Truth Deduction and Computation Logic and

Semantics for Computer Science 1st Edition by RE

Davis ISBN 0716782014 9780716782018 download

https://ebookball.com/product/truth-deduction-and-computation-
logic-and-semantics-for-computer-science-1st-edition-by-re-davis-
isbn-0716782014-9780716782018-19792/

Download more ebook instantly today - Get yours now at ebookball.com

 Get Your Digital Files Instantly: PDF, ePub, MOBI and More
Quick Digital Downloads: PDF, ePub, MOBI and Other Formats

Mathematical Logic for Computer Science 2nd Edition by Mordechai Ben

Ari ISBN 1447103351 9781447103356

https://ebookball.com/product/mathematical-logic-for-computer-
science-2nd-edition-by-mordechai-ben-ari-
isbn-1447103351-9781447103356-19800/

Theory of Computer Science Automata Languages and Computation 3rd

Edition by ISBN KLP Mishra, N Chandrasekaran ISBN 9788120329683

https://ebookball.com/product/theory-of-computer-science-
automata-languages-and-computation-3rd-edition-by-isbn-klp-
mishra-n-chandrasekaran-isbn-9788120329683-15850/

Computer Science Press An Introduction to the Theory of Computation

1st Edition by Eitan Gurari ISBN

https://ebookball.com/product/computer-science-press-an-
introduction-to-the-theory-of-computation-1st-edition-by-eitan-
gurari-isbn-10804/

Logic in Computer Science Modelling and Reasoning about Systems 2nd

Edition by Michael Huth, Mark Ryan ISBN 0511261586 9780511261589

https://ebookball.com/product/logic-in-computer-science-
modelling-and-reasoning-about-systems-2nd-edition-by-michael-
huth-mark-ryan-isbn-0511261586-9780511261589-19806/
Logic in Computer Science Modelling and Reasoning about Systems 2nd
Edition by Michael Huth, Mark Ryan ISBN 0511261586 9780511261589

https://ebookball.com/product/logic-in-computer-science-
modelling-and-reasoning-about-systems-2nd-edition-by-michael-
huth-mark-ryan-isbn-0511261586-9780511261589-19814/

LNCS 2803 Computer Science Logic Frontmatter Pages 1st edition by

Matthias Baaz, Johann Makowsky 3540408010

https://ebookball.com/product/lncs-2803-computer-science-logic-
frontmatter-pages-1st-edition-by-matthias-baaz-johann-
makowsky-3540408010-19920/

Computer Science with MATHEMATICA Theory and Practice for Science

Mathematics and Engineering 1st Edition by Roman Maeder ISBN
1595720596 9780521663953

https://ebookball.com/product/computer-science-with-mathematica-
theory-and-practice-for-science-mathematics-and-engineering-1st-
edition-by-roman-maeder-isbn-1595720596-9780521663953-19788/

C for Computer Science and Engineering 1st Edition by Vic Broquard

ISBN 1941415539 9781941415535

https://ebookball.com/product/c-for-computer-science-and-
engineering-1st-edition-by-vic-broquard-
isbn-1941415539-9781941415535-19790/

Digital Logic and Computer Design 1st edition by Morris Mano ISBN
817758409X 978-8177584097

https://ebookball.com/product/digital-logic-and-computer-
design-1st-edition-by-morris-mano-
isbn-817758409x-978-8177584097-17666/
PRINCIPLES OF COMPUTER SCIENCE SERIES
Series Editors
Alfred V. Aho, Bell ??Aephone Laboratories, Murray Hill, New Jersey
Jeffrey
Computer
D. Ullman, Stanford University, Stanford, California
Organization
TRUTH,DEDUCTION,AND
Michael Andrews
Trends in Theoretical Computer Science COMPUTATION
Egon Rorger, Editor
‘B-uth, Deduction, and Computation: Logic and Semantics for Computer Science
R. E. Davis
The luring Omnibus
Logic and Semantics for Computer Science
A. K. Dewdney
Formal Languages and Automata Theory
Vladimir Drobot
Advanced C: Food for the Educated Palate
Narain Gehani
R. E. DAVIS
C: An Advanced Introduction Santa Clara University
Narain Gehani
C: An Advanced Introduction, Ansi C Version
Narain Gehani
C for Personal Computers: IBM PC, AT&T PC 6300, and Compatibles
Narain Gehani
An Introduction to the Theory of Computation
Eitan Gurari
Theory of Relational Databases
David Maier
An Introduction to Solid Modeling
Martti Mantyla
Principles of Computer Design
Leonard R. Marino
UNIX: The Minimal Manual
Jim Moore
A Logical Language for Data and Knowledge Bases
Shamim Naqvi and Shalom Tsur
The Theory of Database Concurrency Control
Christos Papadimitriou
Algorithms for Graphics and Image Processing
Theo Pavlidis
Data Compression Methods and Theory
James A. Storer
The Elements of Artificial Intelligence
Steven lbnimoto
Computational Aspects of VLSI
Jeffrey D. Ullman
hinciples of Database and Knowledge-Base Systems, Volumes I and II
Jeffrey D. Ullman
Algorithmic Studies in Mass Storage Systems
C. K. Wong

OTHER BOOKS OF INTEREST

‘Jewels of Formal Language Theory
Art0 Salomaa
Fussy Sets, Natural Language Computations, and Risk Analysis
Kurt J. Schmucker
LISP An Interactive Approach
Stuart C. Shapiro COMPUTER SCIENCE PRESS
Principles of Database Systems, Second Edition
Jeffrey D. Ullman AN IMPRINT OF W. H. FREEMAN AND COMPANY. NEW YORK
PRINCIPLES OF COMPUTER SCIENCE SERIES
Series Editors
Alfred V. Aho, Bell ??Aephone Laboratories, Murray Hill, New Jersey
Jeffrey
Computer
D. Ullman, Stanford University, Stanford, California
Organization
TRUTH,DEDUCTION,AND
Michael Andrews
Trends in Theoretical Computer Science COMPUTATION
Egon Rorger, Editor
‘B-uth, Deduction, and Computation: Logic and Semantics for Computer Science
R. E. Davis
The luring Omnibus
Logic and Semantics for Computer Science
A. K. Dewdney
Formal Languages and Automata Theory
Vladimir Drobot
Advanced C: Food for the Educated Palate
Narain Gehani
R. E. DAVIS
C: An Advanced Introduction Santa Clara University
Narain Gehani
C: An Advanced Introduction, Ansi C Version
Narain Gehani
C for Personal Computers: IBM PC, AT&T PC 6300, and Compatibles
Narain Gehani
An Introduction to the Theory of Computation
Eitan Gurari
Theory of Relational Databases
David Maier
An Introduction to Solid Modeling
Martti Mantyla
Principles of Computer Design
Leonard R. Marino
UNIX: The Minimal Manual
Jim Moore
A Logical Language for Data and Knowledge Bases
Shamim Naqvi and Shalom Tsur
The Theory of Database Concurrency Control
Christos Papadimitriou
Algorithms for Graphics and Image Processing
Theo Pavlidis
Data Compression Methods and Theory
James A. Storer
The Elements of Artificial Intelligence
Steven lbnimoto
Computational Aspects of VLSI
Jeffrey D. Ullman
hinciples of Database and Knowledge-Base Systems, Volumes I and II
Jeffrey D. Ullman
Algorithmic Studies in Mass Storage Systems
C. K. Wong

OTHER BOOKS OF INTEREST

‘Jewels of Formal Language Theory
Art0 Salomaa
Fussy Sets, Natural Language Computations, and Risk Analysis
Kurt J. Schmucker
LISP An Interactive Approach
Stuart C. Shapiro COMPUTER SCIENCE PRESS
Principles of Database Systems, Second Edition
Jeffrey D. Ullman AN IMPRINT OF W. H. FREEMAN AND COMPANY. NEW YORK
PRINCIPLES OF COMPUTER SCIENCE SERIES
Series Editors
Alfred V. Aho, Bell ??Aephone Laboratories, Murray Hill, New Jersey
Jeffrey
Computer
D. Ullman, Stanford University, Stanford, California
Organization
TRUTH,DEDUCTION,AND
Michael Andrews
Trends in Theoretical Computer Science COMPUTATION
Egon Rorger, Editor
‘B-uth, Deduction, and Computation: Logic and Semantics for Computer Science
R. E. Davis
The luring Omnibus
Logic and Semantics for Computer Science
A. K. Dewdney
Formal Languages and Automata Theory
Vladimir Drobot
Advanced C: Food for the Educated Palate
Narain Gehani
R. E. DAVIS
C: An Advanced Introduction Santa Clara University
Narain Gehani
C: An Advanced Introduction, Ansi C Version
Narain Gehani
C for Personal Computers: IBM PC, AT&T PC 6300, and Compatibles
Narain Gehani
An Introduction to the Theory of Computation
Eitan Gurari
Theory of Relational Databases
David Maier
An Introduction to Solid Modeling
Martti Mantyla
Principles of Computer Design
Leonard R. Marino
UNIX: The Minimal Manual
Jim Moore
A Logical Language for Data and Knowledge Bases
Shamim Naqvi and Shalom Tsur
The Theory of Database Concurrency Control
Christos Papadimitriou
Algorithms for Graphics and Image Processing
Theo Pavlidis
Data Compression Methods and Theory
James A. Storer
The Elements of Artificial Intelligence
Steven lbnimoto
Computational Aspects of VLSI
Jeffrey D. Ullman
hinciples of Database and Knowledge-Base Systems, Volumes I and II
Jeffrey D. Ullman
Algorithmic Studies in Mass Storage Systems
C. K. Wong

OTHER BOOKS OF INTEREST

‘Jewels of Formal Language Theory
Art0 Salomaa
Fussy Sets, Natural Language Computations, and Risk Analysis
Kurt J. Schmucker
LISP An Interactive Approach
Stuart C. Shapiro COMPUTER SCIENCE PRESS
Principles of Database Systems, Second Edition
Jeffrey D. Ullman AN IMPRINT OF W. H. FREEMAN AND COMPANY. NEW YORK
 To John, Michael, Kevin, and Peter

Library of Congress Cataloging-in-Publication Data

Davis, Ruth E.
‘Ruth, deduction, and computation: logic and semantics for
computer science/R. E. Davis.
p. cm.
Includes index.
ISBN o-7167-8201-4
1. Programming languages (Electronic computers&Semantics.
2. Logic, Symbolic and mathematical. 3. Electronic data processing.
I. Title.
QA76.7.D386 1989 89-34922
005.13-dc20 CIP

Copyright 0 1989 by W. H. Freeman and Company

No part of this book may be reproduced by any mechanical,
photographic, or electronic process, or in the form of a phonographic
recording, nor may it be stored in a retrieval system, transmitted,
or
otherwise copied for public or private use, without written permis-
sion
from the publisher.
Printed in the United States of America

Computer Science Press

An imprint of W. H. Freeman and Company
41 Madison Avenue, New York, NY 10010
20 Beaumont Street, Oxford OX1 2NQ, England

1234567890 RRD 8 9
 To John, Michael, Kevin, and Peter

Library of Congress Cataloging-in-Publication Data

Davis, Ruth E.
‘Ruth, deduction, and computation: logic and semantics for
computer science/R. E. Davis.
p. cm.
Includes index.
ISBN o-7167-8201-4
1. Programming languages (Electronic computers&Semantics.
2. Logic, Symbolic and mathematical. 3. Electronic data processing.
I. Title.
QA76.7.D386 1989 89-34922
005.13-dc20 CIP

Copyright 0 1989 by W. H. Freeman and Company

No part of this book may be reproduced by any mechanical,
photographic, or electronic process, or in the form of a phonographic
recording, nor may it be stored in a retrieval system, transmitted,
or
otherwise copied for public or private use, without written permis-
sion
from the publisher.
Printed in the United States of America

