Some properties of solutions to weakly hypoelliptic equations
Abstract.
A linear different operator is called weakly hypoelliptic if any local solution of is smooth. We allow for systems, that is, the coefficients may be matrices, not necessarily of square size. This is a huge class of important operators which cover all elliptic, overdetermined elliptic, subelliptic and parabolic equations.
We extend several classical theorems from complex analysis to solutions of any weakly hypoelliptic equation: the Montel theorem providing convergent subsequences, the Vitali theorem ensuring convergence of a given sequence and Riemann’s first removable singularity theorem. In the case of constant coefficients we show that Liouville’s theorem holds, any bounded solution must be constant and any solution must vanish.
Key words and phrases:
hypoelliptic operators, weakly hypoelliptic operators, hypoelliptic estimate, Montel theorem, Vitali theorem, Liouville theorem, removable singularity2010 Mathematics Subject Classification:
Primary 35H10; Secondary 35B531. Introduction
Hypoelliptic partial differential equations form a huge class of linear PDEs many of which are very important in applications. This class contains all elliptic, overdetermined elliptic, subelliptic and parabolic equations. Recall that a linear differential operator is called hypoelliptic if any solution to is smooth wherever is smooth. The study of hypoelliptic operators was initiated by Hörmander and others, see e.g. [11, 12, 21, 25].
We generalize this class of operators even further by only demanding that any solution to be smooth. We call such operators weakly hypoelliptic. This is not to be confused with partially hypoelliptic operators as introduced by Gårding and Malgrange [9] nor with the almost hypoelliptic operators due to Elliott [6]. We show by example that the class of weakly hypoelliptic operators is strictly larger than that of hypoelliptic operators. The example of a weakly hypoelliptic but nonhypoelliptic operator that we give is defined on and is overdetermined elliptic on . It is of first order and its principal symbol vanishes at . Thus the class of weakly hypoelliptic operators allows for a certain degeneracy of the principal symbol on “small sets” and might be of interest for geometric applications.
Holomorphic functions are the solutions to the CauchyRiemann equations which are elliptic in the case of one variable and overdetermined elliptic in the case of several variables. In any case, they are characterized as solutions to certain hypoelliptic PDEs. We show that the solutions to any weakly hypoelliptic equation share some of the nice properties of holomorphic functions which are familiar from classical complex analysis.
Montel’s theorem says that a locally bounded sequence of holomorphic functions subconverges to a holomorphic function. This does not hold for realanalytic functions. For instance, the sequence is a uniformly bounded sequence of realanalytic functions on , but does not have a convergent subsequence, see e.g. [23, Ex. 1.4.34]. We show that even a slightly stronger version of the Montel theorem holds for solutions to any weakly hypoelliptic equation: Any locally bounded sequence subconverges in the topology to a solution (Theorem 4). The Vitali theorem for holomorphic functions has a similar generalization (Theorem 6). For hypoelliptic equations this has been known for many decades and has motivated the study of socalled Montel spaces in functional analysis.
In case the underlying domain is and the weakly hypoelliptic operator has constant coefficients and satisfies a weighted homogeneity condition, we show that the Liouville theorem holds: any bounded solution must be constant (Theorem 7) and any solution must be zero (Theorem 12). This applies to powers of the Laplace and Dirac operators but also to powers of the heat operator. In the proof we use a simple scaling argument and apply the general Montel theorem.
Finally, we generalize Riemann’s first removable singularity theorem and show that a solution to a weakly hypoelliptic equation can be extended across a submanifold of sufficiently high codimension provided the solution is locally bounded near (Theorem 14 and Corollary 15).
The general setup is such that we consider a linear differential operator acting on sections of vector bundles. So, locally describes a system of linear PDEs with smooth coefficients. These coefficients may be matrices of not necessarily square size. Readers who are not too fond of geometric terminology may simply replace “manifolds” by “open subsets of ” and “sections of vector bundles” by “vectorvalued functions”.
