Get the full ebook with Bonus Features for a Better Reading Experience on ebookgate.

com

Logic and Algebraic Structures in Quantum

Computing 1st Edition Chubb

https://ebookgate.com/product/logic-and-algebraic-
structures-in-quantum-computing-1st-edition-chubb/

OR CLICK HERE

DOWLOAD NOW

Download more ebook instantly today at https://ebookgate.com

Instant digital products (PDF, ePub, MOBI) available
Download now and explore formats that suit you...

Quantum Computing Explained 1st Edition David Mcmahon

https://ebookgate.com/product/quantum-computing-explained-1st-edition-
david-mcmahon/

ebookgate.com

Discrete structures logic and computability 4ed. Edition

Hein

https://ebookgate.com/product/discrete-structures-logic-and-
computability-4ed-edition-hein/

ebookgate.com

An introduction to quantum computing 1st Edition Phillip

Kaye

https://ebookgate.com/product/an-introduction-to-quantum-
computing-1st-edition-phillip-kaye/

ebookgate.com

Classical and Quantum Computing with C and Java

Simulations 1st Edition Yorick Hardy

https://ebookgate.com/product/classical-and-quantum-computing-with-c-
and-java-simulations-1st-edition-yorick-hardy/

ebookgate.com
Formation and Logic of Quantum Mechanics 3 Volume Set Vol
I The Formation of Atomic Models Vol II The Way to Quantum
Mechanics Vol III The Establishment and Logic of Quantum
Mechanics 1st Edition Mituo Taketani
https://ebookgate.com/product/formation-and-logic-of-quantum-
mechanics-3-volume-set-vol-i-the-formation-of-atomic-models-vol-ii-
the-way-to-quantum-mechanics-vol-iii-the-establishment-and-logic-of-
quantum-mechanics-1st-edition-mitu/
ebookgate.com

Maxwell s Demon 2 Entropy Classical and Quantum

Information Computing 1st Edition Harvey Leff

https://ebookgate.com/product/maxwell-s-demon-2-entropy-classical-and-
quantum-information-computing-1st-edition-harvey-leff/

ebookgate.com

Geometry of Time Spaces Non Commutative Algebraic Geometry

Applied to Quantum Theory 1st Edition Audal

https://ebookgate.com/product/geometry-of-time-spaces-non-commutative-
algebraic-geometry-applied-to-quantum-theory-1st-edition-audal/

ebookgate.com

Quantum computing From linear algebra to physical

realizations 1st Edition Mikio Nakahara

https://ebookgate.com/product/quantum-computing-from-linear-algebra-
to-physical-realizations-1st-edition-mikio-nakahara/

ebookgate.com

Computational Intelligence Synergies of Fuzzy Logic Neural

Networks and Evolutionary Computing 1st Edition Nazmul
Siddique
https://ebookgate.com/product/computational-intelligence-synergies-of-
fuzzy-logic-neural-networks-and-evolutionary-computing-1st-edition-
nazmul-siddique/
ebookgate.com
Logic and Algebraic Structures in Quantum Computing

Arising from a special session held at the 2010 North American Annual Meeting of
the ASL, this volume is an international cross-disciplinary collaboration with
contributions from leading experts exploring connections across their respective fields.
Themes range from philosophical examination of the foundations of physics and
quantum logic, to exploitations of the methods and structures of operator theory,
category theory, and knot theory in an effort to gain insight into the fundamental
questions in quantum theory and logic.
The book will appeal to researchers and students working in related fields,
including logicians, mathematicians, computer scientists, and physicists. A brief
introduction provides essential background on quantum mechanics and category
theory, which, together with a thematic selection of articles, may also serve as the
basic material for a graduate course or seminar.

Je n n i f e r Ch u b b is Assistant Professor of Mathematics at the University of San

Francisco, where she teaches a wide range of courses, including quantum computing,
to students in physics, computer science, and mathematics. She has a background in
physics, dynamical systems, and pure and applied math. Her current research focuses
on computable structure theory and algorithmic mathematics.

Al i Es k a n d a r i a n holds the positions of Dean and Professor at The George

Washington University. He is a theoretical physicist and a founding member of the
groups in astrophysics and quantum computing/information. He serves as co-director
of the Center for Quantum Computing, Information, Logic, and Topology.

Va l e n t i na Ha r i z a n ov is a Professor of Mathematics at The George Washington

University, where she also serves as co-director of the Center for Quantum Computing,
Information, Logic, and Topology. She is internationally recognized for her research in
mathematical logic, particularly in computability theory and computable model theory.
 L E C T U R E N OT E S I N L O G I C

A Publication of The Association for Symbolic Logic

This series serves researchers, teachers, and students in the field of symbolic
logic, broadly interpreted. The aim of the series is to bring publications to the
logic community with the least possible delay and to provide rapid
dissemination of the latest research. Scientific quality is the overriding
criterion by which submissions are evaluated.

Editorial Board
Jeremy Avigad
Department of Philosophy, Carnegie Mellon University
Zoe Chatzidakis
DMA, Ecole Normale Supérieure, Paris
Peter Cholak, Managing Editor
Department of Mathematics, University of Notre Dame, Indiana
Volker Halbach
New College, University of Oxford
H. Dugald Macpherson
School of Mathematics, University of Leeds
Slawomir Solecki
Department of Mathematics, University of Illinois at Urbana-Champaign
Thomas Wilke
Institut für Informatik, Christian-Albrechts-Universität zu Kiel

More information, including a list of the books in the series, can be found at
http://www.aslonline.org/books-lnl.html
 L E C T U R E N OT E S I N L O G I C 4 5

Logic and Algebraic Structures in

Quantum Computing

Edited by
JENNIFER CHUBB
University of San Francisco

ALI ESKANDARIAN
George Washington University, Washington DC

VALENTINA HARIZANOV
George Washington University, Washington DC

association for symbolic logic

 University Printing House, Cambridge CB2 8BS, United Kingdom

Cambridge University Press is part of the University of Cambridge.

It furthers the University’s mission by disseminating knowledge in the pursuit of
education, learning, and research at the highest international levels of excellence.

www.cambridge.org
Information on this title: www.cambridge.org/9781107033399
Association for Symbolic Logic
Richard Shore, Publisher
Department of Mathematics, Cornell University, Ithaca, NY 14853
http://www.aslonline.org
© Association for Symbolic Logic 2016
This publication is in copyright. Subject to statutory exception
and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without the written
permission of Cambridge University Press.
First published 2016
A catalogue record for this publication is available from the British Library.
Library of Congress Cataloguing in Publication Data
Names: Chubb, Jennifer. | Eskandarian, Ali. | Harizanov, Valentina S.
Title: Logic and algebraic structures in quantum computing / edited by Jennifer Chubb,
University of San Francisco, Ali Eskandarian, George Washington University, Washington
DC, Valentina Harizanov, George Washington University, Washington DC.
Description: Cambridge : Cambridge University Press, 2016. | Series: Lecture notes in
logic | Includes bibliographical references and index.
Identifiers: LCCN 2015042942 | ISBN 9781107033399 (hardback : alk. paper)
Subjects: LCSH: Quantum computing–Mathematics. | Logic, Symbolic and
mathematical. | Algebra, Abstract.
Classification: LCC QA76.889 .L655 2016 | DDC 006.3/843–dc23 LC record available at
http://lccn.loc.gov/2015042942
ISBN 978-1-107-03339-9 Hardback
Cambridge University Press has no responsibility for the persistence or accuracy of
URLs for external or third-party Internet Web sites referred to in this publication
and does not guarantee that any content on such Web sites is, or will remain,
accurate or appropriate.
 CONTENTS

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
Jennifer Chubb, Ali Eskandarian, and Valentina Harizanov
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Jennifer Chubb and Valentina Harizanov
A (very) brief tour of quantum mechanics, computation, and category
theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Allen Stairs
Could logic be empirical? The Putnam-Kripke debate . . . . . . . . . . . . . . 23
William C. Parke
The essence of quantum theory for computers . . . . . . . . . . . . . . . . . . . . . . 42
Adam Brandenburger and H. Jerome Keisler
Fiber products of measures and quantum foundations . . . . . . . . . . . . . . 71
Samson Abramsky and Chris Heunen
Operational theories and categorical quantum mechanics . . . . . . . . . . . 88
Bart Jacobs and Jorik Mandemaker
Relating operator spaces via adjunctions . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
Andreas Döring
Topos-based logic for quantum systems and bi-Heyting algebras . . . . 151
Bob Coecke
The logic of quantum mechanics – Take II . . . . . . . . . . . . . . . . . . . . . . . . . 174
Dimitri Kartsaklis, Mehrnoosh Sadrzadeh, Stephen Pulman, and Bob
Coecke
Reasoning about meaning in natural language with compact closed
categories and Frobenius algebras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
Louis H. Kauffman
Knot logic and topological quantum computing with Majorana
fermions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337

v
 PREFACE

This project grew out of a Special Session on Logic and the Foundations
of Physics at the 2010 North American Annual Meeting of the Association
for Symbolic Logic1 . Many of the session’s lecturers investigated the role of
algebraic structures in the context of the foundations of quantum physics,
especially in quantum information and computation. In addition to this
session, attendees heard tutorial lectures on quantum computing (given by
Bob Coecke, University of Oxford) and an invited lecture on intuitionistic
quantum logic (by Klaas Landsman, Radboud University, Nijmegen). The
talks were so well-received by conference participants that we felt a volume of
collected works on this subject would be a valuable addition to the literature.
The articles in this volume by mathematicians, philosophers, and scientists
address foundational issues and fundamental abstract structures arising in
highly active areas of theoretical, mathematical, and even experimental physics
relevant to quantum information and quantum computation. We hope that
the present collection advances this worthwhile program of scientific and
mathematical progress.
We would like to thank the authors that contributed to this volume, and
the ASL and Cambridge University Press for publishing it. This project was
partially supported by the George Washington University Centers & Institutes
Facilitating Fund Grant and by the University of San Francisco Faculty
Development Fund. Many thanks also to Bryan Fregoso (a University of San
Francisco student) for his invaluable assistance in assembling this volume.

Jennifer Chubb
Ali Eskandarian
Valentina Harizanov
Summer, 2015, Washington, D.C.

1 The full program is available in the Bulletin of Symbolic Logic, vol. 17 (2011), no. 1, pp. 135–137,

available online at https://www.math.ucla.edu/∼asl/bsl/1701-toc.htm.

vii
 INTRODUCTION

JENNIFER CHUBB, ALI ESKANDARIAN, AND VALENTINA HARIZANOV

In the last two decades, the scientific community has witnessed a surge in
activity, interesting results, and notable progress in our conceptual understand-
ing of computing and information based on the laws of quantum theory. One
of the significant aspects of these developments has been an integration of
several fields of inquiry that not long ago appeared to be evolving, more or less,
along narrow disciplinary paths without any major overlap with each other. In
the resulting body of work, investigators have revealed a deeper connection
among the ideas and techniques of (apparently) disparate fields. As is evident
from the title of this volume, logic, mathematics, physics, computer science
and information theory are intricately involved in this fascinating story. The
inquisitive reader might focus, perhaps, on the marriage of the most unlikely
and intriguing fields of quantum theory and logic and ask: Why quantum logic?
By many, “logic” is deemed to be panacea for faulty intuition. It is often
associated with the rules of correct thinking and decision-making, but not
necessarily in its most sublime role as a deep intellectual subject underlying the
validity of mathematical structures and worthy of investigation and discovery
in its own right. Indeed, within the realm of the classical theories of nature,
one may encounter situations that defy comprehension, should one hold to the
intuition developed through experiencing familiar macroscopic scenarios in
our routine impressions of natural phenomena.
One such example is a statement within the special theory of relativity that
the speed of light is the same in all inertial frames. It certainly defies the
common intuition regarding the observation of velocities of familiar objects in
relative motion. One might be tempted to dismiss it as contrary to observation.
However, while analyzing natural phenomena for objects moving close to
the speed of light and, therefore, unfamiliar in the range of velocities we
are normally accustomed to, logical deductions based on the postulates of
the special relativity theory lead to the correct predictions of experimental
observations.
There exists an undeniable interconnection between the deepest theories of
nature and mathematical reasoning, famously stated by Eugene Wigner as
the unreasonable efficacy of mathematics in physical theories. The sciences,
Logic and Algebraic Structures in Quantum Computing
Edited by J. Chubb, A. Eskandarian and V. Harizanov
Lecture Notes in Logic, 45
c 2016, Association for Symbolic Logic 1
2 JENNIFER CHUBB, ALI ESKANDARIAN, AND VALENTINA HARIZANOV

and in particular physics, have relied on, and benefited from, the economy of
mathematical expressions and the efficacy and rigor of mathematical reasoning
with its underlying logical structure to make definite statements and predictions
about nature. Mathematics has become the de facto language of the quantitative
sciences, particularly scientific theories, and the major discoveries and predictive
statements of these theories (whenever possible) are cast in the language of
mathematics, as it affords them elegance as well as economy of expression.
What happens if the syntax and grammar of such a language become inadequate?
This seems to have been the case when some of the more esoteric predictions
of the then new theory of quantum mechanics began to challenge the scientific
intuition of the times around the turn of the 20th century. This violation
of intuition was so severe that even the most prominent of scientists were
not able to reconcile the dictates of their intuition with the experimentally
confirmed predictions of the theory. The discomfort with some of the features
and predictions of quantum theory were, perhaps, most prominently brought
out in the celebrated work of Einstein, Podolsky, and Rosen (EPR) in the
mid 1930s. EPR fueled several decades of investigations on the foundations
of quantum theory that continue to this day. The main assertion of the
EPR work was that quantum theory had to be, by necessity, incomplete.
Otherwise, long held understanding of what should be taken for granted as
“elements of reality” had to be abandoned. Here, according to EPR, logical
deductions based on primitives that were the very essence of reality and logical
consistency forced the conclusion of the incompleteness of quantum theory;
as if considering quantum theory as complete would question one’s logical
fitness and one’s understanding of reality! Yet, in the decades since, with
increasing sophistication in experimentation, and multiple ways of testing
the theory, quantum theory has consistently outshined the alternatives. In
particular, many predictions relying on the sensibilities of classical theories,
where concepts such as separability, locality, and causality are the seemingly
indispensable factors in our understanding of reality, are found to be entirely
inconsistent with the actual reality around us. Quantum theory has not (as
yet) suffered any such blow.
Confronted with the stark inability to reconcile the predictions of a theory,
which are shown to be correct every time subjected to experimental verification,
and a logical structure that seems to fall short in facilitating correct thinking
and correct decision making (at least, in so far as the behavior of natural
phenomena at the quantum level is concerned), one is forced to consider and
question the validity of the premises on which that logical structure is built, or
to discover alternative structures. Furthermore, the striking applications of
quantum theory in the theory of computation, development of new algorithms,
and the promising prospects for the building of a computing machine operating
on the basis of the laws of quantum theory, necessitate a deeper investigation of
alternative logical structures that encompass the elements of this new quantum
 INTRODUCTION 3

reality. One must then give credence to the argument that, perhaps, the fault is
not with the revolutionary quantum theory; rather, it is with the inadequacies
of logical structures that were insufficient to be expanded and applied to a
world that does not comply with the notions embodied in our understanding
of the macroscopic classical physical theories of nature.
The utility of logical rules is most pronounced when applied to the building
and operation of computing machines. With the advent of computing that
takes advantage of the laws of quantum theory, i.e., quantum computing,
it is only natural to search for those logical and algebraic structures that
underlie the scaffolding of the quantum rules in computations. As obvious
as it is that Boolean logic underlies classical computing and much of classical
reasoning, it is equally obvious that it is not sufficient to express the logic
underlying quantum mechanics or quantum computing. Birkhoff and von
Neumann were among the first to propose a generalization of Boolean logic in
which propositions about quantum systems could be formulated. While their
endeavor was revolutionary, the Birkhoff-von Neumann quantum logic was
not to be the final word on the subject of a logic for quantum mechanics, and
indeed the investigation continues with increasing urgency.
In this volume, we present the work of a select group of scholars with an abid-
ing interest in tackling some of the fundamental issues facing quantum comput-
ing and information theory, as investigated from the perspective of logical and al-
gebraic structures. This selection, no doubt, reflects the intellectual proclivities
and curiosities of the editors, within the reasonable limitations of space and cov-
erage of topics for a volume of this size, and for the purpose of generating ideas
that would fuel further investigation and research in these and related fields.
The first two articles, by Stairs and Parke, address philosophical and histori-
cal issues. Brandenburger and Keisler use ideas from continuous model theory
to explore determinism and locality in quantum mechanical systems. Abramsky
and Heunen, and Jacobs and Mandemaker describe the relationship between
the category-theoretic and operator-theoretic approaches to the foundations
of quantum physics. Döring gives a topos-based distributive form of quantum
logic as an alternative to the quantum logic of Birkhoff and von Neumann.
The papers by Coecke and Kartsaklis et al. use a diagrammatic calculus in
analyzing quantum mechanical systems and, very recently, in computational
linguistics. Kauffman’s article presents an extensive treatment of the prominent
role of algebraic structures arising from topological considerations in quantum
information and computing; the pictorial approach used in knot theory is
closely related to the quantum categorical logic presented in other articles in
this volume.