Computer Science Press

An imprint of W. H. Freeman and Company
41 Madison Avenue, New York, NY 10010
20 Beaumont Street, Oxford OX1 2NQ, England

1234567890 RRD 8 9
 ... Contents Contents
Vlll ix

3.2 !IYuth 49 5.5 With 175

3.2.1 Remarks 53 5.5.1 Some background on lattices 175
Exercises 56 Exercises 187
3.3 Deduction 57 55.2 The Models 187
Exercise 63 5.6 Completeness 193
3.4 Computation 64 Exercises 198
3.4.1 Resolution 64 5.6.1 The h-D-calculus 198
Exercises 81 Exercises 216
3.4.2 Logic Programming 82 5.6.2 Another look at I and J 216
3.4.2.1 Horn Clauses 83 5.6.3 Completeness Results 218
3.4.2.2 An Example 84 Exercises 220
3.4.2.3 Declarative Semantics 85
3.4.2.4 Procedural Semantics 86 Appendix A
3.4.2.5 Nondeterminism 89 SELECTED ANSWERS AND WORKED EXAMPLES 223
3.4.2.6 Invertibility 90 A.1 Preliminary Definitions 223
3.4.2.7 Computing with Partially Defined Objects 92 A.2 Propositional Logic 224
3.4.2.8 Logic Programming vs. Prolog 95 A.3 Predicate Calculus 225
Exercises 96 A.4 Elementary Number Theory 248
3.5 Completeness Results 97 A.5 Lambda Calculus 248

Bibliography 257
4 ELEMENTARY NUMBER THEORY 105
4.1 The Language 105
Index 261
4.2 ‘Ruth 106
4.2.1 Intuitive Number Theory (INT) 107
Exercises 116
4.3 Deduction 117
4.3.1 Consistency of FZVY 122
4.4 Undecidability and Incompleteness 124
4.4.1 Extending Undecidability Results 139
4.4.2 Incompleteness of Fn/l 142
Exercises 144

5 LAMBDA CALCULUS 147

5.1 Introduction 147
5.2 The Language 151
5.3 Deductive System 155
Exercises 166
5.4 Computation 167
Exercise 174

..-. iiaihh- _._..

-- ..--.--_ .-.-_.
 ... Contents Contents
Vlll ix

3.2 !IYuth 49 5.5 With 175

3.2.1 Remarks 53 5.5.1 Some background on lattices 175
Exercises 56 Exercises 187
3.3 Deduction 57 55.2 The Models 187
Exercise 63 5.6 Completeness 193
3.4 Computation 64 Exercises 198
3.4.1 Resolution 64 5.6.1 The h-D-calculus 198
Exercises 81 Exercises 216
3.4.2 Logic Programming 82 5.6.2 Another look at I and J 216
3.4.2.1 Horn Clauses 83 5.6.3 Completeness Results 218
3.4.2.2 An Example 84 Exercises 220
3.4.2.3 Declarative Semantics 85
3.4.2.4 Procedural Semantics 86 Appendix A
3.4.2.5 Nondeterminism 89 SELECTED ANSWERS AND WORKED EXAMPLES 223
3.4.2.6 Invertibility 90 A.1 Preliminary Definitions 223
3.4.2.7 Computing with Partially Defined Objects 92 A.2 Propositional Logic 224
3.4.2.8 Logic Programming vs. Prolog 95 A.3 Predicate Calculus 225
Exercises 96 A.4 Elementary Number Theory 248
3.5 Completeness Results 97 A.5 Lambda Calculus 248

Bibliography 257
4 ELEMENTARY NUMBER THEORY 105
4.1 The Language 105
Index 261
4.2 ‘Ruth 106
4.2.1 Intuitive Number Theory (INT) 107
Exercises 116
4.3 Deduction 117
4.3.1 Consistency of FZVY 122
4.4 Undecidability and Incompleteness 124
4.4.1 Extending Undecidability Results 139
4.4.2 Incompleteness of Fn/l 142
Exercises 144

5 LAMBDA CALCULUS 147

5.1 Introduction 147
5.2 The Language 151
5.3 Deductive System 155
Exercises 166
5.4 Computation 167
Exercise 174

..-. iiaihh- _._..

-- ..--.--_ .-.-_.
PREFACE

I always worked with programming languages because it seemed to

me that until you could understand those, you really couldn’t
understand computers. Understanding them doesn’t really mean
only being able to use them. A lot of people can use them without
understanding them.
Christopher Strachey

The study of programming languages is central to the study of com-

puter science. When studying a programming language, one must
learn both the syntax and the semantics of the language in order to
understand it, rather than simply be able to use it. Everyone knows
and agrees upon what is meant by the syntax of a language, but
what do we mean by “semantics” ? What one requires of a descrip-
tion of the semantics of a language is dependent upon one’s point of
view. A compiler writer is happy with a description of the opera-
tional semantics of a language; someone interested in proving prop-
erties of a given program might prefer to use the axiomatic
semantics of the language. A language designer, being as concerned
with what can be said as with how to say it, may prefer a denota-
PREFACE

I always worked with programming languages because it seemed to

me that until you could understand those, you really couldn’t
understand computers. Understanding them doesn’t really mean
only being able to use them. A lot of people can use them without
understanding them.
Christopher Strachey

The study of programming languages is central to the study of com-

puter science. When studying a programming language, one must
learn both the syntax and the semantics of the language in order to
understand it, rather than simply be able to use it. Everyone knows
and agrees upon what is meant by the syntax of a language, but
what do we mean by “semantics” ? What one requires of a descrip-
tion of the semantics of a language is dependent upon one’s point of
view. A compiler writer is happy with a description of the opera-
tional semantics of a language; someone interested in proving prop-
erties of a given program might prefer to use the axiomatic
semantics of the language. A language designer, being as concerned
with what can be said as with how to say it, may prefer a denota-
 Preface

tional (mathematical) description. The denotational description is mined, but the choice of one move to make is not. Nondeterministic
also most useful when attempting to prove the equivalance of dif- computation is discussed in Section 3.4.2 on logic programming.)
ferent implementations of the same language. We will investigate four languages (formalisms): Propositional
We could argue the advantages and drawbacks of each approach, Logic, Predicate Calculus, Elementary Number Theory, and
or we can realize that each has its place and see what we can learn Lambda Calculus. In each case, the syntax of the language is
about the relationships between these descriptive devices. described first; then we examine separately the truth, deduction,
Because a solid foundation in mathematical logic is also critical to and computation aspects of the language; and finally, discover the
the study of computer science, I have combined an introduction to relationships among these concepts.
formal logic with the study of three approaches to semantics of In Propositional Logic we will find that truth, deduction, and
languages. In particular, I chose to investigate the concepts of truth computation are equivalent and that our computation scheme pro-
(denotational semantics), deduction (axiomatic semantics), and vides a decision procedure for provability in the theory. The model
computation (operational semantics) as they apply to mathematical theory of Propositional Logic is explicated by means of truth tables,
logic and computer science formalisms (i.e., languages) and to and a minimal (in number of axioms) formal theory is presented.
observe how the relationships among these ideas vary as the For computation, the process of complementary literal elimination
expressive power of the formalism increases. (CLE), or ground resolution, which provides a decision procedure
The investigation of each formalism begins with a description of for the theory, is described. An alternative computation scheme is
the syntax of the language. This is simply a description of the sym- offered that is simply a method of generating a truth table for a
bols allowed and the rules by which the symbols may be combined given formula, establishing its truth, and therefore (by complete-
to form legal sentences of the language. ness) its provability.
Truth is an intuitive notion. Since the symbols of a formalism (lan- For Predicate Calculus our computation scheme is expanded to
guage) have no intrinsic meaning, we speak of truth in an interpre- general resolution. An alternative computation scheme for a subset
tation of a formalism. An interpretation gives meaning to the of Predicate Calculus known as Horn clauses is also discussed. This
symbols and sentences of a language by mapping them onto objects is the basis for many logic programming systems, such as Prolog and
in the domain of interpretation. For example, the French map the its descendants. Again, we can show the equivalence of our notions
sequence of symbols “pomme ” onto the object we call “apple.” of truth, deduction, and computation. However, we cannot decide
Deduction involves the formal manipulation of symbols accord- whether or not our computation will yield any results. If a sentence
ing to prescribed rules. A deductive system, or formal theory, speci- is provable, then computation will terminate successfully, and if
fies the rules by which derivations can be made. These derivations computation terminates successfully, then the sentence is prov-
are strictly syntactic, so they have no intrinsic meaning. In particu- able. We cannot, however, put any bound on our computation; we
lar, proof is a syntactic notion; its relationship with truth depends have no way of knowing whether or not the program will
upon the formal theory being investigated. terminate.
Computation is also syntactic in nature and closely related to In Elementary Number Theory we look more closely at what we
deduction. Computation suggests an algorithm; we are interested mean by an algorithm (effectively computable function) and prove
not only in what the legal moves are, but in how we might execute that the formal system is undecidable. We also establish the undeci-
them. Thus a computational scheme must provide some informa- dability of Predicate Calculus (which we had left hanging after
tion to control the use of the rules for rewriting symbols. showing that the procedure we had described did not provide a
(Note: We are concerned with deterministic computation only. decision procedure). We sketch Godel’s incompleteness results,
Thus we always know what to do next. Nondeterministic computa- exhibiting a sentence that is true under the standard interpretation
tion is closer yet to deduction-a set of possible next moves is deter- but not provable.

buL.- -._. _ ._.. l-----.--._.-... --_ -l_-_l.lll_- _-_.__..----

 Preface

tional (mathematical) description. The denotational description is mined, but the choice of one move to make is not. Nondeterministic
also most useful when attempting to prove the equivalance of dif- computation is discussed in Section 3.4.2 on logic programming.)
ferent implementations of the same language. We will investigate four languages (formalisms): Propositional
We could argue the advantages and drawbacks of each approach, Logic, Predicate Calculus, Elementary Number Theory, and
or we can realize that each has its place and see what we can learn Lambda Calculus. In each case, the syntax of the language is
about the relationships between these descriptive devices. described first; then we examine separately the truth, deduction,
Because a solid foundation in mathematical logic is also critical to and computation aspects of the language; and finally, discover the
the study of computer science, I have combined an introduction to relationships among these concepts.
formal logic with the study of three approaches to semantics of In Propositional Logic we will find that truth, deduction, and
languages. In particular, I chose to investigate the concepts of truth computation are equivalent and that our computation scheme pro-
(denotational semantics), deduction (axiomatic semantics), and vides a decision procedure for provability in the theory. The model
computation (operational semantics) as they apply to mathematical theory of Propositional Logic is explicated by means of truth tables,
logic and computer science formalisms (i.e., languages) and to and a minimal (in number of axioms) formal theory is presented.
observe how the relationships among these ideas vary as the For computation, the process of complementary literal elimination
expressive power of the formalism increases. (CLE), or ground resolution, which provides a decision procedure
The investigation of each formalism begins with a description of for the theory, is described. An alternative computation scheme is
the syntax of the language. This is simply a description of the sym- offered that is simply a method of generating a truth table for a
bols allowed and the rules by which the symbols may be combined given formula, establishing its truth, and therefore (by complete-
to form legal sentences of the language. ness) its provability.
Truth is an intuitive notion. Since the symbols of a formalism (lan- For Predicate Calculus our computation scheme is expanded to
guage) have no intrinsic meaning, we speak of truth in an interpre- general resolution. An alternative computation scheme for a subset
tation of a formalism. An interpretation gives meaning to the of Predicate Calculus known as Horn clauses is also discussed. This
symbols and sentences of a language by mapping them onto objects is the basis for many logic programming systems, such as Prolog and
in the domain of interpretation. For example, the French map the its descendants. Again, we can show the equivalence of our notions
sequence of symbols “pomme ” onto the object we call “apple.” of truth, deduction, and computation. However, we cannot decide
Deduction involves the formal manipulation of symbols accord- whether or not our computation will yield any results. If a sentence
ing to prescribed rules. A deductive system, or formal theory, speci- is provable, then computation will terminate successfully, and if
fies the rules by which derivations can be made. These derivations computation terminates successfully, then the sentence is prov-
are strictly syntactic, so they have no intrinsic meaning. In particu- able. We cannot, however, put any bound on our computation; we
lar, proof is a syntactic notion; its relationship with truth depends have no way of knowing whether or not the program will
upon the formal theory being investigated. terminate.
Computation is also syntactic in nature and closely related to In Elementary Number Theory we look more closely at what we
deduction. Computation suggests an algorithm; we are interested mean by an algorithm (effectively computable function) and prove
not only in what the legal moves are, but in how we might execute that the formal system is undecidable. We also establish the undeci-
them. Thus a computational scheme must provide some informa- dability of Predicate Calculus (which we had left hanging after
tion to control the use of the rules for rewriting symbols. showing that the procedure we had described did not provide a
(Note: We are concerned with deterministic computation only. decision procedure). We sketch Godel’s incompleteness results,
Thus we always know what to do next. Nondeterministic computa- exhibiting a sentence that is true under the standard interpretation
tion is closer yet to deduction-a set of possible next moves is deter- but not provable.