The classical proofs of these theorems for holomorphic functions are typically based on special properties of holomorphic functions such as Cauchy’s integral formula. Therefore they may create the misleading impression that these theorems are also very special for holomorphic functions. In the contrary, the above mentioned theorems remain true for all solutions of the largest class of linear PDEs where one could hope for them to hold. Moreover, as we will see, the proofs of the general statements are actually rather simple.
Acknowledgment. The author likes to thank Thomas Krainer and Elmar Schrohe for helpful discussions. Moreover, he thanks Sonderforschungsbereich 647 funded by Deutsche Forschungsgemeinschaft for financial support.
2. Weakly Hypoelliptic Operators
Let be an dimensional differentiable manifold equipped with a smooth positive density . Let and be real vector bundles. If the bundles are complex we simply forget the complex structure and consider them as real bundles. We denote the spaces of smooth sections by and , respectively. Let be a linear differential operator of order . The fact that smooth sections are mapped to smooth sections encodes the smoothness of the coefficients of in local coordinates. The operator restricts to a linear map where stands for compactly supported smooth sections.
Let and be the dual bundles. Given , there is a unique linear differential operator , the formally dual operator, characterized by
for all and such that is compact. Here denote the canonical pairing of and and of and .
We extend to an operator, again denoted by , mapping distributional sections to distributional sections, by
for all and , compare e.g. [3, Sec. 1.1.2]. Here we denote by the evaluation of the distribution on the test section .
The differential operator is called hypoelliptic if for any open subset and any such that is smooth we have that is smooth. Since we will be interested in solutions of only, we make the following definition: The differential operator is called weakly hypoelliptic if for any open subset any satisfying must be smooth. Hörmander’s work [11, Ch. III] (see also the proof of Theorem 2.1 in [24, p. 63]) shows that for operators with constant coefficients over hypoellipticity and weak hypoellipticity are equivalent, at least in the scalar case, i.e. if the coefficients of are scalars rather than matrices. It seems likely that this is also true if the coefficients are constant matrices. In general however, if the coefficients are variable, the class of weakly hypoelliptic operators is strictly larger than that of hypoelliptic operators.
Example 1.
Let , let be the trivial real line bundle and the trivial bundle. The operator is given by
Here are the usual Cartesian coordinates while denote polar coordinates, , . On the operator is overdetermined elliptic (see below) and hence hypoelliptic. But is not hypoelliptic on because is not smooth while is smooth.
We check that is weakly hypoelliptic. Regularity is an issue at the origin only. Let with where is an open disk in centered at the origin. Then is smooth on . From we see that on and shows that does not depend on . Hence on . Subtracting this smooth function we may w.l.o.g. assume that . In this case, is a linear combination of the delta function and its derivatives,
Fix and and choose a test function which coincides with the monomial on a neighborhood of the origin. Then we see
Thus for all and and therefore . To summarize, we have seen that is weakly hypoelliptic, but not hypoelliptic.
For any weakly hypoelliptic operator we denote the kernel of by .
For and any relatively compact measurable subset we define the norm of by
Here we have tacitly equipped and the tangent bundle with a Riemannian metric and a connection . These data induce fiberwise norms and connections on the bundles . Note that is a section of . Since is relatively compact, different choices of metric and connections yield equivalent norms.
If is a compact subset, then we denote by the set of all restrictions to of times continuously differentiable sections, defined on an open neighborhood of . Equipped with the norm , becomes a Banach space.
Similarly, we define the norm for by
Again, different choices of the metric and the volume density yield equivalent norms. The norm extends the norm to the space of all essentially bounded measurable sections. The starting point are the following hypoelliptic estimates (compare [20, p. 331, Prop. 2] for the hypoelliptic case):
Lemma 2 (Hypoelliptic estimates).
Let be weakly hypoelliptic. Then for any , for any compact subset and any open subset containing there is a constant such that
(1) 
for all .
Proof.
