Download the Full Version of textbook for Fast Typing at textbookfull.

com

Foundations of Quantum Programming 1st Edition

Ying

https://textbookfull.com/product/foundations-of-quantum-
programming-1st-edition-ying/

OR CLICK BUTTON

DOWNLOAD NOW

Download More textbook Instantly Today - Get Yours Now at textbookfull.com

 Recommended digital products (PDF, EPUB, MOBI) that
you can download immediately if you are interested.

Mathematical foundations of quantum mechanics Neumann

https://textbookfull.com/product/mathematical-foundations-of-quantum-
mechanics-neumann/

textboxfull.com

Foundations of Programming Languages Second Edition Lee

https://textbookfull.com/product/foundations-of-programming-languages-
second-edition-lee/

textboxfull.com

Foundations of Quantum Field Theory 3rd Edition Klaus

Dieter Rothe

https://textbookfull.com/product/foundations-of-quantum-field-
theory-3rd-edition-klaus-dieter-rothe/

textboxfull.com

Beyond Spacetime The Foundations of Quantum Gravity 1st

Edition Nick Huggett (Editor)

https://textbookfull.com/product/beyond-spacetime-the-foundations-of-
quantum-gravity-1st-edition-nick-huggett-editor/

textboxfull.com
Foundations of Programming Languages Kent D. Lee

https://textbookfull.com/product/foundations-of-programming-languages-
kent-d-lee/

textboxfull.com

Quantum Computing Fundamentals From Basic Linear Algebra

to Quantum Programming 1st Edition Easttom Ii

https://textbookfull.com/product/quantum-computing-fundamentals-from-
basic-linear-algebra-to-quantum-programming-1st-edition-easttom-ii/

textboxfull.com

Foundations of Quantum Theory From Classical Concepts to

Operator Algebras 1st Edition Klaas Landsman

https://textbookfull.com/product/foundations-of-quantum-theory-from-
classical-concepts-to-operator-algebras-1st-edition-klaas-landsman/

textboxfull.com

Foundations of Quantum Theory From Classical Concepts to

Operator Algebras 1st Edition Klaas Landsman

https://textbookfull.com/product/foundations-of-quantum-theory-from-
classical-concepts-to-operator-algebras-1st-edition-klaas-landsman-2/

textboxfull.com

Spectral Theory and Quantum Mechanics Mathematical

Foundations of Quantum Theories Symmetries and
Introduction to the Algebraic Formulation Valter Moretti
https://textbookfull.com/product/spectral-theory-and-quantum-
mechanics-mathematical-foundations-of-quantum-theories-symmetries-and-
introduction-to-the-algebraic-formulation-valter-moretti/
textboxfull.com
Foundations of
Quantum
Programming
 Foundations of
Quantum
Programming

Mingsheng Ying
University of Technology Sydney and Tsinghua University

AMSTERDAM • BOSTON • HEIDELBERG • LONDON

NEW YORK • OXFORD • PARIS • SAN DIEGO
SAN FRANCISCO • SINGAPORE • SYDNEY • TOKYO
Morgan Kaufmann is an imprint of Elsevier
Morgan Kaufmann is an imprint of Elsevier
50 Hampshire Street, 5th Floor, Cambridge, MA 02139, USA

Copyright © 2016 Elsevier Inc. All rights reserved.

No part of this publication may be reproduced or transmitted in any form or by any means, electronic or
mechanical, including photocopying, recording, or any information storage and retrieval system, without
permission in writing from the publisher. Details on how to seek permission, further information about
the Publisher’s permissions policies and our arrangements with organizations such as the Copyright
Clearance Center and the Copyright Licensing Agency, can be found at our website:
www.elsevier.com/permissions.

This book and the individual contributions contained in it are protected under copyright by the Publisher
(other than as may be noted herein).

Notices
Knowledge and best practice in this field are constantly changing. As new research and experience
broaden our understanding, changes in research methods, professional practices, or medical treatment
may become necessary.

Practitioners and researchers must always rely on their own experience and knowledge in evaluating and
using any information, methods, compounds, or experiments described herein. In using such information
or methods they should be mindful of their own safety and the safety of others, including parties for
whom they have a professional responsibility.

To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any
liability for any injury and/or damage to persons or property as a matter of products liability, negligence
or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in
the material herein.

British Library Cataloguing in Publication Data

A catalogue record for this book is available from the British Library

Library of Congress Cataloging-in-Publication Data

A catalog record for this book is available from the Library of Congress

ISBN: 978-0-12-802306-8

For information on all Morgan Kaufmann publications,

visit our website at https://www.elsevier.com/

Publisher: Todd Green

Acquisition Editor: Todd Green
Editorial Project Manager: Lindsay Lawrence
Production Project Manager: Punithavathy Govindaradjane
Designer: Greg Harris

Typeset by SPi Global, India

Preface

“Perhaps the quantum computer will change our everyday lives in this century
in the same radical way as the classical computer did in the last century.”
— excerpt from press release, Nobel Prize in Physics 2012.

Quantum computers promise dramatic advantages over current computers.

Governments and industries around the globe are now investing large amounts
of money with the expectation of building practical quantum computers. Recent
rapid physical experimental progress has made people widely expect that large-
scalable and functional quantum computer hardware will be built within 10–20 years.
However, to realize the super-power of quantum computing, quantum hardware is
obviously not enough, and quantum software must also play a key role. The software
development techniques used today cannot be applied to quantum computers.
Essential differences between the nature of the classical world and that of the quan-
tum world mean that new technologies are required to program quantum computers.
Research on quantum programming started as early as 1996, and rich results
have been presented at various conferences or reported in various journals in the
last 20 years. On the other hand, quantum programming is still a premature subject,
with its knowledge base being highly fragmentary and disconnected. This book is
intended to provide a systematic and detailed exposition of the subject of quantum
programming.
Since quantum programming is still an area under development, the book does not
focus on specific quantum programming languages or techniques, which I believe
will undergo major changes in the future. Instead, the emphasis is placed on the
foundational concepts, methods and mathematical tools that can be widely used
for various languages and techniques. Starting from a basic knowledge of quantum
mechanics and quantum computation, the book carefully introduces various quantum
program constructs and a chain of quantum programming models that can effectively
exploit the unique power of quantum computers. Furthermore, semantics, logics,
and verification and analysis techniques of quantum programs are systematically
discussed.
With the huge investment and rapid progress in quantum computing technology,
I believe that within 10 years more and more researchers will enter the exciting field
of quantum programming. They will need a reference book as the starting point of
their research. Also, a course on quantum programming will be taught at more and
more universities. Teachers and students will need a textbook. So, I decided to write
this book with the two-fold aim:
(i) providing a basis for further research in the area; and
(ii) serving as a textbook for a graduate or advanced undergraduate level course.

ix
x Preface

Quantum programming is a highly interdisciplinary subject. A newcomer and, in

particular, a student is usually frustrated with the requisite knowledge from many
different subjects. I have tried to keep the book as self-contained as possible,
with details being explicitly presented so that it is accessible to the programming
languages community.
Writing this book gave me an opportunity to systemize my views on quantum
programming. On the other hand, topics included in this book were selected and
the materials were organized according to my own understanding of this subject, and
several important topics were omitted in the main body of the book due to my limited
knowledge about them. As a remedy, some brief discussions about these topics are
provided in the prospects chapter at the end of the book.
Acknowledgments

This book has been developed through my research in the last 15 years at the
Quantum Computation and Quantum Information Group of the State Key Laboratory
of Intelligent Technology and Systems, Tsinghua University and the Quantum
Computation Laboratory of the Centre for Quantum Computation and Intelligent
Systems, University of Technology Sydney. I have enjoyed very much collaborations
and discussions with my colleagues and students there. I would like to thank
all of them.
I am particularly indebted to Ichiro Hasuo (University of Tokyo) and Yuan Feng
(University of Technology Sydney) who patiently read the draft of this book and
kindly provided invaluable comments and suggestions. I am very grateful to the
anonymous reviewers for the book proposal; their suggestions were very helpful
for the structure of the book. I also would like to sincerely thank Steve Elliot,
Punithavathy Govindaradjane, Amy Invernizzi, and Lindsay Lawrence, my editors
and project managers at Morgan Kaufmann.
Special thanks go to the Centre for Quantum Computation and Intelligent Sys-
tems, Faculty of Engineering and Information Technology, University of Technology
Sydney for giving me the freedom to pursue my thoughts.
My research on quantum programming has been supported by the Australian
Research Council, the National Natural Science Foundation of China, and the
Overseas Team Program of the Academy of Mathematics and Systems Science,
Chinese Academy of Sciences. All of them are gratefully acknowledged.

xi
 CHAPTER

Introduction
1
“The challenge [of quantum software engineering] is to rework and extend the
whole of classical software engineering into the quantum domain so that program-
mers can manipulate quantum programs with the same ease and confidence that
they manipulate today’s classical programs.”
excerpt from the 2004 report Grand Challenges
in Computing Research [120].

Quantum programming is the study of how to program future quantum comput-

ers. This subject mainly addresses the following two problems:
• How can programming methodologies and technologies developed for current
computers be extended for quantum computers?
• What kinds of new programming methodologies and technologies can
effectively exploit the unique power of quantum computing?
Many technologies that have been very successful in traditional programming will
be broken when used to program a quantum computer, due to the weird nature of
quantum systems (e.g., no cloning of quantum data, entanglement between quantum
processes, and non-commutativity of observables which are all assertions about
program variables). Even more important and difficult is to discover programming
paradigms, models and abstractions that can properly exploit the unique power of
quantum computing – quantum parallelism – but cannot be sourced from knowledge
of traditional programming.