buL.- -._. _ ._.. l-----.--._.-... --_ -l_-_l.lll_- _-_.__..----

xiv Preface Preface xv

Thus in moving from Propositional Logic through Predicate Cal- Elliott Mendelson, for the formal theories of .& 232, and T5Vc and
culus to Elementary Number Theory, we begin with decidability Christopher Wadsworth, for his work on the lambda calculus.
and completeness, first lose decidability, and finally also lose I am also indebted to over six hundred graduate students who
completeness. suffered through various versions of this text over the past several
When studying the X-calculus we run into a few surprises. Intui- years. Many provided comments and occasionally some construc-
tively, we expect to be able to equate termination of computation tive criticism. Donna Potter did an excellent job of providing
with the ability to find a normal form This idea of computation detailed criticism on an early draft.
takes the reduction rules (inference rules) of the formal system and I am grateful to Barbara Friedman of Computer Science Press for
uses them as computational rules. A normal form is reached when her helpfulness and understanding when medical complications
no more reductions are possible. This might lead us to believe that interfered with the timely completion of the manuscript; and I am
terms without normal forms should be different (semantically dis- most grateful to Fred Schlichting for keeping complications from
tinguishable) from those with normal forms. However, this turns turning to tragedy.
out not to be the case. I also wish to thank John R. Allen, and Michael, Kevin, and Peter
This unexpected fact caused C. Wadsworth to look more closely Davis-Allen for putting up with each other and with my absence in
at the idea of computation. A more adequate analysis of computa- the early hours of the morning for so very, very long.
tion involves the concept of a generalized normal form (head nor-
mal form), which gives much more satisfying results with respect to
the models.
We exhibit several examples of terms that are equal in the models
but cannot be proved so in the formalism; thus, the X-calculus is
also incomplete. However, if we restrict our attention to terms with
normal forms, we can establish the equivalence of truth, deduction,
and computation. We also show that if we allow approximate normal
forms, we get completeness in the limit, That is, every term is the
limit of its approximate normal forms, so we can compute the
approximate normal forms of a term to get a closer and closer
approximation of the term itself.
The formalisms presented are arranged in the order of their
expressive power. Anything that can be expressed in one formalism
can be expressed in the one following it. This is one reason that once
we lost decidability and completeness because of statements formal-
izable within the system, they were gone forever.

Acknowledgments
First and foremost I must acknowledge John R. Allen, who pointed
me toward this subject in the first place, insisting that it was prereq-
uisite material for anyone who wanted to pursue study in computer
science. I owe an obvious debt to all my sources, but most notably:

-- -yrq.. _.-._l‘ ----ll.-^------~

xiv Preface Preface xv

Thus in moving from Propositional Logic through Predicate Cal- Elliott Mendelson, for the formal theories of .& 232, and T5Vc and
culus to Elementary Number Theory, we begin with decidability Christopher Wadsworth, for his work on the lambda calculus.
and completeness, first lose decidability, and finally also lose I am also indebted to over six hundred graduate students who
completeness. suffered through various versions of this text over the past several
When studying the X-calculus we run into a few surprises. Intui- years. Many provided comments and occasionally some construc-
tively, we expect to be able to equate termination of computation tive criticism. Donna Potter did an excellent job of providing
with the ability to find a normal form This idea of computation detailed criticism on an early draft.
takes the reduction rules (inference rules) of the formal system and I am grateful to Barbara Friedman of Computer Science Press for
uses them as computational rules. A normal form is reached when her helpfulness and understanding when medical complications
no more reductions are possible. This might lead us to believe that interfered with the timely completion of the manuscript; and I am
terms without normal forms should be different (semantically dis- most grateful to Fred Schlichting for keeping complications from
tinguishable) from those with normal forms. However, this turns turning to tragedy.
out not to be the case. I also wish to thank John R. Allen, and Michael, Kevin, and Peter
This unexpected fact caused C. Wadsworth to look more closely Davis-Allen for putting up with each other and with my absence in
at the idea of computation. A more adequate analysis of computa- the early hours of the morning for so very, very long.
tion involves the concept of a generalized normal form (head nor-
mal form), which gives much more satisfying results with respect to
the models.
We exhibit several examples of terms that are equal in the models
but cannot be proved so in the formalism; thus, the X-calculus is
also incomplete. However, if we restrict our attention to terms with
normal forms, we can establish the equivalence of truth, deduction,
and computation. We also show that if we allow approximate normal
forms, we get completeness in the limit, That is, every term is the
limit of its approximate normal forms, so we can compute the
approximate normal forms of a term to get a closer and closer
approximation of the term itself.
The formalisms presented are arranged in the order of their
expressive power. Anything that can be expressed in one formalism
can be expressed in the one following it. This is one reason that once
we lost decidability and completeness because of statements formal-
izable within the system, they were gone forever.

Acknowledgments
First and foremost I must acknowledge John R. Allen, who pointed
me toward this subject in the first place, insisting that it was prereq-
uisite material for anyone who wanted to pursue study in computer
science. I owe an obvious debt to all my sources, but most notably:

-- -yrq.. _.-._l‘ ----ll.-^------~

 CHAPTER 1

PRELIMINARY
DEFINITIONS

We define here some of the terminology that is used in each of the

following chapters. One of the most important distinctions we must
make is between the notions of truth and proof Most people are fa-
miliar with an informal style of proof, which is best characterized as a
convincing argument. The point of the proof is to convince someone
that a statement is true. (Some people are more easily convinced than
others.) However, we shall be concerned with a different, more formal
type of proof. Rather than concern ourselves with what can be shown
to be true, we simply wish to investigate what follows from what. The
definition of “follows from”, as well as what we have to begin with,
changes from one theory to the next.
A formal theory can be intuitively understood as defining a game
- providing us with a language and rules by which it can be manip-
ulated. Something is provable if it is derivable within the rules of the
game; changing games changes what is provable. There is no intrin-
sic relationship between truth and proof - the first is semantic, the
second syntactic.
Before we define what is meant by a deductive system, or formal
theory, we need to know what it means for a set to be countable.

1
-- ~~__I_ ML---.. ._.-------_. .____i
 CHAPTER 1

PRELIMINARY
DEFINITIONS

We define here some of the terminology that is used in each of the

following chapters. One of the most important distinctions we must
make is between the notions of truth and proof Most people are fa-
miliar with an informal style of proof, which is best characterized as a
convincing argument. The point of the proof is to convince someone
that a statement is true. (Some people are more easily convinced than
others.) However, we shall be concerned with a different, more formal
type of proof. Rather than concern ourselves with what can be shown
to be true, we simply wish to investigate what follows from what. The
definition of “follows from”, as well as what we have to begin with,
changes from one theory to the next.
A formal theory can be intuitively understood as defining a game
- providing us with a language and rules by which it can be manip-
ulated. Something is provable if it is derivable within the rules of the
game; changing games changes what is provable. There is no intrin-
sic relationship between truth and proof - the first is semantic, the
second syntactic.
Before we define what is meant by a deductive system, or formal
theory, we need to know what it means for a set to be countable.

1
-- ~~__I_ ML---.. ._.-------_. .____i
2 Preliminary Definitions 3

DEFINITION: A set is denumerable if it can be put into a one- DEFINITION: An instance of a statement form or schema is a wff
to-one correspondence with the positive integers - in other words, obtained from the statement form by substitution. In particular, we
if there is a way to list (or enumerate) the elements such that each can substitute wffs for all metavariables (statement letters in the state-
element will eventually be listed, but the entire set can never be com- ment form), all occurrences of the same metavariable being replaced
pletely listed (there are always more elements). A set is countable if by the same wff. n
it is either finite or denumerable. n
For example, Theory X1:
DEFINITION: A formal theory 7 consists of the following:
1. The symbols are: A, B, C, 1 , +, ), and (. The symbols 7 and
+ can be read “not” and “implies”, respectively.
1. A countable set, of symbols. (A finite sequence of symbols of 7 is
called an expression of 7.) 2. A, B, and C are wffs, and if P and Q are any wffs, then (+) and
(P -+ Q) are also wffs.
2. A subset of the expressions, called the well-formed formulas
(abbreviated w$s) of 7. The wffs are the legal sentences of the 3. The axioms are:
theory.
b(P-4 1
3. A subset of the wffs called the axioms of 7.
W'-+(Q+P)>>

4. A finite set of relations RI,. . . , & on wffs, called rules of in- 4. There is one rule of inference:
ference. For each .& there is a unique positive integer j such
that for every j wffs and each wff A one can effectively decide Q is a direct consequence of (+Q-+P))
whether the given j wffs are in the relation & to A; if so, A is
called a direct consequence of the given wffs by virtue of &. Theory X has an infinite number of wffs and an Smite number of
For example, the rule modus ponens is a relation on three wffs, axioms. Metavariables P and Q were used to describe both of these
A, A + B, and B, by which B is a direct consequence of A and sets. Some wffs of X are:
A + B. n

A
The set of axioms of a theory is often infinite. It can usually be
specified by providing a finite set, of axiom schemata.

DEFINITION: A schema is a statement form; it provides a tem-

plate showing the form of a wff while leaving some pieces unspecified.
We indicate these unspecified pieces using metavariables. (The pre-
fix meta- indicates that we are not referring to variables that may be
symbols in the theory; we are talking about the theory. Metavariables
stand for wffs in the theory). In the modus ponens example above, A
and B are metavariables. Metavariables are distinguished from sym-
bols of the theory by font. n ‘Remember that a formal theory is just a game. Do not expect that it must
reflect reality, or that something must be true simply because it is provable.
2 Preliminary Definitions 3

DEFINITION: A set is denumerable if it can be put into a one- DEFINITION: An instance of a statement form or schema is a wff
to-one correspondence with the positive integers - in other words, obtained from the statement form by substitution. In particular, we
if there is a way to list (or enumerate) the elements such that each can substitute wffs for all metavariables (statement letters in the state-
element will eventually be listed, but the entire set can never be com- ment form), all occurrences of the same metavariable being replaced
pletely listed (there are always more elements). A set is countable if by the same wff. n
it is either finite or denumerable. n
For example, Theory X1:
DEFINITION: A formal theory 7 consists of the following:
1. The symbols are: A, B, C, 1 , +, ), and (. The symbols 7 and
+ can be read “not” and “implies”, respectively.
1. A countable set, of symbols. (A finite sequence of symbols of 7 is
called an expression of 7.) 2. A, B, and C are wffs, and if P and Q are any wffs, then (+) and
(P -+ Q) are also wffs.
2. A subset of the expressions, called the well-formed formulas
(abbreviated w$s) of 7. The wffs are the legal sentences of the 3. The axioms are:
theory.
b(P-4 1
3. A subset of the wffs called the axioms of 7.
W'-+(Q+P)>>

4. A finite set of relations RI,. . . , & on wffs, called rules of in- 4. There is one rule of inference:
ference. For each .& there is a unique positive integer j such
that for every j wffs and each wff A one can effectively decide Q is a direct consequence of (+Q-+P))
whether the given j wffs are in the relation & to A; if so, A is
called a direct consequence of the given wffs by virtue of &. Theory X has an infinite number of wffs and an Smite number of
For example, the rule modus ponens is a relation on three wffs, axioms. Metavariables P and Q were used to describe both of these
A, A + B, and B, by which B is a direct consequence of A and sets. Some wffs of X are:
A + B. n

A
The set of axioms of a theory is often infinite. It can usually be
specified by providing a finite set, of axiom schemata.

