Textbook Mathematical Analysis and Applications Plenary Lectures Isaac 2017 Vaxjo Sweden Luigi G Rodino Ebook All Chapter PDF

Mathematical Analysis and
Applications Plenary Lectures ISAAC

2017 Växjö Sweden Luigi G. Rodino
Visit to download the full and correct content document:
https://textbookfull.com/product/mathematical-analysis-and-applications-plenary-lectu
res-isaac-2017-vaxjo-sweden-luigi-g-rodino/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...
Analysis Probability Applications and Computation

Proceedings of the 11th ISAAC Congress Växjö Sweden
2017 Karl■Olof Lindahl
https://textbookfull.com/product/analysis-probability-
applications-and-computation-proceedings-of-the-11th-isaac-
congress-vaxjo-sweden-2017-karl%e2%80%90olof-lindahl/
Mathematical Analysis and Applications 1st Edition

Themistocles M. Rassias
https://textbookfull.com/product/mathematical-analysis-and-
applications-1st-edition-themistocles-m-rassias/
Language and Automata Theory and Applications 11th

International Conference LATA 2017 Umeå Sweden March 6
9 2017 Proceedings 1st Edition Frank Drewes
https://textbookfull.com/product/language-and-automata-theory-
and-applications-11th-international-conference-lata-2017-umea-
sweden-march-6-9-2017-proceedings-1st-edition-frank-drewes/
Mathematical Analysis and Applications Selected Topics

1st Edition Michael Ruzhansky
https://textbookfull.com/product/mathematical-analysis-and-
applications-selected-topics-1st-edition-michael-ruzhansky/
Optimal Mean Reversion Trading Mathematical Analysis
and Practical Applications Tim Leung
https://textbookfull.com/product/optimal-mean-reversion-trading-
mathematical-analysis-and-practical-applications-tim-leung/
Current Trends in Mathematical Analysis and Its

Interdisciplinary Applications Hemen Dutta
https://textbookfull.com/product/current-trends-in-mathematical-
analysis-and-its-interdisciplinary-applications-hemen-dutta/
Sustainable Lignin for Carbon Fibers Principles

Techniques and Applications Emmanuel Isaac Akpan
https://textbookfull.com/product/sustainable-lignin-for-carbon-
fibers-principles-techniques-and-applications-emmanuel-isaac-
akpan/
Mathematical Morphology and Its Applications to Signal

and Image Processing 13th International Symposium ISMM
2017 Fontainebleau France May 15 17 2017 Proceedings
1st Edition Jesús Angulo
https://textbookfull.com/product/mathematical-morphology-and-its-
applications-to-signal-and-image-processing-13th-international-
symposium-ismm-2017-fontainebleau-france-
may-15-17-2017-proceedings-1st-edition-jesus-angulo/
An Introduction to Proofs and the Mathematical

Vernacular Martin V. Day
https://textbookfull.com/product/an-introduction-to-proofs-and-
the-mathematical-vernacular-martin-v-day/
Springer Proceedings in Mathematics & Statistics
Luigi G. Rodino
Joachim Toft Editors
Mathematical
Analysis and
Applications—
Plenary Lectures
ISAAC 2017, Växjö, Sweden
Volume 262
This book series features volumes composed of selected contributions from
workshops and conferences in all areas of current research in mathematics and
statistics, including operation research and optimization. In addition to an overall
evaluation of the interest, scientific quality, and timeliness of each proposal at the
hands of the publisher, individual contributions are all refereed to the high quality
standards of leading journals in the field. Thus, this series provides the research
community with well-edited, authoritative reports on developments in the most
exciting areas of mathematical and statistical research today.
More information about this series at http://www.springer.com/series/10533

Luigi G. Rodino Joachim Toft
•
Editors
Mathematical Analysis
and Applications—Plenary
Lectures
ISAAC 2017, Växjö, Sweden
123
Editors
Luigi G. Rodino Joachim Toft
Dipartimento di Matematica Department of Mathematics
Università di Torino Linnaeus University
Turin, Italy Växjö, Sweden
ISSN 2194-1009 ISSN 2194-1017 (electronic)

ISBN 978-3-030-00873-4 ISBN 978-3-030-00874-1 (eBook)
https://doi.org/10.1007/978-3-030-00874-1
Library of Congress Control Number: 2018957643
Mathematics Subject Classification (2010): 35-XX, 46-XX, 60-XX, 32-XX, 47-XX, 65-XX
© Springer Nature Switzerland AG 2018

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made. The publisher remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.
This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface
This volume is a collection of articles devoted to current research topics in

Mathematical Analysis, with emphasis on Fourier analysis and general theory of
partial differential equations. It originates from plenary lectures given at the 11th
International ISAAC Congress, held during 14–18 August 2017 at the University of
Växjö, Sweden.
The papers are authored by six eminent specialists and aim at presenting to a
large audience some challenging and attractive themes of the modern Mathematical
Analysis, namely as follows:
– The contribution of Nils Dencker is devoted to the local solvability of sub-
principal type operators. Dencker proved in 2006 the so-called Nirenberg–
Treves conjecture, expressing a necessary and sufficient condition for the
solvability of the operators of principal type. In this paper, he addresses to
operators with multiple characteristics. Namely, the principal symbol of the
operator is assumed to vanish of at least second order on an involutive manifold.
Precise necessary conditions are expressed for solvability, in terms of the sub-
principal symbol, involving lower order terms.
– In these last years, the fractional-order Laplacian was frequently used in
Mathematical Physics, Differential Geometry, Probability and Finance. The
paper of Gerd Grubb is devoted to this operator and its generalizations, with
attention to homogeneous and non-homogeneous boundary value problems and
corresponding heat equations. The paper represents the first general and rigorous
approach to non-local problems of this type, in the setting of pseudo-differential
operators.
– The paper of Abdelhamid Meziani is devoted to degenerate complex vector
fields in the plane, basic example being the Mizohata operator. The corre-
sponding equations share many properties with the Cauchy–Riemann equation.
Combining with the theory of the hypo-analytic structures of Treves, Meziani
presents an elegant treatment of the subject, including generalizations of the
Riemann–Hilbert problem.
v
vi Preface
– The paper of Alberto Parmeggiani gives a survey on the problem of the lower
bounds. Does the positivity of the symbol imply the positivity of the corre-
sponding operator? A precise answer to this question is largely open. Main
reference is the classical theorem of Fefferman–Phong, stating positivity of the
operator modulo small errors. Here Parmeggiani addresses also to the case of
systems of operators; for them, precise lower bounds represent a largely
unexplored area.
– Following the original idea of Morrey [1], Peetre and others introduced the
function spaces which are nowadays called Morrey spaces. They have a large
number of applications in different contexts, and their definition was extended in
different directions. Yoshihiro Sawano, in his contribution to the volume, pre-
sents a review addressed to non-experts, including the interpolation theory and
the weighted version.
– Localization operators were defined by Berezin [2] in the frame of Quantum
Mechanics, and later used by Daubechies [3] and others in Signal Theory. The
paper of Nenad Teofanov is devoted to this important topic. The definition of
localization operator is given here in terms of the Grossmann–Royer transform,
which simplifies the proof of several known results. Particular emphasis is given
to the action on Gelfand–Shilov and modulation spaces.
Besides plenary talks, about 250 scientific communications were delivered during
the Växjö ISAAC Congress. Their texts are published in an independent volume.
On the whole, the Congress demonstrated, in particular, the relevant role of the
Nordic European countries in several research areas of Mathematical Analysis.
Turin, Italy Luigi G. Rodino

Växjö, Sweden Joachim Toft
July 2018
References
1. Morrey Jr., C.B.: On the solutions of quasi-linear elliptic partial differential equations. Trans.
Am. Math. Soc. 43(1), 126–166 (1938)
2. Berezin, F.A.: Wick and anti-wick symbols of operators. Mat. Sb. (N.S.) 86(128), 578–610
(1971)
3. Daubechies, I.: Time-frequency localization operators: a geometric phase space approach. IEEE
Trans. Inform. Theory 34(4), 605–612 (1988)
Contents
Solvability of Subprincipal Type Operators . . . . . . . . . . . . . . . . . . . . . . 1

Nils Dencker
Fractional-Order Operators: Boundary Problems,
Heat Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
Gerd Grubb
A Class of Planar Hypocomplex Vector Fields: Solvability
and Boundary Value Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
A. Meziani
Almost-Positivity Estimates of Pseudodifferential Operators . . . . . . . . . 109
Alberto Parmeggiani
Morrey Spaces from Various Points of View . . . . . . . . . . . . . . . . . . . . . 139
Yoshihiro Sawano
The Grossmann–Royer Transform, Gelfand–Shilov Spaces,
and Continuity Properties of Localization Operators
on Modulation Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
Nenad Teofanov
vii
Contributors
Nils Dencker Centre for Mathematical Sciences, Lund University, Lund, Sweden
Gerd Grubb Department of Mathematical Sciences, Copenhagen University,
Copenhagen, Denmark
A. Meziani Department of Mathematics, Florida International University, Miami,
FL, USA
Alberto Parmeggiani Department of Mathematics, University of Bologna,
Bologna, Italy
Yoshihiro Sawano Department of Mathematics, Tokyo Metropolitan University,
Hachioji Tokyo, Japan
Nenad Teofanov Department of Mathematics and Informatics, University of Novi
Sad, Novi Sad, Serbia
ix
Solvability of Subprincipal Type
Operators
Nils Dencker
Abstract In this paper we consider the solvability of pseudodifferential operators in

the case when the principal symbol vanishes of order k ≥ 2 at a nonradial involutive
manifold 2 . We shall assume that the operator is of subprincipal type, which means
that the kth inhomogeneous blowup at 2 of the refined principal symbol is of
principal type with Hamilton vector field parallel to the base 2 , but transversal to
the symplectic leaves of 2 at the characteristics. When k = ∞ this blowup reduces
to the subprincipal symbol. We also assume that the blowup is essentially constant
on the leaves of 2 , and does not satisfying the Nirenberg–Treves condition ().
We also have conditions on the vanishing of the normal gradient and the Hessian
of the blowup at the characteristics. Under these conditions, we show that P is not
solvable.
1 Introduction
We shall consider the solvability for a classical pseudodifferential operator P ∈

clm (X ) on a C ∞ manifold X of dimension n. This means that P has an expansion
j
pm + pm−1 + . . . where p j ∈ Shom is homogeneous of degree j, ∀ j, and pm = σ(P)
is the principal symbol of the operator. A pseudodifferential operator is said to be
of principal type if the Hamilton vector field H pm of the principal symbol does not
have the radial direction ξ · ∂ξ on pm−1 (0), in particular H pm = 0. We shall consider
the case when the principal symbol vanishes of at least second order at an involutive
manifold 2 , thus P is not of principal type.
P is locally solvable at a compact set K ⊆ X if the equation
Pu = v (1.1)
N. Dencker (B)
Centre for Mathematical Sciences, Lund University, Box 118, 221 00 Lund, Sweden
e-mail: dencker@maths.lth.se
© Springer Nature Switzerland AG 2018 1

L. G. Rodino and J. Toft (eds.), Mathematical Analysis and Applications—Plenary
Lectures, Springer Proceedings in Mathematics & Statistics 262,
https://doi.org/10.1007/978-3-030-00874-1_1
2 N. Dencker
has a local solution u ∈ D (X ) in a neighborhood of K for any v ∈ C ∞ (X ) in a set of

finite codimension. We can also define microlocal solvability of P at any compactly
based cone K ⊂ T ∗ X , see Definition 2.6.
For pseudodifferential operators of principal type, local solvability is equivalent
to condition () on the principal symbol, see [3, 12]. This condition means that
Im apm does not change sign from − to +

along the oriented bicharacteristics of Re apm (1.2)
for any 0 = a ∈ C ∞ (T ∗ X ). The oriented bicharacteristics are the positive flow of

the Hamilton vector field HRe apm = 0 on which Re apm = 0, these are also called
semibicharacteristics of pm . Condition (1.2) is invariant under multiplication of pm
with nonvanishing factors, and symplectic changes of variables, thus it is invariant
under conjugation of P with elliptic Fourier integral operators. Observe that the sign
changes in (1.2) are reversed when taking adjoints, and that it suffices to check (1.2)
for some a = 0 for which HRe ap = 0 according to [13, Theorem 26.4.12].
For operators which are not of principal type, the situation is more complicated
and the solvability may depend on the lower order terms. Then the refined principal
symbol
i
psub = pm + pm−1 + ∂ξ j ∂ x j p m (1.3)
2 j
is invariantly defined modulo S m−2 under changes of coordinates, see Theorem