1.1 BRIEF HISTORY OF QUANTUM PROGRAMMING

RESEARCH
The earliest proposal for quantum programming was made by Knill in 1996 [139].
He introduced the Quantum Random Access Machine (QRAM) model and proposed
a set of conventions for writing quantum pseudo-code. In the 20 years since then,
research on quantum programming has been continuously conducted, mainly in the
following directions.
Foundations of Quantum Programming.http://dx.doi.org/10.1016/B978-0-12-802306-8.00001-X
Copyright © 2016 Elsevier Inc. All rights reserved.
3
4 CHAPTER 1 Introduction

1.1.1 DESIGN OF QUANTUM PROGRAMMING LANGUAGES

Early research on quantum programming focused on the design of quantum pro-
gramming languages. Several high-level quantum programming languages have
been defined in the later 1990s and early 2000s; for example, the first quantum
programming language, QCL, was designed by Ömer [177], who also implemented
a simulator for this language. A quantum programming language in the style
of Dijkstra’s guarded-command language, qGCL, was proposed by Sanders and
Zuliani [191,241]. A quantum extension of C++ was proposed by Bettelli et al. [39],
and implemented in the form of a C++ library. The first quantum language of the
functional programming paradigm, QPL, was defined by Selinger [194] based on
the idea of classical control and quantum data. A quantum functional programming
language QML with quantum control flows was introduced by Altenkirch and
Grattage [14]. Tafliovich and Hehner [208,209] defined a quantum extension of a
predicative programming language that supports the program development technique
in which each programming step is proven correct when it is made.
Recently, two general-purpose, scalable quantum programming languages,
Quipper and Scaffold, with compilers, were developed by Green et al. [106] and
Abhari et al. [3], respectively. A domain-specific quantum programming language,
QuaFL, was developed by Lapets et al. [150]. A quantum software architecture
LIQUi|>, together with a quantum programming language embedded in F#, was
designed and implemented by Wecker and Svore [215].

1.1.2 SEMANTICS OF QUANTUM PROGRAMMING LANGUAGES

Formal semantics of a programming language give a rigorous mathematical descrip-
tion of the meaning of this language, to enable a precise and deep understanding
of the essence of the language beneath its syntax. The operational or denotational
semantics of some quantum programming languages were already provided when
they were defined; for example, qGCL, QPL and QML.
Two approaches to predicate transformer semantics of quantum programs have
been proposed. The first was adopted by Sanders and Zuliani [191] in designing
qGCL, where quantum computation is reduced to probabilistic computation by the
observation (measurement) procedure, and thus predicate transformer semantics
developed for probabilistic programs can be applied to quantum programs. The
second was introduced by D’Hondt and Panangaden [70], where a quantum predicate
is defined to be a physical observable represented by a Hermitian operator with
eigenvalues within the unit interval. Quantum predicate transformer semantics was
further developed in [225] with a special class of quantum predicates, namely
projection operators. Focusing on projective predicates allows the use of rich
mathematical methods developed in Birkhoff-von Neumann quantum logic [42] to
establish various healthiness conditions of quantum programs.
Semantic techniques for quantum computation have also been investigated in
some abstract, language-independent ways. Abramsky and Coeck [5] proposed a
 1.1 Brief history of quantum programming research 5

category-theoretic formulation of the basic postulates of quantum mechanics, which

can be used to give an elegant description of quantum programs and communication
protocols such as teleportation.
Recent progress includes: Hasuo and Hoshino [115] found a semantic model of
a functional quantum programming language with recursion via Girard’s Geometry
of Interaction [101], categorically formulated by Abramsky, Haghverdi and Scott
[7]. Pagani, Selinger and Valiron [178] discovered a denotational semantics for
a functional quantum programming language with recursion and an infinite data
type using constructions from quantitative semantics of linear logic. Jacobs [123]
proposed a categorical axiomatization of block constructs in quantum programming.
Staton [206] presented an algebraic semantic framework for equational reasoning
about quantum programs.

1.1.3 VERIFICATION AND ANALYSIS OF QUANTUM PROGRAMS

Human intuition is much better adapted to the classical world than the quantum
world. This fact implies that programmers will commit many more faults in designing
programs for quantum computers than in programming classical computers. Thus,
it is crucial to develop verification techniques for quantum programs. Baltag and
Smets [30] presented a dynamic logic formalism of information flows in quantum
systems. Brunet and Jorrand [50] introduced a way of applying Birkhoff-von
Neumann quantum logic in reasoning about quantum programs. Chadha, Mateus and
Sernadas [52] proposed a proof system of the Floyd-Hoare style for reasoning about
imperative quantum programs in which only bounded iterations are allowed. Some
useful proof rules for reasoning about quantum programs were proposed by Feng
et al. [82] for purely quantum programs. A Floyd-Hoare logic for both partial and
total correctness of quantum programs with (relative) completeness was developed
in [221].
Program analysis techniques are very useful in the implementation and opti-
mization of programs. Termination analysis of quantum programs was initiated in
[227], where a measurement-based quantum loop with a unitary transformation as
the loop body was considered. Termination of a more general quantum loop with a
quantum operation as the loop body was studied in [234] using the semantic model
of quantum Markov chains. It was also shown in [234] that the Sharir-Pnueli-Hart
method for proving properties of probabilistic programs [202] can be elegantly
generalized to quantum programs by exploiting the Schrödinger-Heisenberg duality
between quantum states and observables. This line of research has been continued
in [152,153,235,236,238] where termination of nondeterministic and concurrent
quantum programs was investigated based on reachability analysis of quantum
Markov decision processes. Another line of research in quantum program analysis
was initiated by Jorrand and Perdrix [129] who showed how abstract interpretation
techniques can be used in quantum programs.
6 CHAPTER 1 Introduction

1.2 APPROACHES TO QUANTUM PROGRAMMING

Naturally, research on quantum programming started from extending traditional
programming models, methodologies and technologies into the quantum realm.
As stated in Section 1.1, both imperative and functional programming have been
generalized for quantum computing, and various semantic models, verification and
analysis techniques for classical programs have also been adapted to quantum
programming.
The ultimate goal of quantum programming is to fully exploit the power of
quantum computers. It has been well understood that the advantage of quantum
computers over current computers comes from quantum parallelism – superposition
of quantum states – and its derivatives such as entanglement. So, a key issue in
quantum programming is how to incorporate quantum parallelism into traditional
programming models. In my opinion, this issue can be properly addressed in the
following two paradigms of superposition.

1.2.1 SUPERPOSITION-OF-DATA – QUANTUM PROGRAMS WITH

CLASSICAL CONTROL
The main idea of the superposition-of-data paradigm is to introduce new pro-
gram constructs needed to manipulate quantum data, e.g., unitary transformations,
quantum measurements. However, the control flows of quantum programs in such
a paradigm are similar to those of classical programs. For example, in classical
programming, a basic program construct that can be used to define the control flow
of a program is the conditional (if . . . then . . . else . . . fi) statement, or more generally
the case statement:
if (i · Gi → Pi ) fi (1.1)

where for each i, the subprogram Pi is guarded by the Boolean expression Gi , and
Pi will be executed only when Gi is true. A natural quantum extension of statement
(1.1) is the measurement-based case statement:
if (i · M[q] = mi → Pi ) fi (1.2)

where q is a quantum variable and M a measurement performed on q with possible

outcomes m1 , . . . , mn , and for each i, Pi is a (quantum) subprogram. This statement
selects a command according to the outcome of measurement M: if the outcome is
mi , then the corresponding command Pi will be executed. It can be appropriately
called classical case statement in quantum programming because the selection of
commands in it is based on classical information – the outcomes of a quantum
measurement. Then other language mechanisms used to specify the control flow
of quantum programs, e.g., loop and recursion, can be defined based on this case
statement.
The programming paradigm defined here is called the superposition-of-data
paradigm because the data input to and computed by these programs are quantum
 1.2 Approaches to quantum programming 7

data – superposition of data, but programs themselves are not allowed to be

superposed. This paradigm can be even more clearly characterized by Selinger’s
slogan “quantum data, classical control” [194] because the data flows of the programs
are quantum, but their control flows are still classical.
The majority of existing research on quantum programming has been carried out
in the superposition-of-data paradigm, dealing with quantum programs with classical
control.

1.2.2 SUPERPOSITION-OF-PROGRAMS – QUANTUM PROGRAMS

WITH QUANTUM CONTROL
Inspired by the construction of quantum walks [9,19], it was observed in [232,233]
that there is a fundamentally different way to define a case statement in quantum
programming – quantum case statement governed by a quantum “coin”:
qif[c] (i · |i → Pi ) fiq (1.3)

where {|i} is an orthonormal basis of the state Hilbert space of an external “coin”
system c, and the selection of subprograms Pi ’s is made according to the basis states
|i of the “coin” space that can be superposed and thus is quantum information rather
than classical information. Furthermore, we can define a quantum choice:
 
 
[C] |i → Pi = C[c]; qif[c] (i · |i → Pi ) fiq (1.4)
i

Intuitively, quantum choice (1.4) runs a “coin-tossing” program C to create a

superposition of the execution paths of subprograms P1 , . . . , Pn , followed by a
quantum case statement. During the execution of the quantum case statement, each
Pi is running along its own path within the whole superposition of execution paths of
P1 , . . . , Pn . Based on this kind of quantum case statement and quantum choice, some
new quantum program constructs such as quantum recursion can be defined.
This approach to quantum programming can be termed the superposition-of-
programs paradigm. It is clear from the definitions of quantum case statement and
quantum choice that the control flow of a quantum program in the superposition-of-
program paradigm is inherently quantum. So, this paradigm can also be characterized
by the slogan “quantum data, quantum control”1.
I have to admit that this paradigm is still in a very early stage of development,
and a series of fundamental problems are not well understood. On the other hand,
I believe that it introduces a new way of thinking about quantum programming that
can help a programmer to further exploit the unique power of quantum computing.