Let be the kernel of the continuous linear map . Hence is a closed subspace of and thus a Banach space with the norm . Since is weakly hypoelliptic we have . Thus we get the linear restriction map , .
This map is closed. Namely, let with respect to and with respect to . Then we also have with respect to and therefore which proves closedness.
Corollary 3.
Let be a weakly hypoelliptic operator over a compact manifold (without boundary). Then is finitedimensional.
Proof.
Since is compact we can take in Lemma 2. Thus the norm and the norm are equivalent on . By the ArzelàAscoli theorem the embedding is compact. Hence the identity map on is compact, thus is finitedimensional. ∎
A differential operator is called elliptic if the principal symbol is invertible for all nonzero covectors . Elliptic regularity theory implies that all elliptic operators are hypoelliptic, see [14, Thm. 11.1.10] or [19, Ch. III, §5]. The class of elliptic operators contains many examples of high importance for applications such as the Laplace and Dirac operator.
More generally, if the principal symbol is injective instead of bijective for all nonzero covectors , then one calls overdetermined elliptic. In this case is elliptic where denotes the formal adjoint of . Now if is smooth, so is and hence is smooth by elliptic regularity. Therefore overdetermined elliptic operators are hypoelliptic as well.
Another way to generalize elliptic operators within the class of hypoelliptic operators is to consider subelliptic operators with a loss of derivatives where . These operators can also be characterized by a condition on their principal symbol, see [16, Ch. XXVII] for details.
If is parabolic, e.g. if describes the heat equation on a Riemannian manifold, then parabolic regularity using anisotropic Sobolev spaces shows that is hypoelliptic, see e.g. [10, Sec. 6.4].
In contrast, hyperbolic differential operators, e.g. those which describe wave equations, are not hypoelliptic.
Table 1 is a (very incomplete) table of examples for hypoelliptic operators relevant for applications:
type  reference  

, CauchyRiemann  open subset of  trivial line bundle  trivial line bundle  holomorphic functions  elliptic  [7, Sec. I.5] 
complex manifold  holomorphic vector bundle  holomorphic sections  overdet. elliptic  [8, Sec. 0.5]  
, Laplacetype  Riemannian manifold  Riem. vector bundle with connection  harmonic sections  elliptic  [4, Sec. 2.1]  
, biLaplacian  Riemannian manifold  trivial line bundle  trivial line bundle  biharmonic functions  elliptic  
, Dirac  Riemannian spin manifold  spinor bundle  spinor bundle  harmonic spinors  elliptic  [19, Ch. II, §5] 
, heat operator  interval Riemannian manifold  trivial line bundle  trivial line bundle  parabolic  [10, Sec. 6.4]  
, subLaplacian  subRiemannian manifold (bracket generating)  trivial line bundle  trivial line bundle  subelliptic  [13, Thm. 1.1]  
, where complex vector fields (bracket cond.)  trivial line bundle  trivial line bundle  Sobolevsubell.  [18, Thm. C]  
Bismut’s hypo elliptic Laplacian  hypoelliptic  [2, Ch. 3]  
, Kolmogorov  trivial line bundle  trivial line bundle  hypoelliptic  [15, Sec. 22.2] 
Tab. 1 Examples of hypoelliptic differential operators
3. Convergence Results
Let . A family is called locally bounded if for each compact subset we have
Let . By Hölder’s inequality . For we have . Therefore local boundedness implies local boundedness whenever . In particular, local boundedness is the weakest of these boundedness conditions.
We say that a sequence in converges in the topology if the restriction to any compact subset converges in every norm. In other words, the sections and all their derivatives converge locally uniformly.
Theorem 4 (Generalized Montel Theorem).
Let be a weakly hypoelliptic operator. Then any locally bounded sequence has a subsequence which converges in the topology to some .
The experts will notice that this is a direct consequence of Lemma 2 because we have the following
Short proof..