18.1.33 in [13]. In the Weyl quantization the refined principal symbol is given by
pm + pm−1 .
When 2 is not involutive, there are examples where the operator is solvable for
any lower order terms. For example when P is effectively hyperbolic, then even the
Cauchy problem is solvable for any lower order term, see [10, 14, 18]. There are also
results in the cases when the principal symbol is a product of principal type symbols
not satisfying condition (), see [1, 8, 9, 20, 25].
In the case where the principal symbol is real and vanishes of at least second
order at the involutive manifold there are several results, mostly in the case when the
principal symbol is a product of real symbols of principal type. Then the operator is
not solvable if the imaginary part of the subprincipal symbol has a sign change of
finite order on a bicharacteristic of one the factors of the principal symbol, see [7,
19, 22, 23].
This necessary condition for solvability has been extended to some cases when
the principal symbol is real and vanishes of second order at the involutive manifold.
The conditions for solvability then involve the sign changes of the imaginary part of
the subprincipal symbol on the limits of bicharacteristics from outside the manifold,
thus on the leaves of the symplectic foliation of the manifold, see [15–17, 26].
This has been extended to more general limit bicharacteristics of real principal
symbols in [4]. There we assumed that the bicharacteristics converge in C ∞ to a
limit bicharacteristic. We also assumed that the linearization of the Hamilton vector
Solvability of Subprincipal Type Operators 3
field is tangent to and has uniform bounds on the tangent spaces of some Lagrangean
manifolds at the bicharacteristics. Then P is not solvable if condition Lim() is not
satisfied on the limit bicharacteristics. This means that the quotient of the imaginary
part of the subprincipal symbol with the norm of the Hamilton vector field switches
sign from − to + on the bicharacteristics and becomes unbounded when converging
to the limit bicharacteristic. This was generalized in [6] to operators with complex
principal symbols. There we assumed that the normalized complex Hamilton vector
field of the principal symbol converges to a real vector field. Then the limit bichar-
acteristics are uniquely defined, and one can invariantly define the imaginary part of
the subprincipal symbol. Thus condition Lim() is well defined and we proved that
it is necessary for solvability.
In [5] we considered the case when the principal symbol (not necessarily real
valued) vanishes of at least second order at a nonradial involutive manifold 2 .
We assumed that the operator was of subprincipal type, i.e., that the subprincipal
symbol on 2 is of principal type with Hamilton vector field tangent to 2 at the
characteristics, but transversal to the symplectic leaves of 2 . Then we showed that
the operator is not solvable if the subprincipal symbol is essentially constant on the
symplectic leaves of 2 and does not satisfy condition (), which we call Sub().
In the case when the sign change is of infinite order, we also had conditions on the
vanishing of both the Hessian of the principal symbol and the complex part of the
gradient of the subprincipal symbol.
The difference between [5, 6] is that in the first case the Hamilton vector field
of the principal symbol dominates, and in the second the Hamilton vector field of
the subprincipal symbol dominates. In this paper, we shall study the case when
condition () is not satisfied for the refined principal symbol (1.3) which combines
both the principal and subprincipal symbols. We shall assume that the principal
symbol vanishes of at least order k ≥ 2 on at a nonradial involutive manifold 2 .
When k < ∞ then the kth jet of the principal symbol is well defined at 2 , but since
the refined principal symbol is inhomogeneous we make an inhomogeneous blowup,
called reduced subprincipal symbol by Definition 2.1. We assume that the operator
is of subprincipal type, i.e., the reduced subprincipal symbol is of principal type, see
Definition 2.2. We define condition Subk (), which is condition () on the reduced
subprincipal symbol, see Definition 2.3. We assume that the blowup of the refined
principal symbol is essentially constant on the symplectic leaves of 2 , see (2.27).
We also have conditions on the rate of the vanishing of the normal gradient (2.19) and
when k = 2 of the Hessian of the reduced subprincipal symbol (2.21). When k = ∞
all the Taylor terms vanish and condition Sub∞ () reduces to condition Sub()
on 2 from [5]. Under these conditions, we show that if condition Subk () is not
satisfied near a bicharacteristic of the reduced subprincipal symbol then the operator
is not solvable near the bicharacteristic, see Theorem 2.1 which is the main result of
the paper. In the case when the sign change of Subk () is on 2 we get a different
result than in [5], since now we localize the pseudomodes with the the phase function
instead of the amplitude.
The plan of the paper is as follows. In Sect. 2 we make the definitions of the
symbols we are going to use, state the conditions and the main result, Theorem 2.1.
4 N. Dencker
In Sect. 3 we present some examples, and in Sect. 4 we develop normal forms of the
operators, which are different in the case when the principal symbol vanishes of finite
or infinite order at 2 . The approximate solutions, or pseudomodes, are defined in
Sect. 5. In Sect. 6 we solve the eikonal equation in the case when the principal symbol
vanishes of finite order, in Sect. 7 we solve it in the case when the bicharacteristics are
on 2 and in Sect. 8 we solve the transport equations. In order to solve the eikonal and
transport equations uniformly we use the estimates of Lemma 6.1, which is proved
in Sect. 9. Finally, Theorem 2.1 is proved in Sect. 10.
2 Statement of Results
Let σ(P) = p ∈ Shom

m
be the homogeneous principal symbol of P, we shall assume
that
σ(P) vanishes of at least second order at 2 ⊂ T ∗ X \ 0 (2.1)
where
2 is a nonradial involutive manifold of codimension d (2.2)
where 0 < d < n − 1 with n = dim X . Here nonradial means that the radial direction
ξ, ∂ξ is not in the span of the Hamilton vector fields of the manifold, i.e., not equal
to H f on 2 for any f ∈ C 1 vanishing at 2 . Then by a change of local homogeneous
symplectic coordinates we may assume that locally
2 = { η = 0 } (ξ, η) ∈ Rn−d × Rd ξ = 0 (2.3)
for some 0 < d < n − 1, which can be achieved by a conjugation with elliptic Fourier
integral operators.
Now, since p vanishes of at least second order at 2 we can define the order of
p as
2 ≤ κ(w) = min { |α| : ∂ α p(w) = 0 } w ∈ 2 (2.4)
and κ(ω) = minw∈ω κ(w) for ω ⊆ 2 , which is equal to ∞ when p vanishes of

infinite order. This is an upper semicontinuous function on 2 , but since κ(w) is has
values in N ∪ ∞, it attains its minimum κ(ω) on any set ω ⊆ 2 .
If P is of principal type near 2 then, since solvability is an open property, we
find that a necessary condition for P to be solvable at 2 is that condition () for the
principal symbol is satisfied in some neighborhood of 2 . Naturally, this condition
is empty on 2 where we instead have conditions on the refined principal symbol:
i
psub = p + pm−1 + ∂ x j ∂ξ j p (2.5)
2 j
(for the Weyl quantization, the refined principal symbol is given by p + pm−1 ). The
refined principal symbol is invariantly defined as a function on T ∗ X modulo S m−2
under conjugation with elliptic Fourier integral operators, see [13, Theorem 18.1.33]
and [11, Theorem 9.1]. (The latter result is for the Weyl quantization, but the result
easily carries over to the Kohn–Nirenberg quantization for classical operators.) The
subprincipal symbol
i
ps = pm−1 + ∂ x j ∂ξ j p (2.6)
2 j
is invariantly defined on 2 under conjugation with elliptic Fourier integral operators.

Remark 2.1 When 2 = ξ1 = ξ2 = · · · = ξ j = 0 is involutive, the refined prin-
cipal symbol is equal to ps = pm−1 at 2 .
In fact, this follows since ∂ξ p ≡ 0 on 2 . When composing P with an elliptic
pseudodifferential operator C, the value of the refined principal symbol of C P is
equal to cpsub + 2i H p c which is equal to cps at 2 , where c = σ(C). Observe that
the refined principal symbol is complexly conjugated when taking the adjoint of the
operator, see [13, Theorem 18.1.34].
The conormal bundle N ∗ 2 ⊂ T ∗ (T ∗ X ) of 2 is the dual of the normal bundle
T2 T ∗ X/T 2 . The conormal bundle can be parametrized by first choosing local
homogeneous symplectic coordinates so that 2 is given by { η = 0 }. Then the fiber
of N ∗ 2 can be parametrized by η ∈ Rd , d = Codim 2 , so that N ∗ 2 ∼ = 2 × R d
and different parametrizations gives linear transformations on the fiber.
We define the kth jet Jwk ( f ) of a C ∞ function f at w ∈ 2 as the equivalence
class of f modulo functions vanishing of order k + 1 at w. If k = κ(ω) < ∞ is given
by (2.4) for the open neighborhood ω ⊂ 2 then for w = (x, y, ξ, 0) ∈ 2 we find
that Jwk ( p) is a well defined homogeneous function on N ∗ 2 given by
Nw∗ 2 (w, η) → Jwk ( p)(η) = ∂ηk p(w)(η) (2.7)
since ∂ j p ≡ 0 on ω, j < k. Here ∂ηk p(w) is the k-form given by the Taylor term of
order k of p. If κ(ω) = ∞ then of course any jet of p vanishes identically on ω. Here
and in the following, the η variables will be treated as parameters.
Definition 2.1 When k = κ(ω) < ∞ for some open set ω ⊂ 2 we define the
reduced subprincipal symbol by
N ∗ 2 (w, η) → ps,k (w, η) = Jwk ( p)(η) + ps (w) w∈ω (2.8)
which is a polynomial in η of degree k and is given by the blowup of the refined

principal symbol at 2 see Remark 2.4. If κ(ω) = ∞ then we define ps,∞ = ps 2
so we have (2.8) for any k.
6 N. Dencker
Remark 2.2 The reduced subprincipal symbol is well-defined up to nonvanishing

factors under conjugation with elliptic homogeneous Fourier integral operators and
under composition with classical elliptic pseudodifferential operator.
In fact, the reduced subprincipal symbol is equal to the refined principal symbol
modulo terms homogeneous of degree m vanishing at 2 of order k + 1 and terms
homogeneous of degree m − 1 vanishing at 2 . When composing with an elliptic
pseudodifferential operator, both the terms in the refined subprincipal symbol gets
multiplied with the same nonvanishing factor, and the terms proportional to ∂ p vanish

on 2 . Observe that if we multiply psub with c then ps,k gets multiplied with c2 .
Since ps is only defined on 2 , the Hamilton field H ps,k is only well defined
modulo terms that are tangent to the symplectic leaves of 2 , which are spanned by
the Hamilton vector fields of functions vanishing on 2 . Therefore, we shall assume
that the reduced principal symbol essentially is constant on the leaves of 2 for fixed
η by assuming that

dps,k ≤ C0 | ps,k | at ω when |η − η0 | 1 (2.9)
TL
for any leaf L of 2 where ω ⊂ 2 . Since ps,k is determined by the Taylor coefficients
of the refined principal symbol at 2 we find that (2.9) is determined on 2 . When
η = 0 we get condition (2.9) on ps at 2 which was used in [5]. Condition (2.9)
is invariant under multiplication with nonvanishing factors and when dps,k = 0 on
−1
ps,k (0) it is equivalent to the fact that ps,k is constant on the leaves up to nonvanishing
factors by the following lemma.
Lemma 2.1 Assume that f (x, y, ζ) ∈ C ∞ is a polynomial in ζ of degree m for

∞
(x, y) ∈ , such that ∂x f = 0 when
f = 0. Assume that is an open bounded C
domain such that x0 = { x = x0 } is simply connected for all x0 .
Let = { (x, ζ) : ∃ y (x, y) ∈ and |ζ − ζ0 | < c } be the projection on the (x, ζ)
variables and assume that there exist y0 (x) ∈ C ∞ such that (x, y0 (x), ζ) ∈ ×
{ |ζ − ζ0 | < c }, ∀(x, ζ) ∈ . Then
|∂ y f | ≤ C0 | f | in when |ζ − ζ0 | < c (2.10)
for c > 0 implies that
f (x, y, ζ) = c(x, y, ζ) f 0 (x, ζ) in when |ζ − ζ0 | < c (2.11)
where 0 < c0 ≤ c(x, y, ζ) ∈ C ∞ and f 0 (x, ζ) = f (x, y0 (x), ζ) ∈ C ∞ , which

implies that ∂ yα f ≤ Cα | f |, ∀ α, in when |ζ − ζ0 | < c.
If ∂w ps,k = 0 when ps,k = 0 and ps,k satisfies (2.9), then we find from Lemma 2.1
after possibly shrinking ω that ps,k is constant on the leaves of 2 in ω when |η −
η0 | < c0 after multiplication with a nonvanishing factor.
Proof Let 0 = { (x, ζ) ∈ : f (x, y0 (x), ζ) = 0 }. We shall first prove the result
when (x, ζ) ∈ \ 0 . Then f = 0 at (x, y0 (x), ζ) and (2.10) gives that ∂ y log f
is uniformly bounded near (x, y0 (x), ζ), where log f is a branch of the complex
logarithm. Thus, by integrating with respect to y starting at y = y0 (x) in the simply
connected x × { |ζ − ζ0 | < c } we find that log f (x, y, ζ) − log f (x, y0 (x), ζ) ∈
C ∞ is bounded and by exponentiating we obtain
f (x, y, ζ) = c(x, y, ζ) f 0 (x, ζ) in x for |ζ − ζ0 | < c (2.12)
when (x, ζ) ∈ \ 0 . Here f 0 (x, ζ) = f (x, y0 (x), ζ) ∈ C ∞ and 0 < c0 ≤

c(x, y, ζ) ∈ C ∞ is uniformly bounded such that c(x, y0 (x), ζ) ≡ 1. This gives that
f −1 (0) is constant in y when (x, ζ) ∈
/ 0 .
Since ∂x f = 0 when f = 0 we find that 0 is nowhere dense. Let (x0 ,
ζ0 ) ∈ 0 and choose z ∈ C such that ∂x Re z f (x0 , y0 (x0 ), ζ0 ) = 0. Let S± =
{ ± Re z f (x, y0 (x), ζ) > 0 } then
f (x, y, ζ) = c± (x, y, ζ) f (x, y0 (x), ζ) in S± (2.13)
where 0 < c0 ≤ c± (x, y, ζ) ∈ C ∞ is uniformly bounded. By taking the limit of

(2.13) at S = { Re z f (x, y0 (x), ζ) = 0 } we find that c+ = c− when f = 0 at S.
When f = 0 at S then by differentiating (2.13) in x we find that c+ = c− . By
repeatedly differentiating (2.13) in x we obtain by recursion that c± extends to
c0 < c(x, y, ζ) ∈ C ∞ in when |ζ − ζ0 | < c so that (2.11) holds.
If ps,k is constant in y in a neighborhood of the semibicharacteristic, then the

Hamilton field H ps,k will be constant on the leaves and defined modulo tangent
vector to the leaves. Therefore we shall introduce a special symplectic structure
on N ∗ 2 . Recall that the symplectic annihilator to a linear space consists of the
vectors that are symplectically orthogonal to the space. Let T 2σ be the symplectic
annihilator to T 2 , which spans the symplectic leaves of 2 . If 2 = { η = 0 },
(x, y) ∈ Rn−d × Rd , then the leaves are spanned by ∂ y . Let
T σ 2 = T 2 /T 2σ (2.14)
which is a symplectic space over 2 which in these coordinates is parametrized by

T σ 2 = ((x0 , y0 ; ξ0 , 0); (x, 0; ξ, 0)) ∈ T 2 : (x, ξ) ∈ T ∗ Rn−d (2.15)
This is isomorphic to the symplectic manifold T ∗ Rn−d with w ∈ 2 as parameter.