Could logic be empirical? The Putnam-Kripke debate, by Allen Stairs. In

his article in the present volume, Stairs outlines Hilary Putnam’s position that
quantum mechanics provides an empirical basis for a re-evaluation of our
4 JENNIFER CHUBB, ALI ESKANDARIAN, AND VALENTINA HARIZANOV

idea of logic and Saul Kripke’s response, in which he takes issue with the very
idea of a logic that is based on anything empirical. Stairs carefully interprets
their positions, and in the end offers the beginnings of a compromise, which
includes “disjunctive facts,” which can be true even if their disjuncts are not,
and the notion of “l-complementarity,” to describe the relationship between
statements having non-commuting associated projectors. The article wrestles
with the idea of whether and how quantum mechanics should inform our logic
and reasoning processes.
The essence of quantum theory for computers, by William C. Parke. In this
article, Parke provides a thorough yet succinct introduction to the elements
of physical theories, classical and quantum, which are relevant to a deeper
understanding of the mathematical and logical structures underlying (or
derived) from such theories, and important in the appreciation of the more
subtle quandaries of quantum theory, leading to its utilization in computation.
The emphasis has been placed on the physical content of information and
elements of computation from a physicist’s point of view. This includes a
treatment of the role of space-time in the development of physical theories from
an advanced point of view, and the limitations that our current understanding
of space-time imposes on building and utilizing computing machines based
on the rules of quantum theory. The treatment of the principles of quantum
theory is also developed from an advanced point of view, without too much
focus on unnecessary details, but covering the essential conceptual ingredients,
in order to set the stage properly and provide motivation for the work of the
others on logical and algebraic structures.
Fiber products of measures and quantum foundations, by Adam Branden-
burger and H. Jerome Keisler. In this model-theoretic article, the authors use
fiber products of (probability) measures within a framework they construct
for empirical and hidden-variable models to prove determinization theorems.
These objects (fiber products) were conceived by Rae Shortt in a 1984 paper,
and were used recently by Itaı̈ Ben Yaacov and Jerome Keisler in their work on
continuous model theory (2009). Techniques in continuous model theory are
relevant to the notion of models of quantum structures as in that context the
“truth value” of a statement may take on a continuum of values, and can be
thought of as probabilistic. In this case, a technique employed in continuous
model theory is used in the construction of models in proofs of theorems
that assert that every empirical model can be realized by an extension that is
a deterministic hidden-variable model, and for every hidden-variable model
satisfying locality and -independence, there is a realization-equivalent (both
models extend a common empirical submodel) hidden-variable model satisfy-
ing determinism and -independence. The latter statement, together with Bell’s
theorem, precludes the existence of a hidden-variable model in which both
determinism and -independence hold. The notion of -independence was
 INTRODUCTION 5

first formulated by W. Michael Dickson (2005). It says that the choices made
by an entity as to which observable to measure in a system are not influenced
by the process of the determination of the value of a relevant hidden-variable.
Operational theories and categorical quantum mechanics, by Samson Abram-
sky and Chris Heunen. There are two complementary research programs in
the foundations of quantum mechanics, one based on operational theories
(also called general probabilistic theories) and the other on category-theoretic
foundation of quantum theory. Samson Abramsky and Chris Heunen establish
strong and important connections between these two formalisms. Operational
theories focus on empirical and observational content, and quantum mechan-
ics occupies one point in a space of possible theories. The authors define a
symmetric monoidal categorical structure of an operational theory, which they
call process category, and exploit the ideas of categorical quantum mechanics
to obtain an operational theory as a certain representation of this process
category. They lift the notion of non-locality to the general level of operational
category. They further propose to apply a similar analysis to contextuality,
which can be viewed as a broader phenomenon than non-locality.
Relating operator spaces via adjunctions, by Bart Jacobs and Jorik Mande-
maker. By exploiting techniques of category theory, Jacobs and Mandemaker
clarify and present in a unified framework various, seemingly different results
in the foundation of quantum theory found in the literature. They use category-
theoretic tools to describe relations between various spaces of operators on
a finite-dimensional Hilbert space, which arise in quantum theory, including
bounded, self-adjoint, positive, effect, projection, and density operators. They
describe the algebraic structure of these sets of operators in terms of modules
over various semirings, such as the complex numbers, the real numbers, the
non-negative real numbers. The authors give a uniform description of such
modules via the notion of an algebra of the multiset monad. They show how
some spaces of operators are related by free constructions between categories
of modules, while the other spaces of operators are related by a dual adjunction
between convex sets (conveniently described via a monad) and effect modules.
Topos-based logic for quantum systems and bi-Heyting algebras, by Andreas
Döring. Döring replaces the standard quantum logic, introduced by Birkhoff
and von Neumann, which comes with a host of conceptual and interpretational
problems, by the topos-based distributive form of quantum logic. Instead of
having a non-distributive orthomodular lattice of projections, he considers
a complete bi-Heyting algebra of propositions. More specifically, Döring
considers clopen subobjects of the presheaf attaching the Gelfand spectrum to
each abelian von Neumann algebra, and shows that these clopen subojects form
a bi-Heyting algebra. He gives various physical interpretations of the objects
in this algebra and of the operations on them. For example, he introduces two
6 JENNIFER CHUBB, ALI ESKANDARIAN, AND VALENTINA HARIZANOV

kinds of negation associated with the Heyting and co-Heyting algebras, and
gives physical interpretation of the two kinds of negation. Döring considers the
map called outer daseinisation of projections, which provides a link between
the usual Hilbert space formalism and his topos-based quantum logic.
The logic of quantum mechanics – Take II, by Bob Coecke. Schrödinger
maintained that composition of systems is the heart of quantum computing,
and Coecke agrees. He suggests that the Birkhoff-von Neumann formulation
of quantum logic fails to adequately and elegantly capture composition of
quantum systems. The author puts forth a model of quantum logic that is
based on composition rather superposition. He axiomatizes composition
without reference to underlying systems using strict monoidal categories as
the basic structures and explains a graphical language that exactly captures
these structures. Imposing minimal additional structure on these categories
(to obtain dagger compact categories) allows for the almost trivial derivation
of a number of quantum phenomena, including quantum teleportation and
entanglement swapping. This (now widely adopted) formalism has been used
not only to solve open problems in quantum information theory, but has also
provided new insight into non-locality.
Coecke’s framework has been applied both to logic concerned with natural
language interpretations, and to more formal automated reasoning processes.
In this article, the focus is on the former. Coecke applies the graphical language
of dagger compact categories to natural language processing—“from word
meaning to sentence meaning”—implementing Lambek’s theory of grammar
and the notion of words as “meaning vectors.” He argues that sentence
meaning amounts to more than the meanings of the constituent words, but
also the way in which they compose.
In the end, Coecke confesses that dagger compact categories do not capture
all we might want them to, in particular, measurement, observables, and
complementarity are left by the wayside. The model can be expanded (using
spiders!) in such a way that all these are captured. Coecke closes with
speculation about an important question: Where is the traditional logic hiding
in all this?
Reasoning about meaning in natural language with compact closed categories
and Frobenius algebras, by Dimitri Kartsaklis, Mehrnoosh Sadrzadeh, Stephen
Pulman, and Bob Coecke. The authors apply category-theoretic methods to
computational lingustics by mapping the derivations of the grammar logic to
the distributional interpretation via a strongly monoidal functor. Such functors
are structure preserving morphims. Grammatical structure is modeled through
the derivations of pregroup grammars. A pregroup is a partially ordered
monoid with left and right adjoints for every element in the partial order. The
authors build tensors for linguistic constructs with complex types by using
a Frobenius algebra. The Frobenius operations allow them to assign and
 INTRODUCTION 7

compare the meanings of different language constructs such as words, phrases,

and sentences in a single space. The authors present their experimental results
for the evaluation of their model in a number of natural languages.
Knot logic and topological quantum computing with Majorana fermions, by
Louis H. Kauffman. Kauffman presents several topics exploring the relation-
ship between low-dimensional topology and quantum computing. These topics
have been introduced and developed by Kauffman and Samuel J. Lomonaco
over the last ten years. Kauffman uses the diagrammatic approach, and is
particularly interested in models based upon the Temperley-Lieb categories.
He discusses from several different perspectives the Fibonacci model related
to the Temperley-Lieb algebra at fifth roots of unity. Kauffman shows how
knots are related to braiding and quantum operators, as well as to quantum
set-theoretic foundations. For example, the negation can generate the fusion
algebra for a Majorana fermion, which is a particle that interacts with itself
and can even annihilate itself. Thus, Kauffman calls the negation the mark.
He investigates the relationship between knot-theoretic recoupling theory
and topological quantum field theory. Kauffman works with braid groups
and their representations, and produces unitary representations of the braid
groups that are dense in the unitary groups. He describes the Jones polynomial
in terms of his bracket polynomial and applies his approach to design a
quantum algorithm for computing the colored Jones polynomials for knots
and links. Kauffman also gives a quantum algorithm for computing the
Witten-Reshetikhin-Turaev invariant of three manifolds.
DEPARTMENT OF MATHEMATICS
UNIVERSITY OF SAN FRANCISCO
SAN FRANCISCO, CA 94117
E-mail: [email protected]

DEPARTMENT OF PHYSICS
VIRGINIA SCIENCE & TECHNOLOGY CAMPUS
GEORGE WASHINGTON UNIVERSITY
ASHBURN, VIRGINIA 20147
E-mail: [email protected]

DEPARTMENT OF MATHEMATICS
GEORGE WASHINGTON UNIVERSITY
WASHINGTON, D.C. 20052
E-mail: [email protected]
 A (VERY) BRIEF TOUR OF QUANTUM MECHANICS,
COMPUTATION, AND CATEGORY THEORY

JENNIFER CHUBB AND VALENTINA HARIZANOV

This chapter is intended to be a brief treatment of the basic mechanics,

framework, and concepts relevant to the study of quantum computing and
information for review and reference. Part 1 (sections 1– 4) surveys quantum
mechanics and computation, with sections organized according to the com-
monly known postulates of quantum theory. The second part (sections 5–7)
provides a survey of category theory. Additional references to works in this
volume are included throughout, and general references appear at the end.

Part 1: Quantum mechanics & computation

§1. Qubits & quantum states.

Postulate of quantum mechanics: Representing states of systems. The state of
a quantum system is represented by a unit-length vector in a complex Hilbert
space1 , H, that corresponds to that system. The state space of a composite
system is the tensor product of the state spaces of the subsystems.
The Dirac bra-ket notation for states of quantum systems is ubiquitous
in the literature, and we adopt it here. A vector in a complex Hilbert space
representing a quantum state is written as a ket, |, and its conjugate-transpose
(adjoint, or sometimes Hermitian conjugate) is written as a bra, |. In this
notation, a bra-ket denotes an inner product, ϕ|, and a ket-bra denotes an
outer product, |ϕ|.
Each one-dimensional subspace of H corresponds to a possible state of the
system, and a state is usually described as a linear combination in a relevant
orthonormal basis. The basis elements are often thought of as basic states.
Quantum systems can exist in a superposition of more than one basic state: If a
quantum system has access to two basic states, say |α and |, then, in general,
the system’s “current state” can be represented by a linear combination of
these states in complex Hilbert space:
| = c1 |α + c2 |, where ||| = 1.
1A Hilbert space is a complete, normed metric space, where the norm and distance function
are induced by an inner product defined on the space.

Logic and Algebraic Structures in Quantum Computing

Edited by J. Chubb, A. Eskandarian and V. Harizanov
Lecture Notes in Logic, 45
c 2016, Association for Symbolic Logic 8
 QUANTUM MECHANICS & CATEGORY THEORY 9

The complex coefficients, c1 and c2 , of |α and | give classical probabilistic
information about the state. For example, the value |c1 |2 is the probability that
the system would be found to be in state |α upon measurement. The coefficient
itself, c1 , is called the probability amplitude. Two vectors in H represent the
same state if they differ only by a global phase factor: If | = e i |ϕ, then
| and |ϕ represent the same state, and the (real) probabilities described by
the coefficients are the same.
The squared norm of the state vector | is the inner product of | with
itself, i.e., the bra-ket |. The quantity |ϕ||2 is the probability that
upon measurement, | will be found to be in state |ϕ, and ϕ| is the
corresponding probability amplitude. (More about measurement of quantum
systems can be found in Section 3 below.)
1.1. Qubits. A classical bit can be in only one of two states at a given
time, |0 or |1. A quantum bit or qubit may exist in a superposition of these
basic (orthogonal) states, | = c1 |0 + c2 |1, where c1 and c2 are complex
probability amplitudes. More precisely, a qubit is a 2-dimensional quantum
system, the state of which is a unit-length vector in H = C2 . The basic states
for this space are usually thought of as |0 and |1, but at times other bases
are used (for example, {|+, |−} or {| ↑, | ↓}). Basic states are typically
the eigenstates (eigenvectors) of an observable of interest (see discussion of
measurement below).
Any unit vector that is a (complex) linear combination of the basic states
is a pure state and non-trivial linear combinations are superpositions. So-
called mixed states are not proper state vectors, they are classical probabilistic
combinations of pure states and are best represented by density matrices.
The state space of a qubit is often visualized as a point on the Bloch sphere.
The norm of a state vector is always one, and states that differ only by a global
phase factor are identified, so two real numbers,  and φ, suffice to specify a
distinct state via the decomposition
   
 
| = cos |0 + e sin
iφ
|1.
2 2

Respectively, the range of values taken on by  and φ may be restricted

to the intervals [0, ] and [0, 2 ) without any loss of generality, and so the
corresponding distinct states may be mapped uniquely onto the unit sphere in
R3 . In this visualization, the basic vector |0 points up and |1 points down, 
describes the latitudinal angle, and ϕ the longitudinal angle. Orthogonal states
are antipodal on the Bloch sphere. Note that states that differ by a global
phase factor will (by design) coincide in this visualization.
1.2. Composite quantum systems. As described above, a single quantum
system (for example, a single qubit) exists in a pure state that may be a
superposition of basic states. A composition of systems may exist either in a
10 JENNIFER CHUBB AND VALENTINA HARIZANOV

separable or an entangled state. Separable states are states that can be written
as tensor products of pure states of the constituent subsystems. Entangled
states cannot be so written; they are non-trivial (complex) linear combinations
of separable states. In the case of an entangled state, the subsystems cannot be
thought of as existing in states independent of the composed system.
Example 1.1. Suppose
√ we have a system of two qubits, the√first in state
| = (|0 + |1)/ 2 and the second in state |ϕ = (|0 − |1)/ 2. The state
of the combined system is
1
| ⊗ |ϕ = ||ϕ = (|00 − |01 + |10 − |11).
2
Such a state of the composite system that can be written as a tensor product of
pure states is called separable.
Example 1.2. The Bell states of a 2-qubit system are not separable; they are
important and canonical examples of entangled states:
|00 + |11 |00 − |11
√ √
2 2
|01 + |10 |01 − |10
√ √
2 2
Example 1.3. The GHZ states (for Greenberger-Horne-Zeilinger) are ex-
amples of entangled states in composite systems that have three or more
subsystems. The GHZ state for a system with n subsystems is
|0⊗n + |1⊗n
√ .
2
For more on entangled states, see Parke’s article in this volume, or Section 6
of Kauffman’s article.

§2. Transformations and quantum gates.

Postulate of quantum mechanics: Evolution of systems. The time evolution
of a closed quantum system is described by a unitary transformation.
A transformation is unitary if its inverse is equal to its adjoint. Such
transformations preserve inner products and are reversible, deterministic, and
continuous. In quantum computing, algorithms are often described as circuits
in which information (and time) flows from left to right. Quantum gates
represent unitary transformations applied to qubits in such a circuit.
Example 2.1. The Hadamard gate. The 1-qubit Hadamard gate has as input
and output one qubit, as shown in the simple circuit diagram below:
 QUANTUM MECHANICS & CATEGORY THEORY 11
 
Its matrix representation (with respect to the basis |0 = [1 0]T , |1 = [0 1]T )
is:  
1 1 1
H =√ .
2 1 −1
 
1
This transformation applied to the basic state |0 = results in the
  0
1
superposition H |0 = √12 (|0 + |1) = √12 .
1
Example 2.2. The controlled-not gate. Another important quantum gate is
the controlled-not or CNOT gate. The gate requires two inputs, one designated
as the control input (passing through the solid dot) and the other as the target
input:

When the control input is in state |0, the gate does nothing. If the control is
in state |1 (as it is in the diagram above), the gate acts by “flipping” the non-
control (target) input as follows: If the target input is in state | = c0 |0+c1 |1,
then flipping transforms the state to |   = c0 |1 + c1 |0. The gate does not
alter the control bit. Thematrix representation of CNOT is the following (given
 basis |00 = [1 0 0 0] , |01 = [0 1 0 0] , |10 = [0 0 1 0] ,
T T T
with respect to the
|11 = [0 0 0 1] ):
T
⎡ ⎤
1 0 0 0
1⎢ 0 1 0 0 ⎥
CNOT = ⎢ ⎥.
2⎣ 0 0 0 1 ⎦
0 0 1 0
For more on quantum gates and unitary transformations of quantum systems,
see Parke’s and Kauffman’s articles in this volume.