By Lemma 2 the sequence is bounded with respect to the topology (in the sense of topological vector spaces). Since is known to be a Montel space [26, p. 148, Cor. 2] so is the closed subspace (when equipped with the topology). This means that bounded closed subsets are compact, hence has a convergent subsequence. ∎
For those unfamiliar with the theory of Montel spaces we can also provide the following
Elementary proof..
Let be a compact subset. We choose an open, relatively compact subset containing . We fix . Since, by Lemma 2,
the sequence is bounded in the norm. By the ArzelàAscoli theorem there is a subsequence which converges in the norm over . The diagonal argument yields a subsequence which converges in all norms over , .
Now we exhaust by compact sets . We have seen that over each we can pass to a subsequence converging in all norms. Applying the diagonal argument once more, we find a subsequence converging in all norms over all . Thus we found a subsequence which converges in the topology to some .
Since is (sequentially) continuous with respect to the topology, we have , i.e. . ∎
The generalized Montel theorem applies to all examples listed in Table 1. Even in the case of holomorphic functions (in one or several variables) Theorem 4 is a slight improvement over the classical Montel theorem because the classical condition of local boundedness is replaced by the weaker condition of local boundedness. The standard proof of the classical Montel theorem uses the Cauchy integral formula to show equicontinuity and then applies the ArzelàAscoli theorem, see e.g. [23, Sec. 1.4.3].
In the case of harmonic functions on a Euclidean domain the Montel theorem is also classical. One can use estimates based on the Poisson kernel to show equicontinuity and then apply the ArzelàAscoli theorem [1, p. 35, Thm. 2.6].
The Montel theorem provides a criterion for the existence of a convergent subsequence. The next theorem provides a sufficient criterion which ensures that a given sequence converges itself.
Definition 5.
Let be a weakly hypoelliptic operator on . A subset is called a set of uniqueness for if for any the condition implies .
Every dense subset of is a set of uniqueness because .
For holomorphic function of one variable, i.e. , the set is a set of uniqueness if it has an accumulation point in .
Many (but not all) important elliptic operators have the socalled weak unique continuation property. This means that if has nonempty interior, then it is a set of uniqueness provided is connected. Laplace and Diractype operators belong to this class.
Theorem 6 (Generalized Vitali Theorem).
Let be a weakly hypoelliptic operator and let be a set of uniqueness for . Let be a locally bounded sequence. Suppose that the pointwise limit exists for all .
Then converges in the topology to some .
Proof.
By Theorem 4, every subsequence of has a subsequence for which the assertion holds. The limit functions for these subsequences are in and coincide on , hence they all agree. Hence has a unique accumulation point .
If the sequence itself did not converge to then we could extract a subsequence staying outside a neighborhood of . But this subsequence would again have a subsequence converging to , a contradiction. ∎
4. Liouville Property
We now concentrate on the case . All vector bundles over are trivial, so sections can be identified with functions . Hence the coefficients of the differential operator are matrices. If these matrices do not depend on the point we say that has constant coefficients. The Laplace operator , the Dirac operator and the heat operator with constant coefficients. are examples of hypoelliptic operators on
Let be a polynomial in real variables. Here is allowed to have matrixvalued coefficients of fixed size. Let with . We call weighted homogeneous with weight if for some and all , . The corresponding differential operator with constant coefficients
We can now state the following Liouville type theorem.
Theorem 7 (Generalized Liouville Theorem).
Let be a weakly hypoelliptic operator over . Suppose that has constant coefficients in matrices and is weighted homogeneous.
Then each bounded function in must be constant.
Proof.
Let be bounded. For we put . Since is bounded, the family is uniformly bounded. Moreover, so that . By Theorem 4, there is a sequence such that converges locally uniformly to some . We observe and hence .
Fix . For we put . Then and as . Locally uniform convergence yields , hence , so is constant. ∎
Example 8.