DEFINITION: A schema is a statement form; it provides a tem-

plate showing the form of a wff while leaving some pieces unspecified.
We indicate these unspecified pieces using metavariables. (The pre-
fix meta- indicates that we are not referring to variables that may be
symbols in the theory; we are talking about the theory. Metavariables
stand for wffs in the theory). In the modus ponens example above, A
and B are metavariables. Metavariables are distinguished from sym-
bols of the theory by font. n ‘Remember that a formal theory is just a game. Do not expect that it must
reflect reality, or that something must be true simply because it is provable.
 Preliminary Definitions
4 5

Some of the axioms of X are: 1. If a is a subset of Sa and s1 t- A, then $J I- A. (That is, we

can always add extra hypotheses, simply ignoring them in the
(-@+B)) derivation.) This property is sometimes called monotonicity.3

2. a I- A if and only if there is a finite subset s2 of 2 such that

s2 t- A. (That is, since a derivation must be finite, it must require
only a finite number of hypotheses.) This property is sometimes
(-((B-(lA>>-t((-C>-t(Bj(lA))))) called compactness.
Each of the axioms above is constructed by using the templates
provided by the axiom schemata. P and Q in the schemata are con- 3. If s2 I- A, and for each wff B in S2, $J k B, then $1 l- A.
sistently replaced by wffs. (If every wff in s2 can be derived from the set of wffs $1, then
anything that can be derived from s2 can also be derived from
DEFINITION: Let 3 be a set of wffs, and let P be a wff in the La.1
formal theory 7. We say that P is deducible jkom 8 in 7 (denoted
2 I-T P) if there exists a finite sequence of wffs PI,. . . , P, such that Note that a proof is a sequence of wffs with certain restrictions, and
P, = P and for 1 5 i 2 n, Pi is either an axiom, a formula in 5 (called that it is a proof of the last statement in the sequence. Therefore, if
a hypothesis), or a direct consequence of previous Pi’s by virtue of one we consider only the first n elements of the sequence, you have a proof
of the rules of inference. n of the nth statement.
We will investigate the relationships among truth, deduction, and
DEFINITION: The sequence of Pi’s is called a derivation of P computation in each of the languages that we study. When we speak
from 5, or a proof of P from 8. n about a theory, we are interested in various metatheoretic properties.
(Again, meta- indicates that we are speaking about the theory rather
DEFINITION: If P is deducible from the empty set, we write 11. P, than in the theory - wffs being the sentences of the language in the
and say that P is a theorem or P is provable in 7.2 n theory.)

When we say that P is deducible from the empty set, it sounds as DEFINITION: An interpretation supplies a meaning for each of
if we can deduce it from nothing. However, checking the definition of the symbols of a formal theory such that any wff can be understood
deducible from, we see that we always have the set of axioms of the as a statement that is either true or false in the interpretation. n
theory as an implicit set of assumptions.
For example, consider the following proof of A in Theory X: The above is only an intuitive definition of interpretation. We will
be more precise about what is required to specify an interpretation
1. (+A+A)) by the first axiom
when we discuss specific languages.
2. A by 1 and the rule of inference
3All the theories we study in this book are monotonic. There is a wealth of
The following properties of deducibility are consequences of the def-
research on non-monotonic logics. In such systems it is possible that the addition
inition of deducible from: of new information (axioms) can invalidate proofs that were possible prior to the
Let a and SJ be sets of wffs, and A a wff, then addition. Such logics are very useful in modelling common sense reasoning, in
which, for example, we modify conclusions derived from general rules by learning
*We will not continue to use the subscript ‘T on the symbol I-I; the theory we the exceptions. However interesting, these logics are outside the scope of this text.
mean should be clear from context. We refer the interested reader to [l], particularly to [11,27].
 Preliminary Definitions
4 5

Some of the axioms of X are: 1. If a is a subset of Sa and s1 t- A, then $J I- A. (That is, we

can always add extra hypotheses, simply ignoring them in the
(-@+B)) derivation.) This property is sometimes called monotonicity.3

2. a I- A if and only if there is a finite subset s2 of 2 such that

s2 t- A. (That is, since a derivation must be finite, it must require
only a finite number of hypotheses.) This property is sometimes
(-((B-(lA>>-t((-C>-t(Bj(lA))))) called compactness.
Each of the axioms above is constructed by using the templates
provided by the axiom schemata. P and Q in the schemata are con- 3. If s2 I- A, and for each wff B in S2, $J k B, then $1 l- A.
sistently replaced by wffs. (If every wff in s2 can be derived from the set of wffs $1, then
anything that can be derived from s2 can also be derived from
DEFINITION: Let 3 be a set of wffs, and let P be a wff in the La.1
formal theory 7. We say that P is deducible jkom 8 in 7 (denoted
2 I-T P) if there exists a finite sequence of wffs PI,. . . , P, such that Note that a proof is a sequence of wffs with certain restrictions, and
P, = P and for 1 5 i 2 n, Pi is either an axiom, a formula in 5 (called that it is a proof of the last statement in the sequence. Therefore, if
a hypothesis), or a direct consequence of previous Pi’s by virtue of one we consider only the first n elements of the sequence, you have a proof
of the rules of inference. n of the nth statement.
We will investigate the relationships among truth, deduction, and
DEFINITION: The sequence of Pi’s is called a derivation of P computation in each of the languages that we study. When we speak
from 5, or a proof of P from 8. n about a theory, we are interested in various metatheoretic properties.
(Again, meta- indicates that we are speaking about the theory rather
DEFINITION: If P is deducible from the empty set, we write 11. P, than in the theory - wffs being the sentences of the language in the
and say that P is a theorem or P is provable in 7.2 n theory.)

When we say that P is deducible from the empty set, it sounds as DEFINITION: An interpretation supplies a meaning for each of
if we can deduce it from nothing. However, checking the definition of the symbols of a formal theory such that any wff can be understood
deducible from, we see that we always have the set of axioms of the as a statement that is either true or false in the interpretation. n
theory as an implicit set of assumptions.
For example, consider the following proof of A in Theory X: The above is only an intuitive definition of interpretation. We will
be more precise about what is required to specify an interpretation
1. (+A+A)) by the first axiom
when we discuss specific languages.
2. A by 1 and the rule of inference
3All the theories we study in this book are monotonic. There is a wealth of
The following properties of deducibility are consequences of the def-
research on non-monotonic logics. In such systems it is possible that the addition
inition of deducible from: of new information (axioms) can invalidate proofs that were possible prior to the
Let a and SJ be sets of wffs, and A a wff, then addition. Such logics are very useful in modelling common sense reasoning, in
which, for example, we modify conclusions derived from general rules by learning
*We will not continue to use the subscript ‘T on the symbol I-I; the theory we the exceptions. However interesting, these logics are outside the scope of this text.
mean should be clear from context. We refer the interested reader to [l], particularly to [11,27].
6 Preliminary Definitions 7

DEFINITION: An interpretation is a model for a set of wffs 5 if concerned with the relationships between truth and computation, and
every wff in 8 is true in the interpretation. We say that an interpreta- between deduction and computation. So, for each theory we discuss,
tion provides a model for a formal theory ‘7 if it provides a model for we will define precisely what we mean by computation.
the set of theorems of 7. l
DEFINITION: A computation method is complete if for every
Completeness is a metatheoretic concept used to relate truth and sentence S the algorithm will terminate on input S in a finite amount
provability in a formal theory; soundness is the complementary prop of time, indicating whether or not S is true in all interpretations. m
erty.
Besides general completeness, we can speak of completeness with
DEFINITION: A theory is complete if every sentence that is true respect to a particular class of interpretations, narrowing the focus of
in all interpretations is provable in the theory. n
our discussion from truth in all interpretations to truth in some cho-
sen interpretations. This idea surfaces in Elementary Number Theory
DEFINITION: A theory is sound if every provable sentence is true (ENI), where we show that the theory is incomplete with respect to
in all interpretations. m the standard model (Intuitive Number Theory or 2X7), and is useful
in the study of programming formalisms.
Our example theory X is complete. It is possible to prove every Two other metatheoretic concepts of interest are the decidability
sentence that is true in all interpretations. In fact, it is possible to and consistency of a theory.
prove every sentence in theory X. We gave a sample proof of the wff
consisting of the single proposition letter A. Since the axioms and the DEFINITION: A formal theory is decidable if there exists an ef-
rule of inference are schemata we could use the form of that proof, fective procedure that will determine, for any sentence of the theory,
instantiated with different specific wffs to prove any wff. For example, whether or not that sentence is provable in the theory. n
we will prove (A+A).
We will occasionally wish to speak of the decidability of a property
1. (+(A+A)-+(A-+A))) by the first axiom
in a more general sense.
2. (A-+A) by 1 and the rule of inference
DEFINITION: Any property is said to be decidable if there exists
But theory X is not sound. If a wff A is true in all interpretations, an effective procedure (i.e., terminating algorithm) that will determine
then certainly its negation, TA, must be false. However, both of these whether or not the property holds. m
sentences are provable in X. Thus, something that is not true in all
interpretations is provable; therefore, X is unsound. This definition is purposefully vague definition since we may wish
A theory in which we can prove false results is not of much use, so to apply it to properties of symbols (is x a symbol of 7?), strings
we will insist that every theory we use be sound. It is easy to legislate, of symbols (is x a wff?), sets of wffs (is P deducible from set x?),
but we cannot ignore enforcement; although it is obvious that we want sequences of wffs (is x a proof?), and so on.
our theories to be sound, it is not always so obvious how we can prove
that they are sound. Soundness must be demonstrated, not simply DEFINITION: A theory is consistent if it contains no wff such
assumed. that both the wff and its negation are provable. m
If a theory is both sound and complete, we will say truth and de-
duction are equivalent and write “truth w deduction.” We will also be An inconsistent theory is usually not very useful, alt\ugh there are