We define the symplectic structure of N ∗ 2 by lifting the structure of 2 to
the fibers, so that the leaves of N ∗ 2 are given by L × { η0 } where L is a leaf
2 for η0 ∈ R . In the chosen
d
of coordinates, these leaves are parametrized by
(x0 , y; ξ0 , 0) × { η0 } : y ∈ Rd . The radial direction in N ∗ 2 will be the radial
direction in 2 , i.e. ξ, ∂ξ , lifted to the fibers. Similarly, a vector field V ∈ T (N ∗ 2 )
is parallel to the base of N ∗ 2 if it is in T 2 , which means that V η = 0.
8 N. Dencker
If ps,k is constant in y then H ps,k coincides with the Hamilton vector field of ps,k
−1
on ps,k (0) ⊂ N ∗ 2 with respect the symplectic structure on the symplectic manifold
N ∗ 2 . In fact, in the chosen coordinates we obtain from (2.9) that
H ps,k = ∂ξ ps,k ∂x − ∂x ps,k ∂ξ (2.16)
modulo ∂ y , which is nonvanishing if ∂x,ξ ps,k = 0. Thus H ps,k is well-defined modulo

terms containing ∂ y making it well defined on T σ 2 × Rd. Now, if ps,k = 0 then
by (2.9) we find that dps,k T 2 vanishes on T 2σ so dps,k T 2 is well defined on
T σ 2 . We may identify T (N ∗ 2 ) with T 2 × Rd since the fiber η is linear.
Definition 2.2 We say that the operator P is of subprincipal type on N ∗ 2 if the

following hold when ps,k = 0 on N ∗ 2 : H ps,k is parallel to the base,

dps,k T σ 2 = 0 (2.17)
and the corresponding Hamilton vector field H ps,k of (2.17) does not have the radial
direction. The (semi)bicharacteristics of ps,k with respect to the symplectic structure
of N ∗ 2 are called the subprincipal (semi)bicharacteristics.
Clearly, if coordinates are chosen so that (2.3) holds, then (2.17) gives that
∂x,ξ ps,k = 0 when ps,k = 0 and the condition that the Hamilton vector field does
not have the radial direction means that ∂ξ ps,k = 0 or ∂x ps,k ∦ ξ when ps,k = 0.
Because of (2.17) we find that H ps,k is transversal to the foliation of N ∗ 2 and
by (2.9) it is parallel to the base at the characteristics. The semibicharacteristic of
ps,k can be written = 0 × { η0 } ⊂ T (N ∗ 2 ), where 0 ⊂ 2 is transversal to the
leaves of 2 and η0 is fixed. The definition can be localized to an open set ω ⊂ N ∗ 2 .
It is a generalization of the definition of subprincipal type in [5], which is the special
case when η = 0. When P is of subprincipal type and satisfies (2.9), then we find
from Lemma 2.1 that ps,k is constant on the leaves of 2 near a semibicharacter-
istic after multiplication with a nonvanishing factor. We can now state a condition
corresponding to () on the reduced subprincipal symbol.
Definition 2.3 If k = κ(ω) for an open set ω ⊂ N ∗ 2 , then we say that P satisfies
condition Subk () if Im aps,k does not change sign from − to + when going in
the positive direction on the subprincipal bicharacteristics of Re aps,k in ω for any
0 = a ∈ C ∞ .

Observe that when k < κ(ω) or k = κ(ω) = ∞ then ps,k = ps 2 on ω and
Subk () means that the subprincipal symbol ps satisfies condition () on T σ 2 ,
which is condition Sub() in [5]. In general, we have that condition Subk () is
condition () given by (1.2) on the reduced subprincipal symbol ps,k with respect
to the symplectic structure of N ∗ 2 . But it is equivalent to the condition () on the
reduced subprincipal symbol ps,k with respect to the standard symplectic structure.
In fact, condition Subk () means that condition () holds for ps,k η=η0 for any η0 .
By using Lemma 2.1 we may assume that ps,k is independent of y after multiplying
with 0 = a ∈ C ∞ . In that case, the conditions are equivalent and both are invariant
under multiplication with nonvanishing smooth factors.
By the invariance of condition () given by [13, Theorem 26.4.12] it suffices
to check condition Subk () for some a such that HRe aps,k = 0. We also find that
condition Subk () is invariant under symplectic changes of variables, thus it is
invariant under conjugation of the operator by elliptic homogeneous Fourier integral
operators. Observe that the sign change is reversed when taking the adjoint of the
operator.
Next, we assume that condition Subk () is not satisfied on a semibicharacteris-
tic of ps,k , i.e., that Im aps,k changes sign from − to + on the positive flow of
HRe aps,k = 0 for some 0 = a ∈ C ∞ , where η is constant on . Thus, by Lemma 2.1
we may assume that ps,k is constant on the leaves in a neighborhood ω of , and
by multiplying with a we may assume that a ≡ 1 and that y is constant on the
semibicharacteristic.
Definition 2.4 Let p be of subprincipal type on N ∗ 2 and a subprincipal
semibicharacteristic of p. We say that a C ∞ section of spaces L ⊂ T (N ∗ 2 ) is
gliding for if L is symplectic of maximal dimension 2n − 2(d + 1) ≥ 2 so that L
is the symplectic annihilator of T and the foliation of 2 , which gives L ⊂ T 2
since η is constant on L. We say that a C ∞ foliation of N ∗ 2 with symplectic leaves
M is gliding for if the section of tangent spaces T M is a gliding section for .
Actually, he gliding foliation M for a subprincipal semibicharacteristic is
uniquely defined at , since it is determined by the unique annihilator T M and
is transversal to the foliation of 2 when p = 0 by (2.17). This definition can be
localized to a neighborhood of a subprincipal semibicharacteristic.
Example 2.1 Let p be of subprincipal type on N ∗ 2 . Assume that 2 = { η = 0 },
∂ y p = { η, p } = 0 and ∂x,ξ spans T M of the gliding foliation M of N ∗ 2 for the
bicharacteristics of HRe p = 0. Then we may complete x, ξ, τ = Re p and η to a
symplectic coordinate system (t, x, y; τ , ξ, η) so that the foliation M is given by
intersection of the level sets of τ , t, y and η. In fact, in that case we have ∂ Re p = 0
but ∂x Re p = ∂ξ Re p = 0.
In the case when η0 = 0 and k = κ(ω) < ∞ we will have estimates on the
rate of vanishing of ∂η ps,k on the subprincipal semibicharacteristic. Recall that the
semibicharacteristic can be written × { η0 }. Observe that
∂η ps,k = J k−1 (∂η p) = J k−1 (∂ p) (2.18)
since p vanishes of at least order k at 2 and that the normal derivatives ∂η is well-
defined modulo nonvanishing factors at η = 0. Let ω ⊂ 2 be a neighborhood of the
subprincipal semibicharacteristic and let M be the local C ∞ foliation of N ∗ 2 at
ω which is gliding for the semibicharacteristics. When η0 = 0 we shall assume that
there exists ε > 0 so that

V1 · · · V ∂η ps,k ≤ C | ps,k |1/k+ε on ω when |η − η0 | 1 (2.19)
10 N. Dencker
for any vector fields V j ∈ T M, 0 ≤ j ≤ and any . Condition (2.19) gives that
V1 · · · V ∂η ps,k vanishes when ps,k = 0. This definition is invariant under symplec-
tic changes of coordinates and multiplication with nonvanishing factors. Observe that
V1 · · · V ∂η ps,k = 0 when η0 = 0 since then p = ∂η p = 0 and V j ∈ T M ⊂ T 2 .
Condition (2.19) with = 0 gives that η → | ps,k (w, η)|(k−1)/k−ε is Lipschitz con-
tinuous, thus η → ps,k (w, η) vanishes at η0 of order 3 when k = 2 and order 2 when
k > 2.
In the case k = κ(ω) = 2 we shall also have a similar condition on the rate of
vanishing of ∂η2 ps,k on the subprincipal semibicharacteristic. Then

∂η2 ps,k = J 0 (∂ 2 p) = Hess p 2 (2.20)
is the Hessian of the principal symbol p at 2 , which is well defined on the normal
bundle N 2 since it vanishes on T 2 . Since p = ∂η p = 0 on 2 , we find that
Hess p is invariant modulo nonvanishing smooth factors under symplectic changes
of variables and multiplication of P with elliptic pseudodifferential operators. With
the gliding C ∞ foliation M of N ∗ 2 for the semibicharacteristics we shall assume
that there exists ε > 0 so that
V1 · · · V Hess p ≤ C | ps,k |ε on ω (2.21)
for any vector fields V j ∈ T M, 0 ≤ j ≤ and any . This definition is invariant under
symplectic changes of coordinates and multiplication with nonvanishing factors.
Remark 2.3 Conditions (2.19) and (2.21) are well defined and invariant under mul-
tiplication with elliptic pseudodifferential operators and conjugation with elliptic
Fourier integral operators.
Examples 3.1–3.3 show that conditions (2.19) and (2.21) are essential for the
necessity of Subk () when k = 2.
Example 2.2 If Re ps,k = τ , 2 = { η = 0 }, T M is spanned by ∂x,ξ and t → Im ps,k
vanishes of order 3 ≤ < ∞ at t = t0 (y, η) ∈ C ∞ then (2.19) and (2.21) hold. If
t0 (y) is independent of η then conditions (2.19) and (2.21) hold for any finite > 0.
In fact, if 0 < < ∞ then we can write Im ps,k = a(t − t0 (y, η)) with a = 0.
α
If > k−1 k
then for any α we find that ∂x,ξ ∂η Im ps,k vanishes of order − 1 > /k
α
at t = t0 , and if > 2 then ∂x,ξ ∂η Im ps,k vanishes of order − 2 > 0 at t = t0 . If t0
2
α
is independent of η then ∂x,ξ ∂ηj Im ps,k vanishes of order for any j and α.
Since ∂η ps,k is homogeneous of degree k − 1 in η, we find from Euler’s
identity that ∂η ps,k (w, η) = (k − 1) η · Hess p(w, η). Thus (2.21) implies that
|V1 · · · V ∂η ps,k | | ps,k |ε when η = 0, but we shall only use condition (2.21)
when (2.19) holds, see Theorem 2.1. Here a b means a ≤ Cb for some constant
C, and similarly for a b.
Now, by (2.9) we have assumed that the reduced subprincipal symbol ps,k is
constant on the leaves of 2 near up to multiplication with nonvanishing factors,
but when κ < ∞ we will actually have that condition on the following symbol.
Definition 2.5 If k = κ(ω) < ∞ is the order of p on an open set ω ⊆ 2 then we

define the extended subprincipal symbol on N ∗ ω by
N ∗ 2 (w, η) → qs,k (w, η, λ) = λJw2k−1 ( p)(η/λ1/k )

∼ ps,k (w, η) + O(λ−1/k )
+ J k−1 ( ps )(η/λ1/k ) = (2.22)
w
which is a weighted polynomial in η of degree 2k − 1. When κ(ω) = ∞ we define

qs,∞ ≡ ps .
By the invariance of p and ps , the extended subprincipal symbol transforms as jets

under homogeneous symplectic changes of coordinates. It is well defined up to non-
vanishing factors and terms proportional to the jet Jwk−1 (∂η p)(η/λ1/k ) ∼
= λ1/k−1 ∂ηk p
−1
modulo O(λ ) under multiplication with classical elliptic pseudodifferential oper-
ators. The extended and the reduced subprincipal symbols are complexly conjugated
when taking adjoints.
Remark 2.4 The extended subprincipal symbol (2.22) is given by the blowup of the
reduced principal symbol at η = 0 so that
λ2−m psub (x, y; λξ, λ1−1/k η) ∼

= λqs,k (x, y; ξ, η, λ)
= λ ps,k (x, y; ξ, η) + O(λ1−1/k ) modulo O(1) (2.23)
We also have that
λ2−m ∂η psub (x, y; λξ, λ1−1/k η) ∼

= λ1/k ∂η qs,k (x, y; ξ, η, λ)
= λ1/k ∂η ps,k (x, y; ξ, η) + O(1) (2.24)
modulo O(λ1/k−1 ) and
λ2−m ∂η2 psub (x, y; λξ, λ1−1/k η) ∼

= λ2/k−1 ∂η2 qs,k (x, y; ξ, η, λ)
= λ2/k−1 ∂η2 ps,k (x, y; ξ, η) + O(λ1/k−1 ) (2.25)

modulo O(λ2/k−2 ). Observe that if P is of subprincipal type then dqs,k T σ 2 = 0
when qs,k = 0 for λ 1 since this holds for ps,k .
In fact, dqs,k ∼
= dps,k modulo O(λ−1/k ) and since |dps,k | = 0 the distance between
−1 −1
qs,k (0)and ps,k (0) is O(λ−1/k ) for λ 1. Observe that composition of the operator
P with elliptic pseudodifferential operators gives factors proportional to Jwk−1 (∂η p)
(η/λ1/k ) which we shall control with (2.19).
By (2.19) we have that ∂η ps,k = 0 when ps,k = 0 at ω. We shall also assume this
for the next term in the expansion of qs,k ,
∂η qs,k = O(λ−2/k ) when ps,k = 0 at ω for |η − η0 | < c0 and λ 1 (2.26)