§3. Measurement.
Postulate of quantum mechanics: Measurement. The notion of measurement
is described in terms of observables represented by Hermitian (self-adjoint)
matrices. (It should be noted that not all such matrices describe physically
meaningful measurements.)
A Hermitian matrix has all real eigenvalues, and these represent the possible
values obtained upon measurement of the observable. Moreover, distinct
eigenvalues yield orthogonal eigenvectors. These matrices are often described
in terms of their spectral decompositions. Upon measurement, a system’s
12 JENNIFER CHUBB AND VALENTINA HARIZANOV

state (or wave function) experiences a “collapse” and is not preserved. After
measurement, the state of the system is the eigenvector corresponding to the
eigenvalue that was the result of the measurement.
Example 3.1. If the matrix A corresponding to an observable A has (real)
eigenvalue a and corresponding unit-length eigenvector |va , then the proba-
bility that measuring A on state |ϕ will yield the value a is given by |va |ϕ|2 .
If a is the result of the measurement of A on |ϕ, the system is left in state
|va . If we consider the result of such a measurement as a random variable,
the expected value (expectation value) of that quantity is given by ϕ|A|ϕ.
Very briefly, if the matrices representing two different observables are non-
commuting, then the observables are often referred to as complementary
and measurements of these observables are subject to uncertainty limits.
Complementary observables suffer from necessarily limited precision when
measured simultaneously as a result of the Heisenberg Uncertainty Principle.

§4. No-go theorems and teleportation.

4.1. No cloning. In classical computation, it is possible to implement error
correction by simply duplicating the classical data as needed. This is not the
case in quantum computations.
Let | be an arbitrary state in state space H, and |e be an ancillary state
(independent of |) in an identical state space. To “clone” the state |,
we would need to have a unitary transformation that when applied to ||e
replaces the ancillary state with a copy of |, yielding ||.
Theorem 4.1 (No-cloning theorem). There is no unitary operator U so that
for all states | and ancillary states |e,
U ||e = ||.
To see why, consider the possibility that there does exist such an operator U .
As U must be unitary, it must preserve inner products, hence for any  and ϕ,
we must have the following:
ϕ| = e|ϕ||e = e|ϕ|U † U ||e = ϕ|ϕ|| = (ϕ|)2 .
We see that ϕ| must be either 0 or 1 in order for this equality to hold, and
so such a U preserves inner product only selectively—the states |ϕ and |
must be identical or orthogonal.
4.2. The EPR paradox, hidden variables, and Bell’s Theorem. In 1935, Ein-
stein, Podolsky, and Rosen (EPR) questioned the completeness of quantum
mechanics in the form of a thought experiment involving the measurement of
one part of a 2-particle entangled system. According to EPR, two mutually
exclusive conclusions may be reached regarding quantum mechanics: either
quantum mechanics is incomplete, or the physical quantities associated with
two non-commuting operators cannot have simultaneous reality. Subsequently,
 QUANTUM MECHANICS & CATEGORY THEORY 13

building on the behavior of a two-component system under the laws of quantum

theory, EPR argue for the incompleteness of quantum theory.
The following scenario captures the idea of the quandary they posed. Imagine
that two particles, A and B, interact and then part ways. If one measures
the momentum of particle A, he may compute the momentum of particle B
exactly due to entanglement. If he subsequently measures the momentum
of particle B, the result will be exactly that computed value. Similarly, the
particles’ positions may be observed, computed, and checked. However, the
measurement operators corresponding to these observables (position and
momentum) do not commute, and hence an exact knowledge of position
entails some uncertainty in the value of momentum. The EPR argument makes
a case for being able to assign two different wave functions (or states) to the
same reality (particle B), by judicious choice of measurements on particle A,
which leads to the conclusion that quantum mechanics must be incomplete.
A related question is this: How does particle B “know” to have a precisely
defined momentum and an uncertain position when particle A’s momentum
is measured? According to the principle of locality, a physical process occur-
ring in one place should not be able to affect a physical process in another
location (outside the light cone of the first process). This scenario seems to
entail either superluminal transmission of information between the particles
(violating locality), or some “hidden variable” or “element of reality” encoding
the information as yet unaccounted for by quantum mechanics (assuming
determinism or realism). This is the idea underlying the famous EPR paradox.
In 1964, John Stewart Bell formalized (mathematically) the notions of
locality and realism, and gave a set of inequalities that would provide a test
of quantum mechanics against a local hidden variable theory. In the 1970s
and 1980s, physical experiments (carried out most famously by Alain Aspect)
demonstrated in favor of the former. What is known as Bell’s Theorem is the
summary of all this, asserting that no locally realistic theory can make the
predictions of quantum mechanics.
Another related theorem is the Kochen-Specker Theorem, which says that a
non-contextual hidden variable theory (one in which the value of an observable
in a system is independent of the apparatus used to measure it) is unable to
make the predictions of quantum mechanics.
4.3. Quantum teleportation. It would be difficult to overstate the importance
of entanglement in quantum computing and the difficulty in representing and
interpreting this phenomenon in possible quantum logics. A basic illustration
of the power of entanglement is in the quantum teleportation protocol: An EPR
pair, that is, a pair of qubits in a (entangled) Bell state, are prepared. One qubit
is in the possession of entity A (Alice) and the other is in the possession of
entity B (Bob). Alice also has a qubit, |, which she would like to send to Bob.
To do this, Alice applies a CNOT transformation to her two qubits, using | as
the control, followed by an application of the Hadamard transformation to |.
14 JENNIFER CHUBB AND VALENTINA HARIZANOV

She then measures both of her qubits2 (they are destroyed in the process), and
(classically) communicates to Bob the (classical) information that results of
her measurements. Upon receiving this information, Bob preforms one of four
corresponding transformations, T , resulting in the transformation of his qubit
into the state |, which Alice wished to transmit to him.

Note that this protocol does not violate the no-cloning theorem (Alice’s copy
is destroyed), nor Bell’s Theorem (classical information must be transmitted
subluminally).
For alternative formulations of the quantum teleportation protocol in a
graphical language and another (similar) formulation in quantum topology,
see Coecke’s and Kauffman’s (respectively) articles in this volume.
For more detailed exposition on all these ideas and topics, the following
texts may be useful:
Textbooks at the undergraduate level
• Quantum Computing for Computer Scientists, by Noson S. Yanofsky and
Mirco A. Mannucci, Cambridge University Press, 2008.
• An Introduction to Quantum Computing, by Phillip Kaye, Raymond
Laflamme, and Michele Mosca, Oxford University Press, 2007.
• Quantum Computing: A Gentle Introduction, by Eleanor Rieffel and
Wolfgang Polak, MIT Press, 2011.
• Quantum Computer Science, by N. David Mermin, Cambridge University
Press, 2007.
At the graduate or research level
• Quantum Computation and Quantum Information, by Michael A. Nielsen
and Isaac L. Chuang, Cambridge University Press, 2011.

Part 2: Category theory for quantum computing

In physics, in the 1970s, Penrose used graphical language to represent

linear operators, their products, and tensor products: boxes for operators,
incoming wires for superscripts, and outgoing wires for subscripts. These
diagrams represented various categories, which are of importance in physics
2 This entire process is sometimes called a Bell measurement.
 QUANTUM MECHANICS & CATEGORY THEORY 15

and quantum computing. Of particular importance are tensor categories,

also called monoidal categories, which have been used by S. Abramsky and
B. Coecke as a framework for quantum theory. Their categorical quantum
mechanics can be also viewed as a suitable quantum logic. We will give a brief
survey of monoidal categories. For more details see [3] and [1].

§5. Basic category theory. A category C consists of a class of objects, ob(C),

and a class of morphisms, hom(C), also called maps or arrows with specific
abstract properties. For every pair of objects, A and B, there is a class of
morphisms denoted by homC (A, B), or simply hom(A, B) when the category
is clear from the context. A morphism f has a domain dom(f) (also
called source) and a codomain cod(f) (also called target), which we write
f : dom(f) → cod(f). The morphisms are equipped with composition
◦, which is an associative operation that respects domain and codomain
information. That is,
(i) (f ◦ g) ◦ h = f ◦ (g ◦ h),
where f : A → B, g : D → A, and h : C → D. For every object A, the set
hom(A, A) contains the identity morphism idA such that for every f : A → B,
we have
(ii) f ◦ idA = f
and
(iii) idB ◦ f = f.
The equations (i)–(iii) can be viewed as the axioms for the categories. The
opposite category (also called dual category) of C is formed by reversing the
morphisms, that is, by interchanging the domain and the codomain of each
morphism. It is denoted by C op . A category C is called small if both ob(C) and
hom(C) are sets, and it is called locally small is for every pair of objects A, B,
the class hom(A, B) is a set.
A morphism f : A → B is called a monomorphism or monic if f ◦ g1 = f ◦ g2
implies g1 = g2 for all morphisms g1 , g2 : C → A. A morphism f : A → B has
a left inverse, also called a retraction of f, if there is a morphism g : B → A such
that g ◦ f = idA . Clearly, a morphism with a left inverse is a monomorphism.
The converse may not be true. A morphism f : A → B is called an epimorphism
or epic if g1 ◦ f = g2 ◦ f implies g1 = g2 for all morphisms g1 , g2 : B → C . A
morphism f : A → B has a right inverse, also called a section of f, if there is a
morphism g : B → A such that f ◦ g = idB . A morphism with a right inverse
is an epimorphism, but the converse may not be true. If a morphism has both a
left inverse and a right inverse, then the two inverses are equal. Hence we have
the following definition. A morphism f : A → B is called an isomorphism if
there exists a morphism g : B → A such that f ◦ g = idB and g ◦ f = idA . If
16 JENNIFER CHUBB AND VALENTINA HARIZANOV

it exists, g is unique and is called the inverse of f, and hence f is the inverse
of g.
Examples of well-known categories include the category of sets as objects
with functions as morphisms, the category of vector spaces as objects with
linear maps as morphisms, and the category of Hilbert spaces as objects with
unitary transformations as morphisms. In the graphical representation, object
variables label edges (“wires”) and morphism variables label nodes (“boxes”).
The composition is represented by connecting the outgoing edge of one diagram
to the incoming edge of another, while the identity morphism is represented as
a continuing edge.
Functors capture the notion of a homomorphism between two categories.
They preserve identity morphisms and composition of morphisms. More
precisely, a functor Φ from a category C to a category D is a function that maps
every object A of C to an object Φ(A) of D, as well as every morphism of C to
a corresponding morphism of D such that the following is satisfied. For every
pair A, B of objects from C, each morphism f ∈ hom(A, B) in C is mapped to
a morphism Φ(f) ∈ hom(Φ(A), Φ(B)) in D such that

Φ(g ◦ h) = Φ(g) ◦ Φ(h) ∧ Φ(idA ) = idΦ(A) .

A functor from C to D is also called a covariant functor, in order to distinguish

it from a contravariant functor, which reverses the order of composition. A
contravariant functor Ψ from C to D is a map that associates to each object A
in C an object Ψ(A) in D, and associates to each morphism f ∈ hom(A, B) in
C a morphism Ψ(f) ∈ hom(Ψ(B), Ψ(A)) in D such that

Ψ(g ◦ h) = Ψ(h) ◦ Ψ(g) ∧ Ψ(idA ) = idΨ(A) .

A functor Φ between locally small categories C and D is called faithful if it is

injective when restricted to each set of morphisms that have a given domain
and codomain. That is, for every pair A, B of objects in C, the induced function

ΦA,B : homC (A, B) → homD (Φ(A), Φ(B))

is injective. On the other hand, a faithful functor may not be injective on

objects or morphisms. A functor Φ is called full if the induced functions ΦA,B
are surjective.
Natural transformations capture the notion of a homomorphism between
two functors. That is, given two categories, C and D, and two functors from C
to D, Φ and Ψ, a natural transformation N : Φ → Ψ consists of the family of
morphisms for every object A of C, A : Φ(A) → Ψ(A), such that for every
f ∈ homC (A, B), we have

Ψ(f) ◦ A = B ◦ Φ(f).
 QUANTUM MECHANICS & CATEGORY THEORY 17

The content of the equation is captured by the following diagram.

A
Φ(A) > Ψ(A)
Φ(f) Ψ(f)
∨ ∨
B
Φ(B) > Ψ(B)

§6. Monoidal categories. A monoidal category captures the notion of a

tensor product as a binary operation of objects, A ⊗ B, and of morphisms,
f ⊗ g. The domain of f ⊗ g is the tensor product of the domains of f and g,
and the codomain of f ⊗ g is the tensor product of the codomains of f and g.
The tensor product of objects is associative in the sense that for every triple
(A, B, C ) of objects, there is an isomorphism
αA,B,C : (A ⊗ B) ⊗ C → A ⊗ (B ⊗ C ).
The tensor product is a bifunctor, which means that it satisfies the following
equations for morphisms:
(f1 ⊗ f2 ) ◦ (f3 ⊗ f4 ) = (f1 ◦ f3 ) ⊗ (f2 ◦ f4 )
and
idA⊗B = idA ⊗ idB .

(See Coecke’s article in this volume for a wire diagram representation of this
equation.)
A monoidal category also has a constant unit object denoted by I . For every
object A, there is an isomorphism (left)
A : I ⊗ A → A

and an isomorphism (right)

A : A ⊗ I → A.
For morphisms f : A → A , g : B → B  , h : C → C  , we have
(f ⊗ (g ⊗ h)) ◦ αA,B.C = αA ,B  .C  ◦ (f ⊗ g) ⊗ h),
f ◦ A = A ◦ (idI ⊗ f),
f◦ A = A ◦ (f ⊗ idI ).
In addition, the following triangle axiom is satisfied for every pair of objects
A, B:

A ⊗ idB = (idA ⊗ B ) ◦ αA,I,B .

18 JENNIFER CHUBB AND VALENTINA HARIZANOV

Both sides map (A ⊗ I ) ⊗ B to A ⊗ B. This equation is captured in the

following diagram.
αA,I,B
(A ⊗ I ) ⊗ B > A ⊗ (I ⊗ B)

A ⊗idB > idA ⊗B

<
A⊗B

Also, the following pentagon axiom is satisfied for every quadruple of objects
A, B, C, D:
(idA ⊗ αB,C,D ) ◦ (αA,B⊗C,D ◦ (αA,B,C ⊗ idD )) = αA,B.C ⊗D ◦ αA⊗B,C,D .
Both sides map ((A ⊗ B) ⊗ C ) ⊗ D to A ⊗ (B ⊗ (C ⊗ D)). This relationship
is visualized in the following diagram.
αA,B⊗C,D
(A ⊗ (B ⊗ C )) ⊗ D > A ⊗ ((B ⊗ C ) ⊗ D)
∧
αA,B,C ⊗idD idA ⊗αB,C,D
∨
((A ⊗ B) ⊗ C ) ⊗ D A ⊗ (B ⊗ (C ⊗ D))
>

αA⊗B,C,D > αA,B,C ⊗D

(A ⊗ B) ⊗ (C ⊗ D)

In the graphical language, the tensor product of objects is represented

by parallel wires (input or output) from the bottom to the top, and the
unit object is represented by no wire. Tensor product of morphisms is
represented by stacking their diagrams. Examples of monoidal categories
are vector spaces, or Hilbert spaces, with either direct sum or tensor product,
as well as sets with direct products or disjoint unions. When no additional
properties are assumed for a monoidal category, we often call it planar monoidal
category.
Joyal and Street [2] established a coherence theorem for planar monoidal
categories, which captures the correspondence between the formal language
and the graphical language we described. The formal language of categories
uses object variables and morphism variables, and object constants (such
as I ) and morphism constants (such as idA ), and operation symbols (such
as ◦ and ⊗). These are used to form terms and equations (formulas). The
coherence theorem of Joyal and Street states that an equation in the language
of monoidal categories follows from the axioms of monoidal categories if and
only if it holds in the graphical language, up to planar equivalence. Roughly
speaking, here, a diagram D1 is planar equivalent to a diagram D2 if it is
possible to transform D1 to D2 by continuously moving the boxes and wires
of D1 (without crossing or cutting). Other coherence theorem for special
 QUANTUM MECHANICS & CATEGORY THEORY 19

categories are of the similar nature. The part of a coherence theorem that states
that an equation following from the axioms holds in the graphical language
is called a soundness theorem, and its converse is called a completness theorem.
Soundness is guaranteed by assuring that the axioms hold in the graphical
language.
A braided monoidal category is a monoidal category with a family of
isomorphisms for every pair of objects A, B,
A,B : A ⊗ B → B ⊗ A.
−1
Hence A,B exists, where
−1
A,B : B ⊗ A → A ⊗ B.
Two hexagon axioms are satisfied for every triple of objects A, B, C :
(idB ⊗ A,C ) ◦ αB,A,C ◦ ( A,B ⊗ idC ) = αB,C,A ◦ A,B⊗C ◦ αA,B,C
and
−1 −1 −1
(idB ⊗ C,A ) ◦ αB,A,C ◦ ( B,A ⊗ idC ) = αB,C,A ◦ B⊗C,A ◦ αA,B,C .
The first of these axioms is captured in the diagram below.
αB,A,C
(B ⊗ A) ⊗ C > B ⊗ (A ⊗ C )
∧
A,B ⊗idC idB ⊗ A,C
∨
(A ⊗ B) ⊗ C B ⊗ (C ⊗ A)
∧
αA,B,C αB,C,A
∨
A ⊗ (B ⊗ C ) > (B ⊗ C ) ⊗ A
A,B⊗C

It follows that
−1
A,B ◦ A,B = idA⊗B .
Graphical language is extended to picture braiding A,B and is represented by
an under- (over-) crossing.