We directly recover the classical Liouville theorems for holomorphic and for harmonic functions. In the case of bounded harmonic functions Nelson gave a particularly short proof based on the mean value property [22]. In fact, for harmonic functions it suffices to assume that they are bounded from below (or from above) [1, Thm. 3.1]. This cannot be deduced from Theorem 7 but the theorem also applies to biharmonic functions on or to solutions of higher powers of . The function is biharmonic, bounded from below and nonconstant. Hence unlike for harmonic functions we need to assume boundedness from above and from below to conclude that a biharmonic function is constant.
Similarly, bounded harmonic spinors on must be constant.
Remark 9.
Here is a silly argument why all bounded polynomials on must be constant. Given such a polynomial choose larger than half the degree of . Then and Theorem 7 applies.
Example 10.
Theorem 7 also applies to the heat operator. Bounded solutions to the heat equation on must be constant. Note that there do exist nontrivial solutions on which vanish for [17, pp. 211213]. They are unbounded on for each however.
Moreover, Theorem 7 applies to powers of the heat operator. So, for instance, bounded solutions to
must be constant.
Remark 11.
Theorem 12 (Generalized Liouville Theorem, version).
Let . Let be a weakly hypoelliptic operator over . Suppose that has constant coefficients in matrices and is weighted homogeneous. Then
Proof.
Remark 13.
In the case of scalar constant coefficient hypoelliptic operators Theorems 7 and 12 can also be seen as follows: If the polynomial had a zero , then would vanish along the curve by homogeneity. This would violate Hörmander’s hypoellipticity criterion [11, Thm. 3.3.I] for . Thus is the only zero of . Now [5, Thm. 2.28] (whose proof is a simple application of the Fourier transform) says that any solution must be a polynomial. If it is in or in it must be constant or vanish, respectively.
5. Removable Singularities
Let be a Riemannian manifold and denote the Riemannian distance of by . For a subset let . For we denote by
the closed neighborhood of with removed.
Theorem 14.
Let be an embedded submanifold of codimension . Let be a weakly hypoelliptic operator of order over . Let . Suppose that for each compact subset there exists a constant such that for all sufficiently small
Then extends uniquely to some .
Proof.
Uniqueness of the extension is clear because is dense in . To show existence let be a smooth function such that

on ;

on ;

everywhere.
For We define by
Given a compact subset the function is smooth in a neighborhood of provided is small enough. This is true because the function is smooth on an open neighborhood of with removed.
We extend to a distribution : Let be a test section. The compact support of is denoted by . For sufficiently small . We put
The limit exists because for
We check that is a distribution. Fix a compact subset . Then, choosing sufficiently small, we obtain
Here the first summand is finite because of the assumption in the theorem and the second because is a compact subset of . Hence we find for all with :
so is continuous in .
It remains to show that solves in the distributional sense. For we compute
where is a linear differential operator of order for each fixed . It is obtained from the general Leibniz rule. The coefficients of depend linearly on and its derivatives up to order . Since is constant outside the coefficients of are supported in . For this reason the integration by parts above is justified; there are no boundary terms. We find
Hence , i.e., in the distributional sense. By weak hypoellipticity of the extension must be smooth and solves in the classical sense. ∎
Corollary 15.
Let be a weakly hypoelliptic operator of order over . Let be an embedded submanifold of codimension . Let be locally bounded near .
Then extends uniquely to some .
Proof.
Since is locally bounded near we have for any compact subset
as . Since we get and therefore as . Theorem 14 yields the claim. ∎
Example 16.
Let be an open subset and a complex submanifold of complex codimension . Then any holomorphic function on , locally bounded near , extends uniquely to a holomorphic function on . This is Corollary 15 with and . It is classically known as Riemann’s first removable singularity theorem [23, Thm. 4.2.1].
Note that being a complex submanifold is actually irrelevant; any real submanifold of real codimension will do. Moreover, by Theorem 14 one can relax the condition that be locally bounded near . A local estimate of the form with is sufficient. This criterion is sharp because for , , and we have a solution of on which satisfies but does not extend across .