1 The slogan “quantum data, quantum control” was used in [14] and in a series of its continuations
to describe a class of quantum programs for which the design idea is very different from that
introduced here.
8 CHAPTER 1 Introduction

1.3 STRUCTURE OF THE BOOK

This book is a systematic exposition of the theoretical foundations of quantum
programming, organized along the line from superposition-of-data to superposition-
of-programs. The book focuses on imperative quantum programming, but most ideas
and techniques introduced in this book can also be generalized to functional quantum
programming.
The book is divided into four parts:
• Part I consists of this introductory chapter and Chapter 2, Preliminaries. The
prerequisites for reading this book are knowledge of quantum mechanics and
quantum computation and reasonable familiarity with the theory of
programming languages. All prerequisites for quantum mechanics and quantum
computation are provided in Chapter 2. For theory of programming languages, I
suggest the reader consult the standard textbooks, e.g., [21,158,162,200].
• Part II studies quantum programs with classical control in the
superposition-of-data paradigm. This part contains three chapters. Chapter 3
carefully introduces the syntax and the operational and denotational semantics of
quantum programs with classical control (case statement, loop and recursion).
Chapter 4 presents a logical foundation for reasoning about correctness of
quantum programs with classical control. Chapter 5 develops a series of
mathematical tools and algorithmic techniques for analysis of quantum programs
with classical control.
• Part III studies quantum programs with quantum control in the
superposition-of-programs paradigm. This part consists of two chapters.
Chapter 6 defines quantum case statement and quantum choice and their
semantics, and establishes a set of algebraic laws for reasoning about quantum
programs with the constructs of quantum case statement and quantum choice.
Chapter 7 illustrates how recursion with quantum control can be naturally
defined using quantum case statement and quantum choice. It further defines the
semantics of this kind of quantum recursion with second quantization – a
mathematical framework for dealing with quantum systems where the number of
particles may vary.
• Part IV consists of a single chapter designed to give a brief introduction to
several important topics from quantum programming that have been omitted in
the main body of the book and to point out several directions for future research.
The dependencies of chapters are shown in Figure 1.1.
• Reading the Book: From Figure 1.1, we can see that the book is designed to be
read along the following three paths:
• Path 1: Chapter 2 → Chapter 3 → Chapter 4. This path is for the reader who
is mainly interested in logic for quantum programs.
• Path 2: Chapter 2 → Chapter 3 → Chapter 5. This path is for the reader who
is interested in analysis of quantum programs.
• Path 3: Chapter 2 → Chapter 3 → Chapter 6 → Chapter 7. This path is for
the reader who would like to learn the basic quantum program constructs in
 1.3 Structure of the book 9

FIGURE 1.1
Dependencies of chapters.

not only the superposition-of-data but also the superposition-of-programs

paradigms.
Of course, only a thorough reading from the beginning to the end of the book
can give the reader a full picture of the subject of quantum programming.
• Teaching from the Book: A short course on the basics of quantum programming
can be taught based on Chapters 2 and 3. Furthermore, Parts I and II of this book
can be used for a one- or two-semester advanced undergraduate or graduate
course. A one-semester course can cover one of the first two paths described
previously. Since the theory of quantum programming with quantum control (in
the superposition-of-programs paradigm) is still at an early stage of its
development, it is better to use Chapters 6 and 7 as discussion materials for a
series of seminars rather than for a course.
• Exercises: The proofs of some lemmas and propositions are left as exercises.
They are usually not difficult. The reader is encouraged to try all of them in order
to solidify understanding of the related materials.
• Research Problems: A couple of problems for future research are proposed at
the end of each chapter in Parts II and III.
• Bibliographic Notes: The last sections of Chapters 2 through 7 are bibliographic
notes, where citations and references are given, and recommendations for further
reading are provided. The complete bibliography is provided in a separate
section at the end of the book, containing the alphabetized list of both cited
references and those recommended for further reading.
• Errors: I would appreciate receiving any comments and suggestions about this
book. In particular, if you find any errors in the book, please email them to:
[email protected] or [email protected].
 CHAPTER

Preliminaries
2
This chapter introduces the basic concepts and notations from quantum mechanics
and quantum computation used throughout the book.
• Of course, quantum programming theory is built based on quantum mechanics.
So, Section 2.1 introduces the Hilbert space formalism of quantum mechanics,
which is exactly the mathematical knowledge base of this book.
• Quantum circuits are introduced in Section 2.2. Historically, several major
quantum algorithms appeared before any quantum programming language was
defined. So, quantum circuits usually serve as the computational model in which
quantum algorithms are described.
• Section 2.3 introduces several basic quantum algorithms. The aim of this section
is to provide examples for quantum programming rather than a systematic
exposition of quantum algorithms. Thus, I decided not to include more
sophisticated quantum algorithms.
In order to allow the reader to enter the core of this book – quantum program-
ming – as quickly as possible, I tried to make this chapter minimal. Thus, the
materials in this chapter are presented very briefly. Total newcomers to quantum
computation can start with this chapter, but at the same time I suggest that they read
the corresponding parts of Chapters 2, 4, 5, 6 and 8 of book [174] for more detailed
explanations and examples of the notions introduced in this chapter. On the other
hand, for the reader who is familiar with these materials from a standard textbook
such as [174], I suggest moving directly to the next chapter, using this chapter only
for fixing notations.

2.1 QUANTUM MECHANICS

Quantum mechanics is a fundamental physics subject that studies phenomena at
the atomic and subatomic scales. A general formalism of quantum mechanics can
be elucidated based on several basic postulates. We choose to introduce the basic
postulates of quantum mechanics by presenting the mathematical framework in
which these postulates can be properly formulated. The physics interpretations of
Foundations of Quantum Programming.http://dx.doi.org/10.1016/B978-0-12-802306-8.00002-1
Copyright © 2016 Elsevier Inc. All rights reserved.
11
12 CHAPTER 2 Preliminaries

these postulates are only very briefly discussed. I hope this provides the reader a
short cut towards a grasp of quantum programming.

2.1.1 HILBERT SPACES

A Hilbert space usually serves as the state space of a quantum system. It is defined
based on the notion of vector space. We write C for the set of complex numbers.
For each complex number λ = a + bi ∈ C, its conjugate is λ∗ = a − bi. We adopt
the Dirac notation which is standard in quantum mechanics: |ϕ, |ψ, . . . stands for
vectors.
Definition 2.1.1. A (complex) vector space is a nonempty set H together with two
operations:
• vector addition + : H × H → H
• scalar multiplication · : C × H → H
satisfying the following conditions:
(i) + is commutative: |ϕ + |ψ = |ψ + |ϕ for any |ϕ, |ψ ∈ H.
(ii) + is associative: |ϕ + (|ψ + |χ) = (|ϕ + |ψ) + |χ for any
|ϕ, |ψ, |χ ∈ H.
(iii) + has the zero element 0, called the zero vector, such that 0 + |ϕ = |ϕ for
any |ϕ ∈ H.
(iv) each |ϕ ∈ H has its negative vector −|ϕ such that |ϕ + (−|ϕ) = 0.
(v) 1|ϕ = |ϕ for any |ϕ ∈ H.
(vi) λ(μ|ϕ) = λμ|ϕ for any |ϕ ∈ H and λ, μ ∈ C.
(vii) (λ + μ)|ϕ = λ|ϕ + μ|ϕ for any |ϕ ∈ H and λ, μ ∈ C.
(viii) λ(|ϕ + |ψ) = λ|ϕ + λ|ψ for any |ϕ, |ψ ∈ H and λ ∈ C.
To define the notion of Hilbert space, we also need the following:
Definition 2.1.2. An inner product space is a vector space H equipped with an
inner product; that is, a mapping:
·|· : H × H → C

satisfying the following properties:

(i) ϕ|ϕ ≥ 0 with equality if and only if |ϕ = 0;
(ii) ϕ|ψ = ψ|ϕ∗ ;
(iii) ϕ|λ1 ψ1 + λ2 ψ2  = λ1 ϕ|ψ1  + λ2 ϕ|ψ2 
for any |ϕ, |ψ, |ψ1 , |ψ2  ∈ H and for any λ1 , λ2 ∈ C.
For any vectors |ϕ, |ψ ∈ H, the complex number ϕ|ψ is called the inner
product of |ϕ and |ψ. Sometimes, we write (|ϕ, |ψ) for ϕ|ψ. If ϕ|ψ = 0, then
we say that |ϕ and |ψ are orthogonal and write |ϕ ⊥ |ψ. The length of a vector
|ψ ∈ H is defined to be

||ψ|| = ψ|ψ.

A vector |ψ is called a unit vector if ||ψ|| = 1.

 2.1 Quantum mechanics 13

The notion of limit can be defined in terms of the length of a vector.

Definition 2.1.3. Let {|ψn } be a sequence of vectors in H and |ψ ∈ H.
(i) If for any  > 0, there exists a positive integer N such that ||ψm − ψn || <  for
all m, n ≥ N, then {|ψn } is called a Cauchy sequence.
(ii) If for any  > 0, there exists a positive integer N such that ||ψn − ψ|| <  for
all n ≥ N, then |ψ is called a limit of {|ψn } and we write |ψ = limn→∞ |ψn .
Now we are ready to present the definition of Hilbert space.
Definition 2.1.4. A Hilbert space is a complete inner product space: that is, an
inner product space in which each Cauchy sequence of vectors has a limit.
A notion that helps us to understand the structure of a Hilbert space is its basis.
In this book, we only consider finite-dimensional or countably infinite-dimensional
(separable) Hilbert space.
Definition 2.1.5. A finite or countably infinite family {|ψi } of unit vectors is
called an orthonormal basis of H if
(i) {|ψi } are pairwise orthogonal: |ψi  ⊥ |ψj  for any i, j with i = j;
(ii) {|ψi } span the whole
space H: each |ψ ∈ H can be written as a linear
combination |ψ = i λi |ψi  for some λi ∈ C and a finite number of |ψi .
The numbers of vectors in any two orthonormal bases are the same. This is
called the dimension of H and written as dim H; in particular, if an orthonormal
basis contains infinitely many vectors, then H is infinite-dimensional and we write
dim H = ∞.
Infinite-dimensional Hilbert spaces are required in quantum programming theory
only when a data type is infinite, e.g., integers. If it is hard for the reader to
understand infinite-dimensional Hilbert spaces and associated concepts (e.g., limits
in Definition 2.1.3, closed subspaces in Definition 2.1.6 following), she/he can simply
focus on finite-dimensional Hilbert spaces, which are exactly the vector spaces that
were learned in elementary linear algebra; in this way, the reader can still grasp an
essential part of this book.
Whenever H is finite-dimensional, say dim H = n, and we consider a fixed
orthonormal basis {|ψ1 , |ψ2 , . . . , |ψn }, then each vector |ψ = ni=1 λi |ψi  ∈ H
can be represented by the vector in Cn :
⎛ ⎞
λ1
⎝ ... ⎠
λn