A symmetric monoidal category is a braided monoidal category where the

−1
braiding A,B is the inverse B,A . It is called symmetry and is graphically
represented by a crossing.
For monoidal categories C and D, a functor Φ : C → D is called a monoidal
functor if there are also morphisms φA,B : Φ(A) ⊗ Φ(B) → Φ(A ⊗ B) and
20 JENNIFER CHUBB AND VALENTINA HARIZANOV

φ : ID → Φ(IC ), which preserve the tensor structure as follows. For every

triple of objects A, B, C of C,
Φ(αA,B,C ) ◦ φA⊗B,C ◦ (φA,B ⊗ idΦ(C ) ) = φA,B⊗C ◦ (idΦ(A) ⊗ φB,C ) ◦
αΦ(A),Φ(B),Φ(C ) ,

Φ(A) = Φ( A) ◦ φA,I ◦ (idΦ(A) ⊗ φ),

Φ(A) = Φ(A ) ◦ φI,A ◦ (φ ⊗ idΦ(A) ).
For example, the last equation has the diagram:
Φ(A)
I ⊗ Φ(A) > Φ(A)
∧
φ⊗idΦ(A) Φ(A )
∨
φI,A
Φ(I ) ⊗ Φ(A) > Φ(I ⊗ A)

If the maps φA,B and φ are also invertible (isomorphisms), the functor is called
a strong monoidal functor; if they are the identity maps, the functor is called a
strict monoidal functor.
Given two monoidal categories, C and D, and two strong monoidal functors
from C to D, Φ with φ Φ and Ψ with φ Ψ, a natural transformation N : Φ → Ψ
with morphisms A : Φ(A) → Ψ(A) is a monoidal natural transformation if for
every pair of objects A, B of C, we have
A⊗B ◦ φA,B
Φ Ψ
= φA,B ◦( A ⊗ B ).

For braided monoidal categories C and D, a monoidal functor Φ : C → D is

called a braided monoidal functor if it is compatible with braiding as follows.
For every pair of objects A, B of C,
Φ( A,B ) ◦ φA,B = φB,A ◦ Φ(A),Φ(B) .

An example of a symmetric monoidal category is the category of sets with

functions as morphisms, with Cartesian product, and symmetry given by
A,B (x, y) = (y, x). Another example of a symmetric monoidal category is the
category of vector spaces with linear maps as morphisms, with tensor product,
and symmetry given by A,B (x ⊗ y) = y ⊗ x.
A monoidal category C is called right autonomous if every object A of C
has a right dual, denoted by A∗ , and there are two morphisms, the unit
∗ ∗
A : I → A ⊗ A and the counit A : A ⊗ A → I , which satisfy the following
adjunction triangle equalities:
idA = (A ⊗ idA ) ◦ (idA ⊗ A ),
(idA∗ ⊗ A ) ◦ ( A ⊗ idA∗ ) = idA∗ .

A∗ , A , A and the first triangle equality are graphically represented as follows:

 QUANTUM MECHANICS & CATEGORY THEORY 21

idA ⊗
> A ⊗ A∗ ⊗ A
A
A
εA ⊗idA
> ∨
idA
A

A left autonomous monoidal category is defined dually and a left dual of A is

denoted by ∗ A. A monoidal category is autonomous if it is both right and left
autonomous. In a braided right autonomous category, a right dual of A is also
a left dual of A, so the category is autonomous. A compact closed category
is a right autonomous symmetric monoidal category. A category of sets with
binary relations as morphisms and direct product as tensor product and where
A∗ = A is a compact closed category. The category of finite dimensional vector
spaces (or finite dimensional Hilbert spaces) with tensor product and with A∗
being the dual space of A is a compact closed category. On the other hand,
if we allow infinite dimensional vector spaces, the categories of vector spaces
and of Hilbert spaces are not autonomous.

§7. Dagger categories. A dagger category is a category C equipped with a

contravariant functor † : C → C, which is identity on the objects and involutive
on the morphisms. More specifically, to each morphism f : A → B a
morphism f † : B → A is assigned such that
†
(f † )† = f ∧ idA = idA ,
and for every morphism g : B → C ,
(g ◦ f)† = f † ◦ g † .
Morphism f † is called the adjoint of f. The adjoint is diagrammatically
represented by reversing the location but not the direction of the wires and by
marking the upper right corner (in contrast to the upper left corner) in the
box. In general, the adjoint of a diagram is its mirror image.
The category of sets with binary relations as morphisms is a dagger category
with relational inverse R† as adjoint of R. The category of Hilbert spaces with
bounded linear maps is a dagger category with the usual adjoints. A morphism
f is called Hermitian if it is self-adjoint: f † = f. A morphism f is called
unitary if it is an isomorphism and f −1 = f † . A dagger functor Φ between two
dagger categories C and D is a functor that satisfies the following additional
22 JENNIFER CHUBB AND VALENTINA HARIZANOV

equality for every morphism f in C:

Φ(f † ) = (Φ(f))† .
A dagger monoidal category C is a category that is both monoidal and dagger
and the two structures are compatible in the sense that the morphisms from
the monoidal structure, αA,B,C , A , A , are unitary and the following equality
is satisfied for every pair of morphisms f, g:
(f ⊗ g)† = f † ⊗ g † .
A dagger symmetric monoidal category is a dagger braided monoidal category
such that its symmetry (braiding) is unitary. A dagger compact closed category
C, also simply called dagger compact category, is a dagger symmetric monoidal
category that is also compact closed, together with a relation to connect the
dagger structure to the compact structure. Specifically, the dagger is used to
connect the unit to the counit so that for all objects A in C, we have:
A = A⊗A∗ ◦ A† .
Dagger compact categories are of great importance for foundations of
quantum information and computing. Selinger [4] proved a completeness
and hence coherence result for dagger compact closed categories. That is, he
established that an equation follows from the axioms of dagger compact closed
categories if and only if it holds in finite dimensional Hilbert spaces. Thus,
this coherence theorem allows us to use the diagrammatic calculus of dagger
compact categories to precisely express and verify some fundamental quantum
information notions and protocols.

REFERENCES

[1] S. Abramsky and B. Coecke, A categorical semantics of quantum protocols, Proceedings of

the 19th Annual IEEE Symposium on Logic in Computer Science, IEEE, 2004, pp. 415– 425.
[2] A. Joyal and R. Street, The geometry of tensor calculus I, Advances in Mathematics, vol. 88
(1991), pp. 55–112.
[3] P. Selinger, A survey of graphical languages for monoidal categories, New Structures for
Physics (B. Coecke, editor), Lecture Notes in Physics, vol. 813, Springer, 2011, pp. 289–355.
[4] , Finite dimensional Hilbert spaces are complete for dagger compact closed categories,
Logical Methods in Computer Science, vol. 8 (2012), pp. 1–12.

DEPARTMENT OF MATHEMATICS
UNIVERSITY OF SAN FRANCISCO
SAN FRANCISCO, CA 94117
E-mail: [email protected]

DEPARTMENT OF MATHEMATICS
GEORGE WASHINGTON UNIVERSITY
WASHINGTON, D.C. 20052
E-mail: [email protected]
COULD LOGIC BE EMPIRICAL? THE PUTNAM-KRIPKE DEBATE

ALLEN STAIRS

Abstract. Not long after Hilary Putnam published “Is Logic Empirical,” Saul Kripke presented
a critique of Putnam’s argument in a lecture at the University of Pittsburgh. Kripke criticized both
the substance of Putnam’s version of quantum logic and the idea that one could “adopt” a logic
for empirical reasons. This paper reviews the debate between Putnam and Kripke. It suggests
the possibility of a “middle way” between Putnam and Kripke: a way in which logic could be
broadly a priori but in which empirical considerations could still bear on our views about the
logical structure of the world. In particular, considerations drawn from quantum mechanics might
provide an example.

Some years ago, Hilary Putnam published a paper called “Is Logic Empiri-
cal?” [7] in which he argued that quantum mechanics provides an empirical case
for revising our views about logic. (The paper was republished in his collected
works as “The Logic of Quantum Mechanics”. Page references will be to the
reprinted version.) In 1974, Saul Kripke presented a talk at the University
of Pittsburgh called “The Question of Logic,” offering a detailed rebuttal of
Putnam’s case. As of this writing, almost 40 years later, Kripke’s paper still
hasn’t appeared in print and apart from my 1978 dissertation and a paper I
published 28 years later [9], very little has been written on the disagreement
between Putnam and Kripke. This is unfortunate; the issues are well worth
investigating. In my 2006 paper [9], I adopted the device of writing about Paul
Kriske and Prof. Tupman out of deference to the fact that there is no published
version of Kripke’s talk. Here I’ll simply write directly about Putnam and
Kripke. If I get Kripke wrong, I hope he’ll let us know.
As for the plan of the paper, we begin by reviewing Putnam’s arguments;
after that we move to Kripke’s rebuttal. This will lead to a larger discussion of
what logic and the empirical might have to do with one another.

§1. Putnam on quantum logic. We think of logical truths as a special case

of necessary truths, but Putnam reminds us that we now reject certain claims
about geometry that once seemed necessary. We would once have said that
if two lines are straight and a constant distance apart over some portion
of their span, they can’t converge elsewhere. For anyone not familiar with
Logic and Algebraic Structures in Quantum Computing
Edited by J. Chubb, A. Eskandarian and V. Harizanov
Lecture Notes in Logic, 45
c 2016, Association for Symbolic Logic 23
24 ALLEN STAIRS

non-Euclidean geometry, Putnam claims that this seems as intuitively clear

as saying that there are no married bachelors, or that nothing can be scarlet
all over and bright green all over at the same time. In the case of the lines,
however, we’ve come to believe not just that the claim might be false but that
in some instances it is false.
We might say that what Putnam describes applies to geodesics, but “geodesic”
doesn’t mean “straight line.” However, Putnam insists that this won’t do. On
our intuitive conception, shortest paths are straightest and conversely. The
notion of a geodesic preserves this, and lines that depart from geodesics will
not seem straighter. One way to put it: if we say that geodesics behaving as
Putnam describes aren’t straight lines, we’ll have to say that there can be points
with no straight line between them. Putnam thinks we miss the significance of
relativity if we represent its geometrical claims as mere changes of meaning.
He writes:
The important point is that [‘straight line’] does not ‘change meaning’
in the trivial way one might at first suspect. Once one appreciates
that something that was formerly literally unimaginable has indeed
happened, then one also appreciates that the usual ‘linguistic’ moves
only help to distort the nature of the discovery and not to clarify it.
(p. 177)
Putnam argues that we’ve made a similar discovery about logic itself. We pair
statements about quantum quantities with subspaces of Hilbert space and we
can extend this map from simple statements to compound ones by associating
“or” with subspace span (p ∨ q), “and” with subspace intersection (p ∧ q), and
“not” with orthocomplement (p ⊥ ). If we take the mapping seriously, however,
we have a conflict with classical logic. Suppose the quantity A has two possible
values a1 and a2 , associated with rays α1 and α2 . Suppose, likewise, that B
has two values b1 and b2 associated with rays 1 and 2 . Now consider the
expressions
(A = a1 or A = a2 ), (B = b1 or B = b2 )
and associate them with the subspaces
(α1 ∨ α2 ), (1 ∨ 2 ).
Quantum mechanics, read as Putnam reads it, gives us cases where both
disjunctions are true. That means the conjunction
(A = a1 or A = a2 ) and (B = b1 or B = b2 )
is also true. However each of the following pick out the null subspace of
Hilbert space
(α1 ∧ 1 ), (α1 ∧ 2 ), (α2 ∧ 1 ), (α2 ∧ 2 )
and so the corresponding conjunctions are false. Hence
(A = a1 ∧B = b1 )∨(A = a1 ∧B = b2 )∨(A = a2 ∧B = b1 )∨(A = a2 ∧B = b2 )
 COULD LOGIC BE EMPIRICAL? THE PUTNAM-KRIPKE DEBATE 25

is false but this discrepancy between the distributed and undistributed formulas
is impossible classically. Putnam writes:
Conclusion: the mapping is nonsense—or, we must change our logic.
(p. 179)
On the other hand, if we do “adopt the heroic course of changing our logic”
there’s a straightforward way to proceed:
. . . just read the logic off from the Hilbert space H (S). (p. 179)
The advantage, says Putnam, is that
all so-called anomalies in quantum mechanics come down to the
non-standardness of the logic. (p. 179)
and the anomalies go away if we change our logic.
Putnam offers several illustrations. “Complementarity,” understood as
the failure of quantum mechanics to specify joint values for noncommuting
quantities comes down to logical incompatibility in quantum logic; the com-
plementary quantities don’t share eigenspaces. He also argues that quantum
logic accounts for the two-slit experiment. To derive the incorrect classical
probabilities, we have to distribute a proposition R about where the photon
hits the screen over a disjunction of propositions A1 and A2 about which
hole the photon passes through. If we treat ‘R ∧ (A1 ∨ A2 )’ as equivalent to
‘(R ∧ A1 ) ∨ (R ∧ A2 )’, then we end up with the wrong probabilities.
Putnam also claims that if we analyze barrier penetration quantum logically,
we avoid explaining the effect by appeal to a supposedly mysterious ‘disturbance
by the measurement.’ (p. 182) In fact, the account he gives (on p. 183) can’t be
right for any finite population of atoms (exercise for the reader; look especially
at statement (8) and Putnam’s comment on it) but let that pass. In classical
physics, the state provides a complete description relative to the terms of
the theory, of the system. In quantum theory, there are “states” or “state
descriptions,” but Putnam writes that
A system has no complete description in quantum mechanics; such
a thing is a logical impossibility (p. 185)
Quantum states are “logically strongest consistent statements” but they aren’t
states “in the sense of statements which imply every true proposition about
S” (p. 185) This might suggest that quantum states are like statistical states in
classical mechanics, and that their failure to provide a complete list of all the
truths is a reflection of our epistemic situation. However, this isn’t Putnam’s
view. Rather, he tells us that a quantum system has, e.g., a position by virtue of
the truth of a disjunction of position statements and it also has a momentum by
virtue of the truth of disjunction of momentum statements1 . Here is Putnam
articulating what we will call the value-definiteness thesis:
1 Putnam knows that strictly, there are no position and momentum eigenstates; the oversimplifi-

cation is merely for illustration.

26 ALLEN STAIRS

1. For any such question as ‘what is the value of M (S) now?’ where
M is a physical magnitude, there exists a statement Ui which
was true of S at t0 such that had I known Ui was true at t0 , I
could have predicted the value of M (S) now, but
2. It is logically impossible to possess a statement Ui which was
true of S at t0 from which one could have predicted the value
of every magnitude M now.
We can predict any one magnitude, if we make an appropriate
measurement, but we can’t predict them all.
The advantage of giving up classical logic, according to Putnam, is this:
These examples makes the principle clear. The only laws of classical
logic that are given up in quantum logic are distributive laws . . . and
every single anomaly vanishes once we give these up. (p. 184)
Putnam’s argument for adopting quantum logic is that if we do, the interpre-
tive puzzles of the theory dissolve. If we insist on classical logic, we have to say
such supposedly objectionable things as that measurements create the values
of the quantities measured or that there is a “cut between the observer and the
observed” or that there are undetectable hidden variables. But Putnam says
. . . I think it is more likely that classical logic is wrong than that there
are either hidden variables or “cuts between the observer and the
system”, etc.
This completes the analogy with geometry. We could preserve Euclidean
geometry, but only by paying the high intellectual price of admitting gratuitous
universal forces. Likewise for classical logic: we can preserve it only by paying
an unacceptable price in the coin of untoward claims about quantum systems.