References
 [1] S. Axler, P. Bourdon and W. Ramey: Harmonic Function Theory, 2. ed., SpringerVerlag, New York, 2001
 [2] J.M. Bismut: The hypoelliptic Laplacian on the cotangent bundle, J. Amer. Math. Soc. 18 (2005), 379–476
 [3] C. Bär, N. Ginoux and F. Pfäffle: Wave equations on Lorentzian manifolds and quantization, ESI Lectures in Mathematics and Physics. Europ. Math. Soc., Zürich, 2007
 [4] N. Berline, E. Getzler and M. Vergne: Heat kernels and Dirac operators, Grundlehren der Mathematischen Wissenschaften, 298. SpringerVerlag, Berlin, 1992
 [5] Y. V. Egorov and M. A. Shubin: Linear partial differential equations. Foundations of the classical theory, Encyclopaedia Math. Sci., 30, SpringerVerlag, Berlin, 1992
 [6] R. Elliott: Almost hypoelliptic differential operators, Proc. London Math. Soc. 19 (1969), 537–552
 [7] E. Freitag and R. Busam: Complex analysis, Universitext. SpringerVerlag, Berlin, 2005
 [8] P. Griffiths and J. Harris: Principles of algebraic geometry, Wiley Classics Library. John Wiley & Sons, New York, 1994
 [9] L. Gårding and B. Malgrange: Opérateurs différentiels partiellement hypoelliptiques et partiellement elliptiques, Math. Scandinavica 9 (1961), 5–21
 [10] A. Grigor’yan: Heat kernel and analysis on manifolds, AMS/IP Studies in Advanced Mathematics, 47. Amer. Math. Soc., International Press, Boston, 2009
 [11] L. Hörmander: On the theory of general partial differential operators, Acta Math. 94 (1955), 161–248
 [12] L. Hörmander: Hypoelliptic differential operators, Ann. Inst. Fourier 11 (1961), 477–492
 [13] L. Hörmander: Hypoelliptic second order differential equations, Acta Math. 119 (1967), 147–171
 [14] L. Hörmander: The analysis of linear partial differential operators II. Differential operators with constant coefficients, Grundlehren der mathematischen Wissenschaften, 257. SpringerVerlag, Berlin, 1983
 [15] L. Hörmander: The analysis of linear partial differential operators III. Pseudodifferential operators, reprint of the 1994 edition. Classics in Mathematics. SpringerVerlag, Berlin, 2007
 [16] L. Hörmander: The analysis of linear partial differential operators IV. Fourier integral operators, Grundlehren der mathematischen Wissenschaften, 275. SpringerVerlag, Berlin, 1985
 [17] F. John: Partial differential equations, Fourth edition. Applied Mathematical Sciences, 1. SpringerVerlag, New York, 1982
 [18] J. J. Kohn: Hypoellipticity and loss of derivatives, Ann. of Math. (2) 162 (2005), 943–986
 [19] H. B. Lawson and M.L. Michelsohn: Spin geometry, Princeton Mathematical Series, 38. Princeton University Press, Princeton, 1989
 [20] B. Malgrange: Existence et approximation des solutions des équations aux dérivées partielles et des équations de convolution, Annales de l’institut Fourier 6 (1956), 271–355
 [21] B. Malgrange: Sur une classe d’opérateurs différentiels hypoelliptiques, Bull. Soc. Math. France 85 (1957), 283–306
 [22] E. Nelson: A proof of Liouville’s Theorem, Proc. of the Amer. Math. Soc. 12 (1961), 995
 [23] V. Scheidemann: Introduction to Complex Analysis in Several Variables, Birkhäuser, Basel, 2005
 [24] M. Taylor: Pseudodifferential Operators, Princeton Univ. Press, Princeton, 1981
 [25] F. Trèves: Opérateurs différentiels hypoelliptiques, Ann. Inst. Fourier 9 (1959), 1–73
 [26] F. Trèves: Topological vector spaces, distributions and kernels, Unabridged republication of the 1967 original, Dover Publications, Mineola, 2006