12 N. Dencker
−1
Actually, we only need this where ps,k ∧ d p s,k vanishes of infinite order at ps,k (0)
in ω, where dps,k ∧ d p s,k is the complex part of dps,k .
We shall also assume a condition similar to (2.9) on the extended subprincipal
symbol

dqs,k ≤ C0 |qs,k | at ω when |η − η0 | < c0 and λ 1 (2.27)
TL
for any leaf L of N ∗ 2 where ω ⊂ 2 . By letting λ → ∞ we obtain that (2.9)

= ps,k modulo O(λ−1/k ). Also, multiplication of psub by a ∼
holds, since qs,k ∼ = a0
+ a−1 + . . . with a j homogeneous of degree j and a0 = 0 gives that qs,k gets multi-
= a0 psub + a−1 p ∼
plied by the expansion of η → a0 (x, y; ξ, η/λ1/k ) since apsub ∼ =
a0 psub modulo terms in S m−1
vanishing of order k at 2 . Thus, condition (2.27) is
invariant under multiplication of psub with classical elliptic symbols. Also, (2.27)
is invariant under changes of homogeneous symplectic coordinates that preserves
2 = { η = 0 } and T L. Now, we have ∂x,ξ qs,k = 0 when qs,k = 0 and λ 1 since
P is of subprincipal type.
Remark 2.5 Since the semibicharacteristic is transversal to the leaves of 2 and
condition (2.27) holds near the semibicharacteristic, Lemma 2.1 gives that
qs,k (x, y; ξ, η, λ) = c(x, y; ξ, η, λ)

qs,k (x; ξ, η, λ) λ1 (2.28)
for |η − η0 | < c0 near the semibicharacteristic. Here qs,k (x; ξ, η, λ) is the value of
qs,k at the intersection of the semibicharacteristic and the leaf. In fact, the proof of the
lemma extends to symbols depending uniformly on the parameter 0 < λ−1/k 1.
Condition (2.27) is not invariant under multiplication of P with elliptic pseudod-
ifferential operators or conjugation with elliptic Fourier integral operators. In fact, if
A has symbol a then the refined principal symbol of the composition A P is equal to
apsub + 2i1 { a, p } which adds 2i λ1/k−1 ∂ y a∂η ps,k to qs,k . But (2.26) is invariant, since
the term containing the factor ∂η ps,k is O(λ−2/k ) when k > 2 and has vanishing η
−1
derivative at ps,k (0) by (2.21) when k = 2.
This is one reason why we have to control the terms with ∂η ps,k with (2.19). When
k < ∞, qs,k is a polynomial in η/λ1/k of degree 2k − 1 and c in (2.28) is an analytic
function in η/λ1/k on ω when |η − η0 | < c0 . Actually, it suffices to expand c in η/λ1/k
up to order k in order to obtain (2.28) modulo O(λ−1 ). If C(x, y; ξ, η/|ξ|1/k ) =
c(x, y; ξ, η, |ξ|) in (2.28) obtain that C psub is constant in y modulo S m−2 in
then we1−1/k

ω when η − η0 |ξ| 1−1/k
< c0 |ξ| .
In the case when the principal symbol p is real, a necessary condition for solv-
ability of the operator is that the imaginary part of the subprincipal symbol does
not change sign from − to + when going in the positive direction on a C ∞ limit
of normalized bicharacteristics of the principal symbol p at 2 , see [4]. When p
vanishes of exactly order k on 2 = { η = 0 } and the localization

η → ∂ηα p(x, y; 0, ξ)η α /α!
|α|=k
is of principal type when η = 0 such limit bicharacteristics are tangent to the leaves
of 2 . In fact, then |∂η p(x, y; ξ, η)| ∼ = |η|k−1 and |∂x,y,ξ p(x, y; ξ, η)| = O(|η|k ),
which gives H p = ∂η p∂ y + O(|η| ). Thus the normalized Hamilton vector field is
k
equal to A∂ y , A = 0, modulo terms that are O(|η|), so the normalized Hamilton

vector fields have limits that are tangent to the leaves. That the η derivatives dominates
∂ p can also be seen from Remark 2.4. When the principal symbol is proportional to
a real valued symbol, this gives examples of nonsolvability when the subprincipal
symbol is not constant on the leaves of 2 , see Example 3.4 and [4] in general. Thus
condition (2.27) is natural for the the study of the necessity of Subk () if there are
no other conditions on the principal symbol.
We shall study the microlocal solvability of the operator, which is given by the
following definition. Recall that H(s)loc
(X ) is the set of distributions that are locally in
the L 2 Sobolev space H(s) (X ).
Definition 2.6 If K ⊂ S ∗ X is a compact set, then we say that P is microlocally

an integer N so that for every f ∈ H(N ) (X ) there exists
loc
solvable at K if there exists

u ∈ D (X ) such that K WF(Pu − f ) = ∅.
Observe
that solvability at a compact set K ⊂ X is equivalent to solvability at
S ∗ X K by [13, Theorem 26.4.2], and that solvability at a set implies solvability at
a subset. Also, by [13, Proposition 26.4.4] the microlocal solvability is invariant
under conjugation by elliptic Fourier integral operators and multiplication by elliptic
pseudodifferential operators. We can now state the main result of the paper.
Theorem 2.1 Assume that P ∈ clm (X ) has principal symbol that vanishes of at
least second order at a nonradial involutive manifold 2 ⊂ T ∗ X \ 0. We assume
that P is of subprincipal type, satisfies conditions (2.26) and (2.27) but does not
satisfy condition Subk () near the subprincipal semibicharacteristic × { η0 } in
N ∗ 2 where ⊂ ω ⊂ 2 and k = κ(ω).
In the case when η0 = 0 we assume that P satisfies conditions (2.19) and when
k = 2 we also assume condition (2.21) for a gliding symplectic foliation M of N ∗ 2
for the subprincipal semibicharacteristics near .
In the case η0 = 0 and k = 2 we assume condition (2.21) for a gliding symplectic
foliation M of N ∗ 2 for the subprincipal semibicharacteristics near , and when
k > 2 we assume no extra condition.
Under these conditions, P is not locally solvable near ⊂ 2 .
Examples 3.1–3.3 show that conditions (2.19) and (2.21) are essential for the
necessity of Subk () when k = 2. Due to the results of [4], condition (2.27) is
natural if there are no other conditions on the principal symbol, see Example 3.4.
Observe that for effectively hyperbolic operators, which are always solvable, 2 is
not an involutive manifold, see Example 3.7.
Remark 2.6 It follows from the proof that we don’t need condition (2.26) in the case
when condition (2.27) holds on the leaves of 2 that intersect the semibicharacteristic.
In the case when η = 0 on the subprincipal semibicharacteristics, condition (2.21)
14 N. Dencker
only involves Hess p at 2 . This gives a different result than Theorem 2.7 in [5],
since in that result condition (2.21) is not used, condition (2.27) only involves ps but
we also have conditions on |dps ∧ d p s | and Hess p on 2 .
Now let S ∗ X ⊂ T ∗ X be the cosphere bundle where |ξ| = 1, and let u(k) be the
L Sobolev norm of order k for u ∈ C0∞ . In the following, P ∗ will be the L 2 adjoint
2
of P. To prove Theorem 2.1 we shall use the following result.
Remark 2.7 If P is microlocally solvable at ⊂ S ∗ X , then Lemma 26.4.5 in [13]

gives that for any Y X such that ⊂ S ∗ Y there exists an integer ν and a pseudod-
ifferential operator A so that WF(A) ∩ = ∅ and
u(−N ) ≤ C(P ∗ u(ν) + u(−N −n) + Au(0) ) u ∈ C0∞ (Y ) (2.29)
where N is given by Definition 2.6.
We shall prove Theorem 2.1 in Sect. 10 by constructing localized approximate

solutions to P ∗ u ∼
= 0 and use (2.29) to show that P is not microlocally solvable at .
3 Examples
Example 3.1 Consider the operator
P = Dt + ia(t) y (x, y) ∈ Rn−d × Rd (3.1)
where 0 < d < n, a(t) is real and has a sign change from − to +. This operator is
equal to the Mizohata operator when a(t) = t. We find that P is of subprincipal type,
k = 2 and ps,2 (t, τ , η) = τ + ia(t)|η|2 is constant on the leaves of 2 = { η = 0 }.
Condition (2.27) hold but Sub2 () does not hold since t → a(t)|η|2 changes sign
from − to + when η = 0. Since |∂η ps,2 | ∼
= Hess ps,2 ∼
= |a(t)| when η = 0 and ps,2
is independent of (x, ξ) we find that conditions (2.19) and (2.21) hold. Theorem 2.1
gives that P is not locally solvable.
Example 3.2 The operator
P = Dt + i(Dx1 Dx2 + t Dx22 ) x ∈ Rn n ≥ 3 (3.2)
is solvable, see [2]. We find that P is of subprincipal type, k = 2, 2 = { ξ1 = ξ2 = 0 }

and ps,2 (t, τ , ξ) = τ + i(ξ1 ξ2 + tξ22 ). Condition Sub2 () does not hold since t →
ξ1 ξ2 + tξ22 changes sign from − to + when ξ1 = −tξ2 and ξ2 = 0. Since |∂ξ ps,2 | ∼ =
Hess ps,2 ∼= 1 | ps,2 | ∼
= |t| when ξ2 = 0 and τ = ξ1 = 0, we find that condi-
tions (2.19) and (2.21) do not hold.
Example 3.3 Consider the following generalization of Example 3.2 given by

P = Dt + i(Dx1 Dx2 + t 2 j+1 Dx22 ) + i(2 j 2 + j)t 2 j−1 x12 x ∈ Rn (3.3)
for j > 0 and n ≥ 3. We find that P is of subprincipal type, k = 2, 2 =

{ ξ1 = ξ2 = 0 } and ps,2 (t, τ , ξ) = τ + i(ξ1 ξ2 + t 2 j+1 ξ22 ). Thus Sub2 () does not
hold since t → ξ1 ξ2 + t 2 j+1 ξ22 changes sign from − to + when ξ1 = −t 2 j+1 ξ2
and ξ2 = 0. Since |∂ξ ps,2 | ∼
= Hess ps,2 ∼
= 1 and | ps,2 | ∼= |t|2 j+1 when τ = ξ1 = 0
and ξ2 = 0, we find that conditions (2.19) and (2.21) do not hold. By choosing
x2 − t 2 j+1 x1 as new x2 coordinate we obtain that

P = Dt + i Dx1 + i(2 j + 1)t 2 j x1 Dx2 + i(2 j 2 + j)t 2 j−1 x12 (3.4)
Then by conjugating P with e(2 j+1)t x1 /2

2j 2
we obtain P = Dt + i Dx1 Dx2 which has
constant coefficients and is solvable.
P = Dt + f (t, y, Dx ) + i y (x, y) ∈ Rn−2 × R2 (3.5)
where f (t, y, ξ) ∈ Shom1

is real and y = ∂ y1 ∂ y2 is the wave operator in y ∈ R2 . We
find that P is of subprincipal type, k = 2, 2 = { η = 0 } and ps,2 (t, y, τ , ξ, η) =
τ + f (t, y, ξ) − iη1 η2 so (2.27) is not satisfied if ∂ y f = 0. Since −i P = y −
i Dt − i f (t, y, Dx ) it follows from Theorem 1.2 in [16] that P is not solvable if
∂ y f = 0.
P = Dt + i f (t, x, Dx ) + B(t, x, D y ) (x, y) ∈ Rn−d × Rd (3.6)
where 0 < d < n, f (t, x, ξ) ∈ Shom 1

is real and B(t, x, η) ∈ Shom
2
vanishes of degree
k ≥ 2 at 2 = { η = 0 }. Then ps,k = τ + i f (t, x, ξ) + Bk (t, x, η) where Bk is the
kth Taylor term at 2 of the principal symbol of B, so (2.27) is satisfied everywhere.
Assume that B(t, η) is independent of x and the sign change in t → f (t, x, ξ) +
Im Bk (t, η) is from − to + of order < ∞ at t = t0 . If t → ∂η Bk (t, η) vanishes
of order greater than /k at t = t0 then (2.19) holds. If k = 2 and t → ∂η2 Bk (t, η)
vanishes at t = t0 then (2.21) holds. Then P is not solvable by Theorem 2.1 and
Remark 2.6.
If Im B(x, η) = 0 is constant in t and k is odd with Im Bk (x, η) ≷ 0, ∀ x, then con-
dition Subk () implies that t → f (t, x, ξ) is nonincreasing. In fact, Sard’s theorem
gives for almost all values f 0 of f that there exists (t, x, ξ) so that f (t, x, ξ) = f 0
and ∂t f (t, x, ξ) = 0. Then one can choose η so that f (t, x, ξ) + Im Bk (x, η) = 0
so Subk () gives ∂t f (t, x, ξ) ≤ 0.
If t → f (t, x, ξ) is nonincreasing, B(x, η) is constant in t and Re B ≡ 0, then
P is solvable. In fact, then [P ∗ , P] = 2i[Re P, Im P] = 2∂t f ≤ 0 so Re Pu2
Pu2 + P ∗ u2 P ∗ u2 + u2 and u Re Pu if |t| 1 in the support
of u ∈ C0∞ .
16 N. Dencker
Example 3.6 The linearized Navier–Stokes equation

∂t u + a j (t, x)∂x j u + x u = f a j (x) ∈ C ∞ (3.7)
j
is of subprincipal type. The symbol is

iτ + i a j (t, x)ξ j − |ξ|2 (3.8)
j
so P is of subprincipal type, k = 2, 2 = { ξ = 0 } and ps,2 (τ , ξ) = iτ − |ξ|2 . Thus

