/4$, not $0$, so condition (2.3) excludes it and the theorem predicts that any solution of this equation satisfying (2.3) must be $u\equiv0$. A direct way to break the theorem would be to find any equation with $\operatorname{Re}\lambda(0,0)<0$ carrying a nonzero solution that satisfies (2.3) or (4.6); in the irregular case the paper reports no example showing whether condition (4.6) is sharp.

Watch

Extended reading notes

Core claim

The central discovery is that uniqueness near a singular point is controlled by a double limit rather than by a single decay rate in $t$: a solution is unique if it is $o(R^2)$ on the disks $D_R$ after all sufficiently small positive times are taken into account. In Case 1 (Briot–Bouquet type), Theorem 2.2 states that with $\operatorname{Re}\lambda(0,0)<0$, every $C^1$ solution holomorphic in $x$ that satisfies (2.3) is equal to $u_0$ on $(0,T_1)\times D_{R_1}$. Theorems 3.2 and 4.2 extend the same statement to totally characteristic equations and to irregular singular equations on sectors $S(\theta,R)$, using the sectorial analogue (4.6). The proof works by subtracting $u_0$, writing the difference as a linear transport equation, and integrating along characteristic curves that reach $t=0$; the weak smallness condition supplies the estimates that keep the characteristics inside the domain. Theorems 2.10 and 3.5 apply this uniqueness to show that local holomorphic solutions satisfying the condition can be continued analytically to the origin.

Load-bearing premise

The proof stands on sign conditions: $\operatorname{Re}\lambda(0,0)<0$ must keep the decay coefficient negative along characteristics, and in Cases 2 and 3 the coefficient $c$ in $\partial F/\partial v=b(t)+x^{p+1}c(t,x)$ must keep non-positive real part (or $c(0,0)<0$), so that characteristics remain in the domain until $t=0$; without these signs the uniqueness conclusion is false in general.

Editorial extensions

If this is right

With $\operatorname{Re}\lambda(0,0)<0$, any solution that is uniformly small as $t\to0$ on a fixed disk is unique: Corollaries 2.3, 3.3, and 4.3.
Local holomorphic solutions of Briot–Bouquet type equations satisfying the double-limit condition extend holomorphically to a neighborhood of the origin (Theorem 2.10); the same holds for totally characteristic equations (Theorem 3.5).
The new uniqueness condition is weaker than earlier rates such as $u=O(\mu(t)^\varepsilon)$, so the theorem applies to solutions whose decay in $t$ is slower than any fixed power as long as their size on shrinking $x$-disks is $o(R^2)$.
In the irregular singular case, uniqueness holds on sectorial domains with arbitrarily small aperture for solutions satisfying the sectorial condition (4.6), not only for solutions that vanish uniformly as $t\to0$.

Reading between the lines

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

The $o(R^2)$ threshold looks sharp: since the explicit solution $u=x^2/4$ sits exactly at limit $1/4$, the theorem suggests that quadratic-in-$x$ growth is what permits non-uniqueness and that the condition cannot be relaxed to $O(R^2)$ without adding hypotheses.
The same characteristic-decay proof may extend to higher-order equations by differentiating enough times to close the system, but the paper explicitly leaves that extension open.
Read as a removable-singularity principle, the theorem predicts that for this class a solution which is sufficiently small on every $R$-disk near $t=0$ has no genuine singularity at the origin; only solutions with at least quadratic growth in $x$ can escape uniqueness.
A natural test in the irregular case is to search for nontrivial solutions satisfying (4.6) for equations like $t u_t=-u-x^2u_x+t(xu_x)^2$; the paper reports no example showing whether that condition is sharp.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Bottom line: this is a competent and honest uniqueness paper in a narrow area. The new thing is the double-limit growth condition (2.3)/(3.5)/(4.6), which is genuinely weaker than the earlier O(µ(t)^epsilon) or O(|t|^epsilon) hypotheses, and the sectorial Case 3 theorem, which I don't think exists before. The counterexamples in Remarks 2.4, 3.4, and 4.4 show the sign condition on Re lambda(0,0) is sharp, and the x^2/4 example is a nice calibration for the R^2 scaling.

The proof method is a characteristic-based argument, and for Cases 1 and 3 it is written out in enough detail that I could follow the estimates. The sign conditions Re c <= 0 in Case 2 and c(0,0) < 0 in Case 3 do real work; without them the proof would not close, and the paper says so itself. The cited existence theorems are used as black boxes to get the reference solution u0; that is fine, since uniqueness is not assumed in the proof. Self-citation is heavy but mostly to prior published results by the group, so I don't see circularity.