§2. Kripke on Putnam. Kripke’s critique of Putnam has two parts. One
deals with the particulars of Putnam’s argument. There Kripke’s case is strong.
However, granting that Kripke is right about Putnam’s particular quantum
logical proposal wouldn’t show that logic isn’t empirical, nor would it show
that quantum mechanics doesn’t give us a reason to change our views about
logic. In the second part of his critique, Kripke argues that the very idea of
changing logic for empirical reasons is confused.
In what follows, I quote at length from my partial transcript of Kripke’s talk.
The indirect debate between Kripke and Putnam was an important episode,
and the reader will get a better sense of it if s/he reads Kripke’s own words. To
be sure, there is a matter of propriety here, but Kripke’s own words do a better
job than my paraphrase would of spelling out his view and therefore, it seems
fairest to him to use those words.
2.1. Quantum logic and simple arithmetic. The first part of Kripke’s argu-
ment is intended to show that if we follow Putnam, we have to agree to the
 COULD LOGIC BE EMPIRICAL? THE PUTNAM-KRIPKE DEBATE 27

untoward conclusion that 2 × 2 ≥ 5. According to Putnam, if M has possible

values m1 , m2 , . . . , mn then there is a true statement ascribing one of these
values to M . The statement

M = m1 ∨ M = m2 ∨ · · · ∨ M = mn

is true, Putnam would say, and the summary of the value-definiteness thesis
above makes clear what he means: one of the disjuncts really is true, and if we
knew which, we could predict the outcome of an M -measurement. However,
the logically strongest statement about the system may not tell us which disjunct
is true.
Is it really clear that Putnam meant this? Here’s a passage that would be
hard to make sense of otherwise. Sz is a position state and T1 , T2 , etc. are
momentum states. (Substitute eigenstates of different spin components if you
prefer.) We suppose Sz to be known.
The idea that momentum measurement ‘brings into being’ the value
found arises very naturally if one does not appreciate the logic being
employed in quantum mechanics. If I know that Sz is true, then I
know that for each Tj the conjunction Sz · Tj is false. It is natural to
conclude (smuggling in classical logic) that Sz · (T1 ∨ T2 ∨ · · · ∨ TR )
is false, and hence that we must reject (T1 ∨ T2 ∨ · · · ∨ TR )—i.e.,
we must say ‘the particle has no momentum’. Then one measures
momentum, and one gets a momentum—so the measurement must
have ‘brought it into being’. However, the error was in passing
from the falsity of Sz · T1 ∨ Sz · T2 ∨ · · · ∨ Sz · TR to the falsity of
Sz · (T1 ∨ T2 ∨ · · · ∨ TR ). This latter statement is true (assuming
Sz ) and so it is true that ‘the particle has a momentum’ . . . and
the momentum measurement merely finds this momentum (while
disturbing the position); it does not create it, or disturb it in any
way. It is as simple as that. (p. 186)
“Simple” or not, Kripke draws out an untoward consequence. Suppose we’re
given two quantities, A and B, each with two possible values 1 and 2. Thus the
set {1, 2} is the set of possible values of A and also of B. Putnam will say that
1. A = 1 ∨ A = 2
2. B = 1 ∨ B = 2
are both true. However, he will also say that each of the following are false:
1. A=1∧B =1
2. A=1∧B =2
3. A=2∧B =1
4. A=2∧B =2
But Kripke argues:
28 ALLEN STAIRS

The usual mathematical definition of multiplication is this: suppose

we have two sets with two elements. Then the cardinality of their
product is the cardinality of the Cartesian product of the two sets
. . . where x comes from the first set and y comes from the second
set, so where x comes from {a, b} and y comes from {c, d } where
{a, b} and {c, d } are our two two-element sets. We want to consider
how many ordered pairs there are. So the classical arithmetician
says “There are four, namely a, c, a, d , b, c, b, d  . . . ” But we
can all see the fallacy in any conclusion that these are the only pairs.
The “fallacy” is that if x comes from {a, b}, then we have the disjunction
x =a∨x =b
and similarly we have
y = c ∨ y = d.
Now suppose that the set {a, b} is the set of possible values of the quantity A
above (i.e., {a, b} = {1, 2}) and {c, d } is the set of values of the quantity B
(i.e., {c, d } = {1, 2} as well.) We’ll let Kripke pick up the story:
Now I claim that there is a fifth pair A, B where these are the
two quantities mentioned by Putnam. Remember that Putnam
does not think these are funny pseudo-numbers. The idea is that A
was already one of the two numbers 1 and 2 [and] B was already
one of the numbers 1 and 2. So A is certainly in the first set [i.e.,
{a, b} = {1, 2}—AS] because A is equal either to 1 or to 2. B is
certainly in the second set [i.e., {c, d } = {1, 2}— AS] because B is
either equal to 1 or to 2, though we may not have measured which.
So the pair A, B is in our Cartesian product. But certainly we
cannot say that A, B equals 1, 1 if we adopt the usual criterion
of identity of ordered pairs because that would mean that A = 1 and
B equals 1, and that contradicts [the falsity of (3)]. Also, A, B
does not equal 1, 2 because it is false that A equals 1 and B equals
2. And A, B does not equal 2, 1 [and] A, B does not equal
2, 2 . . . So there is a fifth and hitherto overlooked, I might say,
ordered pair in the Cartesian product of these two finite sets.
Kripke’s point, of course, is that this is absurd, but that it’s where we end up if
we follow Putnam.
There may be various ways Putnam could respond, but Kripke insists that
one obvious rejoinder won’t do: it won’t do to accuse Kripke of begging the
question. Kripke insists: he’d only be begging the question if he had assumed
a premise that Putnam rejects. However, the distributive law isn’t a premise in
his argument. Kripke simply reasons from premises that Putnam accepts to
the conclusion that if none of the pairs 1, 1, 1, 2, 2, 1, 2, 2 gives the joint
values for A and B, then joint values require that there be another pair in the
 COULD LOGIC BE EMPIRICAL? THE PUTNAM-KRIPKE DEBATE 29

Cartesian product. Since Putnam would claim that none of the four ordered
pairs gives the joint values, and would also claim that both quantities really do
have values, the untoward (and absurd) conclusion follows. As Kripke puts it
in connection with a closely-related example
. . . if you say that I am begging the question then you yourself, I
think, are begging the question, because only if my reasoning was
invalid did I need any extra premise which I have begged against
Putnam.
2.2. The impossibility of “adopting” a logic. Kripke is right, I believe: there’s
no convincing quantum-logical defense of the value-definiteness thesis. (See [8]
for more discussion) and in what follows, we will assume that value-definiteness
doesn’t hold. Kripke’s larger point is that there is a problem at the core of
Putnam’s view. Putnam, he thinks, believes that we could somehow decide
to “adopt” a logic; Kripke insists that this is incoherent. We misunderstand
logic if we think there are “logics” among which we could somehow choose.
There is reasoning. Specific formal systems may or may not adequately capture
aspects of correct reasoning. But there is no neutral place outside logic from
which to decide what “logic” to adopt.
Whether Putnam really holds the view of logic that Kripke attributes to him
isn’t clear. That said, it’s a useful foil for making Kripke’s own view of logic
clearer. Therefore, while we won’t ignore the question of how well Kripke’s
criticisms fit Putnam, the exegetical question won’t be our main concern.
Putnam remarked on our intuitive sense of contradiction when faced with
his geometrical example. Kripke reads Putnam this way:
Just as in the case of non-Euclidean geometry we throw intuition to
the wind and adopt an axiomatic system as supposedly describing
the real physical world . . . so on every other domain we cannot
rely on intuition. Once one has a rival system of axioms, the mere
fact that an old system struck us as the only intuitively acceptable
one should be given little weight. Once alternative geometries are
under consideration, we abandon any mere intuitive preference
for Euclidean geometry, and once alternative logics are under
consideration, we abandon any mere intuitive preference for a
particular system of logic.
Kripke thinks there is a deep confusion here. Formal systems are not logic.
Formal systems may or may not faithfully reflect correct principles of reasoning,
but we have no alternative to using “intuition,” by which Kripke means
reasoning, to assess the formal systems. If changing our formal system is
supposed to entail changing the way we reason, then we have no place to stand
outside of reasoning from which to do this. “Logics” qua formal systems aren’t
logic. As Kripke sees it, Putnam’s fundamental error lies in missing this point.
30 ALLEN STAIRS

Once we grasp this, the idea that we could change our logic in response to
empirical considerations makes no sense.
Even if we grant that there’s no place to stand outside reasoning, there’s a
more general phenomenon here. What William Alston called doxastic practices
(see his [1, Ch. 4], for instance) typically have the sort of self-supposing quality
that Kripke’s point relies on. We can reconsider how to evaluate beliefs based
on sensory input; when we do, we’ll need to rely on at least some such beliefs
and hence on the practice of forming beliefs based on sense evidence. We can
consider what memory can and can’t teach us; we can’t avoid relying on at least
some memories when we do. Equally important, these practices aren’t insulated
from one another. In considering what weight to give memory, for example,
we’ll make use of claims that we’ve accepted on the basis of the implicit and
explicit rules/practices we use for assessing other kinds of empirical claims.
We can also reason about how to reason, as Kripke would be the first to insist.
Putnam may seem to be saying that we can evaluate logic without relying on
logic broadly conceived (i.e., on logic qua reasoning) but it’s not clear that he
means this or needs to say it. In order to rebut a measured version of Putnam’s
view, Kripke would have to show that reasoning is the one doxastic practice
to which the deliverances of other doxastic practices are irrelevant. Putnam’s
larger point would be made if sometimes what we discover empirically can
properly enter into our deliberations about how to reason.
Be that as it may, Putnam’s main argument seems to be that if we give up the
distributive law, we’ll be blocked from drawing untoward conclusions. Thus,
we won’t be able to argue that the probabilities in the two-slit experiment must
fit a crude application of the law of total probability, and we won’t need to
say that measurement creates the values that it records. However, this is too
quick. We might be able to avoid any number of unwelcome conclusions if we
simply refused to reason in certain ways; that hardly makes a case for merely
opportunistic “revisions” of logic. And while Putnam might judge that a failure
of the distributive law is “more likely” than hidden variables or “cuts between
observer and observed,” Kripke can reply that without something more than
a mere and tendentious cost-benefit analysis, we haven’t been presented with
an intelligible alternative. The distributive law seems to be a correct way to
reason. Putnam hasn’t shown us any deep problems with the idea that there are
Bohmian-style hidden variables; he merely tells us that he finds them unlikely.
He objects to the idea that measurement might bring the values it yields into
being. However, his main objection seems to be that this is a strange notion of
“measurement.” This threatens to turn the argument into a mere quibble. The
idea that the interactions we call measurements bring new states of affairs into
being might be a reason to pick or invent a different word, but it doesn’t count
against the possibility that things really work this way.
We leave the vexed issue of measurement (or “measurement”) aside and turn
to a different part of Kripke’s reply: his case that the very idea of “adopting a
 COULD LOGIC BE EMPIRICAL? THE PUTNAM-KRIPKE DEBATE 31

logic” makes no sense. Kripke takes his cue from Lewis Carroll’s “What the
Tortoise Said to Achilles” and from Quine’s “Truth by Convention”. He says:
The basic problem is this: if logical truths are mere hypotheses . . .
and one can adopt them as one will, how, unless one has a logic in
advance, can one possibly deduce anything from them?
Kripke develops the example of universal instantiation at greatest length.
Imagine someone who doesn’t see that from a universal claim, each instance
follows. Imagine further that our poor reasoner is willing to accept Kripke’s
authority that all ravens are black and is also willing to accept Kripke’s authority
in more general logical matters. There’s a raven, J , out of our subject’s sight,
but he doesn’t see that believing this and accepting that all ravens are black
commits him to accepting that J is black. Kripke tells the tale charmingly:
So I say to him “Oh. You don’t see that. Well let me tell you: from
every universal statement, each instance follows.” He will say “Oh.
Yes. I believe you.” So now I say to him, “Ah. So ‘All ravens
are black’ is a universal statement and ‘This raven is black’ is an
instance.” “Yes. Yes” He agrees. So I say to him “All universal
statements imply their instances. This particular statement that all
ravens are black implies this particular instance.” “Well, hmm, I’m
not entirely sure,” he will say. “I don’t really see that I’ve got to
accept that!”
The problem is clear. As Kripke puts it
If he was not able to make the simple inference “All ravens are black,
therefore J is black” where J is a particular raven, then giving him
some super-premise like “Every universal statement implies each
instance” won’t help him either.
It won’t help because he would already have to be in command of the principle
to apply it; the idea that he could adopt it is incoherent. Kripke makes similar
points about non-contradiction, adjunction and Lewis Carroll’s modus ponens
example. We can embody these principles in formal systems, but there’s no
sense to the idea that someone, so to speak, standing outside these principles
could adopt them.
These are all cases where we couldn’t adopt a particular principle unless
we already grasped it intuitively. Perhaps that doesn’t apply to all logical
principles, and in any case Putnam’s example had to do with giving up rather
than adopting a principle. However, Kripke thinks this would miss the point.
Here’s what he says:
. . . I don’t really mean that we adopt as basic just those things to
which we can figure out that this argument applies, What I mean is
this: you can’t undermine intuitive reasoning in the case of logic and
try to get everything on a much more rigorous basis. One has just
32 ALLEN STAIRS

to think not in terms of some formal set of postulates but intuitively.

That is, one has to reason. One can’t just adopt a formal system
independently of any reasoning about it because if one tried to do
so one wouldn’t understand the directions for setting up the system
itself. And so any comparison of logic to geometry which says that
in the case of logic as in the supposed case of geometry, intuition can
be thrown to the dogs—that is, any reasoning outside the system of
postulates can be thrown to the dogs—must be wrong. One can only
reason as we always did, independently of any special set of rules
called “logic,” in setting up a formal system or in doing anything
else. And if proof by cases was part of our intuitive apparatus then
there is no analogy to geometry which says that this should not be
respected.
Kripke is surely right: logic isn’t just a matter of formal systems. We can
also agree that questions about how to reason have a special status among the
various kinds of questions we can ask. We can agree further that for at least
some logical principles there’s no sense to be made of the idea that we might
“adopt” them, and we can even concede that nothing could count as adopting
a logic wholesale. Whether this scuttles the idea that empirical considerations
could bear on logic is less clear, however.
2.3. Rejecting subalternation as a case of change in logic. To make progress,
we need to look at what Kripke concedes about changes in logic and how he
accounts for them. The most useful place to begin is with what he says about
the principle of subalternation for universal categoricals—in particular, that
“All P are Q” implies “Some P are Q.”2 Logicians once accepted this principle,
and yet we no longer do. What Kripke says is surely right: if we accept
subalternation, we overlook the case where ‘P’ is empty. ‘All deserters will be
shot on sight’ may be true, and that may be exactly the reason why there are no
deserters. But if there are no deserters, it would be very odd to say that some
deserters are shot. Intuitive reflection makes clear that something has to give.
If we overlook empty terms, we’ll be tempted to think subalternation is valid.
We correct ourselves by mere reflection—by ordinary reasoning. However, we
can ask if this is always so. When we discover that we’ve overlooked a case and
accepted an incorrect logical principle as a result, is this always a matter of
ordinary reasoning, or do empirical considerations sometimes come into play?