Sub2 () holds since −|ξ|2 does not change sign when t changes.
Example 3.7 Effectively hyperbolic operators P are weakly hyperbolic operators for
which the fundamental matrix F has two real eigenvalues, here F = J Hess p 2
with p = σ(P) and J (x, ξ) = (ξ, −x) is the symplectic involution. Then P is solv-
able for any subprincipal symbol by (see [14, 18]) but in this case 2 is not an
involutive manifold.
4 The Normal Form
We are going to prepare the operator microlocally near the semibicharacteristic. We

have assumed that P ∗ has the symbol expansion pm + pm−1 + . . . where p j ∈ Shom
j
is homogeneous of degree j. By multiplying P ∗ with an elliptic classical pseu-

dodifferential operator, we may assume that m = 2 and p = p2 . By choosing local
homogeneous symplectic coordinates (x, y; ξ, η) we may assume that X = Rn and
2 = { η = 0 } ⊂ T ∗ Rn \ 0 with the symplectic foliation by leaves spanned by ∂ y .
If p vanishes of order k < ∞ at ω ⊂ 2 we find that

p(x, y; ξ, η) = Bα (x, y; ξ, η)η α /α! (x, y, ξ) ∈ ω (4.1)
|α|=k
where Bα is homogeneous of degree 2 − k, and Bα (x, y; ξ, 0) ≡ 0 for some |α| = k

and some (x, y, ξ, 0) ∈ ω. When p vanishes of infinite order we get (4.1) for any k.
We shall first consider the case when k = κ(ω) < ∞. Recall the reduced sub-
principal symbol ps,k (w, η) = Jwk ( p)(η) + ps (w), w ∈ 2 , by Definition 2.1, and
the extended subprincipal symbol qs,k (w, η, λ) = λJw2k−1 ( p)(η/λ1/k ) + Jwk−1 ( ps )
(η/λ1/k ) by Definition 2.5. Observe that these are invariantly defined and are the
complex conjugates of the corresponding symbols of P by Remark 2.4. We also find
from Remark 2.4 that
psub (x, y; ξ,η/|ξ|1/k ) ∼

= |ξ|qs,k (x, y; ξ0 , η0 , |ξ|)
= |ξ| ps,k (x, y; ξ0 , η0 ) + O(|ξ|1−1/k ) modulo O(1) (4.2)
where (ξ0 , η0 ) = |ξ|−1 (ξ, η). We also have
∂η psub (x, y; ξ, η/|ξ|1/k ) ∼

= |ξ|1/k ∂η ps,k (x, y; ξ0 , η0 ) modulo O(1) (4.3)
and
∂η2 psub (x, y; ξ, η/|ξ|1/k )

∼
= |ξ|2/k−1 ∂ 2 ps,k (x, y; ξ0 , η0 ) modulo O(|ξ|1/k−1 ) (4.4)
η
When k < ∞ we shall localize with respect to the metric
gk (d x, dy; dξ, dη) = |d x|2 + |dy|2 + |dξ|2 / 2

+ |dη|2 / 2−2/k
(4.5)
where = (|ξ|2 + 1)1/2 . If g,δ is the metric corresponding to the symbol classes
m
S,δ we find that
g1,0 ≤ gk ≤ g1−1/k,0
When k = ∞ we shall let g∞ = g1,0 which is the limit metric when k → ∞.

We shall use the Weyl calculus symbol notation S(m, gk ) where m is a weight for
gk , one example is m = (|ξ|2 + 1)m/2 . Observe that we have the usual asymptotic
j
expansion when composing S( m , gk ) with S,δ when > 0 and δ < 1 − k1 .
Remark 4.1 If k < ∞, f is homogeneous of degree m and vanishes of order j at

2 then f ∈ S( m− j/k , gk ) when |η| |ξ|1−1/k .
One example is p = σ(P ∗ ) ∈ S( , gk ) in ω when |η| |ξ|1−1/k for k = κ(ω).

In fact, when |η| |ξ|1−1/k we have | f | |ξ|m− j |η| j |ξ|m− j/k . Differentiation in
x or y does not change this estimate, differentiation in ξ lowers the homogeneity
by one and when taking derivatives in η we may lose a factor η j = O(|ξ|1−1/k ). We
shall prepare the symbol in domains of the type

= (x, y, λξ, λ1−1/k η) : (x, y, ξ, η) ∈ ⊂ S ∗ Rn , λ > 0
(4.6)
which is a gk neighborhood consisting of the inhomogeneous rays through .

Now for k < ∞ we shall use the blowup mapping
χ : (x, y; ξ, η) → (x, y; ξ, η/|ξ|1/k ) (4.7)
which is a bijection when |ξ| = 0. The pullback by χ maps symbols in S( m , gk )

where |η| |ξ|1−1/k to symbols in S1,0
m
where |η| |ξ|, see for example (4.2). Also
Taylor expansions in η where |η| |ξ|1−1/k get mapped by χ∗ to polyhomogeneous
18 N. Dencker
expansions, and a conical neighborhood is mapped by χ to the gk neighborhood

.

The blowup
psub ◦ χ(x, y; ξ, η) = q(x, y; ξ, η) = |ξ|qs,k (x, y; ξ/|ξ|, η/|ξ|, |ξ|)

∼
= |ξ| ps,k (x, y; ξ/|ξ|, η/|ξ|) ∈ S 1 modulo S
1−1/k
(4.8)
1,0 1,0
is a sum of terms homogeneous of degree 1 − j/k for j ≥ 0 by Definition 2.5. We

shall prepare the blowup qs,k and get it on a normal form after multiplication with
pseudodifferential operators and conjugation with elliptic Fourier integral operators.
We have assumed that P is of subprincipal type and does not satisfy condi-
tion Subk () near a subprincipal semicharacteristic × { η0 } ⊂ N ∗ 2 , which is
transversal to the leaves of N ∗ 2 . By changing and η0 we may obtain that Im aps,k
changes sign from + to − on the bicharacteristic × { η0 } of Re aps,k for some
0 = a ∈ C ∞ . The differential inequality (2.27) in these coordinates means that
|∂ y qs,k | ≤ C|qs,k | when |ξ| 1 (4.9)
in a conical neighborhood ω in N ∗ 2 containing × { η0 }. By shrinking ω we may

obtain that the intersections of ω and the leaves of 2 are simply connected. Then
by putting |ξ| = λ we obtain from Remark 2.5 that
qs,k (x, ξ, η) ∼
= c(x, y, ξ, η)qs,k (x, y, ξ, η) at ω when |ξ| 1 (4.10)
0
modulo S1,0 . Here
qs,k is the value of qs,k at the intersection of the semibicharacteristic
and the leaf. Here 0 = c ∈ S1,0 0
is a sum of terms homogeneous of degree − j/k for
j ≥ 0 such that |c| > 0 when |ξ| 1. In fact, c has an expansion in η/|ξ|1/k and
0
it suffices to take terms up to order k in c to get (4.10) modulo S1,0 . Thus the term
homogeneous of degree 0 in c is nonvanishing in the conical neighborhood ω. By
cutting off the coefficients of the lower order terms of c where |ξ| 1, we may
assume that c = 0 in ω.
By multiplying P with a pseudodifferential operator with symbol C = c ◦ χ−1 ∈
S(1, gk ) when |η| |ξ|1−1/k , we obtain by Remark 2.5 the refined principal symbol
1

psub + ∂ y C∂η psub modulo S(1, gk ) in χ(ω) (4.11)
2i
= C P. Here
for P psub = C psub is constant on the leaves of N ∗ 2 modulo S(1, gk )
in χ(ω). We have that
∂η psub = C −1 ∂η
psub + ∂η C −1
psub
where ∂η C −1 ∈ S( 1/k−1 , gk ) when |η| |ξ|1−1/k . Thus by multiplying P with a

pseudodifferential operator with symbol 1 − 2i1 ∂η C −1 ∂ y C ∈ S(1, gk ), we obtain the
refined principal symbol
psub + c0 ∂η
psub modulo S(1, gk ) in χ(ω) (4.12)
for some c0 ∈ S(1, gk ) which may depend on y. Then we find that
∂ y ( psub ) ∼
psub + c0 ∂η = ∂ y c0 ∂η
psub modulo S(1, gk )
By putting q = psub ◦ χ obtain that ∂ y q = 0 in ω when |ξ| 1. We shall control the

term proportional to ∂η psub ◦ χ = |ξ|1/k ∂η q ∈ S 1/k by using condition (2.19), see
Lemma 6.1. Observe that q ∼ = q1 = ps,k ◦ χ modulo S 1−1/k , which is homogeneous
and independent of y near ω. By the invariance of the condition, we may assume that
a is independent of y. Then the semibicharacteristics are constant in η so we may
choose a independent of (y, η).
Observe that changing a changes and η0 by the invariance, but we may
assume that × { η0 } is arbitrarily close to the original semibicharacteristic by
[13, Theorem 26.4.12]. Since Im aq1 changes sign on × { η0 } there is a maximal
semibicharacteristic × { η0 } on which Im aq1 = 0 and because of the sign change
we may shrink so that it is not a closed curve. Here could be a point, which is
always the case if the sign change is of finite order. By continuity, ∂x,ξ Re aq1 = 0
near × { η } for η close to η0 and we may extend a to a nonvanishing symbol that is
homogeneous of degree 0 near . Multiplying P with an elliptic pseudodifferential
operator with symbol a = a ◦ χ−1 we may assume that a ≡ 1.
Recall that conditions (2.19) (and (2.21) when k = 2) holds in some neighborhood
of × { η0 } with the gliding foliation M of N ∗ 2 . By using Darboux’ theorem we
can choose local coordinate functions (x, ξ) such that T M is spanned by ∂x and ∂ξ for
the leaves of M. Now 0 = HRe q1 is tangent to × { η0 }, transversal to the symplectic
foliation of 2 , constant in y and in the symplectic annihilator of T M. Since T M
is symplectic, this gives that Re q1 and η are constant on the leaves M. Now take
τ = Re q1 when η = η0 and extend it is so that τ is independent of η. Then we can
complete τ , y and η to a homogeneous symplectic coordinate system (t, x, y; τ , ξ, η)
in a conical neighborhood ω of in 2 so that (x, ξ)η=η0 is preserved. Since the
change of variables preserves the (y, η) variables, it preserves 2 = { η = 0 } and
its symplectic foliation and the fact that ∂ y q = { η, q } = 0. When η = η0 we have
that Re q1 = τ and the leaves T M of the foliation M are spanned by ∂x = Hξ and
∂ξ = −Hx modulo ∂ y when η = η0 . Since q is independent of y we may assume that
V is in the span of ∂x,ξ in (2.19) and (2.21). Since the η variables are preserved, the
blowup map χ and the inhomogeneous rays are preserved and the coordinate change
is an isometry with respect to the metric gk .
By conjugating with elliptic Fourier integral operators in the variables (t, x) inde-
pendently of y microlocally near ⊂ 2 , we obtain that Re q1 = τ in a conical
neighborhood of when η = η0 . This gives q1 = τ + (t, x, τ , ξ, η) in a neigh-
borhood of × { η0 }, where is homogeneous and Re ≡ 0 when η = η0 . Since
this is a change of symplectic variables (t, x; τ , ξ) for fixed (y, η) we find by the
20 N. Dencker
invariance that t → Im (t, x, 0, ξ, η0 ) changes sign from + to − near . Observe

that the reduced principal symbol is invariant under the conjugation by Remark 2.2
so the condition that ∂ y q = { η, q } = 0 is preserved, but we may also have a term
c∂η q ∈ S 1/k where c could depend on y.
Next, we shall use the Malgrange preparation theorem on q1 . Since ∂τ q1 = 0 near
× { η0 } we obtain that
τ = c(t, x, τ , ξ, η)q1 (t, x, τ , ξ, η) + r (t, x, ξ, η) (4.13)
locally for η close to η0 when |ξ| = 1, and by a partition of unity near × { η0 },

which can be extended by homogeneity so that c is homogeneous of degree 0 and
r is homogeneous of degree 1. Observe that this gives that ∂η q1 = ∂η c−1 (τ − r ) −
c−1 ∂η r by (4.13). By taking the τ derivative of (4.13) using that q1 = 0 and ∂τ q1 = 1
at × { η0 } we obtain that c = 1 on × { η0 }. Multiplying the operator P ∗ with
a pseudodifferential operator with symbol c ◦ χ−1 ∈ S(1, gk ) when |η| |ξ|1−1/k
we obtain that q1 (t, x, τ , ξ, η) = τ − r (t, x, ξ, η) in a conical neighborhood of ×
{ η0 }.
Writing r = r1 + ir2 with r j real, we may complete τ − r1 (t, x, ξ, η0 ), t, y and η
to a homogeneous
symplectic coordinate system (t, x, y; τ , ξ, η) near × { η0 } so

that (x, ξ) t=0 is preserved. This is a change of coordinates in (t, x; τ , ξ) which as
before is independent of the variables (y, η). We find that ∂τ r = { r, t } = 0 and ∂ y r =
{ η, r } = 0 are preserved and Re r η=η ≡ 0. By the invariance we find that t →
0
r2 (t, x, ξ, η0 ) changes sign from + to − near . As before, the blowup map χ and
the inhomogeneous rays are preserved and the coordinate change is an approximate
isometry with respect to the metric gk .
By conjugating with elliptic Fourier integral operators in (t, x) which are constant
in y microlocally near ⊂ 2 , the calculus gives as before that
q1 (t, x, τ , ξ, η) = τ + f (t, x, ξ, η) (4.14)
where f = −r . When η = η0 we have Re f ≡ 0 and Im f has a sign change from