-I

_ll.----.--- -
_._. --- .- -_
___-._..
_---

--
6 Preliminary Definitions 7

DEFINITION: An interpretation is a model for a set of wffs 5 if concerned with the relationships between truth and computation, and
every wff in 8 is true in the interpretation. We say that an interpreta- between deduction and computation. So, for each theory we discuss,
tion provides a model for a formal theory ‘7 if it provides a model for we will define precisely what we mean by computation.
the set of theorems of 7. l
DEFINITION: A computation method is complete if for every
Completeness is a metatheoretic concept used to relate truth and sentence S the algorithm will terminate on input S in a finite amount
provability in a formal theory; soundness is the complementary prop of time, indicating whether or not S is true in all interpretations. m
erty.
Besides general completeness, we can speak of completeness with
DEFINITION: A theory is complete if every sentence that is true respect to a particular class of interpretations, narrowing the focus of
in all interpretations is provable in the theory. n
our discussion from truth in all interpretations to truth in some cho-
sen interpretations. This idea surfaces in Elementary Number Theory
DEFINITION: A theory is sound if every provable sentence is true (ENI), where we show that the theory is incomplete with respect to
in all interpretations. m the standard model (Intuitive Number Theory or 2X7), and is useful
in the study of programming formalisms.
Our example theory X is complete. It is possible to prove every Two other metatheoretic concepts of interest are the decidability
sentence that is true in all interpretations. In fact, it is possible to and consistency of a theory.
prove every sentence in theory X. We gave a sample proof of the wff
consisting of the single proposition letter A. Since the axioms and the DEFINITION: A formal theory is decidable if there exists an ef-
rule of inference are schemata we could use the form of that proof, fective procedure that will determine, for any sentence of the theory,
instantiated with different specific wffs to prove any wff. For example, whether or not that sentence is provable in the theory. n
we will prove (A+A).
We will occasionally wish to speak of the decidability of a property
1. (+(A+A)-+(A-+A))) by the first axiom
in a more general sense.
2. (A-+A) by 1 and the rule of inference
DEFINITION: Any property is said to be decidable if there exists
But theory X is not sound. If a wff A is true in all interpretations, an effective procedure (i.e., terminating algorithm) that will determine
then certainly its negation, TA, must be false. However, both of these whether or not the property holds. m
sentences are provable in X. Thus, something that is not true in all
interpretations is provable; therefore, X is unsound. This definition is purposefully vague definition since we may wish
A theory in which we can prove false results is not of much use, so to apply it to properties of symbols (is x a symbol of 7?), strings
we will insist that every theory we use be sound. It is easy to legislate, of symbols (is x a wff?), sets of wffs (is P deducible from set x?),
but we cannot ignore enforcement; although it is obvious that we want sequences of wffs (is x a proof?), and so on.
our theories to be sound, it is not always so obvious how we can prove
that they are sound. Soundness must be demonstrated, not simply DEFINITION: A theory is consistent if it contains no wff such
assumed. that both the wff and its negation are provable. m
If a theory is both sound and complete, we will say truth and de-
duction are equivalent and write “truth w deduction.” We will also be An inconsistent theory is usually not very useful, alt\ugh there are

-I

_ll.----.--- -
_._. --- .- -_
___-._..
_---

--
8 preliminary Definitions
9

exceptions, particularly in the field of artificial intelligence; hence, all Well, no one said that every interpretation has to make sense. How-
the theories we study in detail will be consistent. However, it is impor- ever, interpretations are our means of connecting the purely syntactic
tant to realize that not all theories are consistent; consistency must be world of formal theories with the world of semantics.
demonstrated, particularly when dealing with programming languages.
We defined above what is meant by a decidable theory - that there
Exercises 1 .O
exists an effective procedure capable of determining whether or not any
given well-formed formula is a theorem. So if, within a given theory, 1. Review the definitions of sound, complete, consistent, and decid-
we can find some rule by which we can always tell whether a given wff able. Is it conceivable that a theory might be:
is provable or not, our theory is decidable. The example theory X is
decidable. In fact, the decision procedure is very simple: given any a. sound, but not complete?
wff, terminate immediately with the answer “Yes, it is provable.” b. sound, but not consistent?
We also know that X is inconsistent. We proved (A + A) as an
c. consistent, but not sound?
example above. We know that (l(A -+ A)) is also provable, as is any
axiom. Thus we have exhibited a wff such that both it and its negation d. complete, but not decidable?
are provable, implying that the theory is not consistent. (Aside: Look e. consistent and complete, but not sound?
again at the definitions of “sound” and “consistent.” The terms are
closely related, but not the same. For example, one might be able to Give a rationale or an example or both justifying each answer.
conceive of a theory in which one could prove only false statements.
Such a theory would be consistent but not sound.)
Wffs are strictly syntactic objects; they have meaning only when
an interpretation is given for the symbols. However, our definition of
consistency has assumed something about the meaning of 1. We have
said that it means negation. We shall fix the meaning of some of our
symbols - for example, the logical connectives 1 and + with which
we build complicated wffs from the atomic wffs. The meaning assigned
to these symbols will be defined precisely in Section 2.2.
The meaning attached to wffs of a theory is determined by the
meaning of the connectives, together with an interpretation. To supply
an interpretation for our theory X, we must specify the domain of
interpretation, and a relation over that domain for each of the symbols
A, B, and C. For example:
domain = integers
A means “2 is an even number”
B means “3 is an even number”
C means “21 is divisible by 7”
Under this interpretation, the wff (A --t (B ---t (X))) means: If 2 is
an even number, then (3 is an even number implies that it is not the
case that 21 is divisible by 7).
8 preliminary Definitions
9

exceptions, particularly in the field of artificial intelligence; hence, all Well, no one said that every interpretation has to make sense. How-
the theories we study in detail will be consistent. However, it is impor- ever, interpretations are our means of connecting the purely syntactic
tant to realize that not all theories are consistent; consistency must be world of formal theories with the world of semantics.
demonstrated, particularly when dealing with programming languages.
We defined above what is meant by a decidable theory - that there
Exercises 1 .O
exists an effective procedure capable of determining whether or not any
given well-formed formula is a theorem. So if, within a given theory, 1. Review the definitions of sound, complete, consistent, and decid-
we can find some rule by which we can always tell whether a given wff able. Is it conceivable that a theory might be:
is provable or not, our theory is decidable. The example theory X is
decidable. In fact, the decision procedure is very simple: given any a. sound, but not complete?
wff, terminate immediately with the answer “Yes, it is provable.” b. sound, but not consistent?
We also know that X is inconsistent. We proved (A + A) as an
c. consistent, but not sound?
example above. We know that (l(A -+ A)) is also provable, as is any
axiom. Thus we have exhibited a wff such that both it and its negation d. complete, but not decidable?
are provable, implying that the theory is not consistent. (Aside: Look e. consistent and complete, but not sound?
again at the definitions of “sound” and “consistent.” The terms are
closely related, but not the same. For example, one might be able to Give a rationale or an example or both justifying each answer.
conceive of a theory in which one could prove only false statements.
Such a theory would be consistent but not sound.)
Wffs are strictly syntactic objects; they have meaning only when
an interpretation is given for the symbols. However, our definition of
consistency has assumed something about the meaning of 1. We have
said that it means negation. We shall fix the meaning of some of our
symbols - for example, the logical connectives 1 and + with which
we build complicated wffs from the atomic wffs. The meaning assigned
to these symbols will be defined precisely in Section 2.2.
The meaning attached to wffs of a theory is determined by the
meaning of the connectives, together with an interpretation. To supply
an interpretation for our theory X, we must specify the domain of
interpretation, and a relation over that domain for each of the symbols
A, B, and C. For example:
domain = integers
A means “2 is an even number”
B means “3 is an even number”
C means “21 is divisible by 7”
Under this interpretation, the wff (A --t (B ---t (X))) means: If 2 is
an even number, then (3 is an even number implies that it is not the
case that 21 is divisible by 7).
 CHAPTER 2

PROPOSITIONAL LOGIC

2.1 The Language

The first two requirements for a description of a formal theory are

precisely what we need to describe the syntax of the language. Thus,
we describe the language of Propositional Logic by giving the symbols
of the language and the rules for generating well-formed formulas of
the language.

1. A countable set of symbols:

a. Proposition (or predicate) letters: A, B, C, . . . , Z and the

same letters subscripted by positive integers: Al, Br, . . . ,
Z1; A2, B2, . . . , Z2; As, . . .
b. Logical connectives: --t and 1.
c. Two auxiliary symbols: ) and (.

2. The class of well-formed formulas consists of:

a. Atomic formulas - that is, the proposition letters A, B, . . .

b. Compound Formulas - if P and Q are wffs, then (+) and
(P-t&) are also wffs. (For ease in reading, we can anticipate
the interpretation of 1 as negation and + as implication,

JiilmLe-__-- . 11 -__I
-~
 CHAPTER 2

PROPOSITIONAL LOGIC

2.1 The Language

The first two requirements for a description of a formal theory are

precisely what we need to describe the syntax of the language. Thus,
we describe the language of Propositional Logic by giving the symbols
of the language and the rules for generating well-formed formulas of
the language.

1. A countable set of symbols:

a. Proposition (or predicate) letters: A, B, C, . . . , Z and the

same letters subscripted by positive integers: Al, Br, . . . ,
Z1; A2, B2, . . . , Z2; As, . . .
b. Logical connectives: --t and 1.
c. Two auxiliary symbols: ) and (.

2. The class of well-formed formulas consists of:

a. Atomic formulas - that is, the proposition letters A, B, . . .

b. Compound Formulas - if P and Q are wffs, then (+) and
(P-t&) are also wffs. (For ease in reading, we can anticipate
the interpretation of 1 as negation and + as implication,

JiilmLe-__-- . 11 -__I
-~
12 Propositional Logic 2.1 The Language 13

thus 4’ may be read as “not P” and P+Q as “if P then Q” to express all formulas. When we consider computation we will use a
or “P implies Q.“) different set of sufficient connectives than the one we chose for our
formal system of Propositional Logic.

DEFINITION: If P is a wff in which no proposition letters other DEFINITION: The scope of a connective is the wff or wffs to which
than those in the set {Ai, . . . , An} occur, then P is said to be a formula the connective applies. n
in the proposition letters Al, . . . , A,. n

To make complicated expressions easier to read, we will adopt some

For example, each of the following is a formula in the proposition conventions for elimination of parentheses. First, we will omit the
letters A, B, C, D: outer pair of parentheses; thus, (A-B) may be written as A-B, and
((A+(~C))-tWB)) (-A) as 1A.
Second, when a form contains only one binary connective, paren-
(+WB+(C--+D))))
theses are omitted by association to the left. Thus,
((b+W(C+W
A+B-+A--+C stands for (( (A-+B)-tA)+C)
(D-D)
We use the symbols A, B, C, . . . , 2, subscripted as needed, for Last, the connectives are ordered from greatest to least scope, or
metavariables - that is, they stand for arbitrary wffs as opposed to lowest to highest precedence as follows: s-, +, V, A, 1. Parentheses
the individual proposition letters. We rarely need to be concerned with are eliminated according to the rule that 5 has the greatest scope and
specific proposition letters of the language. Most results are stated us- 1 has the least, or, if you prefer to think in terms of precedence, 1 has
ing metavariables and therefore apply to any wff in which the metavari- the highest precedence and G has the least. That is, 7 applies to the
ables have been consistently replaced. smallest wff following it, then A connects the smallest wffs surrounding
We introduce three more propositional connectives as abbreviations: it, then V, +, and finally E.
For example, TA V B G A-B V 4’ A A stands for the fully paren-
(A A II) for c-w+(+9)) thesized formula (((TA) V B) z (A+(B V ((4’) A A)))). Omitting the
(AV B) for ((lA)+B) unnecessary parentheses in the wffs given above,
(A E B) for ((A-W A (B-4)
((A+(X))+(D+B)) becomes A--C+(D+B)
Any formula can be seen as a logical function; the proposition let- (l(A-+(B+(C+D)))) becomes T(A*(B-~(C+D)))
ters appearing in the formula are the arguments, and the value of the (((lA)+B)-+(C+D)) becomes -A-+B+(C-tD)
function for a given set of arguments is either true or false. Any for- (D-+D) becomes D+D
mula written using any of the five connectives can be reexpressed as a
formula using only the connectives 1 and +. Exercises 2.1

DEFINITION: 8 is a suficient set of connectiues if any logical 1. Remove all redundant pairs of parentheses (i.e., those that are
function can be expressed in terms of connectives in the set S. n not needed due to the conventions described above) from each of
the following wffs:
The set (1, -} is a sufficient set of connectives. Sometimes it is
more convenient to use one set of connectives than another. It makes no a. (((A A bW-0)) V (A A (~-+A))))+(C A WW
difference, formally, as long as the set of connectives chosen is sufficient b. (((B+(C-+A))-B)-(C-A))
12 Propositional Logic 2.1 The Language 13

thus 4’ may be read as “not P” and P+Q as “if P then Q” to express all formulas. When we consider computation we will use a
or “P implies Q.“) different set of sufficient connectives than the one we chose for our
formal system of Propositional Logic.