2 Kripke talks briefly about cases where we see that an argument we once accepted is invalid.

Here we change our beliefs about logic, but we do so simply and straightforwardly by reasoning.
He also offers a cursory discussion of intuitionism. Here he claims that the intuitionists introduced
new connectives, defined in terms of provability, and so the intuitionist’s apparent rejection, e.g.,
of excluded middle isn’t really in competition with the classical principle. Whether that’s the best
reading of intuitionists such as Brouwer I will leave to others to decide.
 COULD LOGIC BE EMPIRICAL? THE PUTNAM-KRIPKE DEBATE 33

2.4. Future contingents, bivalence and the empirical. Consider a debate that
Kripke doesn’t mention but that has a long history: whether propositions
about the future provide reasons to give up bivalence. Two sorts of views
suggest that the answer might be yes. One is that some propositions about the
future (e.g. “There will be a sea battle tomorrow” or “This atom will decay an
hour from now”) are contingent in a more-than-merely-logical sense. Another
is the view that only the present exists, usually called “presentism” and the
“growing block” view, according to which the present and the past but not the
future are real. The difference between presentism proper and the growing
block won’t matter for our purposes, we’ll use “presentism” for both.
Neither future contingent propositions nor presentism alone make the case
against bivalence. Suppose some propositions about the future are contingent
in the sense of not being determined by facts about the past. Suppose, that
is, that determinism is false. If the so-called “block universe view” is correct,
all events, past present and future, are ontologically on a par. In that case,
the facts about the events making up the block entire settle the truth or falsity
of future contingent propositions even if determinism is false. An event’s
being undetermined is a matter of its relationship to other events and to the
laws of nature; whether we live in a block universe and whether the laws are
deterministic are independent questions. On the other hand, suppose that
presentism is correct. Then even though future states of affairs don’t exist,
deterministic laws plus the facts about the present could suffice to settle the
truth or falsity of propositions about the future.
What, then, if presentism is true and determinism false? Perhaps bivalence
about future contingent propositions can still be defended, though it’s not clear
how or why. What if it can’t? One response is to abandon excluded middle—to
claim that when ‘P’ is indeterminate, ‘P ∨ not-P’ is likewise indeterminate
(a view usually associated with Łukasiewicz.) However, there’s a plausible
objection: if ‘P’ is indeterminate, then it’s not true that P, hence ‘not-P’ is true.
If so, then even if ‘P’ is indeterminate, ‘P ∨ not-P’ will be true by virtue of its
second disjunct.3
Another familiar account of future contingents appeals to branching time
and supervaluation (see [12]). On this approach, a statement about the future
is true at the present moment just in case it holds on each branch or history
passing through this moment, and false if it is false on each such branch.
Contingent statements about the future will therefore be neither true nor
false. However, this permits true disjunctions with no true disjuncts. Suppose
{P1 , P2 , . . . , Pn } is a set of future contingent propositions that are mutually
exclusive, not logically exhaustive, but such that on each branch passing through

3 Scope matters here; using ‘F ’ as a future-tense operator, the claim is that when ‘F (X )’ is

indeterminate, ‘not-F (X )’ is true, even though ‘F (not-X )’ is indeterminate. See Bourne [3, pp. 82
ff.] for useful discussion.
34 ALLEN STAIRS

the present, one of them is true. An artificial example: suppose a coin will
be tossed, that the outcome isn’t determined, but that on each branch the
outcome is either Heads or Tails. Then
The coin will come up heads or the coin will come up tails
is true at the present moment even though neither disjunct is.
Supervaluation preserves excluded middle and non-contradiction. Whether
it preserves all classical logical truths might be more of an accounting issue
than a substantive one. Even with excluded middle intact, the possibility of a
true disjunction with no true disjuncts isn’t part of logical business as usual.
The novelty seems at first to sit comfortably with Kripke’s view. Our belief that
true disjunctions require true disjuncts came from overlooking a (complex)
possibility: the combination of presentism and future contingents. However,
further thought may seem to favor Putnam. This particular case for true
disjunctions without true disjuncts depends on assumptions about the world:
that the block universe view and determinism are both false. The overlooked
possibility is a substantive one, and reasoning alone won’t tell us if it holds.
This suggests that matters of logic depend on the way things are, as Putnam’s
view would maintain.
The status of determinism is a contingent, empirical matter. However, as we
noted above, even if determinism is false, this wouldn’t be enough to undermine
bivalence. The crucial additional assumption is presentism, and it might be
argued that this is not an empirical matter; certainly the debate has often
proceeded as though it’s not. However, there are able defenders of the coherence
of presentism and of the block universe. If both views are indeed coherent,
empirical considerations plausibly bear on which is correct. Indeed, Putnam
himself famously invoked special relativity to argue against presentism (albeit
not under that name) in “Time and Physical Geometry.” [6] His argument that
past, present and future are equally “real” don’t rest on general philosophical
considerations; it depends on the structure of Minkowski space-time. It may
be, then, that whether presentism is true depends on the facts about space-time.
If so, it suggests that assessing the need for the logical revisions at issue in
the debate over future contingents depends on contingent, empirical facts
about the world.4 The broad issue is whether claims about reality could have
consequences for logic. Future contingent propositions give rise to a dilemma:
if bivalence holds for such propositions, it’s because of something about the
world: the falsity of presentism or the truth of determinism. If bivalence fails,
it’s because presentism is true and determinism false. In either case, the claim is
empirical. The question of determinism is certainly empirical and the question
of presentism is at least arguably so. Thus, whether bivalence holds is an
empirical matter, and that, it seems, is enough to make Putnam’s larger point.
4 Of course, not everyone agrees that Putnam’s arguments are sound. See, for example, Stein [11]

and Bourne [3]. To repeat, the point here is not to take sides in this debate.
 COULD LOGIC BE EMPIRICAL? THE PUTNAM-KRIPKE DEBATE 35

The arguments above are skeletal and open to challenge, but suppose we
grant them. There’s a plausible Kripkean reply. Whether bivalence holds
might be an empirical matter, but if so the correct conclusion is that bivalence
is not a principle of logic. Furthermore, the conclusion that bivalence isn’t a
correct principle of logic is not an empirical one. We come to it by reflecting
on the possibilities, and we discover that there is a genuine possibility we had
overlooked: the possibility that there are no facts to ground the truth or falsity
of certain propositions about the future. That this is possible remains so even
if the possibility isn’t realized.
2.5. Détente? What’s just been said concedes something important to
Kripke, but suggests a possibility for détente. Logic writ large (let’s write
bold-face ‘logic’ for that) would remain a matter of reasoning, broadly under-
stood. The logic of the actual world could still be a contingent matter. The
analogy with geometry helps here. Suppose (unlikely, but science sometimes
takes strange turns) we became convinced that the world is Euclidean after
all. We would still know that the scenario Putnam describes is possible in a
broad sense. It would just be that it’s never actualized. The question of what
the detailed geometry of a world could be would remain, broadly speaking, a
priori; the question of what it is in fact would be empirical. That the world
could be pseudo-Riemannian is not empirical knowledge. That it is or isn’t
pseudo-Riemannian is an empirical claim. Likewise, that bivalence could fail is
arguably not empirical knowledge. That it does (or does not) fail in a particular
way is arguably empirical. And though we won’t try to give a general account
of what counts as a question of logic, questions about the status of bivalence
plausibly count.
This raises two questions. The first is whether there’s a case of this sort to be
made by appeal to quantum mechanics. We’ll take that up in the following
section. The second question will be raised but no more than raised: in light
of what quantum mechanics teaches us, is it quite so clear that logic really is
something we can know by a priori?

§3. Quantum logic reconsidered. Putnam’s quantum logical proposal offered

a formal structure and some interpretative principles and rationales. The
structure is the lattice of subspaces of Hilbert space, but the beginnings of the
disagreement with Kripke come from the interpretive overlay. Let ‘Sz = + 12 ’
say that the electron has spin +1/2 in direction z, and similarly for ‘Sx = + 12 ’
and ‘Sx = − 12 .’ Putnam, as we know, would say that when
 
1 1 1
Sz = + ∧ Sx =+ ∨ S x =−
2 2 2
is true, one of the disjuncts in parentheses really is true, but that it’s logically
impossible for us to know which. However, there’s another approach: treat
36 ALLEN STAIRS

‘(Sx = + 12 ∨ Sx = − 12 )’ as a disjunctive fact—as a case of a disjunction that’s

true in spite of not having a true disjunct. We’ve already seen reasons of one
sort for taking the idea of disjunctive facts seriously. Quantum mechanics gives
reasons of a different sort.
What follows is intended merely as a sketch, and if the reader finds it hand-
waving, that’s because the author is waving his hands. The goal isn’t to defend
a view in detail (indeed, I am by no means certain that the view is correct) but
simply to make its outlines clear enough to consider.
The paradigmatically curious quantum example is the case of two quantities—
call them P and Q—that share no eigenstates. This is the heart of what Bohr
called complementarity and it has two characteristic features. First, there’s no
arrangement that measures P and Q at the same time. Second, if we’re certain
what outcome a P-measurement would yield, we are not certain what outcome
a Q-measurement would yield; all values of Q have at least some positive
probability. The goal in this section is to see how we might move from here
to something more clearly relevant to logic, and to do it in a way that doesn’t
stray far from what a typical physicist would find plausible. Note that we aren’t
following Putnam’s approach. Putnam argued that if we adopt a strong set
of logical claims, we solve the interpretive problems of quantum mechanics.
There’s no such goal here. We’re trying to see what quantum mechanics might
teach us about logic if we start from things that many physicists already believe.
The first point is simple: quantum mechanical quantities can have values. A
system can have an energy or a spin in a particular direction. Few physicists
would disagree.5 The second point goes beyond ordinary common sense but not
beyond the common sense of most physicists. Stick with our complementary
quantities P and Q. When P has a value, Q does not. Thus: if there’s a true
statement
P = pi
then there is no true statement
Q = Qj .
No doubt most physicists believed this before no-hidden-variable proofs became
widely known, but those proofs provide another reason. If we accept a handful
of plausible constraints, then it’s impossible for all quantum quantities to have
values at the same time. Those constraints aren’t beyond challenge, but our
purpose isn’t to make an iron-clad case. It’s to make it plausible that quantum
mechanics has consequences for logic.
The third point starts with a piece of physics common sense and then moves
a bit beyond. It’s that there are purely disjunctive truths about quantum

5 Though few would disagree, this isn’t the same as saying none would. Quantum Bayesians

such as Carleton Caves, Christopher Fuchs and Rüdiger Schack are exceptions. See, for example,
their [4]. For some relevant discussion see Stairs [10]
 COULD LOGIC BE EMPIRICAL? THE PUTNAM-KRIPKE DEBATE 37

mechanical systems. To see why this is plausible, start with a special case of
our first point: degenerate quantities can have values. For example: energy is
often degenerate; the subspace that goes with
E=e
for some values e of the energy may not be one-dimensional. In spite of
this, there’s nothing strange about saying that the system really can have
energy e—that ‘E = e’ can be true. With that in mind, consider a simple but
instructive example: a spin-one system whose z-spin is 0. The state |z0  is a
superposition of |x+  and |x− . On any orthodox account, the statement
Sx = 0
is definitely false; |z0  and |x0  are orthogonal. It’s also of a piece with our
second point to say that the system doesn’t have a definite x-spin. Neither
‘Sx = +1’ nor ‘Sx = −1’ is true. But consider the degenerate quantity (Sx)2 —
the square of the spin in the x direction. Again, on any orthodox account, this
quantity has a value: +1. Few physicists would be shocked to be told that
‘(Sx)2 = +1’ is true when ‘Sz = 0’ is true. But if the square of the spin is +1,
it would be gratuitously peculiar to say that ‘Sx = +1’ and ‘Sx = −1’ are both
false. Instead, we can say that for (Sx)2 to take the value +1 and for
Sx = +1 ∨ Sx = −1
to be true are one and the same fact: ‘Sx = +1 ∨ Sx = −1’ is true even
though neither disjunct is. In short, bivalence fails, though for different
reasons than in the case of future contingents, and we have a true disjunction
with indeterminate disjuncts.6 ‘Sx = +1’ and ‘Sx = −1’ stand in a different
relationship to ‘Sz = 0’ than ‘Sx = 0’ does. ‘Sz = 0’ excludes ‘Sx = 0’ in an
old-fashioned classical way: the two are contraries. The relationship between
‘Sz = 0’ on the one hand and ‘Sx = +1’ and ‘Sx = −1,’ is not found in
classical physics. For the states that go with these statements, the term is
superposition, but there’s no standard word for the relationship between the
statements themselves. For present purposes, I propose l-complementarity.
In the language of Hilbert space, propositions are l-complementary when
their associated projectors don’t commute. But while that picks out the sorts
of cases we’re interested in, it doesn’t make a connection with logic. It’s also
too restrictive: in principle l-complementarity is more general than Hilbert
space non-commutativity. Kochen and Specker’s [5] partial Boolean algebra
approach is a better way to characterize l-complementarity formally. When
X and Y are l-complementary they do not belong to a common Boolean

6 Note that even if someone insisted that each disjunct is false, we’d still have a true disjunction

with no false disjunct. Why anyone would insist on any such thing, however, is unclear to say the
least.
38 ALLEN STAIRS

subalgebra of the partial Boolean algebra.7 However, this leaves the logical
point unclear. The proposal on offer is that l-complementarity goes with
a particular kind of failure of bivalence: if propositions X and Y are l-
complementary, then there are possible states of affairs in which X is true but
Y is neither true nor false.
With this in mind, consider distribution. In particular, consider
Sz = 0 ∧ (Sx = +1 ∨ Sx = −1).
The proposal is that both conjuncts are true, but neither disjunct of the
disjunction is true. That’s why we can’t distribute. The expression
(Sz = 0 ∧ Sx = +1) ∨ (Sz = 0 ∧ Sx = −1)
either fails to pick out an element of the algebra of propositions (on the partial
Boolean algebra approach) or picks out a statement that can’t be true (on a
lattice approach.) The distributive law fails, but not in a way that threatens
looming arithmetical catastrophe; Kripke’s “missing pair” is nowhere in the
neighborhood.
This isn’t what Putnam would say. He would say that the x-spin has a
definite value, either +1 or -1 but that it’s logically impossible to state this
value along with the z-spin value. However, once we recognize the possibility
of disjunctive facts, it’s clear that Putnam’s picture goes beyond saying that
‘Sx = +1 ∨ Sx = −1’ is true. We can assert the disjunction without accepting
the value-definiteness thesis.
The proposal under consideration includes these points:
1. Quantum mechanical quantities sometimes have values, though not all
quantities have values at once.
2. Bivalence fails; some statements about quantum systems are neither true
nor false;
3. Disjunctions can be true even though none of their disjuncts are.
4. Unrestricted distribution of “and” over “or” fails.
Perhaps (1)–(4) fit quantum systems; perhaps not. What I hope to have made
plausible is that they aren’t shocking. A full discussion would call for much
more detail (see Stairs [9] for some additional thoughts) but we turn to a
different question: how well does the proposal meet Kripke’s worries?
First, there’s no question of “standing outside logic” and choosing a logic.
This is a case of revision in light of finding an overlooked possibility: the possi-
bility of l-complementary propositions. On the one hand, if l-complementarity
is a genuine possibility, it’s one that we came to by way of quantum mechanics,
and quantum mechanics was an empirical discovery. However, grasping the

7 A partial Boolean algebra is a family of Boolean algebras that share a common 0 and 1. X ∨ Y

and X ∧ Y are only defined when X and Y belong to a common member of the family.
 COULD LOGIC BE EMPIRICAL? THE PUTNAM-KRIPKE DEBATE 39

implications for logic comes from reasoning about the theory and the con-
clusions about logic, and it would survive a change of physics. Recall the
case of geometry. We can (dimly) imagine discovering that the best theory of
space-time is Euclidean after all. However even if non-Euclidean geometry
didn’t fit this world, non-Euclidean space-time would be a genuine, albeit
unrealized possibility. Reasoning won’t tell us the actual geometrical structure
of the world, but empirical discoveries won’t tell us what the geometrical
possibilities are. Similarly, for all we can say for sure, we’ll find that the correct
account of quantum phenomena is some version of Bohmian mechanics. If
we do, physics would give us no reason to believe that the world exhibits
l-complementarity, nor disjunctive facts, nor failures of distributivity. How-
ever, this wouldn’t undermine the possibility of l-complementarity, nor the
possibility of disjunctive facts, nor the possibility of a world where distributivity
fails. The analogy with geometry is still apt: the possible structures, logical or
geometrical, go beyond the actual. Empirical findings may prompt us to have
thoughts we wouldn’t have had otherwise, but the discovery that something is
a non-actual possibility is not an empirical discovery. However, the structure
that the world actually instantiates—logical or geometrical—is something we
can only discover empirically.

§4. Coda: Some loose ends and some thoughts on logic and the limits of
thought. A question that often comes up in discussions of quantum logic is
whether it’s meant to apply universally, so to speak—whether quantum logic
is ‘the true logic,’ to use the phrase in Bacciagaluppi’s “Is Logic Empirical?”
[2] The point of view of this paper is that this is a misleading question. The
proposal, rather, is that if quantum mechanics is true, the world embodies
a logical relationship that hadn’t been noticed before: the one we’ve called
l-complementarity. If so, not all propositions are bivalent and distributivity
fails in certain special circumstances. Even if l-complementarity is a genuine
possibility, however, it doesn’t apply to every set of propositions. Compare:
suppose failures of bivalence are possible because it’s possible that determinism
and the block universe picture both fail. That admission wouldn’t call for
treating all propositions as neither true nor false, nor for saying that bivalence
fails in every domain.8 The point, rather, is that something we might have
taken to hold in all cases—as a matter of logic—holds only in some.
What’s been said also doesn’t take issue with the idea that our knowledge
of logic is a matter of reasoning. That’s not because this is beyond dispute.
It’s because a central aim of the paper was to see where things stand if we
concede to Kripke that what we’ve labeled logic is a matter of reasoning. We
have argued that even if Kripke is right and logic is not empirical, there’s still a

8 In particular, to take one important example, it gives no reason at all to think that mathematical

propositions are non-bivalent.