The notion of subspace is also important for understanding the structure of a

Hilbert space.
Definition 2.1.6. Let H be a Hilbert space.
(i) If X ⊆ H, and for any |ϕ, |ψ ∈ X and λ ∈ C,
(a) |ϕ + |ψ ∈ X;
(b) λ|ϕ ∈ X,
then X is called a subspace of H.
14 CHAPTER 2 Preliminaries

(ii) For each X ⊆ H, its closure X is the set of limits limn→∞ |ψn  of sequences
{|ψn } in X.
(iii) A subspace X of H is closed if X = X.
For any subset X ⊆ H, the space spanned by X:
⎧ ⎫
⎨ n ⎬
spanX = λi |ψi  : n ≥ 0, λi ∈ C and |ψi  ∈ X (i = 1, . . . , n) (2.1)
⎩ ⎭
i=1

is the smallest subspace of H containing X. In other words, spanX is the subspace of

H generated by X. Moreover, spanX is the closed subspace generated by X.
We defined orthogonality between two vectors previously. It can be further
defined between two sets of vectors.
Definition 2.1.7. Let H be a Hilbert space.
(i) For any X, Y ⊆ H, we say that X and Y are orthogonal, written X ⊥ Y, if
|ϕ ⊥ |ψ for all |ϕ ∈ X and |ψ ∈ Y. In particular, we simply write |ϕ ⊥ Y if
X is the singleton {|ϕ}.
(ii) The orthocomplement of a closed subspace X of H is

X ⊥ = {|ϕ ∈ H : |ϕ ⊥ X}.

The orthocomplement X ⊥ is also a closed subspace of H, and we have (X ⊥ )⊥ =

X for every closed subspace X of H.
Definition 2.1.8. Let H be a Hilbert space, and let X, Y be two subspaces of H.
Then

X ⊕ Y = {|ϕ + |ψ : |ϕ ∈ X and |ψ ∈ Y}

is called the sum of X and Y. 

This definition can be straightforwardly generalized to the sum ni=1 Xi of more
than two subspaces
 Xi of H. In particular, if Xi (1 ≤ i ≤ n) are orthogonal to each
other, then ni=1 Xi is called an orthogonal sum.
With the above preparation, we can present:
• Postulate of quantum mechanics 1: The state space of a closed (i.e., an
isolated) quantum system is represented by a Hilbert space, and a pure state of
the system is described by a unit vector in its state space.

A linear combination |ψ = ni=1 λi |ψi  of states |ψ1 , . . . , |ψn  is often called
their superposition, and the complex coefficients λi are called probability amplitudes.
Example 2.1.1. A qubit – quantum bit – is the quantum counterpart of a bit. Its
state space is the two-dimensional Hilbert space:

H2 = C2 = {α|0 + β|1 : α, β ∈ C}.

 2.1 Quantum mechanics 15

The inner product in H2 is defined by

(α|0 + β|1, α  |0 + β  |1) = α ∗ α  + β ∗ β 

for all α, α  , β, β  ∈ C. Then {|0, |1} is an orthonormal basis of H2 , called the

computational basis. The vectors |0, |1 themselves are represented as
   
1 0
|0 = , |1 =
0 1

in this basis. A state of a qubit is described by a unit vector |ψ = α|0 + β|1 with
|α|2 + |β|2 = 1. The two vectors:
   
|0 + |1 1 1 |0 − |1 1 1
|+ = √ = √ , |− = √ = √
2 2 1 2 2 −1

form another orthonormal basis. Both of them are superpositions of |0 and |1. The
two-dimensional Hilbert space H2 can also be seen as the quantum counterpart of
the classical Boolean data type.
Example 2.1.2. Another Hilbert space often used in this book is the space of
square summable sequences:
 ∞ ∞

H∞ = αn |n : αn ∈ C for all n ∈ Z and |αn |2 < ∞ ,
n=−∞ n=−∞

where Z is the set of integers. The inner product in H∞ is defined by

 ∞ ∞
 ∞
αn |n, α  |n = αn∗ αn
n=−∞ n=−∞ n=−∞

for all αn , αn ∈ C (−∞ < n < ∞). Then {|n : n ∈ Z} is an orthonormal basis,
and H∞ is infinite-dimensional. This Hilbert space can be seen as the quantum
counterpart of the classical integer data type.
Exercise 2.1.1. Verify that the inner products defined in the previous two
examples satisfy conditions (i)–(iii) in Definition 2.1.2.

2.1.2 LINEAR OPERATORS

We studied the static description of a quantum system, namely its state space as a
Hilbert space, in the previous subsection. Now we turn to learning how to describe
the dynamics of a quantum system. The evolution of and all operations on a quantum
system can be depicted by linear operators in its state Hilbert space. So, in this
subsection, we study linear operators and their matrix representations.
Definition 2.1.9. Let H and K be Hilbert spaces. A mapping

A:H→K
16 CHAPTER 2 Preliminaries

is called an (a linear) operator if it satisfies the following conditions:

(i) A(|ϕ + |ψ) = A|ϕ + A|ψ;
(ii) A(λ|ψ) = λA|ψ
for all |ϕ, |ψ ∈ H and λ ∈ C.
An operator from H to itself is called an operator in H. The identity operator
in H that maps each vector in H to itself is denoted IH , and the zero operator in
H that maps every vector in H to the zero vector is denoted 0H . For any vectors
|ϕ, |ψ ∈ H, their outer product is the operator |ϕψ| in H defined by

(|ϕψ|)|χ = ψ|χ|ϕ

for every |χ ∈ H. A class of simple but useful operators are projectors. Let X be
a closed subspace of H and |ψ ∈ H. Then there exist uniquely |ψ0  ∈ X and
|ψ1  ∈ X ⊥ such that

|ψ = |ψ0  + |ψ1 .

The vector |ψ0  is called the projection of |ψ onto X and written |ψ0  = PX |ψ.
Definition 2.1.10. For each closed subspace X of H, the operator

PX : H → X, |ψ → PX |ψ

is called the projector onto X. 

Exercise 2.1.2. Show that PX = i |ψi ψi | if {|ψi } is an orthonormal basis
of X.
Throughout this book, we only consider bounded operators, as defined in the
following:
Definition 2.1.11. An operator A in H is said to be bounded if there is a constant
C ≥ 0 such that

A|ψ ≤ C · ψ

for all |ψ ∈ H. The norm of A is defined to be the nonnegative number:

A = inf{C ≥ 0 : ||A|ψ|| ≤ C · ||ψ|| for all ψ ∈ H}.

We write L(H) for the set of bounded operators in H.

All operators in a finite-dimensional Hilbert space are bounded.
Various operations of operators are very useful in order to combine several opera-
tors to produce a new operator. The addition, scalar multiplication and composition of
operators can be defined in a natural way: for any A, B ∈ L(H), λ ∈ C and |ψ ∈ H,

(A + B)|ψ = A|ψ + B|ψ,

(λA)|ψ = λ(A|ψ),
(BA)|ψ = B(A|ψ).
 2.1 Quantum mechanics 17

Exercise 2.1.3. Show that L(H) with addition and scalar multiplication forms a
vector space.
We can also define positivity of an operator as well as an order and a distance
between operators.
Definition 2.1.12. An operator A ∈ L(H) is positive if for all states |ψ ∈ H,
ψ|A|ψ is a nonnegative real number: ψ|A|ψ ≥ 0.
Definition 2.1.13. The Löwner order  is defined as follows: for any A, B ∈
L(H), A  B if and only if B − A = B + (−1)A is positive.
Definition 2.1.14. Let A, B ∈ L(H). Then their distance is

d(A, B) = sup ||A|ψ − B|ψ|| (2.2)

|ψ

where |ψ traverses all pure states (i.e., unit vectors) in H.

Matrix Representation of Operators:
Operators in a finite-dimensional Hilbert space have a matrix representation,
which is very convenient in applications. After reading this part, the reader should
have a better understanding of those abstract notions defined previously through a
connection from them to the corresponding notions that she/he learned in elementary
linear algebra.
If {|ψi } is an orthonormal basis of H, then an operator A is uniquely determined
by the images A|ψi  of the basis vectors |ψi  under A. In particular, when dim H = n
is finite and we consider a fixed orthonormal basis {|ψ1 , . . . , |ψn }, A can be
represented by the n × n complex matrix:
⎛ ⎞
  a11 ... a1n
A = aij n×n = ⎝ ... ⎠
an1 ... ann

where
aij = ψi |A|ψj  = (|ψi , A|ψj )
n
for every i, j = 1, . . . , n. Moreover, the image of a vector |ψ
 = i=1 αi |ψi  ∈ H
under operator A is represented by the product of matrix A = aij n×n and vector |ψ:
⎛ ⎞ ⎛ ⎞
α1 β1
A|ψ = A ⎝ . . . ⎠ = ⎝ . . . ⎠
αn βn

where βi = nj=1 aij αj for every i = 1, . . . , n. For example, IH is the unit matrix,
and 0H is the zero matrix. If
⎛ ⎞ ⎛ ⎞
α1 β1
|ϕ = ⎝ . . . ⎠ , |ψ = ⎝ . . . ⎠ ,
αn βn
 
then their outer product is the matrix |ϕψ| = aij n×n with aij = αi βj∗ for every
i, j = 1, . . . , n. Throughout this book, we do not distinguish an operator in a finite-
dimensional Hilbert space from its matrix representation.
18 CHAPTER 2 Preliminaries

Exercise 2.1.4. Show that in a finite-dimensional Hilbert space, addition, scalar

multiplication and composition of operators correspond to addition, scalar multipli-
cation and multiplication of their matrix representations, respectively.