Soft spots are real but minor in aggregate. Case 2's proof is explicitly a sketch: the analogue of Lemma 2.6 is asserted, not proved. I believe it fills in, because the added c term has negative real part and the x(dc/dx) term is controlled by taking R small, but the reader has to do the work. Lemma 4.6 has a harmless algebraic slip: the final displayed bound is not what the preceding line gives, but the preceding line already yields the needed o(eta^{m-1}) estimate, so the theorem is unaffected. More important, Remark 4.4(3) admits that condition (4.6) is not known to be natural: no example shows the scaling is right in the sectorial case. That does not break the theorem, but it means the main result of Section 4 is less motivated than Sections 2 and 3. Generalization to higher order is also explicitly left open.

Who it is for: researchers working on singular first-order PDEs, Briot-Bouquet equations, and totally characteristic or irregular singular equations. It is not a broad-audience paper. It deserves a serious referee; I would send it out. If I refereed, I would ask the author to expand the Case 2 proof, fix Lemma 4.6, and say a bit more about whether (4.6) has supporting examples. None of this is fatal. The central claim holds up.

Referee Report

0 major / 4 minor

Summary. The paper studies uniqueness for nonlinear singular first-order PDEs of the form t u_t = F(t,x,u,u_x) with t ∈ R and x ∈ C, under a very weak double-limit growth condition (2.3). The author splits the analysis into three cases according to the structure of ∂F/∂v at (0,0): Case 1 (Briot–Bouquet type), Case 2 (regular singularity in x), and Case 3 (irregular singularity). In each case, under the sign condition Re λ(0,0) < 0 (plus Re c ≤ 0 in Case 2 and c(0,0) < 0 in Case 3), the only solution is the distinguished solution u0 from the corresponding existence theorem. The proofs use a characteristic method: after subtracting u0, the difference w satisfies a first-order linear equation along characteristics; estimates on the characteristic flow show that w(τ) decays like (t1/τ)^a, and the limit t1→0 yields w = 0. Applications to analytic continuation of local holomorphic solutions are given in Sections 2.3 and 3.3.

Significance. If the results hold, they are a significant contribution: condition (2.3) is far weaker than the previously known O(μ(t)^ε) decay conditions, and the counterexamples in Remarks 2.4, 3.4, and 4.4 show the sharpness of the sign of Re λ(0,0) and of the R^2 scaling. The characteristic estimates in Case 1 (Proposition 2.5) and Case 3 (Proposition 4.5) are written out in considerable detail, and the use of Nagumo's lemma in sectorial domains is appropriate. The paper is self-contained modulo three prior existence theorems from [1], [2], and [11], used as black boxes; this is acceptable because the uniqueness question concerns a single solution already known to exist. The main weakness is that the proof of Theorem 3.2 in Case 2 is presented only as a sketch; the missing estimate is, however, standard and easily supplied.

minor comments (4)

[§3.2, proof of Theorem 3.2] The proof of Theorem 3.2 is only sketched. In the analogue of Lemma 2.8 one must additionally use Re c ≤ 0 to obtain d|x|/dt ≥ −|b(t,x)|/t along the characteristic (3.8), which gives the same bound on |x(t1)| as in Case 1; please add this one-line argument so that the reader is not left to infer it.
[§4.2, Lemma 4.6] In the final displayed inequality of the proof of Lemma 4.6, the right-hand side should be (2^m ε/θ) (η/2)^{m−1} rather than the printed expression. In addition, the estimate is proved on S((η/2)θ, (η/2)R); to cover the full sector S(ηθ, ηR) one should apply the result with η replaced by 2η. Both points are harmless for the conclusion, but should be corrected for clarity.
[§2.2, Lemma 2.6] In the proof of Lemma 2.6, the passage from (2.10) to (2.11) via Cauchy's integral formula implicitly uses the supremum of w on a slightly larger disk than D_R (for instance D_{2R}); this should be stated explicitly.
[Throughout] There are a few typographical slips: 'Theroem 2.1' in the proof of Theorem 2.2 should be 'Theorem 2.1', and in Remark 4.4(1) the symbol x2 should be understood as x^2. These do not affect the mathematics.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the uniqueness proofs reduce an arbitrary solution to the reference solution via characteristic estimates, with no fitted input or definitional identification.