40 ALLEN STAIRS

place for empirical considerations in thinking about logic. The empirical is not
about what the logical possibilities are, but about which ones are realized.
That leaves a perplexing possibility that we’ll raise but not unravel. The
quantum logical story sketched here sees what we’ve called l-complementarity
as a feature of the world. The world, so this story goes, has logical structure
just as surely as it has geometrical structure; a bit too cutely, logic is empirical
even if logic isn’t. However, if this is correct it has an interesting implication:
we might not be capable of grasping all of what logic encompasses. This, in
turn could have the consequence that we are incapable of grasping the full
logical structure of the actual world.
Go back to the case of geometry. Suppose space-time indeed has the
structure of a pseudo-Riemannian manifold. In order to figure this out, we
needed the capacity to grasp the relevant concepts. That wasn’t inevitable; after
all, there are individual people who lack that capacity. Even if we had all been
unable to think the right thoughts, the world would still be pseudo-Riemannian.
The same goes if the l-complementarity-based account of quantum mechanics
gets the character of the world right. We are, collectively, lucky enough to be
able to grasp the relevant structures and concepts; collective truth though this
may be, it doesn’t apply to everyone and need not have applied to anyone.
However, it might be that the actual geometrical structure or the actual
logical structure of the world isn’t what we think it is. And it might be that
whatever that structure is lies beyond our cognitive reach. Logic would come
unpinned from reasoning in a different way than the one Kripke argued against.
One might dismiss this as a silly kind of skepticism. That would be fair if
the suggestion were that we might be deeply and radically ignorant about logic.
However, that’s not the thought. On the contrary (though we haven’t discussed
this) a full explication of l-complementarity assumes that propositions are
sometimes related exactly as classical logic says they are. (A partial Boolean
algebra, after all, is a family of Boolean algebras. Similar remarks apply to
orthomodular lattices.)
The point, rather, is this. What quantum mechanics may well represent is a
case in which we stumbled on a surprising exception to logical business as usual.
However, a full account of l-complementarity calls for positing relationships
among properties that we don’t grasp easily. Studying, for example, partial
Boolean algebras, as abstract mathematical structures is, of course, not the
issue. The difficulty is in grasping what it means for states of affairs in the
world to mirror that structure. One might fairly say that the persistent difficulty
in understanding quantum mechanics has been understanding what it means
for the world to have the structure that the mathematics seems to attribute to it.
In light of this, the possibility that there might be yet more esoteric exceptions
to business as usual doesn’t seem quite so silly. A proper modesty suggests
that there’s no guarantee that we’ll find them even if they exist. And a healthy
suspicion about our limitations suggests there’s no guarantee we would be
 COULD LOGIC BE EMPIRICAL? THE PUTNAM-KRIPKE DEBATE 41

able to recognize them even if they’re there. Logic in its fullness just might be
beyond our grasp.

REFERENCES

[1] W. Alston, Perceiving God, Cornell University Press, Ithaca New York, 1991.
[2] G. Bacciagaluppi, Is logic empirical?, Handbook of Quantum Logic (D. Gabbay,
D. Lehmann, and K. Engesser, editors), Elsevier, Amsterdam, 2009, pp. 49–78.
[3] C. Bourne, A Future for Presentism, Clarendon Press, Oxford, 2006.
[4] C. M. Caves, C. A. Fuchs, and R. Schack, Subjective probability and quantum certainty,
Studies in History and Philosophy of Modern Physics, vol. 38 (2007), p. 255.
[5] S. Kochen and E. Specker, The problem of hidden variables in quantum mechanics, Journal
of Mathematics and Mechanics, vol. 17 (1967), pp. 59–87.
[6] H. Putnam, Time and physical geometry, Journal of Philosophy, vol. 64 (1967), pp. 240–247.
Reprinted in Mathematics, Matter and Method, Cambridge University Press, 1975, pp. 198-205.
[7] , Is logic empirical?, Boston Studies in the Philosophy of Science (Robert S. Cohen
and Marx W. Wartofsky, editors), vol. 5, D. Reidel, Dordrecht, 1968, pp. 216–241. Reprinted as
The logic of quantum mechanics in Mathematics, Matter and Method, Cambridge University Press,
1975, pp. 174-197.
[8] A. Stairs, Quantum logic, realism and value-definiteness, Philosophy of Science, vol. 50
(1983), pp. 578–602.
[9] , Kriske, Tupman and Quantum Logic: the quantum logician’s conundrum, Physical
Theory and its Interpretation (W. Demopoulos and I. Pitowsky, editors), Springer, 2006.
[10] , A loose and separate certainty: Caves, Fuchs and Schack on quantum probability
one, Studies in History and Philosophy of Modern Physics, vol. 42 (2011), pp. 158–166.
[11] H. Stein, On Einstein-Minkowski space-time, The Journal of Philosophy, vol. 65 (1968),
pp. 5–23.
[12] R. H. Thomason, Indeterminist time and truth value gaps, Theoria, vol. 36 (1970), pp. 264–
281.

DEPARTMENT OF PHILOSOPHY
UNIVERSITY OF MARYLAND
COLLEGE PARK, MD 20742
E-mail: [email protected]
Random documents with unrelated
content Scribd suggests to you:
 CHAPTER XXXI.
DARREL’S RESOLVE.

On the afternoon which witnessed Merriwell’s and Clancy’s

disastrous experiences near Camp Hawtrey, Ellis Darrel had been laid
up nearly a week with his broken arm. He had been taken to
Dolliver’s because the Ophir lads knew that the ranch offered more
comforts than could possibly be had in the camp at Tinaja Wells.
Dolliver, too, had telephone connection with Ophir, and but little time
had been lost in getting a doctor.
Darrel was young and, at the time of his injury, in perfect physical
health. A year of roughing it in the West, all the way from British
Columbia to Mexico, had put a keen edge on his powers of
endurance. For him, therefore, a broken arm did not cause the
mischief which would have been the case in one less hardened and
robust.
In three or four days he was out of bed, and sitting around
Dolliver’s with his arm in a sling. Enforced idleness worried and
fretted him. On the very day Frank and Owen had saved the coyote
dog, Darrel had begun contemplating a return to Tinaja Wells.
The one thing in all the world which Darrel desired with a full
heart was to prove his innocence in the forgery matter. He felt that
he could not rest easy a moment until he had probed that forgery to
the bottom and had unmasked the person who had written the
name of Alvah Hawtrey on the five-hundred-dollar check.
The colonel, after considering the circumstantial evidence, had
reached the conclusion that Darrel was the forger. He had therefore
turned the boy from his door and would have nothing more to do
with him. To wipe that blot from his name was Darrel’s one purpose
in life. Merriwell had promised his help, but Darrel believed that it
was his duty to do most of the work for himself.
After supper, in the evening of the day so many important events
had happened at Camp Hawtrey, Darrel was sitting with the rancher
in front of the house.
The cloudless Arizona sky was never more beautiful. When the
sun sets in the Southwest, it drops out of sight suddenly, and night
falls as swiftly as a drop curtain. One moment it is day; then, almost
the next moment, the clear-cut stars are glittering overhead.
The entrance to Mohave Cañon was but a little distance away and
facing the front of Dolliver’s house. The opening yawned like a huge
black cavity on the sky line, stretching into the far distance amid
ominous shadows.
With dreamy eyes young Darrel stared across the trail and into the
gloomy gulch. Somehow the last year of his life resembled that
cañon as he saw it then. That forgery had flung him into a black and
forbidding path, through which he had wandered—and was still
wandering—aimlessly. Would he never be able to fight his way out of
the gloom and the dishonor and regain his rightful place in his
uncle’s esteem, and in the eyes of honest men?
While Dolliver, a man of few words, like all who live much by
themselves, sat silently and smoked his short black pipe, and while
Darrel still gazed reflectively into the black mouth of the cañon, two
figures slowly disentangled themselves from the shadows and bore
down on the ranch.
“Some ’un from up the gulch,” Dolliver roused to remark,
“mebbyso from Tinaja Wells.”
But they were not from the Wells. As the riders came close and
halted, Darrel discovered that they were two whom he knew—
Bleeker and Hotchkiss.
 “Great jumpin’ sandhills!” exclaimed the voice of Hotchkiss. “That
you, Darrel?”
“Sure,” laughed Darrel. “Why not?”
“We reckoned you would still be in bed, El,” spoke up Bleeker.
“Must be pulling along in fine shape, eh?”
“How long do you think a busted arm ought to keep a fellow
down, anyhow?”
“Depends a heap on the fellow, El. Between you, and me, and the
gatepost, I don’t believe anything’ll keep you down very long.”
“Can’t you get off and stop a while?” Darrel asked.
“No. We’re bound for Gold Hill. Been kicked out of Camp Hawtrey.”
“Kicked out? Great Scott! What do you mean by that, Bleek?”
“Down at the bottom of it, we’re friends of yours, and Jode don’t
want us around. Something happened up at the camp, this
afternoon, that brought matters to a show-down.”
“What was that?”
Bleeker crooked one knee around the saddle horn and rested
easily while he told about the trouble over the coyote dog.
“That’s what happened,” said he, when the recital was finished,
“and I’ll bet a pound of prunes against a toothpick that Jode’s laying
to unload a little of the trouble onto you.”
“How could he do that?” queried Darrel.
“Why, by making his uncle believe that your unholy influence sent
Merriwell and Clancy to our camp to kick up a row. Parkham has
already been sent to the Hill after the colonel. He’ll be out here,
bright and early, to-morrow morning; then Jode will sing his little
song and make the colonel believe just what he wants him to. The
friendly relations of the two clubs have had a knock-out blow.
There’ll be nothing doing, in an athletic way, for some time to come.
Pretty tough on Merriwell. But he’ll come out all right, for that’s a
way he has. Get well as quick as you can, pard, and then come on
to Gold Hill. There are a lot of us there that are ready to fight for
you. Buenas noches!”
Bleeker straightened around in his saddle and rattled his spurs.
Presently he and Hotchkiss were clattering away along the main trail
in the direction of home.
These revelations came to Darrel like a blow. He felt, and perhaps
he was right, that Merriwell’s friendship for him had made an enemy
of Jode.
“What do you think of that, Dolliver?” asked Darrel, appealing to
the rancher.
“Why,” was the answer, “I opine that half brother o’ yourn is about
as onnery as they make ’em.”
“I’m the one who is at the bottom of Merriwell’s trouble with
Jode.”
“You can’t help it if ye are. Better hit the hay, son. I reckon you’ve
been up a heap too long as it is.”
Darrel went to bed that night pondering the subject of Merriwell’s
failure to inspire a friendly spirit in the dealings between the two
athletic clubs.
“He could have succeeded,” was Darrel’s bitter conclusion, “if it
hadn’t been for his friendship for me. What will Jode be trying next,
I wonder? Where is that fiendish temper of his going to land him, if
something isn’t done to curb it?”
Long into the night Darrel canvassed the unpleasant problem in
his mind. As a consequence, he went to sleep about midnight and
woke up with the sun at least two hours’ high.
 “Has my uncle passed on his way to Camp Hawtrey, Dolliver?”
were his first words when he found the rancher.
“All of an hour ago,” was the reply.
“I wanted to talk with him,” muttered Darrel.
“A heap o’ palverin’ you’d ‘a’ done with him,” grunted Dolliver. “The
kunnel ain’t eager for no conversation with you, son.”
Darrel realized that, but it did not alter his determination to see if
he could not talk with his uncle and try to make things easier for
Merriwell.
The morning passed slowly, Darrel deciding one moment that duty
called him to Tinaja Wells and Merriwell, and again that his proper
course was to ride to Camp Hawtrey and interview the colonel.
Noon came, and Darrel ate little of the food Dolliver had set out
on the kitchen table.
“If ye don’t eat,” grumbled Dolliver, “ye can’t expect to git around
very soon.”
Darrel’s mind was on something else besides his dinner.
“I wish you’d saddle up a horse for me, Dolliver,” he said. “I’m
going to take a ride.”
“More’n likely ye’ll fall off before ye’ve gone fur. Where ye goin’ to
ride?”
“Camp Hawtrey.”
“Take a fool’s advice, son, and don’t.”
“I’m going to talk with the colonel. If you won’t put the gear on a
horse for me, I reckon I can manage it myself.”
“Oh, I’ll do it, if ye’re bound ter ride. But wait a couple o’ hours.
It’s plumb in the heat o’ the day, and ridin’ ’ll come a heap harder for
you now than it will later.”
An hour or two would make little difference, and Darrel laid down
on his bed for a short rest before taking the ride. He fell asleep
almost immediately, and was awakened by a familiar voice trying to
get some one over the telephone. It was his uncle, there in the room
with him, asking for Bradlaugh’s office. Bradlaugh was not in,
evidently.
“Tell him,” said Colonel Hawtrey, “that I’ll talk with him from here
late this afternoon. This is mighty important—don’t neglect to tell
him that.”
Colonel Hawtrey had just ridden down the cañon after his talk
with Merriwell. He was still red and wrathful. As he whirled from the
telephone, he was confronted by Darrel.
The boy’s face was as white as the bandage that swathed his arm,
but he stood resolutely between his uncle and the open outside
door.
“Colonel,” he began, “I want you to listen to me. I’m not talking
for myself, but for Merriwell. Don’t think that I——”
“Not a word,” snapped the colonel. “You haven’t anything to say
that I care to hear.”
He strode around Darrel. The boy stepped forward to lay a
detaining hand on his arm. Roughly the colonel shook him off,
hurried from the house, vaulted into the saddle of his waiting horse,
and spurred for the cañon. He did not so much as look back.
“Nice way for an uncle to treat his nephew!” exclaimed Dolliver,
from a place outside the house near the door. “But I told ye how it
’u’d be,” he added.
“He can’t shake me like that!” cried Darrel. “I’m going to do what I
can to straighten out this trouble of Merriwell’s. Get the horse for
me, Dolliver, and I’ll hike right after him.”
 “Ye’ve got plenty o’ nerve, son, but blame’ poor jedgment,”
growled the rancher.
“Why didn’t you call me,” demanded Darrel, “when you saw him
coming?”
“Didn’t see him comin’. Didn’t have a notion anybody had dropped
in till I saw the strange hoss at the hitchin’ pole.”
“Will you get the horse for me, Mr. Dolliver?”
The “mister” was pretty formal. The fact that Darrel used it proved
that he was on edge and would not take “no” for an answer.
Dolliver got the horse and helped Darrel into the saddle. He
wished him luck, too, although in the same breath he declared that
the boy was running a big risk and would have his trouble for
nothing.
Darrel’s pale face was set resolutely as he urged the horse into a
gallop and disappeared through the mouth of the cañon.
 CHAPTER XXXII.
THE LEDGE AT THE GULCH.