2.1.3 UNITARY TRANSFORMATIONS

The postulate of quantum mechanics 1 introduced in Subsection 2.1.1 provides the
static description of a quantum system. In this subsection, we give a description of
the dynamics of a quantum system, with the mathematical tool prepared in the last
subsection.
The continuous-time dynamics of a quantum system are given by a differential
equation, called the Schrödinger equation. But in quantum computation, we usually
consider the discrete-time evolution of a system – a unitary transformation. For any
operator A ∈ L(H), it turns out that there exists a unique (linear) operator A† in H
such that
 
(A|ϕ, |ψ) = |ϕ, A† |ψ

for all |ϕ, |ψ ∈ H. The operator A† is called the adjoint of A. In particular,
 if an
operator in an n-dimensional Hilbert space is represented by the matrix A = aij n×n ,
then its adjoint is represented by the transpose conjugate of A:
 
A† = bij n×n

with bij = a∗ji for every i, j = 1, . . . , n.

Definition 2.1.15. An (bounded) operator U ∈ L(H) is called a unitary
transformation if the adjoint of U is its inverse:
U † U = UU † = IH .

All unitary transformations U preserve the inner product:

(U|ϕ, U|ψ) = ϕ|ψ

for any |ϕ, |ψ ∈ H. The condition U † U = IH is equivalent to UU † = IH when H

is finite-dimensional. If dim H = n, then a unitary operator in H is represented by an
n ×n unitary matrix U; i.e., a matrix U with U † U = In , where In is the n-dimensional
unit matrix.
A useful technique for defining a unitary operator is given in the following:
Lemma 2.1.1. Suppose that H is a (finite-dimensional) Hilbert space and K is a
closed subspace of H. If linear operator U : K → H preserves the inner product:
(U|ϕ, U|ψ) = ϕ|ψ

for any |ϕ, |ψ ∈ K, then there exists a unitary operator V in H which extends U;
i.e., V|ψ = U|ψ for all |ψ ∈ K.
Exercise 2.1.5. Prove Lemma 2.1.1.
 2.1 Quantum mechanics 19

Now we are ready to present:

• Postulate of quantum mechanics 2: Suppose that the states of a closed
quantum system (i.e., a system without interactions with its environment) at
times t0 and t are |ψ0  and |ψ, respectively. Then they are related to each other
by a unitary operator U which depends only on the times t0 and t,

|ψ = U|ψ0 .

To help the reader understand this postulate, let us consider two simple examples.
Example 2.1.3. One frequently used unitary operator on a qubit is the Hadamard
transformation in the two-dimensional Hilbert space H2 :
 
1 1 1
H= √
2 1 −1

It transforms a qubit in the computational basis states |0 and |1 into their
superpositions:
   
1 1 1
H|0 = H = √ = |+,
0 2 1
   
0 1 1
H|1 = H = √ = |−.
1 2 −1

Example 2.1.4. Let k be an integer. Then the k-translation operator Tk in the

infinite-dimensional Hilbert space H∞ is defined by

Tk |n = |n + k

for all n ∈ Z. It is easy to verify that Tk is a unitary operator. In particular, we write

TL = T−1 and TR = T1 . They move a particle on the line one position to the left and
to the right, respectively.
More examples will be seen in Section 2.2, where unitary transformations are
used as quantum logic gates in a quantum circuit.

2.1.4 QUANTUM MEASUREMENTS

Now that we understand both the static and dynamic descriptions of a quantum sys-
tem, observation of a quantum system is carried out through a quantum measurement,
which is defined by:
• Postulate of quantum mechanics 3: A quantum measurement on a system with
state Hilbert space H is described by a collection {Mm } ⊆ L(H) of operators
satisfying the normalization condition:
†
Mm Mm = IH , (2.3)
m
20 CHAPTER 2 Preliminaries

where Mm are called measurement operators, and the index m stands for the
measurement outcomes that may occur in the experiment. If the state of a
quantum system is |ψ immediately before the measurement, then for each m,
the probability that the result m occurs in the measurement is
†
p(m) = ||Mm |ψ||2 = ψ|Mm Mm |ψ (Born rule)

and the state of the system after the measurement with outcome m is
Mm |ψ
|ψm  = √ .
p(m)

It is easy to see that the normalization

 condition (2.3) implies that the probabilities
for all outcomes sum up to m p(m) = 1.
The following simple example should help the reader to understand this postulate.
Example 2.1.5. The measurement of a qubit in the computational basis has two
outcomes defined by measurement operators:

M0 = |00|, M1 = |11|.

If the qubit was in state |ψ = α|0 + β|1 before the measurement, then the
probability of obtaining outcome 0 is
†
p(0) = ψ|M0 M0 |ψ = ψ|M0 |ψ = |α|2 ,

and in this case the state after the measurement is

M0 |ψ
√ = |0.
p(0)

Similarly, the probability of outcome 1 is p(1) = |β|2 and in this case the state after
the measurement is |1.
Projective Measurements:
A specially useful class of measurements is defined in terms of Hermitian
operators and their spectral decomposition.
Definition 2.1.16. An operator M ∈ L(H) is said to be Hermitian if it is self-
adjoint:

M † = M.

In physics, a Hermitian operator is also called an observable.

It turns out that an operator P is a projector; that is, P = PX for some closed
subspace X of H, if and only if P is Hermitian and P2 = P.
A quantum measurement can be constructed from an observable based on
the mathematical concept of spectral decomposition of a Hermitian operator.
Due to the limit of space, we only consider spectral decomposition in a finite-
dimensional Hilbert space H. (The infinite-dimensional case requires a much heavier
 2.1 Quantum mechanics 21

mathematical mechanism; see [182], Chapter III.5. In this book, it will be used only
in Section 3.6 as a tool for the proof of a technical lemma.)
Definition 2.1.17
(i) An eigenvector of an operator A ∈ L(H) is a non-zero vector |ψ ∈ H such
that A|ψ = λ|ψ for some λ ∈ C, where λ is called the eigenvalue of A
corresponding to |ψ.
(ii) The set of eigenvalues of A is called the (point) spectrum of A and denoted
spec(A).
(iii) For each eigenvalue λ ∈ spec(A), the set

{|ψ ∈ H : A|ψ = λ|ψ}

is a closed subspace of H and it is called the eigenspace of A corresponding

to λ.
The eigenspaces corresponding to different eigenvalues λ1 = λ2 are orthogonal.
All eigenvalues of an observable (i.e., a Hermitian operator) M are real numbers.
Moreover, it has the spectral decomposition:

M= λPλ
λ∈spec(M)

where Pλ is the projector onto the eigenspace corresponding to λ. Then it defines

a measurement {Pλ : λ ∈ spec(M)}, called a projective measurement because all
measurement operators Pλ are projectors. Using the Postulate of quantum mechanics
3 introduced earlier, we obtain: upon measuring a system in state |ψ, the probability
of getting result λ is
†
p(λ) = ψ|Pλ Pλ |ψ = ψ|P2λ |ψ = ψ|Pλ |ψ (2.4)

and in this case the state of the system after the measurement is
Pλ |ψ
√ . (2.5)
p(λ)

Since all possible outcomes λ ∈ spec(M) are real numbers, we can compute the
expectation – average value – of M in state |ψ:

Mψ = p(λ) · λ
λ∈spec(M)

= λψ|Pλ |ψ
λ∈spec(M)

= ψ| λPλ |ψ

λ∈spec(M)
= ψ|M|ψ.
22 CHAPTER 2 Preliminaries

We observe that, given the state |ψ, probability (2.4) and post-measurement state
(2.5) are determined only by the projectors {Pλ } (rather than M itself). It is easy to
see that {Pλ } is a complete set of orthogonal projectors; that is, a set of operators
satisfying the conditions:

Pλ if λ = δ,
(i) Pλ Pδ =
0H otherwise;

(ii) λ P λ = I H.

Sometimes, we simply call a complete set of orthogonal projectors a projective

measurement. A special case is the measurement in an orthonormal basis {|i} of
the state Hilbert space, where Pi = |ii| for every i. Example 2.1.5 is such a
measurement for a qubit.

2.1.5 TENSOR PRODUCTS OF HILBERT SPACES

Up to now we have only considered a single quantum system. In this section,
we further show how a large composite system can be made up of two or more
subsystems. The description of a composite system is based on the notion of tensor
product. We mainly consider the tensor product of a finite family of Hilbert spaces.
Definition 2.1.18. Let Hi be a Hilbert space with {|ψiji } as an orthonormal basis
for i = 1, . . . , n. We write B for the set having elements of the form:

|ψ1j1 , . . . , ψnjn  = |ψ1j1 ⊗ . . . ⊗ ψnjn  = |ψ1j1  ⊗ . . . ⊗ |ψnjn .

Then the tensor product of Hi (i = 1, . . . , n) is the Hilbert space with B as an

orthonormal basis:

Hi = spanB.
i

It follows from equation (2.1) that each element in i Hi can be written in the
form of

αj1 ,...,jn |ϕ1j1 , . . . , ϕnjn 

j1 ,...,jn

where |ϕ1j1  ∈ H1 , . . . , |ϕnjn  ∈ Hn and αj1 ,...,jn ∈ C for all j1 , . . . , jn . Furthermore,

it can be shown by linearity that the choice of basis {|ψiji } of each factor space Hi
is not essential in the previous definition: for example, if |ϕi  = ji αji |ϕiji  ∈ Hi
(i = 1, . . . , n), then

|ϕ1  ⊗ . . . ⊗ |ϕn  = α1j1 . . . αnjn |ϕ1j1 , . . . , ϕnjn .

j1 ,...,jn

The vector addition, scalar multiplication and inner product in i Hi can be naturally
defined based on the fact that B is an orthonormal basis.
 2.1 Quantum mechanics 23

We will need to consider the tensor product of a countably infinite family of

Hilbert spaces occasionally in this book. Let {Hi } be a countably infinite family of
Hilbert spaces, and let {|ψiji } be an orthonormal basis of Hi for each i. We write B
for the set of tensor products of basis vectors of all Hi :
 

B= |ψiji  .
i

Then B is a finite or countably infinite set, and it can be written in the form of a
sequence of vectors: B = {|ϕn  : n = 0, 1, . . .}. The tensor product of {Hi } can be
properly defined to be the Hilbert space with B as an orthonormal basis:
 

Hi = αn |ϕn  : αn ∈ C for all n ≥ 0 and |αn |2 < ∞ .
i n n

Now we are able to present:

• Postulate of quantum mechanics 4: The state space of a composite quantum
system is the tensor product of the state spaces of its components.
Suppose that S is a quantum system composed of subsystems S1 , . . . , Sn with
state Hilbert space H1 , . . . , Hn . If for each 1 ≤ i ≤ n, Si is in state |ψi  ∈ Hi ,
then S is in the product state |ψ1 , . . . , ψn . Furthermore, S can be in a superposition
(i.e., linear combination) of several product states. One of the most interesting and
puzzling phenomenon in quantum mechanics – entanglement – occurs in a composite
system: a state of the composite system is said to be entangled if it is not a product of
states of its component systems. The existence of entanglement is one of the major
differences between the classical world and the quantum world.
Example 2.1.6. The state space of the system of n qubits is:
⎧ ⎫
⎨ ⎬
H⊗n
2 =C
2n = αx |x : αx ∈ C for all x ∈ {0, 1}n .
⎩ ⎭
x∈{0,1}n

In particular, a two-qubit system can be in a product state such as |00, |1|+ but
also in an entangled state such as the Bell states or the EPR (Einstein-Podolsky-
Rosen) pairs:
1 1
|β00  = √ (|00 + |11), |β01  = √ (|01 + |10),
2 2
1 1
|β10  = √ (|00 − |11), |β11  = √ (|01 − |10).
2 2

Of course, we can talk about (linear) operators, unitary transformations and

measurements in the tensor product of Hilbert spaces since it is a Hilbert space too. A
special class of operators in the tensor product of Hilbert spaces is defined as follows:
24 CHAPTER 2 Preliminaries

Definition2.1.19. Let Ai ∈ L(Hi ) for i =

1,n . . . , n. Then their tensor product is
the operator ni=1 Ai = A1 ⊗ . . . ⊗ An ∈ L i=1 Hi defined by

(A1 ⊗ . . . ⊗ An )|ϕ1 , . . . , ϕn  = A1 |ϕ1  ⊗ . . . ⊗ An |ϕn 

for all |ϕi  ∈ Hi (i = 1, . . . , n) together with linearity.

But other operators rather than tensor products are indispensable in quantum
computation because they can create entanglement.
Example 2.1.7. The controlled-NOT or CNOT operator C in the state Hilbert
space H2⊗2 = C4 of a two-qubit system is defined by

C|00 = |00, C|01 = |01, C|10 = |11, C|11 = |10

or equivalently as the 4 × 4 matrix

⎛ ⎞
1 0 0 0
⎜ 0 1 0 0 ⎟
C=⎜
⎝ 0
⎟.
0 0 1 ⎠
0 0 1 0

It can transform product states into entangled states:

C|+|0 = β00 , C|+|1 = β01 , C|−|0 = β10 , C|−|1 = β11 .

Implementing a General Measurement by a Projective Measurement:

Projective measurements are introduced in subsection 2.1.4 as a special class of
quantum measurements. The notion of tensor product enables us to show that an
arbitrary quantum measurement can be implemented by a projective measurement
together with a unitary transformation if we are allowed to introduce an ancilla
system. Let M = {Mm } be a quantum measurement in Hilbert space H.
• We introduce a new Hilbert space HM = span{|m}, which is used to record the
possible outcomes of M.
• We arbitrarily choose a fixed state |0 ∈ HM . Define operator

UM (|0|ψ) = |mMm |ψ

m

for every |ψ ∈ H. It is easy to check that UM preserves the inner product, and
by Lemma 2.1.1 it can be extended to a unitary operator in HM ⊗ H, which is
denoted by UM too.
• We define a projective measurement M = {M m } in HM ⊗ H with
M m = |mm| ⊗ IH for every m.
Then the measurement M is realized by the projective measurement M together with
the unitary operator UM , as shown in the following:
 2.1 Quantum mechanics 25

Proposition 2.1.1. Let |ψ ∈ H be a pure state.

• When we perform measurement M on |ψ, the probability of outcome m is
denoted pM (m) and the post-measurement state corresponding to m is |ψm .
• When we perform measurement M on |ψ = UM (|0|ψ), the probability of
outcome m is denoted pM (m) and the post-measurement state corresponding to
m is |ψ m .
Then for each m, we have: pM (m) = pM (m) and |ψ m  = |m|ψm . A similar result
holds when we consider a mixed state in H introduced in the next subsection.
Exercise 2.1.6. Prove Proposition 2.1.1.

2.1.6 DENSITY OPERATORS

We have already learned all of the four basic postulates of quantum mechanics. But
they were only formulated in the case of pure states. In this section, we extend these
postulates so that they can be used to deal with mixed states.
Sometimes, the state of a quantum system is not completely known, but we know
 states |ψi , with respective probabilities pi , where
that it is in one of a number of pure
|ψi  ∈ H, pi ≥ 0 for each i, and i pi = 1. A convenient notion for coping with this
situation is the density operator. We call {(|ψi , pi )} an ensemble of pure states or a
mixed state, whose density operator is defined to be

ρ= pi |ψi ψi |. (2.6)

i

In particular, a pure state |ψ may be seen as a special mixed state {(|ψ, 1)} and its
density operator is ρ = |ψψ|.
Density operators can be described in a different but equivalent way.
Definition 2.1.20. The trace tr(A) of operator A ∈ L(H) is defined to be

tr(A) = ψi |A|ψi 

i

where {|ψi } is an orthonormal basis of H.

It can be shown that tr(A) is independent of the choice of basis {|ψi }.
Definition 2.1.21. A density operator ρ in a Hilbert space H is a positive
operator (see Definition 2.1.12) with tr(ρ) = 1.
It turns out that for any mixed state {(|ψi , pi )}, operator ρ defined by equa-
tion (2.6) is a density operator according to Definition 2.1.21. Conversely, for any
density operator ρ, there exists a (but not necessarily unique) mixed state {(|ψi , pi )}
such that equation (2.6) holds.
The evolution of and a measurement on a quantum system in mixed states can be
elegantly formulated in the language of density operators:
• Suppose that the evolution of a closed quantum system from time t0 to t is
described by unitary operator U depending on t0 and t: |ψ = U|ψ0 , where
26 CHAPTER 2 Preliminaries

|ψ0 , |ψ are the states of the system at times t0 and t, respectively. If the system
is in mixed states ρ0 , ρ at times t0 and t, respectively, then

ρ = Uρ0 U † . (2.7)

• If the state of a quantum system was ρ immediately before measurement {Mm } is

performed on it, then the probability that result m occurs is
 
†
p(m) = tr Mm Mm ρ , (2.8)

and in this case the state of the system after the measurement is
†
Mm ρMm
ρm = . (2.9)
p(m)

Exercise 2.1.7. Derive equations (2.7), (2.8) and (2.9) from equation (2.6) and
Postulates of quantum mechanics 1 and 2.
Exercise 2.1.8. Let M be an observable (a Hermitian operator) and {Pλ : λ ∈
spec(M)} the projective measurement defined by M. Show that the expectation of M
in a mixed state ρ is

Mρ = p(λ) · λ = tr(Mρ).

λ∈spec(M)

Reduced Density Operators:

Postulate of quantum mechanics 4 introduced in the last subsection enables us
to construct composite quantum systems. Of course, we can talk about a mixed
state of a composite system and its density operator because the state space of the
composite system is the tensor product of the state Hilbert spaces of its subsystems,
which is a Hilbert space too. Conversely, we often need to characterize the state of a
subsystem of a quantum system. However, it is possible that a composite system
is in a pure state, but some of its subsystems must be seen as in a mixed state.
This phenomenon is another major difference between the classical world and the
quantum world. Consequently, a proper description of the state of a subsystem of
a composite quantum system can be achieved only after introducing the notion of
density operator.
Definition 2.1.22. Let S and T be quantum systems whose state Hilbert spaces
are HS and HT , respectively. The partial trace over system T

trT : L(HS ⊗ HT ) → L(HS )

is defined by

trT (|ϕψ| ⊗ |θζ |) = ζ |θ · |ϕψ|

for all |ϕ, |ψ ∈ HS and |θ , |ζ  ∈ HT together with linearity.

Other documents randomly have
different content
 PATTIES À LA PONTIFE. (ENTRÉE.)

(A fast day, or Maigre dish.)

Mince, but not very small, the yolks of six fresh hard-boiled eggs;
mince also and mix with them a couple of fine truffles,[120] a large
saltspoonful of salt, half the quantity of mace and nutmeg, and a
fourth as much of cayenne. Moisten these ingredients with a
spoonful of thick cream, or béchamel maigre (see page 109), or with
a dessertspoonful of clarified butter; line the patty-moulds, fill them
with the mixture, cover, and bake them from twelve to fifteen minutes
in a moderate oven. They are excellent made with the cream-crust of
page 347.
120. The bottled ones will answer well for these.
Yolks hard-boiled eggs, 6; truffles, 2 large; seasoning of salt,
mace, nutmeg, and cayenne; cream, or béchamel maigre, 1
tablespoonful, or clarified butter, 1 dessertspoonful: baked moderate
oven, 12 to 15 minutes.
Obs.—A spoonful or two of jellied stock or gravy, or of good white
sauce, converts these into admirable patties: the same ingredients
make also very superior rolls or cannelons. For Patties à la
Cardinale, small mushroom-buttons stewed as for partridges,
Chapter XIII., before they are minced, must be substituted for
truffles; and the butter in which they are simmered should be added
with them to the eggs.
 EXCELLENT MEAT ROLLS.

Pound, as for potting (see page 305), and with the same
proportion of butter and of seasonings, some half-roasted veal,
chicken, or turkey. Make some forcemeat by the receipt No. 1,
Chapter VI., and form it into small rolls, not larger than a finger; wrap
twice or thrice as much of the pounded meat equally round each of
these, first moistening it with a teaspoonful of water; fold them in
good puff-paste, and bake them from fifteen to twenty minutes, or
until the crust is perfectly done. A small quantity of the lean of a
boiled ham may be finely minced and pounded with the veal, and
very small mushrooms, prepared as for a partridge (page 329), may
be substituted for the forcemeat.
 SMALL VOLS-AU-VENTS, OR PATTY-CASES.