+ to − as t increases near by the invariance of condition (). In fact, the reduced
principal symbol is invariant under the conjugation so the condition that ∂ y q =
{ η, q } = 0 is preserved, but we may also have a term c∂η q ∈ S 1/k where c could
depend on y. Observe that τ is invariant under the blowup mapping χ given by (4.7).
By the invariance, we find that (2.19) (and (2.21) if k = 2) holds for q1 with T M
spanned by ∂x and ∂ξ when η = η0 . In fact, since y, η and t are independent of (x, ξ)
we find that the span of ∂x,ξ is invariant modulo terms proportional to ∂τ . Since ∂η q1
is independent of τ by (4.14) we may take V j in the span of ∂x,ξ in (2.19) (and (2.21)
if k = 2) when η = η0 . Thus, by putting τ = − Re f in (2.19) we obtain that there
exists ε > 0 so that
α
|∂x,ξ ∂η f | | Im f |1/k+ε ∀ α |η − η0 | < c (4.15)
Another random document with
no related content on Scribd:
CONCLUSION
Constantine’s accession proved to be, like the coming of Alexander,

a turning-point in the history of the world. His so-called conversion
put into the hands of the Catholic Church a weapon for the
suppression of all rivalry, of which she was not slow to make use.
Already in his reign many of the heathen temples were torn
down[1217], and under the rule of his morose and gloomy successor,
Constantius, the work of demolition went on apace[1218]. The
accession of the philosophic Julian gave the worshippers of other
gods than Christ a short respite, and even allowed some of the
temples destroyed in the former reigns to be restored by or at the
expense of the Christians[1219]. Julian’s heroic death in Persia again
threw the crown into the hands of a Christian emperor, whose reign
of seven months gave him little time, as he perhaps had small
inclination, for persecution[1220]; but under his successors Valentinian
and Valens, heathen sacrifices were forbidden under severe
penalties. The end came under Gratian, when the temple estates
were confiscated, the priests and vestals deprived of the stipends
which they had hitherto received from the public treasury, and the
heathen confraternities or colleges were declared incapable of
receiving legacies[1221]. Only a few rich men like the Vettius Agorius
Praetextatus whom we have seen among the worshippers of
Mithras, or the Quintus Aurelius Symmachus, whose learned and
patriotic life has been so well described by Sir Samuel Dill[1222], could
henceforth venture to practise, even with maimed rites, the faiths
condemned by the Court and the Church.
As for the Gnostic sects, which since Hadrian’s time had striven with
such success as we have seen to combine magic and other ancient
beliefs with Christianity, they found but short shrift at the hands of the
triumphant Church. By an edict issued by Constantine before his
own reception into the Church, all their “houses of prayer” were
confiscated for the benefit of the Catholic Church, their meeting even
in private forbidden, and their books seized and burned[1223].
“Thus,” says Eusebius, “were the lurking places of the heretics

broken up by the emperor’s command, and the savage beasts
they have harboured (I mean the chief authors of their impious
doctrines) driven to flight. Of those whom they had deceived,
some, intimidated by the emperor’s threats, disguising their real
feelings, crept secretly into the Church. For since the law
directed that search should be made for their books, those of
them who practised evil and forbidden arts were detected, and
these were ready to secure their own safety by dissimulation of
every kind[1224].”
Throughout the length and breadth of the Roman Empire all but a
very few Roman nobles thus professed the faith of Christ. In the
words of the dying Julian, the Galilaean had conquered.
From this time until our own, Christianity has reigned in the West
with no serious rival. In the VIIth century, when Mahommed’s Arabs,
flushed with the enthusiasm of a new faith which owed something at
least to the relics of Gnosticism, poured in upon an Empire wearied
out alike by perpetual war against the barbarians and by its own civil
and religious dissensions, the Church was compelled to abandon to
them her conquests in Africa and the East. In Europe, however, she
continued in unchecked supremacy, gathering to herself and
assimilating the barbarians who at one time seemed likely to
extinguish all civilization; and she thus became a bond uniting many
nations and languages in one community of faith and thought. She
even succeeded in keeping alive the remains of that Greek art and
learning which still form our best and proudest intellectual
possession, and if during her reign many of the precious monuments
of antiquity perished, the fault was not entirely hers. In every respect,
her rule was supreme; and such enemies as she had in Europe were
those of her own household. The Manichaeans who, as has been
said, once bid fair to deprive her of some of her fairest provinces,
never dared to make open war upon her, and their secret defection
was punished by an unsparing use of the secular arm. The German
Reformation of the XVIth century has probably left her stronger than
before, and the few losses that she has suffered in the Old World
have been more than compensated by the number of lieges she has
succeeded in attaching to herself in the New.
In the days of her infancy, and before she thus came into her
inheritance, Christianity borrowed much from the rivals over which
she was in the long run to reign supreme. Her outward observances,
her ritual, and the organization of her hierarchy, are perhaps all due
to the associations that she finally overcame. The form of her
sacraments, the periods of her fasts and festivals, and institutions
like monachism, cannot be explained without reference to those
religions from whose rivalry she so long suffered. That, in such
matters, the Church should take what was useful to her was, as said
above, part of her consciously expressed policy, and doubtless had
much to do with her speedy triumph. To show that her dogmas also
took many things from the same source would involve an invasion
into the domain of professional theology, for which I have neither
authority nor desire. But if, at some future time, investigation should
show that in this respect also Christianity owes something to her
forerunners and rivals, the argument against her Divine origin would
not thereby be necessarily strengthened. That, in the course of her
development, she acquired characteristics which fitted her to her
environment would be in strict conformity with the laws which appear
to govern the evolution of all institutions; and if the Power ruling the
universe chooses to work by law rather than by what seems to us
like caprice, such a choice does not show Him to be lacking either in
wisdom or benevolence.
As was said at the outset, everyone must be left to place his own
interpretation on the facts here attempted to be set forth. But if, per
impossibile, we could approach the study of the origins of
Christianity with the same mental detachment and freedom from
prejudice with which we might examine the worship of the Syrian
Jupiter Dolichenus or the Scandinavian Odin, we should probably
find that the Primitive Church had no need of the miraculous powers
which were once assigned as the reason for her gradual and steady
advance to all but universal dominion. On the contrary, it may be that
Christianity would then appear as a link—although a most important
and necessary link—in a regular chain of events which began more
than three centuries before she emerged from her birthplace in
Palestine into that Roman world which in three centuries more was
to be hers of right. No sooner had Alexander’s conquests made a
world-religion possible, than there sprang up, as we have seen, in
his own city of Alexandria, a faith with a far higher and purer idea of
Divinity than any that had until then been known in the West. Then
the germs already present in small fraternities like those of the
Orphics and the Essenes blossomed forth into the fantastic and
unwholesome growths, as we must needs think them, of that
Gnosticism which marked the transition of the ancient world from
Paganism to Christianity. Lastly there came in from the countries
under the influence of Rome’s secular enemy, Persia, the heresy of
Marcion, the religion of Mithras, and the syncretistic policy of Manes
and his continuators. Against all these in turn, Christianity had to
struggle in a contest where the victory was not always on her side:
and if in time she overthrew them all, it can only be because she was
better fitted to the needs of the world than any of her predecessors
or contemporaries.
INDEX
Abel, Ophite story of, ii. 52;