In a great many ways Merriwell had shown his friendship for Ellis
Darrel. From the very first, when Darrel had reached the camp at
Tinaja Wells as the “boy from Nowhere,” Merriwell had believed in
him and had befriended him.
As he rode toward Camp Hawtrey, Darrel recalled how cleverly
Merriwell had defended him against the charge of robbing the
colonel’s safe. So successful was the defense that even the stern old
colonel was forced to admit that Darrel was innocent.
And again, at the time the rope had given way and Darrel had
fallen on the cliff, it was Merriwell who had risked his neck to climb
to the ledge where Darrel lay unconscious, had fastened a rope
about him, and had lowered him to safety. It was Merriwell, too, who
had played “a game within a game” on the football field and had
taken from Lenning certain evidence of Lenning’s scoundrelly work.
As a slight repayment for all this loyalty and friendship, Darrel felt
that he should do what he could to straighten out the
misunderstanding between the colonel and Merriwell.
Even if he could get the colonel’s attention, Darrel was doubtful of
his ability to sway the colonel toward Merriwell’s side. It was a time,
however, when Darrel was resolved to give himself the benefit of
every doubt, in the hope of being of some service to his friend.
If Jode was successful in making the colonel believe that Darrel’s
influence had caused the trouble between him and Merriwell, then
Darrel would do his utmost to set his uncle right on that point. This,
very likely, would put an altogether different complexion on the clash
about the coyote dog.
If convinced that Darrel had nothing to do with the actions of
Merriwell and Clancy, the colonel might be in a receptive mood so far
as evidence against Jode was concerned. This, at least, was what
Darrel hoped.
A mile or so from the mouth of the cañon the right-hand wall was
broken into by the mouth of a gulch. This gulch was the one in
which the Gold Hill Boys had pitched their camp.
Years before, a mining company had thrown a dam across the
mouth of the gulch. This dam had backed up the water for several
miles.
Darrel turned his horse into the gulch and followed a bridle path
that led onward close to the water’s edge. Rapidly, as he advanced,
the gulch widened out. The slopes on either side of the stream
became less steep, pine trees began to show themselves, and
flaming poppies, in irregular beds, made the slopes look like terraced
gardens.
“First time I ever knew there was a place like this holed away
among these hills,” muttered the boy, staring around him with all the
delight aroused by a new and pleasant discovery. “It’s a mighty fine
place, and no mistake. Where’s that camp, I wonder?”
Pulling the horse to a halt, he lifted himself in the stirrups and
peered ahead. He could not see the gleam of the tents, but he did
see something else which caused him to utter an exclamation of
surprise and disappointment.
In the distance two figures were moving in his direction, on foot.
One of them was the colonel, as he could see plainly, and the other
was Jode.
“Beastly luck!” grumbled Darrel. “How can I talk with the colonel if
Jode’s around? I’ll just leave the horse in the brush and watch them,
for a spell. Maybe Jode will leave the colonel, and I’ll get my
chance.”
Quickly turning the horse from the trail, Darrel spurred up the
slope of the gulch wall for a short distance and rode into a chaparral
of mesquite. Here he dismounted, hitched the horse to a scraggly
paloverde, and crept back to the edge of the bushes to watch.
He had had no exercise to amount to anything for nearly a week,
and he was astonished to find how his exertions tired him. He half
reclined as he stared out of the thicket, resting as he watched the
trail for the colonel and Jode to appear.
It was plain that the two could not be going far from the camp.
Had they been traveling any considerable distance, they would have
brought their mounts.
Not many minutes passed before the two hove in sight. Only a
little way from the place where Darrel had turned from the trail, the
colonel and Jode altered their course and began climbing the slope.
The colonel was carrying a small package wrapped in brown paper.
It seemed evident to Darrel that the two from the camp would
pass within a few yards of the chaparral. What if they discovered the
horse? The boy compressed his lips sternly. If that happened, then
he would show himself at once and talk to the colonel, in spite of
Jode. But he hoped the horse would not be seen, and that he could
watch his chances and have the colonel all to himself for a few
minutes.
The climb must have tired the colonel, for he halted and sat down
on a convenient bowlder for a brief rest. Jode dropped to the ground
at his side. They were not more than twenty feet from Darrel.
“It won’t take me ten minutes to load the hole and set off the
charge, Jode,” the colonel was saying, “and then we’ll see what sort
of rock we uncover. There’s a vein there—I’m too old a hand at the
business to be fooled—but whether it amounts to much or not
remains to be seen.”
“You’re mighty clever at this sort of business, Uncle Al,” returned
Jode admiringly. “I wish I knew as much about dips, angles, and
formations as you do.”
“It won’t be necessary for you to work along that line, my boy,”
said the colonel affectionately. “You’re to educate yourself for
commercial work, and learn to take care of what I shall one day
leave you.”
“I hope,” observed Jode, “that it will be a long time before I shall
be called on to do that. There’s no chance, you think, of patching up
our differences with the Ophir fellows?”
“No chance—at least, not so long as Merriwell has anything to do
with the Ophir team. I’ve cancelled the Thanksgiving Day game.”
“That’s pretty tough! I think, uncle, we could play Ophir, even with
Merriwell in their crowd, and show them that we can be square and
let bygones be bygones.”
“What you say, Jode, does you a lot of credit. Our boys are
gentlemen, however, and not hoodlums. I could not sanction your
playing with a team where such a spirit as Merriwell and Clancy
showed yesterday is liable to crop out at any moment.”
“Whatever you say goes, Uncle Al. But I wish the thing could be
patched up in some way.”
“Well, I don’t see how it can. Mr. Bradlaugh has placed Merriwell in
charge of the Ophir eleven, and a team is bound to reflect the spirit
of the coach. There’ll be no more exhibitions of petty partisanship
between the two clubs if I can help it.” The colonel got up and
stooped to lay hold of the bundle he had been carrying. “What’s the
matter?” he asked, starting quickly erect.
Jode had given a jump and uttered a startled exclamation.
 “I—I thought I saw that coyote dog among the rocks, up toward
the ledge,” he answered, in a smothered voice.
“What if you did?”
“Why, I heard—some one in the camp told me—that a coyote dog
always lays for the fellow who tries to hurt him or——”
“Stuff and nonsense!” scoffed the colonel. “You ought to be above
such superstitious notions, Jode. Never mind if you did catch a
glimpse of the dog. Come on and we’ll go up to the ledge and do
our work there.”
“I wish I’d brought my revolver,” said Jode, as he again began
climbing at his uncle’s side.
“You’ll not need your revolver.”
Contrary to Darrel’s fears, the two passed well to the side of the
chaparral. The colonel’s mind was busy with the work that lay ahead
of him, and Jode was still plainly experiencing a few qualms on the
score of the coyote dog. As he climbed, Jode’s shifty eyes were fixed
on the rocks where he believed he had caught sight of the skulking
animal.
What Darrel had overheard pass between his half brother and the
colonel gave him a queer feeling of regret for the part he was
playing. It seemed almost as though he was a spy and an
eavesdropper. The colonel’s affection for Jode was deep and sincere,
there could not be the slightest doubt; but Jode’s manner, his very
talk, to Darrel’s mind, lacked all that the colonel’s so frankly
expressed.
“What business is it of mine?” thought Darrel bitterly. “So long as I
am under a cloud I have no right to criticize Jode. I wish he’d clear
out and give me a chance at the colonel.”
Some twenty or thirty feet above the chaparral, and forty or fifty
feet to the left of it, was a ledge of rock standing straight out from
the sloping gulch wall. A mass of loose bowlders overhung the
ledge.
This was the spot toward which the colonel and Jode were
climbing. Observing this, Darrel quietly forced his way upward along
one side of the patch of mesquite. At the upper edge of the
chaparral he found a rift in the slope. It was like a trench, deep
enough to hide a man, and ran straight toward the crest of the gulch
wall.
Still watching and hoping for an opportunity to speak a few words
in private with the colonel, Darrel crawled into the trench and made
his way to a point that was on a level with the top of the ledge.
When he finally halted and peered over the edge of the rift, he
found that some thirty feet of rough ground separated him from the
colonel and Jode.
The colonel was on his knees, carefully opening the parcel he had
brought with him. A small coil of fuse and a couple of sticks of
dynamite were presently taken from the package.
“There were three sticks here when I wrapped up the package in
Gold Hill,” said the colonel, lifting his eyes to Jode’s. “What’s become
of the rest of the dynamite?”
“Are you sure?” Jode answered. “Some one must have taken out
one of the sticks.”
“Of course I may be mistaken,” muttered the colonel.
Cutting off a length of fuse, he trimmed it with a pocket knife;
then, taking a cap from his pocket, he pushed it over the trimmed
end. Next, he picked up one of the sticks of giant powder, slit it
lengthwise on four sides, and dropped it into a hole that had been
drilled in the shelf. The other stick was pushed down on the first,
and both were gently tamped down on the cap, which was in the
bottom of the hole.
 “Now, clear out, Jode,” said the colonel. “It’s only a two-minute
length of fuse, and I shall have to scramble for safety when I touch
it off.”
Jode jumped from the ledge and hurried to get away among a lot
of bowlders at a safe distance. The colonel lighted a match, touched
it to the fuse, and Darrel flattened himself out in the bottom of the
rift.
The next moment he heard a crash, but it was not the crash of an
explosion. A startled cry came from the colonel, and Darrel, thrilled
with a weird premonition of disaster, rose to his knees and again
looked out over the top of the rift. What he saw, there on the ledge
of the gulch wall, caused him to gasp and close his eyes to shut out
the horror of it.
 CHAPTER XXXIII.
FOLLOWING DARREL.

Frank and his chums, in riding from Tinaja Wells to Dolliver’s,

passed the mouth of the gulch only a few moments after Darrel had
ridden into it. Had Frank encountered Darrel, there is no doubt but
that he would have persuaded him against going on to Camp
Hawtrey. In that event, some very pretty maneuvers of Fate,
calculated to benefit Darrel, would have been effectually blocked.
But Merry and his two friends missed their new chum by a scant
margin, and galloped on to Dolliver’s. Dolliver, smoking his short
black pipe, was sitting in front of his little establishment, mentally
considering uncles and nephews, and the foolishness of a kid with a
broken arm trying to take a horseback ride before he was well able
to be out of bed.
At sight of Merriwell, Ballard, and Clancy, Dolliver’s reflections
went off at a fresh angle. He now began to concern himself with the
contrariness of human affairs in general.
“Hello, Dolliver!” Frank called, pulling in his black mount, Borak.
“How’s Curly?”
“Plumb locoed,” grunted the rancher.
“You don’t mean to say he’s out of his head?” gasped Frank.
“If he ain’t, then, by the jumpin’ hocus-pocus, I never see a feller
that was.”
“We’ll have to see about this!”
Frank slid from the saddle and started hurriedly into the house.
 “No use lookin’ fer him in the wikiup, Merriwell,” said Dolliver,
“kase he ain’t there.”
“Not in the house?” demanded Frank, recoiling in amazement.
“Where is he, then?”
“Gone to Camp Hawtrey to make the old kunnel talk with him.”
“What do you know about that!” exclaimed Ballard.
“Thunder!” cried the astounded Clancy.
“How long since he left here?” asked Frank.
“Less’n half an hour.”
“Did he ride?”
“Sartain he did. No more business on a hoss than a two-year-old
kid, nuther. He’s wuss to manage than a case o’ the measles,
anyways. Howsumever, he would go. He reckoned he could talk with
the kunnel and smooth things out fer you.”
“How did he know matters had to be smoothed out for me?”
“Bleeker and Hotchkiss dropped in here on their way to the Hill,
and they cut loose about your troubles. That got Darrel all het up.
Right arter dinner, to-day, the kunnel himself blowed in here and
tried to git Mr. Bradlaugh on the telephone. But Bradlaugh was away
on business, I reckon. I wasn’t in the shack at the time, but I heerd
the kunnel sayin’ the business was important and that he’d call up
later this afternoon. Darrel was in the house, though, and tried to
powwow with the kunnel, but the kunnel wouldn’t have it. Runnin’
out, the kunnel climbed his hoss and moseyed up the cañon. Nothin’
’u’d do but Darrel had to mosey arter him.”
“Here’s news, fellows, and no mistake!” breathed Merriwell.
“Curly wasn’t able to take such a ride,” growled Ballard, “and
that’s a cinch.”
 “What does he think he can do, anyhow?” asked Clancy. “He’s not
on the colonel’s visiting list.”
“Have you any idea what he intended to do, Dolliver?” Merry went
on.
“Palaver with that grouchy old uncle o’ his,” replied the rancher.
“Jode’s tryin’ to make the kunnel believe Darrel set you up to act like
you done. I allow that Darrel wants to disabuse his mind, thinkin’
that if he’s out o’ it you’ll have less trouble comin’ to an
understandin’ with Hawtrey.”
“Foolish!” muttered Merriwell. “He couldn’t make the colonel
believe any such thing, and it wouldn’t help if he could. I wish we’d
get here in time to head Darrel off. What’ll happen to him when he
gets to Camp Hawtrey?”
“I don’t opine he’ll ever git there,” and Dolliver shook his head
dubiously. “He wa’n’t able to sit a hoss, not noways.”
Frank hurried to Borak and leaped into the saddle.
“Only one thing to do, fellows,” he announced, “and that’s for us
to ride for Camp Hawtrey.”
“Bully!” exulted the red-headed chap. “That gang will sure
welcome us with open arms.”
“They will that,” agreed Dolliver. “Say, if you go to the kunnel’s
camp, jest now, ye’ll have the time o’ your lives.”
“All right,” answered Frank, “I don’t care how hot a time they give
us providing we can do something to help Darrel. Come on, fellows!”
He pointed Borak for the mouth of the cañon, and set off at
speed. Clancy and Ballard made after him.
The cañon trail was narrow and the riders were obliged to proceed
in single file. When they turned into the gulch, however, they were
able to ride stirrup to stirrup.
 “I don’t like the prospect a little bit,” said Frank. “Now that Bleeker
and Hotch have left the Gold Hill camp, there isn’t a fellow there
that’s at all friendly toward Darrel.”
“Hawtrey’s there,” suggested Ballard. “Don’t forget that, Chip.
Hawtrey won’t have anything to do with Curly, but you can bet he
won’t let Jode rough things up with him.”
“That’s right, Pink. Darrel must be a little hazy in his mind to start
for the Gold Hill camp at such a time as this.”
“He’s trying to do you a good turn, Chip,” suggested Clancy.
“Sure he is—I give him credit for that—but the crazy old lobster
can’t do me any good, or himself, either. He ought to stay in the
house for another week yet.”
“Bosh!” returned Clancy. “Curly is all rawhide and India rubber. A
broken wing hadn’t ought to bother him much more than a mild case
of the mumps. You’ll notice we haven’t run across him lying along
the road.”
“He’ll stick it out, you can bank on that,” said Ballard. “He’s
probably in Camp Hawtrey this minute. That bunch would be pretty
yellow if they didn’t treat him right.”
Clancy had a sudden thought.
“Say, Chip,” said he, “we’re taking this hike to help Curly, but I
don’t think we’ll do him much good if we plunge full tilt into the
camp. They’re a suspicious lot, and they might think it a frame-up of
Curly’s. Suppose we reconnoiter a little before we show ourselves?”
“How’ll we reconnoiter, Clan?” asked Merry.
“The top of the gulch wall, about where we were yesterday, is a
good place for that.”
“I guess you’ve got the right end of the stick, Clan. If we’re to
climb the bank we’d better begin right here. Strikes me this is as
good a place as we’ll find, and it’s far enough this side of the camp
so we can make the climb without being seen.”
The slope was not steep, but it was easier for the boys to walk up
the incline and lead their horses. In perhaps ten minutes they had
reached the crest, and were able to take a comprehensive survey of
the gulch below.
“Jove!” exclaimed Merry. “There are two fellows on a bowlder
down there. See them? They are just below that chaparral of
mesquite. One of them looks like the colonel to me. Wonder if the
other is Darrel?”
“Not on your life!” murmured Clancy. “The other is Jode.”
“Sure enough!” agreed Ballard. “We’d better lead our horses back
from the rim, and drop down on the rocks. If the colonel and Jode
happened to look up here, they’d see us.”
Ballard’s suggestion was carried out at once; then, on their knees,
the lads continued to peer downward. Presently the colonel and Jode
got up and began climbing. They passed well to the left of the
chaparral, angled across the face of the slope, and stepped upon a
ledge that jutted out from the gulch side.
“I’m next to what’s going on down there,” said Merry. “Remember
what Bleek told us, Clan, when I asked him where Jode got that
dynamite for the cartridge?”
“He said something about Hawtrey stumbling on a ‘prospect,’” was
the answer, “and that Jode was to fill a hole, and the colonel was to
load it and set it off.”
“That’s what the colonel is about to do. Let’s move down the gulch
a little way and find a place directly over the ledge.”
A hundred yards carried the boys to a spot above the ledge.
Masses of splintered granite and loose bowlders covered the slope
between the ledge and the crest of the gulch wall. The boys were
able to look over the intervening rocks, however, and get a clear
view of the ledge level.
Colonel Hawtrey, on his knees, was at work capping a fuse and
ramming dynamite into the hole where the blast was to be set off.
“You’re right about it, Chip,” said Clancy. “The colonel’s going to
have a little blow-up, down there, and probably he’ll make a ‘strike.’
How many poor prospectors, do you suppose, have passed that
‘prospect’ by? That’s the way things work out, in this world. Here’s
the colonel, with more mines and money than he knows what to do
with, just falling right over a good thing. Now——”
“Look!” broke in Ballard, grabbing Frank’s arm and pointing
downward and to the left of the ledge. “See that long break in the
gulch wall, running from the top right down to that bunch of
chaparral? Who’s that looking out of it?”
“Darrel!” murmured Merriwell, astounded.
“Curly, as sure as you’re a foot high!” fluttered Clancy. “Now, what
the deuce do you suppose he’s up to?”
It was a surprising situation, and no mistake. Darrel, screened in
the rift, was cautiously looking out and keeping track of the
movements of the colonel and Jode.
“Curly wants to talk with the colonel,” said Frank, after a moment’s
thought, “and he’s waiting for Jode to get out of the way.”
“I could slip down that chute,” suggested Ballard, “and slide right
into Darrel. We could bring him up here, with us, and——”
“Wait till after the blast,” cut in Merry. “The colonel’s just touching
it off.”
“See Jode scramble for the tall rocks!” chuckled Clancy. “He’s not
going to take any chances on being knocked over by flying stones.”
 “Neither is Curly,” added Ballard. “He has ducked down into the
bottom of that hole of his.”
“Two sticks of dynamite will lift a pretty big chunk out of that
ledge,” said Merriwell, “and before it lets go we’d better push back a
little. The charge——”
The words died on Merry’s lips. A bowlder, just above the ledge,
had slipped from its moorings and was rolling over and over,
grinding and crashing toward the ledge. The colonel had just risen
from lighting the fuse. He saw the bowlder, and tried frantically to
get out of the way of it. In his haste, he slipped and fell prone upon
the ledge. The next moment the bowlder was upon him!
Welcome to Our Bookstore - The Ultimate Destination for Book Lovers
Are you passionate about books and eager to explore new worlds of
knowledge? At our website, we offer a vast collection of books that
cater to every interest and age group. From classic literature to
specialized publications, self-help books, and children’s stories, we
have it all! Each book is a gateway to new adventures, helping you
expand your knowledge and nourish your soul
Experience Convenient and Enjoyable Book Shopping Our website is more
than just an online bookstore—it’s a bridge connecting readers to the
timeless values of culture and wisdom. With a sleek and user-friendly
interface and a smart search system, you can find your favorite books
quickly and easily. Enjoy special promotions, fast home delivery, and
a seamless shopping experience that saves you time and enhances your
love for reading.
Let us accompany you on the journey of exploring knowledge and
personal growth!

ebookgate.com