DEFINITION: If P is a wff in which no proposition letters other DEFINITION: The scope of a connective is the wff or wffs to which
than those in the set {Ai, . . . , An} occur, then P is said to be a formula the connective applies. n
in the proposition letters Al, . . . , A,. n

To make complicated expressions easier to read, we will adopt some

For example, each of the following is a formula in the proposition conventions for elimination of parentheses. First, we will omit the
letters A, B, C, D: outer pair of parentheses; thus, (A-B) may be written as A-B, and
((A+(~C))-tWB)) (-A) as 1A.
Second, when a form contains only one binary connective, paren-
(+WB+(C--+D))))
theses are omitted by association to the left. Thus,
((b+W(C+W
A+B-+A--+C stands for (( (A-+B)-tA)+C)
(D-D)
We use the symbols A, B, C, . . . , 2, subscripted as needed, for Last, the connectives are ordered from greatest to least scope, or
metavariables - that is, they stand for arbitrary wffs as opposed to lowest to highest precedence as follows: s-, +, V, A, 1. Parentheses
the individual proposition letters. We rarely need to be concerned with are eliminated according to the rule that 5 has the greatest scope and
specific proposition letters of the language. Most results are stated us- 1 has the least, or, if you prefer to think in terms of precedence, 1 has
ing metavariables and therefore apply to any wff in which the metavari- the highest precedence and G has the least. That is, 7 applies to the
ables have been consistently replaced. smallest wff following it, then A connects the smallest wffs surrounding
We introduce three more propositional connectives as abbreviations: it, then V, +, and finally E.
For example, TA V B G A-B V 4’ A A stands for the fully paren-
(A A II) for c-w+(+9)) thesized formula (((TA) V B) z (A+(B V ((4’) A A)))). Omitting the
(AV B) for ((lA)+B) unnecessary parentheses in the wffs given above,
(A E B) for ((A-W A (B-4)
((A+(X))+(D+B)) becomes A--C+(D+B)
Any formula can be seen as a logical function; the proposition let- (l(A-+(B+(C+D)))) becomes T(A*(B-~(C+D)))
ters appearing in the formula are the arguments, and the value of the (((lA)+B)-+(C+D)) becomes -A-+B+(C-tD)
function for a given set of arguments is either true or false. Any for- (D-+D) becomes D+D
mula written using any of the five connectives can be reexpressed as a
formula using only the connectives 1 and +. Exercises 2.1

DEFINITION: 8 is a suficient set of connectiues if any logical 1. Remove all redundant pairs of parentheses (i.e., those that are
function can be expressed in terms of connectives in the set S. n not needed due to the conventions described above) from each of
the following wffs:
The set (1, -} is a sufficient set of connectives. Sometimes it is
more convenient to use one set of connectives than another. It makes no a. (((A A bW-0)) V (A A (~-+A))))+(C A WW
difference, formally, as long as the set of connectives chosen is sufficient b. (((B+(C-+A))-B)-(C-A))
 Propositional Logic 2.2 Truth 15
14

2. Fully parenthesize each of the following wffs according to the con- A B A-B
ventions described above. tt t
tf f
ft t
a. A A lB+C V A A B--A-d7 A TB ff t n

b. B--+C-+A-+B-+C-+A
In informal usage, “if A then B” implies some sort of causal rela-
tionship. However, the (perhaps surprising) truth table entries of lines
3 and 4 can be explained by considering the following:
2.2 Truth
AAB-B (If A and B, then B)
For Propositional Logic, an interpretation consists of an assignment Certainly this implication should be true in all situations; $ we know
of a truth value (t or f, for “true” and “false”, respectively) to each two things, then we must certainly know the second of them.
proposition letter. The interpretation of a wff is determined by the
truth values of the proposition letters in the formula. By giving truth f + t is t: Suppose that A is false and B is true. Then the con-
tables to define the meaning of the connectives, we can describe how junction A A B is false, and the conclusion B is true.
the truth value of a wff is built from the truth values of its component
wffs. In constructing a truth table we exhibit all possible interpreta- f --f f is t: If B is false, we have that false implies false is true.
tions of a wff. Each line of the truth table assigns a unique combination
Intuitively, we simply do not want to consider cases in which the hy-
of truth values to the proposition letters of the wff, thus determining
pothesis of an implication is false as counterexamples to the truth of
an interpretation; we consider all possible interpretations by listing all
the implication. We define the implication to be true unless we can
possible assignments. In the following truth tables we use metavari-
show it to be false by exhibiting a case in which the hypothesis is
ables to emphasize that the definitions apply to arbitrarily complex
satisfied but the conclusion does not follow.
wffs.
If we know the meaning of the proposition letters, we can always
determine the meaning of a complicated wff by repeated use of the
DEFINITION: The logical connective “not,” written 1, represents
truth tables given above. Note that an interpretation is free to change
negation; TA is true iff A is false, as shown by the following truth only the meaning of individual proposition letters. The meaning of the
table: connectives is fixed by the truth tables just given.
The truth tables for the connectives A, V, and = can be determined
A 1A by considering the formulas they are abbreviations of. Given sepa-
t f rately, they are:
f t n
A B AVB A B AAB A B AGB
tt t tt t tt t
DEFINITION: + is a conditional; A-tB is read “A implies B” or “if
tf t tf f tf f
A then B”. A is the antecedent (or hypothesis) of the implication,
ft t ft f ft f
B is the consequent (or conclusion). The meaning of A-tB is given
ff f ff f ff t
by the following truth table:

--. ------- -- -.-__-..---_-.l .-_I_-- I._--^-

 Propositional Logic 2.2 Truth 15
14

2. Fully parenthesize each of the following wffs according to the con- A B A-B
ventions described above. tt t
tf f
ft t
a. A A lB+C V A A B--A-d7 A TB ff t n

b. B--+C-+A-+B-+C-+A
In informal usage, “if A then B” implies some sort of causal rela-
tionship. However, the (perhaps surprising) truth table entries of lines
3 and 4 can be explained by considering the following:
2.2 Truth
AAB-B (If A and B, then B)
For Propositional Logic, an interpretation consists of an assignment Certainly this implication should be true in all situations; $ we know
of a truth value (t or f, for “true” and “false”, respectively) to each two things, then we must certainly know the second of them.
proposition letter. The interpretation of a wff is determined by the
truth values of the proposition letters in the formula. By giving truth f + t is t: Suppose that A is false and B is true. Then the con-
tables to define the meaning of the connectives, we can describe how junction A A B is false, and the conclusion B is true.
the truth value of a wff is built from the truth values of its component
wffs. In constructing a truth table we exhibit all possible interpreta- f --f f is t: If B is false, we have that false implies false is true.
tions of a wff. Each line of the truth table assigns a unique combination
Intuitively, we simply do not want to consider cases in which the hy-
of truth values to the proposition letters of the wff, thus determining
pothesis of an implication is false as counterexamples to the truth of
an interpretation; we consider all possible interpretations by listing all
the implication. We define the implication to be true unless we can
possible assignments. In the following truth tables we use metavari-
show it to be false by exhibiting a case in which the hypothesis is
ables to emphasize that the definitions apply to arbitrarily complex
satisfied but the conclusion does not follow.
wffs.
If we know the meaning of the proposition letters, we can always
determine the meaning of a complicated wff by repeated use of the
DEFINITION: The logical connective “not,” written 1, represents
truth tables given above. Note that an interpretation is free to change
negation; TA is true iff A is false, as shown by the following truth only the meaning of individual proposition letters. The meaning of the
table: connectives is fixed by the truth tables just given.
The truth tables for the connectives A, V, and = can be determined
A 1A by considering the formulas they are abbreviations of. Given sepa-
t f rately, they are:
f t n
A B AVB A B AAB A B AGB
tt t tt t tt t
DEFINITION: + is a conditional; A-tB is read “A implies B” or “if
tf t tf f tf f
A then B”. A is the antecedent (or hypothesis) of the implication,
ft t ft f ft f
B is the consequent (or conclusion). The meaning of A-tB is given
ff f ff f ff t
by the following truth table:

--. ------- -- -.-__-..---_-.l .-_I_-- I._--^-

16 Propositional Logic 2.2 Truth 17

Remember, each row of a truth table corresponds to one possible For example, A V 47 G B has the following truth table:
interpretation. If a wff P is made up of the proposition letters Pr, . . . ,
A B C -2 (Ah17) (AVS=B)
P, and propositional connectives, there are 2n possible interpretations
t t t f t t
assigning truth values to the proposition letters.
t t f t t t
The propositional connectives are logical functions in that they take t ft f t f
truth-valued arguments and define a truth-valued result. The connec- t f f t t f
tives defined above are unary (-), or binary (A, V, +, z), but we can ftt f f f
easily define a function of any number of arguments by supplying a f t f t t t
truth table that specifies the value of the function for all possible val- fft f f t
ues of its arguments, as we used truth tables to define the connectives f f f t t f
1, +, V, A, and G. If a function takes n arguments, then to define the
values of the function under all possible combinations of values of its
Reading across each line with a t in the last column, it is easy to
arguments will require a truth table with 2n lines.
see that A V 4’ G B is equivalent to
We can demonstrate that (1, A, V} is a sufficient set of connectives
by showing that any logical function f(zi, . . . , 2,) can be expressed in
terms of 1, A, and V.

DEFINITION: The principal connective of a wff is the one that

THEOREM 2.2.1 (1, A, V} is a sufficient set of connectives. is applied last in constructing the wff (i.e., the one that has the great-
est scope). n
Proof: Consider an arbitrary logical function f(zr, . . . ,Xn) of n
arguments. We represent each argument by a proposition letter Pi Truth tables can be abbreviated by writing the wff once, listing
and construct a truth table exhibiting the value of the function under the possible assignments of truth values under each propositionletter,
all possible combinations of values of the arguments. (If there are n and then, step by step, entering the truth values of each component
proposition letters, then there are 2n lines in the truth table.) sentence under the principal connective of that component.
For each line of the truth table that contains the value t as value of Thus, we have the alternative form shown below for the truth table
the function, construct the conjunction above (we have numbered the columns to indicate the order in which
they are filled in).
P’1 A P’2 A.. . A P:,
(A v (1 c>>= B
t t f t t t
t t t f t t
t t f t f f
t t t f f f
f f f t f t
f t t f t t
f f f t t f
f t t f f f
1 3 2 1 4 1
16 Propositional Logic 2.2 Truth 17

Remember, each row of a truth table corresponds to one possible For example, A V 47 G B has the following truth table:
interpretation. If a wff P is made up of the proposition letters Pr, . . . ,
A B C -2 (Ah17) (AVS=B)
P, and propositional connectives, there are 2n possible interpretations
t t t f t t
assigning truth values to the proposition letters.
t t f t t t
The propositional connectives are logical functions in that they take t ft f t f
truth-valued arguments and define a truth-valued result. The connec- t f f t t f
tives defined above are unary (-), or binary (A, V, +, z), but we can ftt f f f
easily define a function of any number of arguments by supplying a f t f t t t
truth table that specifies the value of the function for all possible val- fft f f t
ues of its arguments, as we used truth tables to define the connectives f f f t t f
1, +, V, A, and G. If a function takes n arguments, then to define the
values of the function under all possible combinations of values of its
Reading across each line with a t in the last column, it is easy to
arguments will require a truth table with 2n lines.
see that A V 4’ G B is equivalent to
We can demonstrate that (1, A, V} is a sufficient set of connectives
by showing that any logical function f(zi, . . . , 2,) can be expressed in
terms of 1, A, and V.