full rationale

The paper's central claim is the uniqueness of solutions satisfying the double-limit condition (2.3), (3.5), or (4.6). The reference solution u0 is supplied by prior existence theorems (Lope–Roque–Tahara [11], Bacani–Tahara [1], Bacani–Tahara [2]); these are used only to fix a candidate, not to assume that every solution in the larger class equals it. The proof then forms w = u − u0, derives linearized transport-type equations (2.8), (3.6), and (4.9) with explicit coefficients, and uses the sign conditions Re λ(0,0) < 0, Re c ≤ 0 (Case 2), and c(0,0) < 0 (Case 3) together with the growth condition to estimate w along characteristics and conclude w = 0. The growth condition (2.3) is an assumption on the solution u, not a consequence of the existence theorem; it enters only through the smallness estimates in Lemma 2.6 and the analogous lemmas. No parameter is fitted to the target, no quantity is defined in terms of the conclusion, and the cited existence results have stated assumptions that do not include the target uniqueness under (2.3). The paper even flags its own open points (e.g., Remark 4.4(3) and the higher-order generalization in Section 2.1), which are limitations of scope rather than circular dependencies. Therefore the derivation is self-contained relative to the cited existence theorems, and no circular step is present.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

No fitted parameters, empirical constants, or new objects are introduced. The paper is a pure analytical proof; all assumptions are explicit in the theorem statements.

assumptions (4)

domain assumption Existence and uniqueness of the reference solution u0 with |u0| and |du0/dx| bounded by M*mu(t) (Theorem 2.1 from [11] for Case 1; Theorem 3.1 from [1] for Case 2; Theorem 4.1 from [2] for Case 3).
The uniqueness statements compare any solution to this u0; if no such u0 exists, the conclusion 'u = u0' is empty. These theorems are cited from prior work by the same research group and used as black boxes.
domain assumption Weight function mu(t) positive, increasing, integrable with integral_0^{T0} mu(s)/s ds < infinity, and |F(t,x,0,0)| and |dF/dv(t,x,0,0)| both O(mu(t)) as t -> 0 (assumption A2).
Defines the scale of the reference solution and the admissible growth near t=0; the integrability is used to bound phi(sigma).
ad hoc to paper Sign conditions on the coefficient c: Re c(t,x) <= 0 in Case 2 (3.4) and c(0,0) < 0 in Case 3 (4.5).
These hypotheses are not derived from the equation; they guarantee that the transport term keeps the real part of the effective coefficient negative along characteristics, which drives the decay estimates in Lemmas 2.7, 4.8 and 4.10.
standard math Standard analytic tools: Cauchy's integral formula, Nagumo's lemma in sectorial domains (cited from Lemma 4.2 of [2]), and continuation theorems for ODEs (Theorem 4.1 of Coddington-Levinson [7]).
Used in Lemmas 2.6, 4.6 and Corollary 2.9 to convert sup estimates into derivative estimates and to extend characteristic solutions.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Uniqueness of the solution of nonlinear singular first order partial differential equations." pith.science (2026). https://pith.science/paper/HTER257B

@misc{pith2026190808182,
  author       = {Pith},
  title        = {Pith review of: Uniqueness of the solution of nonlinear singular first order partial differential equations},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HTER257B}},
  note         = {Machine review of arXiv:1908.08182}
}

abstract

This paper deals with nonlinear singular partial differential equations of the form $t \partial u/\partial t=F(t,x,u,\partial u/\partial x)$ with independent variables $(t,x) \in \mathbb{R} \times \mathbb{C}$, where $F(t,x,u,v)$ is a function continuous in $t$ and holomorphic in the other variables. Under a very weak assumption we show the uniqueness of the solution of this equation. The results are applied to the problem of analytic continuation of local holomorphic solutions of equations of this type.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

15 extracted references · 15 canonical work pages

[1]

D. B. Bacani, H. Tahara, Existence and uniqueness theore m for a class of singular nonlinear partial diﬀerential equations, Publ. Res. Inst. Math. Sci., 48 (2012), no. 4, 899-917

work page 2012
[2]

D. B. Bacani, H. Tahara, Unique solvability of some nonli near partial diﬀerential equations with Fuchsian and irregular singular ities, J. Math. Soc. Japan, 66 (2014), no. 3, 1017-1042. 23

work page 2014
[11]

J. E. C. Lope, M. P. Roque, H. Tahara, On the unique solvab ility of cer- tain nonlinear singular partial diﬀerential equations, Z. A nal. Anwend., 31 (2012), no. 3, 291-305

work page 2012
[3]

M. S. Baouendi, C. Goulaouic, Singular nonlinear Cauchy problems, J. Diﬀerential Equations, 22 (1976), no. 2, 268-291

work page 1976
[4]

Briot, J

Ch. Briot, J. Cl. Bouquet, Recherches sur les propri´ et´ es des fonctions d´ eﬁnies par des ´ equations diﬀ´ erentielles, J. Ecole Polytech. 21 (1856), 133-197

work page
[5]