and Manichaean, ii. 304
Aberamenthôu, name used in Magic Papyri and Pistis Sophia, i. 102.
See Jesus, Texts of Saviour
Abiuth, receiver of Ariel in Texts of Saviour, ii. 186
Abraham, named in Mag. Pap., i. 106 n. 6; ii. 34;
an astrologer apud Artapanus, i. 173;
inspired by Ialdabaoth, ii. 53;
Bosom of, in Marcion’s system, ii. 211
Abraxas, in system of Basilides, ii. 90, 92
Abydos, gods of, i. 33 n. 1;
excavations at, i. 36
Achaea, worship of Goddesses Twain in, i. 135;
Cilician pirates deported to, ii. 229
Achaemenides, Persian religion under, i. 122; ii. 234
Achamoth, Sophia of Ophites, ii. 45 n. 1;
called the Mother by Valentinus, ii. 112 n. 3;
the Sophia Without of Valentinus, ii. 117 n. 2;
baptism in name of, by Marcus, ii. 189 n. 1.
See Sophia (2)
Acheron, Isis shining in, i. 60
Achilles, his horror of Hades, i. 59, 150;
his flattery of Zeus, i. 95;
his purification by Ulysses, i. 121 n. 4
Achrammachamari, name of Great Propator in Texts of Saviour and
Mag. Pap., ii. 142 n. 2
Acropolis, sacred things of Eleusis lodged in, i. 39;
Serapeum built opposite, i. 52
Acrostics, use of, in Jewish, Greek and Christian literature, i. 169 n.
1;
in Valentinian epitaph, ii. 129 n. 3
Adam, the protoplast, Ophite story of, ii. 52, 58, 70;
and Manichaean, ii. 299;
and neo-Manichaean, ii. 329
Adam or Adamos, god of Samothrace, i. 139 n. 1; ii. 54 n. 6
Adamas, the Ophite, the First Man or Great Light, ii. 38;
gives birth to Second Man or Son, ibid.;
called Father-and-Son, ii. 39;
androgyne, ii. 40;
forms triad with Holy Spirit, ii. 41 nn. 2, 3;
all things except matter contained in, ii. 44 n. 2, 64;
all light returns to, ii. 65, 80;
called Caulacau, ii. 94 n. 3.
See First Man, Caulacau, Hades
Adamas, king of the Twelve Aeons in Pistis Sophia, his rebellion, ii.
48 n. 4, 152 n. 1;
place of, ii. 137 n. 3;
ruler of Zodiac, ii. 152;
delays redemption of souls, ii. 153;
sends demon in shape of flying arrow, ii. 156;
probably Diabolos or Cosmocrator of Valentinus, ii. 163.
See Sabaoth Adamas
Adamas of the Light, in neo-Manichaeism, ii. 325;
slayer of monster, ii. 329
Adonai, epithet of Zeus in Mag. Pap., i. 106;
in Coptic, ii. 46 n. 3;
son of Ophite Sophia, ii. 47;
ruler of planetary sphere in Diagram, ii. 69;
meaning of name of, ii. 71 n. 1;
address of soul to, ii. 72
Adonis, wailed for in Athens, temp. Alcibiades, i. 16;
Dying God of Mediterranean, i. 37;
Asiatic form of Dionysos, i. 47;
identified with Osiris, i. 55;
identified with Dionysos by Orphics, i. 137, 145;
identified with Dionysos at Eleusis, i. 139 n. 1;
androgyne, i. 185;
Ophites attend mysteries of, ii. 21, 54;
identified with Phrygian god, ii. 31;
fiend in hell in Texts of Saviour, ii. 186
Advent, the. See Parusia
Aegean, islands of, birthplace of gods, i. 16, 52;
early worship of Alexandrian gods in, i. 52;
and of Eleusinian, ii. 135
Aeinous or Aionios (Everlasting), member of Valentinian Dodecad, ii.
101
Aelius Aristides, quoted, i. 55 n. 2, 58, 60, 64 n. 3; ii. 66 n. 2
Aeon, Thirteenth, highest place of Left in P.S., ii. 143, 150;
Authades would-be ruler of, ii. 151, 153;
first dwelling-place of Pistis Sophia, ii. 155;
place below it made for Pistis Sophia, ii. 155, 156;
Pistis Sophia restored to, ii. 157
Aeons, the Twelve, described, ii. 143, 152, 153;
souls made from tears of rulers of, ii. 153;
Jesus takes away part of their power, ii. 154;
divided into repentant and unrepentant, ii. 182;
the mystery of, in Bruce Papyrus, ii. 195.
See Zodiac
Aerodios, power mentioned in Bruce Papyrus, ii. 191
Aeschines, son of Glaucothea, i. 22;
Demosthenes’ invective against, quoted, i. 138.
See Sabazius
Aeschylus, quoted, i. 48, 55, 123
Aether, offspring of Time ap. Orphics, i. 123
Afghanistan, included in Persian Empire, i. 1
Africa, political power of priesthoods in, i. 31;
Mithraism in Northern, ii. 230;
christianized Manichaeism of, ii. 339
Agape or Love, supreme God of Diagram, ii. 68, 123 n. 3;
supreme God of Valentinus, ii. 98 n. 1;
feminine member of Valentinian Dodecad, ii. 101;
supreme God of Marcion, ii. 210;
seal of Azrua in neo-Manichaeism, ii. 341.
See Eros
Agdistis, name of androgyne Cybele, ii. 39, 40
Ageratos or Never-ageing, member of Valentinian Decad, ii. 101
Agla, cabalistic word used in mediaeval magic, ii. 139 n. 1
Agra on the Ilissus, mysteries of, i. 41
Agrestius, a clarissimus and high priest of Mithras, ii. 239
Ahnas-el-Medineh or Heracleopolis, mentioned in magic spell, i. 98
Ahriman, Areimanios, or Arimanius, in Bundahish slayer of
Gayômort, i. 126 n. 3;
not entirely evil till Sassanid times, ii. 232, 253;
Magi sacrifice to, ii. 234;
son of Zervan Akerene (Cumont), ii. 236, 252;
altars dedicated to, ii. 239;
Mithras superior to, ii. 240;
in Bundahish slayer of bull Goshurun, ii. 246, 254;
ruler of earth in Mithraism, ii, 255, 256;
modified worship of, in Mithraism, ii. 278;
likeness of representation of, to Manichaean Satan, ii. 291.
See Goshurun
Ahura Mazda, the Omniscient Lord, i. liii;
father of Gayômort, i. lxi;
Supreme Being of Yashts, ii. 231;
his relations to Amshaspands, ii. 232;
in Behistun inscription, ii. 233;
not mentioned in Mithraic monuments, ii. 239;
in Bundahish, ii. 246;
replaced by Jupiter O.M. in Mithraism, ii. 246;
worship of, restored by Ardeshîr, ii. 284
Ailoaios or Eloaeus, ruler of planetary sphere in Diagram, ii. 69, 70 n.
2, 74 n. 3;
address to, ii. 73;
sphere of Venus, 74 n. 1
Akae, cryptographic name in Book of Enoch, i. 169, 170
Akinetos or Immovable, member of Valentinian Decad, ii. 101
Albigenses, successors of Manichaeans, ii. 357
Al-Bîrûnî, quoted, ii. 279, 280, 283, 284, 286 n. 1, 307
Alcibiades, goes to Susa, i. 7;
Adonis wailed for when Sicilian expedition of, sails, i. 16
Alcmaeon of Crotona, calls stars gods, i. 186 n. 2
Aldabeim, name of sun in Mag. Pap., ii. 46 n. 3
Aletheia, member of 1st Valentinian syzygy, ii. 98
Alexander of Abonoteichos, his impostures, i. 24;
comes to Rome under Marcus Aurelius, ii. 203
Alexander, King of Epirus, Asoka’s mission to, i. 20
Alexander the Great, his conquests hellenize Mediterranean Basin, i.
lviii;
the world before and after, i. 1 sqq.;
greatest individual in history, i. 4, 12 sqq.;
his aims and achievements, i. 5-8, 26-27;
his deification explained, i. 18;
religious associations follow conquests of, i. 22-26 sqq.;
his work in Egypt, i. 29, 44;
his probable plans for universal religion, i. 30;
breaks down national barriers, i. 54, 107;
makes world-religions possible, i. 111;
his conduct towards Jews, i. 150;
re-settles Samaria, i. 177;
son of Zeus in serpent form, ii. 49;
his effect on cosmology and ethics, ii. 86;
consoled by Anaxarchus for death of Clitus, ii. 87.
See India
Alexander, bishop of Lycopolis, quoted, ii. 294 n. 2, 295 n. 2.
Alexander Severus, the Emperor, gods in lararium of, i. 82;
his success against Persians, ii. 226
Alexander the Valentinian, leader of Anatolic School, ii. 119
Alexandria, its foundation by Alexander, i. 5;
its importance not at first recognized, i. 28;
Sema of Alexander at, i. 30;
a Greek city, i. 44;
Serapeum of, i. 48, 51, 58 n. 1;
oracle of Serapis at, i. 77;
worship of Serapis at, i. 82 n. 2, 86;
destruction of temples at, by Theodosius, i. 83, 84;
Hadrian’s opinion of, i. 86;
early Gnostics start from, i. 111; ii. 8;
Orphics plentiful at, i. 156;
Simon Magus’ doctrines at, i. 198; ii. 89;
intellectual centre of Roman world, ii. 88;
Basilides teaches at, ii. 90
Alfenius Julianus Kumenius, clarissimus and priest of Mithras, ii. 268
Allat, the goddess, Ereshkigal an epithet of, i. 100
Alleius Craeonius, author on magic, i. 105
Amazons, the story of, suggests bisexual deity, ii. 40
Ambrose of Milan, convert from Valentinianism, i. 112 n. 1; ii. 21 n. 5;
his date, ii. 132 n. 2
Amélineau, E., translates Pistis Sophia, ii. 13;
translates Bruce Papyrus, ii, 190;
his date for Bruce Papyrus, ii. 194;
quoted, ii. 178, 191, 192, 193, 195
Amen of Thebes, the god, father of Alexander, i. 18;
priesthood of, i. 23, 31 sqq.;
Ptolemies raise temples to, i. 52
Amenhotep IV, King of Egypt, failure of monotheistic teaching of, i.
11;
priests of Amen crush heresy of, i. 31
Amens, the Three, powers mentioned in P.S., ii. 142;
and in Bruce Papyrus, ii. 193
Amens, the Seven, powers mentioned in P.S., ii. 141;
and in Bruce Papyrus, ii. 193
Amenti, the god called Lord of, i. 33;
Osiris the bull of, i. 45, 102;
Jesus the conqueror of, i. 102 n. 1;
a hell in Texts of Saviour, ii. 182, 186;
horrors of Egyptian, ii. 196.
See Aberamenthôu, Jesus, Khent-Amentit
Ameretât or Immortality, one of the Amshaspands, i. 181 n. 1; ii. 324
n. 4, 355.
See Appellant and Respondent
Amitrochates, son of Chandragupta, his desire for Greek learning, i.
8 n. 3
Amon. See Amen of Thebes
Amos, the Prophet, inspired by Ialdabaoth ap. Ophites, ii. 81 n. 2
Amshaspands, the Seven, and the Seven Planets, i. 117;
names of, i. 181, n. 1; ii. 103 n. 3;
likeness of, to “roots” of Simon Magus, ii. 103 n. 3;
and to Aeons of Valentinus, ii. 103 n. 3;
in Avesta, ii. 232;
absent from early Manichaeism, ii. 327 n. 4;
mention of, in neo-Manichaeism, ii. 330, 355
Anat, the goddess, assessor of Yahweh, ii. 32 n. 4
Anatolia, its religious peculiarities temp. Christ, ii. 28 sqq., 77;
its worship of double axe, 67 n. 3
Anaxarchus the Atomist philosopher, consoles Alexander after death
of Clitus, ii. 87
Ancient of Days, name of Valentinian Ialdabaoth, ii. 107 n. 2
Andrew the Apostle, Saint, name of, shows predilection of Jews for
Greek names, i. 173 n. 2;
mentioned in Pistis Sophia, ii. 157
Anebo, letter of Porphyry to, for threats of Egyptian magicians to
gods, i. 104 n. 3
Angels, Essenes sworn to preserve the names of, i. 153, 157;
no names of, in O.T., until Daniel, i. 158;
rulers over tribes of demons, ibid.;
sinning, cast into abyss of fire (Baruch), i. 165;
Ennoia produces world-making, ap. Simon M., i. 187;
patterns after which worlds made (Philo), i. 187 n. 3;
world to be freed from rule of, ap. Simon, i. 196;
Simonians say God of Jews one of world-making (Epiphanius), i.
199;
seven heavens are also, ap. Valentinians (Irenaeus), ii. 107 n. 4;
are Logoi sent into soul by Jesus and Sophia, ii. 110;
souls after death, brides of, ibid.;
terror of angels at speech of man (Valentinus), ii. 112 n. 3;
Archons of Adamas in Texts of Saviour beget, ii, 152 n. 1;
Splenditenens and Atlas of Manichaeism, ii. 297, 298.
See Enoch, Gabriel, Great Council, Michael, Tertullian
Annu or On, Egyptian name of Heliopolis and chief seat of worship of
Râ, i. 31
Anthesteria, ceremonies of, show resurrection and marriage of
Dionysos, i. 42
Anthropos, member of 3rd Valentinian syzygy, ii. 98
Antigonus Monophthalmos, King of Syria, his retort when hailed as a
god, i. 19;
Phrygia occupied alternately by him and Lysimachus, ii. 29
Antigonus Gonatas, King of Macedonia, mission of Asoka to, i. 20.
See Tarn
Antinous, death of, fixes date of Hadrian’s letter to Servian, i. 86 n. 5
Antioch, worship of Serapis at, i. 35;
birthplace of Carpocrates, i. 111;
and of Saturninus, ii. 9, 89
Antiochus I Soter, King of Syria, mission of Asoka to, i. 20
Antiochus III the Great, King of Syria, seizes Palestine, i. 151;
transports Jewish families to Anatolia, ii. 28
Antiochus IV Epiphanes, King of Syria, attempts to hellenize Jews, i.
151, 156, 162, 163;
Book of Daniel written temp., i. 158;
caught between Romans and Parthians, i. 160;
his mystic antagonist Taxo, i. 170;
Samaritans accept reforms of, i. 177
Antitheses, the. See Marcion
Antonines, the, Isis-worship at its apogee temp., i. 54, 81
Antoninus Pius, the Emperor, Simonians in Rome temp., i. 199
Anubis, the god, son of Osiris and Nephthys, i. 35;
tribal deity of jackal totem, i. 36;
his seeking for Osiris in Rome, i. 70;
in procession at Cenchreae, i. 72;
mask of, used as disguise, i. 78.
See Marcus Volusius
Apelles, the Marcionite, his tenets, ii. 218
Apep, the serpent, enemy of the sun-god Ra, ii. 78
Aphrodite, the goddess, worshipped under other names by
confraternities, i. 25;
and Adonis, i. 37; ii. 31;
daughter of Zeus, i. 124 n. 3;
identified by Orphics with Isis and others, i. 137 n. 1;
Orphic hymn to, i. 142 n. 2;
called Cytheraea, i. 143;
the Mother of the Gods in Cyprus, ii. 40;
called Mother of All Living in Asia, ii. 135 n. 3;
on Mithraic monuments, ii. 238.
See Venus
Apis, the “life” of Osiris, i. 32, 45, 49
Apocalypse of St John, the, its date, ii. 26 n. 3;
quoted, i. 145 n. 1, 158, 169, 182 n. 4; ii. 4 n. 1, 25
Apocatastasis, return of the worlds to God, an Ophite doctrine, ii. 42,
57
Apollo, the god, his birthplace, i. 16;
identified with Horus, i. 48, 63;
his contempt for mankind, i. 57;
his place in Orphic legend, i. 125, 147;
on Mithraic monuments, ii. 238;
distinct from Helios, i. 240;
worship of, under Julian, i. 269
Apollonius of Tyana, image of, in Alexander Severus’ lararium, i. 82
Apophasis of Simon Magus, the, described, i. 179;
quoted, i. 182, 188, 189, 193, 194; ii. 90 n. 5
Apostles, demand only faith from converts, i. lvii;
do not borrow from earlier creeds, i. 88;
their meeting with Simon Magus, i. 176, 177;
in Clementines, i. 178;
intolerance of, due to Jewish origin (Bouché-Leclercq), ii. 10;
souls of, in P.S. drawn from Treasure-house, ii. 137, 147
Apostolical Constitutions, their date, ii. 7 n. 2;
quoted, i. 87 n. 1; ii. 7 nn. 2, 3, 219 n. 2
Appellant and Respondent gods, the, in neo-Manichaeism, ii. 302 n.
1, 324, 343, 354, 355
Apuat, the god, “opener of the ways,” i. 33
Apuleius of Madaura, quoted, i. 56, 59, 60, 61, 62, 63, 64, 65, 66, 67,
68, 71, 73-74, 75, 77, 86 n. 3, 101 n. 2
Aramati, the Vedic goddess, identified with Spenta Armaiti of the
Avesta, ii. 45 n. 1, 300 n. 2
Ararat or Ararad, Mt, Books of Jeû hidden in, ii. 147 n. 5
Arbela, Greek troops on Persian side at, i. 7;
Alexander’s pursuit after, i. 13
Arcadia, Eleusinian triad worshipped in, i. 135
Arcadius, the Emperor, Church dedicated to, in place of Serapeum, i.
84
Archelai Acta. See Hegemonius
Archimedes, his calculation of places of stars sinful (Hippolytus), i.
112 n. 2
Architect of the Universe. See Demiurge
Archon, the Great, of Basilides, the Demiurge, ii. 91;
likeness of, to Ialdabaoth, ii. 94
Archontics, the, a sect related to the Ophites, ii. 77
Arctinus of Miletus, first Greek author to mention purification, i. 121
n. 4
Arctos, the Great Bear, in Mithraism, ii. 266
Arda viraf namak, the, quoted, ii. 264 n. 5
Ardeshîr, the Shah, restorer of Persian nationality, ii. 226, 282;
his son Peroz converted to Manichaeism, ii. 281;
restores worship of Ahura Mazda, ii. 284
Ares, the god, identified with Roman Mars, i. 17;
Homeric or Orphic hymn to, i. 141 n. 2, 142 n. 2;
on Mithraic monuments, ii. 238
Argolis, the, Eleusinian triad worshipped in, i. 135
Ariel, a fiend in Texts of Saviour, ii. 186
Arimaspi, the, fables concerning, i. 2 n. 1
Aristaeus, pro-Jewish writer, i. 173
Aristides. See Aelius Aristides
Aristides, Christian apologist, ii. 203, 204 n. 1
Aristion, Athenian courtezan member of religious confraternity, i. 22
Aristophanes, quoted, i. 17 n. 1, 40 n. 4, 124, 137;
scholiast on, i. 17 n. 1
Aristotle, his monotheism, i. 10;
says that religion follows form of government, i. 12, 15;
that Orpheus did not exist, i. 121 n. 1
Armageddon, covers name of Rome, i. 170 n. 5
Armenia, Ophites in, ii. 76;
kings of, claim descent from Persian heroes, ii. 225 n. 1;
Marcionites and Bardesanites in, ii. 283;
invasion of, by Mihr Nerses, ii. 285
Arnobius, adv. Gentes, quoted, i. 124 n. 3; ii. 39 nn. 2, 4, 264 n. 5
Arrian, Anabasis, quoted, i. 4 n. 1
Arsaces, founder of Parthian kingdom, ii. 224
Arsinoe, wife of Ptolemy Philadelphus, i. 18
Artapanus, On the Jews, quoted, i. 173
Artemis, the goddess, the Ephesian, i. lvi, 40;
birthplace of Greek, i. 16;
Indian worship of, i. 17;
Orthia, i. 100 n. 2;
priestesses of Ephesian, called bees, i. 143 n. 4;
Phrygian, ii. 67 n. 3;
See Diana
Aryans, their dealings with lower races, i. 3, 92
Asar-hapi, Osiris as Apis, i. 49
Asha Vashishta or Truth, the Amshaspand, i. 181 n. 1
Asia, before Alexander, i. 1;
made Greek by Alexander, i. 5;
rush of Greeks to, i. 7;
Greek spoken throughout, i. 8;
cruelty of Assyrian domination in, i. 12;
returns to Persian ways, ii. 225
Asia Minor, native religions of, i. lviii, 37, 126; ii. 29, 36, 49, 67 n. 3;
gods of, coalesce with Greek, i. 17;
home of Dionysos worship, i. 43 n. 3;
Alexandrian gods worshipped in, i. 53;
Vedic gods worshipped in, i. 122 n. 3;
Eleusinian gods worshipped in, i. 136;
Orphics in, i. 141, 156; ii. 236;
priestesses called bees in, i. 143 n. 4;
Jewish atrocities in, temp. Trajan, i. 173 n. 1;
Ophite heresy probably native to, ii. 26, 76;
Jewish settlements in, ii. 28;
Jewish magicians in, temp. Apostles, ii. 33;
matriarchate in, ii. 40;
Babylonian culture in, ii. 48;
serpent worship in, ii. 49, 77, 78;
reverts to Persian ways, ii. 225;
Mithraism in, ii. 229, 232, 268
Askew, Dr, sells Pistis Sophia to British Museum, ii. 134
Asklepios or Aesculapius, the god, Alexander of Abonoteichos priest
of, i. 24;
Serapis statue that of, i. 48 n. 3, 78 n. 2;
identified with Serapis, i. 78, 87
Aso, the Ethiopian queen, enemy of Osiris, i. 33, 37 n. 1
Asoka, his missions to Greek kings, i. 20
Assur-bani-pal, King of Assyria, his library at Kuyunjik, i. 94, 114
Assyria, penitential psalms of, i. 115;
Jews tributary to, i. 160 n. 4
Assyrians, the, tyranny of, i. 3;
suzerains of Hebrews, i. 150;
name used for Syrians in Christian times, ii. 53 n. 4, 54 n. 6;
worship of Mylitta by (Herodotus), ii. 234
Astaphaios or Astaphaeus, ruler of planetary sphere in Diagram, ii.
47;
name derived from magic (Origen), ii. 47, 48;
once called Astanpheus, ii. 47, 69 n.
lord of third gate, ii. 70 n. 2, 73, 74 n. 3;
address to, ii. 73
Astarte, the goddess, worship of, brought into Greece, i. 17;
worshipped by Greek confraternity, i. 25;
Phoenician form of earth goddess, i. 126;
dove, totem-animal of, ii. 135 n. 3;
Mater Viventium, ibid.
Astrampsuchos, name of Roman writer on magic, i. 107;
name of celestial guard in Bruce Papyrus, i. 107 n. 1;
power worshipped by the Peratae, ibid.
Astrology, origin of, in Chaldaea, i. 113;
fundamental idea of, i. 114;
system of correspondences results from, i. 115, 116;
impulse given to, by Greek mathematics, i. 116, 117;
all religions in Graeco-Roman world take note of, i. 117, 118;
gives new life to Gnosticism, i. 119;
Ophites mix astrological ideas with Orphic teaching, ii. 78;
first prominent in Gnosticism in Excerpta Theodoti, ii. 158 n. 1;
its great vogue in Rome under Severi, ibid.;
reprobated in Pistis Sophia, ii. 185;
part of scheme of punishments and salvation in Texts of Saviour,
ii. 185 n. 2;
its importance in Mithraism, ii. 235, 276.
See Babylonia
Atargatis or Dea Syria, favourite deity of Nero, ii. 31;
her Anatolian name and identification with other goddesses, ii. 31
n. 1;
homonym of Derketo (Garstang), ii. 40 n. 1;
her identity with the Mother of the Gods, ii. 299 n. 1;
Manichaean Mother of Life derived from, ii. 300 n. 2
Athamas the Pythagorean, his doctrine of “roots,” i. 197
Athanasius, Saint, creed of, i. 89
Athena, the goddess, identified with Minerva, i. 17;
her part in Eleusinian Mysteries, i. 39;
the Homeric, i. 57, 95, 124 n. 3;
statue of Helena of Tyre as, i. 198;
See Minerva, Pallas
Athenagoras, quoted, i. lvii n. 1, 63 n. 5, 64 n. 3; ii. 18 n. 2
Athens, foreign worships in, i. 16, 17 n. 1, 137;
accepts deification of Alexander, i. 18;
gathering in, for Eleusinian Mysteries, i. 38-41;
Alexandrian religion in, i. 52, 76;
Orphic myths brought into, by Epimenides, i, 121;
Orphic gold plates in Museum at, i. 132
Athos Mt, Philosophumena discovered at, ii. 11
Atlas. See Corybas, Omophorus
Attis or Atys, the god, his worship brought into Greece, i. 17, 136;
his legend, i. 37; ii. 39;
identified with Sun, i. 118;
and with Dionysos, Adonis and Osiris, i. 137 n. 1, 145; ii. 17;
and with Sabazius, i. 138, 139;
androgyne, i. 185;
Gnostics attend mysteries of, ii. 21;
Phrygia, home of worship of, ii. 28, 67 n. 3;
to Ophites, type of world-soul, ii. 65 n. 3
Augustine of Hippo, Saint, convert from Manichaeism, i. 112 n. 1;
well informed about Manichaeans, ii. 352;
quoted, i. 103 n. 4; ii. 10 n. 1, 12 n. 4, 25, 261, 298 n. 1, 317, 319,
331, 332, 343, 346, 349 n. 4, 350
Augustus, the Emperor, Samaria’s capital named Sebaste in honour
of, i. 177;
Galatians become Roman temp., ii. 28;
Parthians’ terror of (Horace), ii. 225
Aurelian, the Emperor, his worship of sun-god, i. 119 n. 1; ii. 228;
position of Christianity under, ii. 23;
restores Roman arms in the East, ii. 226;
gives up Dacia to Goths, ii. 271
Authades, the Proud God of the Pistis Sophia, last member of Triad
of the Left, ii. 151;
his disobedience, ii. 152;
his envy of Pistis Sophia, ii. 155;
sends demon in shape of flying arrow, ii. 156;
his place given to Pistis Sophia, ii. 162
Autogenes, power mentioned in Bruce Papyrus, ii. 192
Autophyes or Self-produced, member of Valentinian Decad, ii. 101
Avebury, Lord, quoted, i. 91, 99 n. 1
Avesta, the Zend, Seven Amshaspands of, i. 117;
emanation doctrine in, ii. 35;
First Man in, ii. 38 n. 3;
Supreme Being in, ii. 231;
Ormuzd and Ahriman in, ii. 236;
bull Goshurun in, ii. 243;
denounces magic, ii. 275 n. 2;
doubtful about eternity of evil, ii. 289;
quoted, ii. 310, 311.