DEFINITION: The principal connective of a wff is the one that

THEOREM 2.2.1 (1, A, V} is a sufficient set of connectives. is applied last in constructing the wff (i.e., the one that has the great-
est scope). n
Proof: Consider an arbitrary logical function f(zr, . . . ,Xn) of n
arguments. We represent each argument by a proposition letter Pi Truth tables can be abbreviated by writing the wff once, listing
and construct a truth table exhibiting the value of the function under the possible assignments of truth values under each propositionletter,
all possible combinations of values of the arguments. (If there are n and then, step by step, entering the truth values of each component
proposition letters, then there are 2n lines in the truth table.) sentence under the principal connective of that component.
For each line of the truth table that contains the value t as value of Thus, we have the alternative form shown below for the truth table
the function, construct the conjunction above (we have numbered the columns to indicate the order in which
they are filled in).
P’1 A P’2 A.. . A P:,
(A v (1 c>>= B
t t f t t t
t t t f t t
t t f t f f
t t t f f f
f f f t f t
f t t f t t
f f f t t f
f t t f f f
1 3 2 1 4 1
18 Propositional Logic

DEFINITION: A wff that is always true, no matter what the truth 2. Determine whether each of the following wffs is unsatisfiable, a
values of its statement letters (proposition letters) may be, is called a tautology, both, or neither.
tautology. Such a statement P is said to be logically valid or true
in all interpretations, is denoted + P, and is read “P is logically a. (lA-+B)+(7B+A)
valid.” n b. (A-+B)-+(B--+A)
c. l(A-+(B+A))
DEFINITION: A wff is satisfiable if it is true in some interpreta-
d. l(A-+B)-+A
tion (i.e., there exists an assignment of truth values to the proposition
letters that makes the wff true). n e. Y(A+B)+Y(B+A)
3. Show that V, and A are commutative and associative.
If P is true in all models, then TP is false in all models; that is,
there does not exist an interpretation in which TP is true. a. commutativity:
AVBGBVA
DEFINITION: P is unsatisfiable if TP is logically valid. An un- AABEBAA
satisfiable wff is also called a contradiction. n
b. associativity:
The validity of any statement P is decidable. + P if and only if AV(BVC)G(AVB)VC
the column under P in its truth table (or under its main connective in AA(BAC)E(AAB)AC
the abbreviated form) contains only t’s - that is, P is true for every
possible interpretation (assignment of truth values to its proposition
letters). 4. Is + associative?

5. Show that (1, A} is a sufficient set of connectives.

LEMMA 2.2.2 (P A (P+Q))+Q is a tautology.
Proof: 6. A NAND B, written A 1 B, is defined to be ‘(A A B). Show that
; 7 P+Q PA (P-+&j (PA P’-+QWQ {NAND} is a sufficient set of connectives.
t t t 7. A NOR B, written A t B, is defined to be -(A V B). Is {NOR}
tf f f t
a sufficient set of connectives?
ft t f t
ff t f t
2.3 Deduction

The first two requirements for a formal theory - the set of symbols
Exercises 2.2 and the class of well-formed formulas we presented in Section 2.1 -
describe the syntax of the language. It remains to describe the axioms
1. Demonstrate, via the truth table, that the wff and rules of inference of our formal theory for Propositional Logic
(which we shall call C).
((A-(B-C>>j((AjB>j(AjC))) There are many ways of axiomatizing Propositional Logic. The
following set of axioms is a minimal one. If A, B, and C are any wffs
is a tautology. of L, then the following are axioms of C.
18 Propositional Logic

DEFINITION: A wff that is always true, no matter what the truth 2. Determine whether each of the following wffs is unsatisfiable, a
values of its statement letters (proposition letters) may be, is called a tautology, both, or neither.
tautology. Such a statement P is said to be logically valid or true
in all interpretations, is denoted + P, and is read “P is logically a. (lA-+B)+(7B+A)
valid.” n b. (A-+B)-+(B--+A)
c. l(A-+(B+A))
DEFINITION: A wff is satisfiable if it is true in some interpreta-
d. l(A-+B)-+A
tion (i.e., there exists an assignment of truth values to the proposition
letters that makes the wff true). n e. Y(A+B)+Y(B+A)
3. Show that V, and A are commutative and associative.
If P is true in all models, then TP is false in all models; that is,
there does not exist an interpretation in which TP is true. a. commutativity:
AVBGBVA
DEFINITION: P is unsatisfiable if TP is logically valid. An un- AABEBAA
satisfiable wff is also called a contradiction. n
b. associativity:
The validity of any statement P is decidable. + P if and only if AV(BVC)G(AVB)VC
the column under P in its truth table (or under its main connective in AA(BAC)E(AAB)AC
the abbreviated form) contains only t’s - that is, P is true for every
possible interpretation (assignment of truth values to its proposition
letters). 4. Is + associative?

5. Show that (1, A} is a sufficient set of connectives.

LEMMA 2.2.2 (P A (P+Q))+Q is a tautology.
Proof: 6. A NAND B, written A 1 B, is defined to be ‘(A A B). Show that
; 7 P+Q PA (P-+&j (PA P’-+QWQ {NAND} is a sufficient set of connectives.
t t t 7. A NOR B, written A t B, is defined to be -(A V B). Is {NOR}
tf f f t
a sufficient set of connectives?
ft t f t
ff t f t
2.3 Deduction

The first two requirements for a formal theory - the set of symbols
Exercises 2.2 and the class of well-formed formulas we presented in Section 2.1 -
describe the syntax of the language. It remains to describe the axioms
1. Demonstrate, via the truth table, that the wff and rules of inference of our formal theory for Propositional Logic
(which we shall call C).
((A-(B-C>>j((AjB>j(AjC))) There are many ways of axiomatizing Propositional Logic. The
following set of axioms is a minimal one. If A, B, and C are any wffs
is a tautology. of L, then the following are axioms of C.
 20 Propositional Logic 2.3 Deduction 21

Axioms: L1 (A+(B+A)) Clearly, from the definition of proof given in Section 1, we have the
L2 ((A+(B+C))-((A-B)+(A+C))) following lemma.
L3 (((~B)j(-A>>j(((lB>jA)-tB))
LEMMA 2.3.1 l- A, for any axiom A.
Rule of inference: There is only one rule of inference in fZ. It is
modus ponens (MP): B is a direct consequence of A and A+B, We must be careful to distinguish two uses of the word proof. A proof
written in the theory, as defined previously, is a sequence of wffs of the theory
A,A-+BkB such that the last is the statement we wish to prove and each preceding
wff in the sequence is an axiom or follows by a rule of inference from
In other words, assuming A and A-B, we can derive B. We shall two previous members of the sequence. As is the case with the axioms,
abbreviate modus ponens by the initials MP. we can represent an infinite number of actual theorems and proofs by
using metavariables to establish a proof schema. Any instance of a
The axioms are presented as schemata expressed in metavariables theorem schema is a theorem of the theory, just as any instance of an
over ,!Z. They actually describe an infinite number of axioms. Any axiom schema is an axiom.
instance of one of these statement forms is an axiom. We also give proofs of statements about the theory. Lemma 2.3.1 is
such a statement, as are the Deduction Theorem (Theorem 2.3.3) and
DEFINITION: An instance of a statement form is obtained by re- the completeness theorems. These proofs are constructed by the usual
placing each metavariable with a wff consistently throughout the state- mathematical methods and are written in English supplemented by
ment form (that is, each occurrence of the same metavariable must be mathematical symbols and symbols from the theory we are discussing.
replaced by the same wff). n These are actually metatheorems, that is, theorems about the theory
rather than in the theory. However, we will refer only to theorems,
For example, lemmas, etc., without continuing the distinction by using the terms
((C+D)+A-+(C-+D))) metatheorems, metalemmas, and so on. (Note: Each theorem in a
theory must be a well-formed formula of the theory.)
is an instance of axiom Ll, with (C-D) for A, and TA for B.
We shall prove the following lemma in the theory L. That is, we
We have chosen to define our theory using only the connectives 1 will construct a sequence of wffs in L such that the last element of
and +. We have also chosen to use a minimal set of axioms; none the sequence is the statement we wish to prove, and every other wff
of the three can be proven from the other two, and together they are in the sequence is an (instance of an) axiom or follows from previous
sufficient to prove any true statement of Propositional Logic (i.e., fZ is members of the sequence by MP.
complete, as we shall see in Section 2.5). Providing many more axioms,
even though they might be redundant, would make it easier to prove LEMMA 2.3.2 A-A
theorems, so what is the benefit of using a minimal set? Although it
may be more difficult to prove theorems in the theory, a minimal axiom Proof:
set makes it easier to prove theorems about the theory (metatheorems). 1. (A-((A+A)+A))-+((A+(A+A))+(A+A))
For example, suppose we wish to prove something about all theorems an instance of axiom schema L2, with A for A,
of a theory (such as the fact that they are all tautologies). Every (A+A) for B, and A for C
theorem must be derivable from the axioms and rules of inference.
Thus, the smaller the set of axioms and rules of inference the fewer the 2. A+((A+A)+A) axiom schema Llwith A for A,
cases we must consider. (A+A) for B

-----
--- ______.- -- .._. -llll_.-_~. . I”_;

_--
 20 Propositional Logic 2.3 Deduction 21

Axioms: L1 (A+(B+A)) Clearly, from the definition of proof given in Section 1, we have the
L2 ((A+(B+C))-((A-B)+(A+C))) following lemma.
L3 (((~B)j(-A>>j(((lB>jA)-tB))
LEMMA 2.3.1 l- A, for any axiom A.
Rule of inference: There is only one rule of inference in fZ. It is
modus ponens (MP): B is a direct consequence of A and A+B, We must be careful to distinguish two uses of the word proof. A proof
written in the theory, as defined previously, is a sequence of wffs of the theory
A,A-+BkB such that the last is the statement we wish to prove and each preceding
wff in the sequence is an axiom or follows by a rule of inference from
In other words, assuming A and A-B, we can derive B. We shall two previous members of the sequence. As is the case with the axioms,
abbreviate modus ponens by the initials MP. we can represent an infinite number of actual theorems and proofs by
using metavariables to establish a proof schema. Any instance of a
The axioms are presented as schemata expressed in metavariables theorem schema is a theorem of the theory, just as any instance of an
over ,!Z. They actually describe an infinite number of axioms. Any axiom schema is an axiom.
instance of one of these statement forms is an axiom. We also give proofs of statements about the theory. Lemma 2.3.1 is
such a statement, as are the Deduction Theorem (Theorem 2.3.3) and
DEFINITION: An instance of a statement form is obtained by re- the completeness theorems. These proofs are constructed by the usual
placing each metavariable with a wff consistently throughout the state- mathematical methods and are written in English supplemented by
ment form (that is, each occurrence of the same metavariable must be mathematical symbols and symbols from the theory we are discussing.
replaced by the same wff). n These are actually metatheorems, that is, theorems about the theory
rather than in the theory. However, we will refer only to theorems,
For example, lemmas, etc., without continuing the distinction by using the terms
((C+D)+A-+(C-+D))) metatheorems, metalemmas, and so on. (Note: Each theorem in a
theory must be a well-formed formula of the theory.)
is an instance of axiom Ll, with (C-D) for A, and TA for B.
We shall prove the following lemma in the theory L. That is, we
We have chosen to define our theory using only the connectives 1 will construct a sequence of wffs in L such that the last element of
and +. We have also chosen to use a minimal set of axioms; none the sequence is the statement we wish to prove, and every other wff
of the three can be proven from the other two, and together they are in the sequence is an (instance of an) axiom or follows from previous
sufficient to prove any true statement of Propositional Logic (i.e., fZ is members of the sequence by MP.
complete, as we shall see in Section 2.5). Providing many more axioms,
even though they might be redundant, would make it easier to prove LEMMA 2.3.2 A-A
theorems, so what is the benefit of using a minimal set? Although it
may be more difficult to prove theorems in the theory, a minimal axiom Proof:
set makes it easier to prove theorems about the theory (metatheorems). 1. (A-((A+A)+A))-+((A+(A+A))+(A+A))
For example, suppose we wish to prove something about all theorems an instance of axiom schema L2, with A for A,
of a theory (such as the fact that they are all tautologies). Every (A+A) for B, and A for C
theorem must be derivable from the axioms and rules of inference.
Thus, the smaller the set of axioms and rules of inference the fewer the 2. A+((A+A)+A) axiom schema Llwith A for A,
cases we must consider. (A+A) for B