These are quickly and easily made with two round paste-cutters,
of which one should be little more than half the size of the other: to
give the pastry a better appearance, they should be fluted. Roll out
some of the lightest puff-paste to a half-inch of thickness, and with
the larger of the tins cut the number of patties required; then dip the
edge of the small shape into hot water, and press it about half
through them. Bake them in a moderately quick oven from ten to
twelve minutes, and when they are done, with the point of a sharp
knife, take out the small rounds of crust from the tops, and scoop all
the crumb from the inside of the patties, which may then be filled
with shrimps, oysters, lobster, chicken, pheasant, or any other of the
ordinary varieties of patty meat, prepared with white sauce. Fried
crumbs may be laid over them instead of the covers, or these last
can be replaced.
For sweet dishes, glaze the pastry, and fill it with rich whipped
cream, preserve, or boiled custard; if with the last of these put it back
into a very gentle oven until the custards are set.
 ANOTHER RECEIPT FOR TARTLETS.

For a dozen tartlets, cut twenty-four rounds of paste of the usual

size, and form twelve of them into rings by pressing the small cutter
quite through them; moisten these with cold water, or white of egg,
and lay them on the remainder of the rounds of paste, so as to form
the rims of the tartlets. Bake them from ten to twelve minutes, fill
them with preserve while they are still warm, and place over it a
small ornament of paste cut from the remnants, and baked gently of
a light colour. Serve the tartlets cold, or if wanted hot for table put
them back into the oven for one minute after they are filled.
 A SEFTON, OR VEAL CUSTARD.

Pour boiling, a pint of rich, clear, pale veal gravy on six fresh eggs,
which have been well beaten and strained: sprinkle in directly the
grated rind of a fine lemon, a little cayenne, some salt if needed, and
a quarter-teaspoonful of mace. Put a paste border round a dish, pour
in, first two ounces of clarified butter, and then the other ingredients;
bake the Sefton in a very slow oven from twenty-five to thirty
minutes, or until it is quite firm in the middle, and send it to table with
a little good gravy. Very highly flavoured game stock, in which a few
mushrooms have been stewed, may be used for this dish with great
advantage in lieu of veal gravy; and a sauce made of the smallest
mushroom buttons, may be served with it in either case. The mixture
can be baked in a whole paste, if preferred so, or in well buttered
cups; then turned out and covered with the sauce before it is sent to
table.
Rich veal or game stock, 1 pint; fresh eggs, 6; rind, 1 lemon; little
salt and cayenne; pounded mace, 1/4 teaspoonful; butter, 2 oz.:
baked, 25 to 30 minutes, slow oven.
 APPLE CAKE, OR GERMAN TART.

Work together with the fingers, ten ounces of butter and a pound
of flour, until they resemble fine crumbs of bread; throw in a small
pinch of salt, and make them into a firm smooth paste with the yolks
of two eggs and a spoonful or two of water. Butter thickly, a plain tin
cake, or pie mould (those which open at the sides, see plate, page
344, are best adapted for the purpose); roll out the paste thin, place
the mould upon it, trim a bit to its exact size, cover the bottom of the
mould with this, then cut a band the height of the sides, and press it
smoothly round them, joining the edge, which must be moistened
with egg or water, to the bottom crust; and fasten upon them, to
prevent their separation, a narrow and thin band of paste, also
moistened. Next, fill the mould nearly from the brim with the following
marmalade, which must be quite cold when it is put in. Boil together,
over a gentle fire at first, but more quickly afterwards, three pounds
of good apples with fourteen ounces of pounded sugar, or of the
finest Lisbon, the strained juice of a large lemon, three ounces of
fresh butter, and a teaspoonful of pounded cinnamon, or the lightly
grated rind of a couple of lemons: when the whole is perfectly
smooth and dry, turn it into a pan to cool, and let it be quite cold
before it is put into the paste. In early autumn, a larger proportion of
sugar may be required, but this can be regulated by the taste. When
the mould is filled, roll out the cover, lay it carefully over the
marmalade that it may not touch it; and when the cake is securely
closed, trim off the superfluous paste, add a little pounded sugar to
the parings, spread them out very thin, and cut them into leaves to
ornament the top of the cake, round which they may be placed as a
sort of wreath.[121] Bake it for an hour in a moderately brisk oven;
take it from the mould, and should the sides not be sufficiently
coloured put it back for a few minutes into the oven upon a baking
tin. Lay a paper over the top, when it is of a fine light brown, to
prevent its being too deeply coloured. This cake should be served
hot.
121. Or, instead of these, fasten on it with a little white of egg, after it is taken from
the oven, some ready-baked leaves of almond-paste (see page 355), either
plain or coloured.
Paste: flour, 1 lb.; butter, 10 oz.; yolks of eggs, 2; little water.
Marmalade: apples, 3 lbs.; sugar, 14 oz. (more if needed); juice of
lemon, 1; rinds of lemons, 2; butter, 3 oz.: baked, 1 hour.
 TOURTE MERINGUÉE, OR TART WITH ROYAL ICING.[122]
122. The limits to which we are obliged to confine this volume, compel us to omit
many receipts which we would gladly insert; we have, therefore, rejected
those which may be found in almost every English cookery book, for such as
are, we apprehend, less known to the reader: this will account for the small
number of receipts for pies and fruit tarts to be found in the present chapter.

Lay a band of fine paste round the rim of a tart-dish, fill it with any
kind of fruit mixed with a moderate proportion of sugar, roll out the
cover very evenly, moisten the edges of the paste, press them
together carefully, and trim them off close to the dish; spread equally
over the top, to within rather more than an inch of the edge all round,
the whites of three fresh eggs beaten to a quite solid froth and mixed
quickly at the moment of using them with three tablespoonsful of dry
sifted sugar. Put the tart into a moderately brisk oven, and when the
crust has risen well and the icing is set, either lay a sheet of writing-
paper lightly over it, or draw it to a part of the oven where it will not
take too much colour. This is now a fashionable mode of icing tarts,
and greatly improves their appearance.
Bake half an hour.
 A GOOD APPLE TART.

A pound and a quarter of apples weighed after they are pared and
cored, will be sufficient for a small tart, and four ounces more for one
of moderate size. Lay a border of English puff-paste, or of cream-
crust round the dish, just dip the apples into water, arrange them
very compactly in it, higher in the centre than at the sides, and strew
amongst them from three to four ounces of pounded sugar, or more
should they be very acid: the grated rind and the strained juice of
half a lemon will much improve their flavour. Lay on the cover rolled
thin, and ice it or not at pleasure. Send the tart to a moderate oven
for about half an hour. This may be converted into the old-fashioned
creamed apple tart, by cutting out the cover while it is still quite hot,
leaving only about an inch-wide border of paste round the edge, and
pouring over the apples when they have become cold, from half to
three-quarters of a pint of rich boiled custard. The cover divided into
triangular sippets, was formerly stuck round the inside of the tart, but
ornamental leaves of pale puff-paste have a better effect. Well-
drained whipped cream may be substituted for the custard, and be
piled high, and lightly over the fruit.
 TART OF VERY YOUNG GREEN APPLES. (GOOD.)

Take very young apples from the tree before the cores are formed,
clear off the buds and stalks, wash them well, and fill a tart-dish with
them after having rolled them in plenty of sugar, or strew layers of
sugar between them; add a very small quantity of water and bake
the tart rather slowly, that the fruit may be tender quite through. It will
resemble a green apricot-tart if carefully made. We give this receipt
from recollection, having had the dish served often formerly, and
having found it very good.
 BARBERRY TART.

Barberries, with half their weight of fine brown sugar, when they
are thoroughly ripe, and with two ounces more when they are not
quite so, make an admirable tart. For one of moderate size, put into
a dish bordered with paste three quarters of a pound of barberries
stripped from their stalks, and six ounces of sugar in alternate layers;
pour over them three tablespoonsful of water, put on the cover, and
bake the tart for half an hour. Another way of making it is, to line a
shallow tin pan with very thin crust, to mix the fruit and sugar well
together with a spoon before they are laid in, and to put bars of paste
across instead of a cover; or it may be baked without either.[123]
123. The French make their fruit-tarts generally thus, in large shallow pans.
Plums, split and stoned (or if of small kinds, left entire), cherries and currants
freed from the stalks, and various other fruits, all rolled in plenty of sugar, are
baked in the uncovered crust; or this is baked by itself, and then filled
afterwards with fruit previously stewed tender.
 THE LADY’S TOURTE, AND CHRISTMAS TOURTE À LA
CHÂTELAINE.