Textbook Mathematical Analysis and Applications Plenary Lectures Isaac 2017 Vaxjo Sweden Luigi G Rodino Ebook All Chapter PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Textbook Mathematical Analysis and Applications Plenary Lectures Isaac 2017 Vaxjo Sweden Luigi G Rodino Ebook All Chapter PDF

Uploaded by

Copyright:

Available Formats

Mathematical Analysis and

Applications Plenary Lectures ISAAC

Analysis Probability Applications and Computation

Mathematical Analysis and Applications 1st Edition

Language and Automata Theory and Applications 11th

Mathematical Analysis and Applications Selected Topics

Current Trends in Mathematical Analysis and Its

Sustainable Lignin for Carbon Fibers Principles

Mathematical Morphology and Its Applications to Signal

An Introduction to Proofs and the Mathematical

More information about this series at http://www.springer.com/series/10533

ISSN 2194-1009 ISSN 2194-1017 (electronic)

© Springer Nature Switzerland AG 2018

This volume is a collection of articles devoted to current research topics in

Turin, Italy Luigi G. Rodino

Solvability of Subprincipal Type Operators . . . . . . . . . . . . . . . . . . . . . . 1

Abstract In this paper we consider the solvability of pseudodifferential operators in

We shall consider the solvability for a classical pseudodifferential operator P ∈

© Springer Nature Switzerland AG 2018 1

has a local solution u ∈ D (X ) in a neighborhood of K for any v ∈ C ∞ (X ) in a set of

Im apm does not change sign from − to +

for any 0 = a ∈ C ∞ (T ∗ X ). The oriented bicharacteristics are the positive flow of

is invariantly defined modulo S m−2 under changes of coordinates, see Theorem

Let σ(P) = p ∈ Shom

2 = { η = 0 } (ξ, η) ∈ Rn−d × Rd ξ = 0 (2.3)

and κ(ω) = minw∈ω κ(w) for ω ⊆ 2 , which is equal to ∞ when p vanishes of

is invariantly defined on 2 under conjugation with elliptic Fourier integral operators.

Nw∗ 2  (w, η) → Jwk ( p)(η) = ∂ηk p(w)(η) (2.7)

N ∗ 2  (w, η) → ps,k (w, η) = Jwk ( p)(η) + ps (w) w∈ω (2.8)

which is a polynomial in η of degree k and is given by the blowup of the refined

Remark 2.2 The reduced subprincipal symbol is well-defined up to nonvanishing

Lemma 2.1 Assume that f (x, y, ζ) ∈ C ∞ is a polynomial in ζ of degree m for

|∂ y f | ≤ C0 | f | in when |ζ − ζ0 | < c (2.10)

for c > 0 implies that

f (x, y, ζ) = c(x, y, ζ) f 0 (x, ζ) in when |ζ − ζ0 | < c (2.11)

where 0 < c0 ≤ c(x, y, ζ) ∈ C ∞ and f 0 (x, ζ) = f (x, y0 (x), ζ) ∈ C ∞ , which

f (x, y, ζ) = c(x, y, ζ) f 0 (x, ζ) in x for |ζ − ζ0 | < c (2.12)

when (x, ζ) ∈  \ 0 . Here f 0 (x, ζ) = f (x, y0 (x), ζ) ∈ C ∞ and 0 < c0 ≤

f (x, y, ζ) = c± (x, y, ζ) f (x, y0 (x), ζ) in S± (2.13)

where 0 < c0 ≤ c± (x, y, ζ) ∈ C ∞ is uniformly bounded. By taking the limit of

If ps,k is constant in y in a neighborhood of the semibicharacteristic, then the

which is a symplectic space over 2 which in these coordinates is parametrized by

This is isomorphic to the symplectic manifold T ∗ Rn−d with w ∈ 2 as parameter.

H ps,k = ∂ξ ps,k ∂x − ∂x ps,k ∂ξ (2.16)

modulo ∂ y , which is nonvanishing if ∂x,ξ ps,k = 0. Thus H ps,k is well-defined modulo

Definition 2.2 We say that the operator P is of subprincipal type on N ∗ 2 if the

∂η ps,k = J k−1 (∂η p) = J k−1 (∂ p) (2.18)

V1 · · · V Hess p ≤ C | ps,k |ε on ω (2.21)

Definition 2.5 If k = κ(ω) < ∞ is the order of p on an open set ω ⊆ 2 then we

N ∗ 2  (w, η) → qs,k (w, η, λ) = λJw2k−1 ( p)(η/λ1/k )

which is a weighted polynomial in η of degree 2k − 1. When κ(ω) = ∞ we define

By the invariance of p and ps , the extended subprincipal symbol transforms as jets

λ2−m psub (x, y; λξ, λ1−1/k η) ∼

We also have that

λ2−m ∂η psub (x, y; λξ, λ1−1/k η) ∼

modulo O(λ1/k−1 ) and

λ2−m ∂η2 psub (x, y; λξ, λ1−1/k η) ∼

∂η qs,k = O(λ−2/k ) when ps,k = 0 at ω for |η − η0 | < c0 and λ  1 (2.26)

for any leaf L of N ∗ 2 where ω ⊂ 2 . By letting λ → ∞ we obtain that (2.9)

qs,k (x, y; ξ, η, λ) = c(x, y; ξ, η, λ)

equal to A∂ y , A = 0, modulo terms that are O(|η|), so the normalized Hamilton

Definition 2.6 If K ⊂ S ∗ X is a compact set, then we say that P is microlocally

of P. To prove Theorem 2.1 we shall use the following result.

has a local solution u ∈ D (X ) in a neighborhood of K for any v ∈ C ∞ (X ) in a set of

for any 0 = a ∈ C ∞ (T ∗ X ). The oriented bicharacteristics are the positive flow of

2 = { η = 0 } (ξ, η) ∈ Rn−d × Rd ξ = 0 (2.3)

Nw∗ 2 (w, η) → Jwk ( p)(η) = ∂ηk p(w)(η) (2.7)

N ∗ 2 (w, η) → ps,k (w, η) = Jwk ( p)(η) + ps (w) w∈ω (2.8)

where 0 < c0 ≤ c(x, y, ζ) ∈ C ∞ and f 0 (x, ζ) = f (x, y0 (x), ζ) ∈ C ∞ , which

when (x, ζ) ∈ \ 0 . Here f 0 (x, ζ) = f (x, y0 (x), ζ) ∈ C ∞ and 0 < c0 ≤

modulo ∂ y , which is nonvanishing if ∂x,ξ ps,k = 0. Thus H ps,k is well-defined modulo

V1 · · · V Hess p ≤ C | ps,k |ε on ω (2.21)

N ∗ 2 (w, η) → qs,k (w, η, λ) = λJw2k−1 ( p)(η/λ1/k )

∂η qs,k = O(λ−2/k ) when ps,k = 0 at ω for |η − η0 | < c0 and λ 1 (2.26)

qs,k (x, y; ξ, η, λ) = c(x, y; ξ, η, λ)

equal to A∂ y , A = 0, modulo terms that are O(|η|), so the normalized Hamilton

Remark 2.7 If P is microlocally solvable at ⊂ S ∗ X , then Lemma 26.4.5 in [13]

u(−N ) ≤ C(P ∗ u(ν) + u(−N −n) + Au(0) ) u ∈ C0∞ (Y ) (2.29)

P = Dt + ia(t) y (x, y) ∈ Rn−d × Rd (3.1)

where Bα is homogeneous of degree 2 − k, and Bα (x, y; ξ, 0) ≡ 0 for some |α| = k

χ : (x, y; ξ, η) → (x, y; ξ, η/|ξ|1/k ) (4.7)

which is a bijection when |ξ| = 0. The pullback by χ maps symbols in S( m , gk )

|∂ y qs,k | ≤ C|qs,k | when |ξ| 1 (4.9)

in a conical neighborhood ω in N ∗ 2 containing × { η0 }. By shrinking ω we may

where ∂η C −1 ∈ S( 1/k−1 , gk ) when |η| |ξ|1−1/k . Thus by multiplying P with a

By putting q = psub ◦ χ obtain that ∂ y q = 0 in ω when |ξ| 1. We shall control the

invariance that t → Im (t, x, 0, ξ, η0 ) changes sign from + to − near . Observe

locally for η close to η0 when |ξ| = 1, and by a partition of unity near × { η0 },