H. Chen, H. Tahara, On totally characteristic type non-l inear partial diﬀerential equations in the complex domain, Publ. Res. Inst . Math. Sci., 35 (1999), no. 4, 621-636

work page 1999
[6]

H. Chen, Z. Luo, C. Zhang, On the summability of formal sol utions for a class of nonlinear singular PDEs with irregular singul arity. Recent progress on some problems in several complex variables and p artial diﬀerential equations, 53-64, Contemp. Math., 400, Amer. Ma th. Soc., Providence, RI, 2006

work page 2006
[7]

E. A. Coddington, N. Levinson, Theory of ordinary diﬀeren tial equa- tions, McGraw-Hill, 1955

work page 1955

Show all 15 references

[8]

G´ erard, H

R. G´ erard, H. Tahara, Holomorphic and singular solutio ns of nonlinear singular ﬁrst order partial diﬀerential equations. Publ. Re s. Inst. Math. Sci., 26 (1990), no. 6, 979-1000

1990
[9]

G´ erard, H

R. G´ erard, H. Tahara, Singular nonlinear partial diﬀere ntial equations, Aspects of Mathematics, Friedr. Vieweg & Sohn, Braunschwei g, 1996

1996
[10]

Koike, Volevic systems of singular nonlinear partia l diﬀerential equa- tions, Nonlinear Analysis, Theory Method & Applications, 2 4 (1995), no

M. Koike, Volevic systems of singular nonlinear partia l diﬀerential equa- tions, Nonlinear Analysis, Theory Method & Applications, 2 4 (1995), no. 7, 997-1009

1995
[12]

Z. Luo, H. Chen, C. Zhang, Exponential-type Nagumo norm s and summability of formal solutions of singular partial diﬀeren tial equa- tions. Ann. Inst. Fourier, 62 (2012), no. 2, 571-618

2012
[13]

Tahara, Uniqueness of the solution of non-linear sin gular partial diﬀerential equations, J

H. Tahara, Uniqueness of the solution of non-linear sin gular partial diﬀerential equations, J. Math. Soc. Japan, 48 (1996), no. 4, 729-744

1996
[14]

Tahara, Uniqueness of the solution of nonlinear tota lly characteristic partial diﬀerential equations, J

H. Tahara, Uniqueness of the solution of nonlinear tota lly characteristic partial diﬀerential equations, J. Math. Soc. Japan, 57 (2005 ), no. 4, 1045-1065

2005
[15]

Yamazawa, Singular solutions of the Briot-Bouquet t ype partial dif- ferential equations, J

H. Yamazawa, Singular solutions of the Briot-Bouquet t ype partial dif- ferential equations, J. Math. Soc. Japan, 55 (2003), no. 3, 6 17-632. 24

2003

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [1]

D. B. Bacani, H. Tahara, Existence and uniqueness theore m for a class of singular nonlinear partial diﬀerential equations, Publ. Res. Inst. Math. Sci., 48 (2012), no. 4, 899-917

work page 2012

[2] [2]

D. B. Bacani, H. Tahara, Unique solvability of some nonli near partial diﬀerential equations with Fuchsian and irregular singular ities, J. Math. Soc. Japan, 66 (2014), no. 3, 1017-1042. 23

work page 2014

[3] [11]

J. E. C. Lope, M. P. Roque, H. Tahara, On the unique solvab ility of cer- tain nonlinear singular partial diﬀerential equations, Z. A nal. Anwend., 31 (2012), no. 3, 291-305

work page 2012

[4] [3]

M. S. Baouendi, C. Goulaouic, Singular nonlinear Cauchy problems, J. Diﬀerential Equations, 22 (1976), no. 2, 268-291

work page 1976

[5] [4]

Briot, J

Ch. Briot, J. Cl. Bouquet, Recherches sur les propri´ et´ es des fonctions d´ eﬁnies par des ´ equations diﬀ´ erentielles, J. Ecole Polytech. 21 (1856), 133-197

work page

[6] [5]

H. Chen, H. Tahara, On totally characteristic type non-l inear partial diﬀerential equations in the complex domain, Publ. Res. Inst . Math. Sci., 35 (1999), no. 4, 621-636

work page 1999

[7] [6]

H. Chen, Z. Luo, C. Zhang, On the summability of formal sol utions for a class of nonlinear singular PDEs with irregular singul arity. Recent progress on some problems in several complex variables and p artial diﬀerential equations, 53-64, Contemp. Math., 400, Amer. Ma th. Soc., Providence, RI, 2006

work page 2006

[8] [7]

E. A. Coddington, N. Levinson, Theory of ordinary diﬀeren tial equa- tions, McGraw-Hill, 1955

work page 1955