To make this Tourte, which, when filled, is

of pretty appearance, two paste-cutters are
requisite, one the size, or nearly so, of the
inside of the dish in which the entremets is
Lady’s Tourte. to be served, the other not more than an
inch in diameter, and both of them fluted, as
will be seen by the engraving. To make the
paste for it, throw a small half saltspoonful of salt into half a pound of
the finest flour, and break lightly into it four ounces of fresh butter,
which should be firm. Make these up smoothly with cold milk or
water, of which nearly a quarter of a pint will be sufficient, unless the
butter should be very hard, when a spoonful or two more must be
added. Roll the paste out as lightly as possible twice or thrice if
needful, to blend the butter thoroughly with it, and each time either
fold it in three by wrapping the ends over each other, or fold it over
and over like a roll pudding. An additional ounce, or even two, of
butter can be used for it when very rich pastry is liked, but the tourte
will not then retain its form so well. Roll it out evenly to something
more than three-quarters of an inch in thickness, and press the large
cutter firmly through it; draw away the superfluous paste, and lay the
tourte on a lightly floured baking-tin. Roll the remainder of the paste
until it is less than a quarter of an inch thick, and stamp out with the
smaller cutter—of which the edge should be dipped into hot water, or
slightly encrusted with flour—as many rounds as will form the border
of the tourte. In placing them upon it, lay the edge of one over the
other just sufficiently to give a shell-like appearance to the whole;
and with the finger press lightly on the opposite part of the round to
make it adhere to the under paste. Next, with a sharp-pointed knife,
make an incision very evenly round the inside of the tourte nearly
close to the border, but be extremely careful not to cut too deeply
into the paste. Bake it in a gentle oven, from twenty to thirty minutes.
When it is done, detach the crust from the centre, where it has been
marked with the knife, take out part of the crumb, fill the space high
with apricot-jam, or with any other choice preserve, set it again for an
instant into the oven, and serve it hot or cold. Spikes of blanched
almonds, filberts, or pistachio-nuts, may be strewed over the
preserve, when they are considered an improvement; and the border
of the pastry may be glazed or ornamented to the fancy; but if well
made, it will generally please in its quite simple form. It may be
converted into a delicious entrée, by filling it either with oysters, or
sliced sweetbreads, stewed, and served in thick, rich, white sauce,
or béchamel. Lobster also prepared and moulded as for the new
lobster patties of page 359, will form a superior dish even to these.
Obs.—Six ounces of flour, and three of butter, will make sufficient
paste for this tourte, when it is required only of the usual moderate
size. If richer paste be used for it, it must have two or three additional
turns or rollings to prevent its losing its form in the oven.
Christmas Tourte à la Châtelaine.—Make the case for this tourte
as for the preceding one, and put sufficient mincemeat to fill it
handsomely into a jar, cover it very securely with paste, or with two
or three folds of thick paper, and bake it gently for half an hour or
longer, should the currants, raisins, &c., not be fully tender. Take out
the inside of the tourte, heap the hot mincemeat in it, pour a little
fresh brandy over; just touch it with a strip of lighted writing-paper at
the door of the dining-room, and serve it in a blaze; or if better liked
so, serve it very hot without the brandy, and with Devonshire cream
as an accompaniment.[124]
124. Sufficient of cream for this purpose can easily be prepared from good milk.
 GENOISES À LA REINE, OR HER MAJESTY’S PASTRY.

Make some nouilles (see page 5), with the yolks of four fresh
eggs, and when they are all cut as directed there, drop them lightly
into a pint and a half of boiling cream (new milk will answer quite as
well, or a portion of each may be used), in which six ounces of fresh
butter have been dissolved. When these have boiled quickly for a
minute or two, during which time they must be stirred to prevent their
gathering into lumps, add a small pinch of salt, and six ounces of
sugar on which the rinds of two lemons have been rasped; place the
saucepan over a clear and very gentle fire, and when the mixture
has simmered from thirty to forty minutes take it off, stir briskly in the
yolks of six eggs, and pour it out upon a delicately clean baking-tin
which has been slightly rubbed in every part with butter; level the
nouilles with a knife to something less than a quarter of an inch of
thickness, and let them be very evenly spread; put them into a
moderate oven, and bake them of a fine equal brown: should any air-
bladders appear, pierce them with the point of a knife. On taking the
paste from the oven, divide it into two equal parts; turn one of these,
the underside uppermost, on to a clean tin or a large dish, and
spread quickly over it a jar of fine apricot-jam, place the other half
upon it, the brown side outwards, and leave the paste to become
cold; then stamp it out with a round or diamond-shaped cutter, and
arrange the genoises tastefully in a dish. This pastry will be found
delicious the day it is baked, but its excellence is destroyed by
keeping. Peach, green-gage, or magnum bonum jam, will serve for it
quite as well as apricot. We strongly recommend to our readers this
preparation, baked in pattypans, and served hot; or the whole
quantity made into a pudding. From the smaller ones a little may be
taken out with a teaspoon, and replaced with some preserve just
before they are sent to table; or they may thus be eaten cold.
Nouilles of 4 eggs; cream or milk, 1-1/2 pint; butter, 6 oz.; sugar 6
oz.; rasped rinds of lemons, 2; grain of salt: 30 to 40 minutes. Yolks
of eggs, 6: baked from 15 to 25 minutes.
 ALMOND PASTE.

For a single dish of pastry, blanch seven ounces of fine Jordan

almonds and one of bitter;[125] throw them into cold water as they
are done, and let them remain in it for an hour or two; then wipe, and
pound them to the finest paste, moistening them occasionally with a
few drops of cold water, to prevent their oiling; next, add to, and mix
thoroughly with them, seven ounces of highly-refined, dried, and
sifted sugar; put them into a small preserving-pan, or enamelled
stewpan, and stir them over a clear and very gentle fire until they are
so dry as not to adhere to the finger when touched; turn the paste
immediately into an earthen pan or jar, and when cold it will be ready
for use.
125. When these are objected to, use half a pound of the sweet almonds.
Jordan almonds, 7 oz.; bitter almonds, 1 oz.; cold water, 1
tablespoonful; sugar, 7 oz.
Obs.—The pan in which the paste is dried, should by no means be
placed upon the fire, but high above it on a bar or trevet: should it be
allowed by accident to harden too much, it must be sprinkled
plentifully with water, broken up quite small, and worked, as it
warms, with a strong wooden spoon to a smooth paste again. We
have found this method perfectly successful; but, if time will permit, it
should be moistened some hours before it is again set over the fire.
 TARTLETS OF ALMOND PASTE.

Butter slightly the smallest-sized pattypans, and line them with the
almond-paste rolled as thin as possible; cut it with a sharp knife
close to their edges, and bake or rather dry the tartlets slowly at the
mouth of a very cool oven. If at all coloured, they should be only of
the palest brown; but they will become perfectly crisp without losing
their whiteness if left for some hours in a very gently-heated stove or
oven. They should be taken from the pans when two-thirds done,
and laid, reversed, upon a sheet of paper placed on a dish or board,
before they are put back into the oven. At the instant of serving, fill
them with bright-coloured whipped cream, or with peach or apricot
jam; if the preserve be used, lay over it a small star or other
ornament cut from the same paste, and dried with the tartlets. Sifted
sugar, instead of flour, must be dredged upon the board and roller in
using almond paste. Leaves and flowers formed of it, and dried
gradually until perfectly crisp, will keep for a long time in a tin box or
canister, and they form elegant decorations for pastry. When a fluted
cutter the size of the pattypans is at hand, it will be an improvement
to cut out the paste with it, and then to press it lightly into them, as it
is rather apt to break when pared off with a knife. To colour it,
prepared cochineal, or spinach-green, must be added to it in the
mortar.
 FAIRY FANCIES.

(Fantaisies de Fées.)
A small, but very
inexpensive set of
tin cutters must be
had for this pretty
form of pastry,
which is, however,
quite worthy of so slight a cost. The short crust, of page 349,
answers for it better than puff paste. Roll it thin and very even, and
with the larger tin, shaped thus, cut out a dozen or more of small
sheets; then, with a couple of round cutters, of which one should be
about an inch in diameter, and the other only half the size, form four
times the number of rings, and lay them on the sheets in the manner
shown in the engraving. The easier mode of placing them regularly,
is to raise each ring without removing the small cutter from it, to
moisten it with a camel’s hair brush dipped in white of egg, and to lay
it on the paste as it is gently loosened from the tin When all the
pastry is prepared, set it into a very gentle oven, that it may become
crisp and yet remain quite pale. Before it is sent to table, fill the four
divisions of each fantaisie with preserve of a different colour. For
example: one ring with apple or strawberry jelly, another with apricot
jam, a third with peach or green-gage, and a fourth with raspberry
jelly. The cases may be iced, and ornamented in various ways
before they are baked. They are prettiest when formed of white
almond-paste, with pink or pale green rings: they may then be filled,
at the instant of serving, with well-drained whipped cream.
 MINCEMEAT.

(Author’s Receipt.)
To one pound of an unsalted ox-tongue, boiled tender and cut free
from the rind, add two pounds of fine stoned raisins, two of beef
kidney-suet, two pounds and a half of currants well cleaned and
dried, two of good apples, two and a half of fine Lisbon sugar, from
half to a whole pound of candied peel according to the taste, the
grated rinds of two large lemons, and two more boiled quite tender,
and chopped up entirely, with the exception of the pips, two small
nutmegs, half an ounce of salt, a large teaspoonful of pounded
mace, rather more of ginger in powder, half a pint of brandy, and as
much good sherry or Madeira. Mince these ingredients separately,
and mix the others all well before the brandy and the wine are
added; press the whole into a jar or jars, and keep it closely covered.
It should be stored for a few days before it is used, and will remain
good for many weeks. Some persons like a slight flavouring of
cloves in addition to the other spices; others add the juice of two or
three lemons, and a larger quantity of brandy. The inside of a tender
and well-roasted sirloin of beef will answer quite as well as the
tongue.
Of a fresh-boiled ox-tongue, or inside of roasted sirloin, 1 lb.;
stoned raisins and minced apples, each 2 lbs.; currants and fine
Lisbon sugar, each 2-1/2 lbs.; candied orange, lemon or citron rind, 8
to 16 oz.; boiled lemons, 2 large; rinds of two others, grated; salt, 1/2
oz.; nutmegs, 2 small; pounded mace, 1 large teaspoonful, and
rather more of ginger; good sherry or Madeira, 1/2 pint; brandy, 1/2
pint.
Obs.—The lemons will be sufficiently boiled in from one hour to
one and a quarter.
 SUPERLATIVE MINCEMEAT.

Take four large lemons, with their weight of golden pippins pared
and cored, of jar-raisins, currants, candied citron and orange-rind,
and the finest suet, and a fourth part more of pounded sugar. Boil the
lemons tender, chop them small, but be careful first to extract all the
pips; add them to the other ingredients, after all have been prepared
with great nicety, and mix the whole well with from three to four
glasses of good brandy. Apportion salt and spice by the preceding
receipt. We think that the weight of one lemon, in meat, improves
this mixture; or, in lieu of it, a small quantity of crushed macaroons
added just before it is baked.
 MINCE PIES. (ENTREMETS.)

Butter some tin pattypans well, and line them evenly with fine puff
paste rolled thin; fill them with mincemeat, moisten the edges of the
covers, which should be nearly a quarter of an inch thick, close the
pies carefully, trim off the superfluous paste, make a small aperture
in the centre of the crust with a fork or the point of a knife, ice the
pies or not, at pleasure, and bake them half an hour in a well-heated
but not fierce oven: lay a paper over them when they are partially
done, should they appear likely to take too much colour.
1/2 hour.
Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade

Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.

Let us accompany you on the journey of exploring knowledge and

personal growth!

textbookfull.com