-----
--- ______.- -- .._. -llll_.-_~. . I”_;

_--
22 Propositional Logic 2.3 Deduction 23

3. (A-+(A-+A))+(A--+A) by MP on 1 and 2 b. If & is A:

axiom schema Ll, A for A, A for B I- A-B1 Lemma 2.3.2

4. A-t(A-4)
S I-A-B1 property 1 of deducibility
5. A+A by MP on 3 and 4

n Induction step. Assume the theorem is true for i < Ic, then (by the
definition of proof)
We can extend our idea of proof in two ways. First, we allow abbre-
1. BI, is an axiom, or
viations of proofs; that is, we can use any previously proved theorem
in our sequence of statements as an abbreviation for its own proof.
2. BI, is an assumption
Another helpful extension is the common mathematical practice of
proving a statement B on the assumption of some other statement A,
a. & ES, or
and then concluding that we have A implies B. In Propositional Logic
this procedure is justified by the Deduction Theorem. b. BI, is A, or

THEOREM 2.3.3 (Deduction Theorem for L) If 5 is a set of wffs, 3. & follows by MP from Bi and Bj, where i,j < Ic, and Bj is
and A and B are wffs, and ,S, A l- B, then 2 l- A-B. In particular &+Bk.
(when &’ is empty), if A t- B, then k A-+B. (If you can derive B given
the assumption A, then, with no assumptions, you can derive that A The first two cases are treated exactly as in the basis case above.
implies B.) We now consider the third case.
Proof: Let Bl, . . . . B, be a derivation of B from ,S, A; then
B, = B. We shall use induction on i to show that if it is the case that s t- A+(&-+Bk) inductive hypothesis
& A l- Bi, then S l- A+Bi, for every i, 1 5 i 5 n.
2 I-A--+Bi inductive hypothesis
Basis case: i = 1. B1 must be either an axiom or a hypothesis.
!- (A-t(BijBk))j((A-Bi)j(AjBk)) L2
1. If B1 is an axiom:
Zi’ I- ((A+Bi)+(A-+Bk)) MP
I- BI Lemma 2.3.1
t- B1+(A--+B1) Ll &’ t-A+& MP
I- A-+B1 MP on the previous two steps
8 I-A-+B1 property 1 of deducibility (Section 1) The theorem follows as the special case where i = 72: n

2. If B1 is an assumption: Now that we have the Deduction Theorem, the proof of Lemma
2.3.2 is simplified to the point of being trivial. We used Lemma 2.3.2
a. If& ES:
in proving the Deduction Theorem, but we could have simply used the
S FBI def. of “deducible from” (Section 1) proof (without the Deduction Theorem) given earlier to derive what
S k Bp(A-+Bl) Ll we needed without ever making reference to Lemma 2.3.2. Thus the
s. I-A-+Bl MP on the previous two steps following proof, although quite trivial, is not circular.

. “.-- ~I_---
22 Propositional Logic 2.3 Deduction 23

3. (A-+(A-+A))+(A--+A) by MP on 1 and 2 b. If & is A:

axiom schema Ll, A for A, A for B I- A-B1 Lemma 2.3.2

4. A-t(A-4)
S I-A-B1 property 1 of deducibility
5. A+A by MP on 3 and 4

n Induction step. Assume the theorem is true for i < Ic, then (by the
definition of proof)
We can extend our idea of proof in two ways. First, we allow abbre-
1. BI, is an axiom, or
viations of proofs; that is, we can use any previously proved theorem
in our sequence of statements as an abbreviation for its own proof.
2. BI, is an assumption
Another helpful extension is the common mathematical practice of
proving a statement B on the assumption of some other statement A,
a. & ES, or
and then concluding that we have A implies B. In Propositional Logic
this procedure is justified by the Deduction Theorem. b. BI, is A, or

THEOREM 2.3.3 (Deduction Theorem for L) If 5 is a set of wffs, 3. & follows by MP from Bi and Bj, where i,j < Ic, and Bj is
and A and B are wffs, and ,S, A l- B, then 2 l- A-B. In particular &+Bk.
(when &’ is empty), if A t- B, then k A-+B. (If you can derive B given
the assumption A, then, with no assumptions, you can derive that A The first two cases are treated exactly as in the basis case above.
implies B.) We now consider the third case.
Proof: Let Bl, . . . . B, be a derivation of B from ,S, A; then
B, = B. We shall use induction on i to show that if it is the case that s t- A+(&-+Bk) inductive hypothesis
& A l- Bi, then S l- A+Bi, for every i, 1 5 i 5 n.
2 I-A--+Bi inductive hypothesis
Basis case: i = 1. B1 must be either an axiom or a hypothesis.
!- (A-t(BijBk))j((A-Bi)j(AjBk)) L2
1. If B1 is an axiom:
Zi’ I- ((A+Bi)+(A-+Bk)) MP
I- BI Lemma 2.3.1
t- B1+(A--+B1) Ll &’ t-A+& MP
I- A-+B1 MP on the previous two steps
8 I-A-+B1 property 1 of deducibility (Section 1) The theorem follows as the special case where i = 72: n

2. If B1 is an assumption: Now that we have the Deduction Theorem, the proof of Lemma
2.3.2 is simplified to the point of being trivial. We used Lemma 2.3.2
a. If& ES:
in proving the Deduction Theorem, but we could have simply used the
S FBI def. of “deducible from” (Section 1) proof (without the Deduction Theorem) given earlier to derive what
S k Bp(A-+Bl) Ll we needed without ever making reference to Lemma 2.3.2. Thus the
s. I-A-+Bl MP on the previous two steps following proof, although quite trivial, is not circular.

. “.-- ~I_---
 Propositional Logic 2.4 Computation 25
24

LEMMA 2.3.2 A+A 5. A+B 3 and 4, Lemma 2.3.4 Thus lB--A k A-B, and,

Proof: 6. (TB+~A)+(A+B) Deduction Theorem, 1, 5

1. AI-A definition of “deducible from ”

n

2. I- A+A Deduction Theorem

n Exercises 2.3

The following lemma gives us a transitivity rule for implication. Prove each of the following theorems of C:

1.
LEMMA 2.3.4 A-+B, B-C I- A-L’
Proof: 2. A, B I- l(A+lB) (i.e., A, B k A A B)

1. A+B hypothesis 3. A G TTA (i.e., (A-tllA) A (TTA+A))

(First prove A+ ,lA, then use this result together with 1 and
2. B-d’ hypothesis
2.1
3. A hypothesis
4. (A-+B)+B-+-IA)
4. B MP 1,3
5. l(A+lB) t- B (i.e., A A B I- B)
5. c MP 2,4
6. If A, B I- C then +A -rlB) I- C (i.e., A A B k C)
Thus A-tB, B-tC, A t- C, so,
7. A I- ll(A+lB)--t~B (i.e., A I- ‘(A A B)-tlB)
6. AtB, B-C I- A-d’ Deduction Theorem, 3, 5
8. If I- P+A and I- P+B, then I- P-+T(A+YB)
n (i.e., k P+(A A B))

Lemma 2.3.5 states one direction of the equivalence of an implication 9. lA-+(A+B)

with its contrapositive.r
2.4 Computation
LEMMA 2.3.5 (lB+lA)+(A+B)
Proof: Now that we have developed the ideas of truth and deducibility (prov-
ability), we need to know how we can compute in Propositional Logic.
1. YB+TA hypothesis Then we can discuss the relationships among these ideas.
L3 Computation for Propositional Logic will be based on Complemen-
2. (TB+TA)+(-JB+A)-+B)
tary Literal Elimination (CLE), also called ground resoWion. In order
3. A-+B+A) Ll to develop this idea we introduce two new symbols, r and 0 (0 may
be read “box,” or “the empty clause”). These symbols are intended
4. (lB-+A)+B MP 1,2 to represent statements that are always true and always false, respec-
‘The contrapositive of an implication (A-B) is (-B-+-A). tively. In order to relate our language of computation with that of

iHlbb.-.. ..___ ..____.--_ .._- ..- --.-- ---.---.-

 Propositional Logic 2.4 Computation 25
24

LEMMA 2.3.2 A+A 5. A+B 3 and 4, Lemma 2.3.4 Thus lB--A k A-B, and,

Proof: 6. (TB+~A)+(A+B) Deduction Theorem, 1, 5

1. AI-A definition of “deducible from ”

n

2. I- A+A Deduction Theorem

n Exercises 2.3

The following lemma gives us a transitivity rule for implication. Prove each of the following theorems of C:

1.
LEMMA 2.3.4 A-+B, B-C I- A-L’
Proof: 2. A, B I- l(A+lB) (i.e., A, B k A A B)

1. A+B hypothesis 3. A G TTA (i.e., (A-tllA) A (TTA+A))

(First prove A+ ,lA, then use this result together with 1 and
2. B-d’ hypothesis
2.1
3. A hypothesis
4. (A-+B)+B-+-IA)
4. B MP 1,3
5. l(A+lB) t- B (i.e., A A B I- B)
5. c MP 2,4
6. If A, B I- C then +A -rlB) I- C (i.e., A A B k C)
Thus A-tB, B-tC, A t- C, so,
7. A I- ll(A+lB)--t~B (i.e., A I- ‘(A A B)-tlB)
6. AtB, B-C I- A-d’ Deduction Theorem, 3, 5
8. If I- P+A and I- P+B, then I- P-+T(A+YB)
n (i.e., k P+(A A B))

Lemma 2.3.5 states one direction of the equivalence of an implication 9. lA-+(A+B)

with its contrapositive.r
2.4 Computation
LEMMA 2.3.5 (lB+lA)+(A+B)
Proof: Now that we have developed the ideas of truth and deducibility (prov-
ability), we need to know how we can compute in Propositional Logic.
1. YB+TA hypothesis Then we can discuss the relationships among these ideas.
L3 Computation for Propositional Logic will be based on Complemen-
2. (TB+TA)+(-JB+A)-+B)
tary Literal Elimination (CLE), also called ground resoWion. In order
3. A-+B+A) Ll to develop this idea we introduce two new symbols, r and 0 (0 may
be read “box,” or “the empty clause”). These symbols are intended
4. (lB-+A)+B MP 1,2 to represent statements that are always true and always false, respec-
‘The contrapositive of an implication (A-B) is (-B-+-A). tively. In order to relate our language of computation with that of

iHlbb.-.. ..___ ..____.--_ .._- ..- --.-- ---.---.-

26 Propositional Logic 2.4 Computation 27

truth and deduction, we can consider the symbols r and Cl as abbrevi- a. -Q reduces to Q
ations for A V TA and 1A A A, respectively. (It is convenient to assume b. $QI V Q2) reduces to l&r A 7Q2
that our wffs are expressed using only the connectives V, A, and 1.
c. l(Ql A Q2) reduces to l&l V 7Q2
Clearly this is possible since we have already shown (page 16) that
{ 7, A, V } is a sufficient set of connectives.) d. v reduces to •I
e. 10 reduces to 7
DEFINITION: A wff P is in conjunctive normal form, abbrevi-
ated cnf if all of the following are satisfied: 2. The following rules eliminate occurrences of A, 7, or Cl within the
scope of any V.
1. No subformula Q of P whose principal connective is 1 contains
another occurrence of 1, or an occurrence of V, A, 7, or Cl. a. &I V (Q2 A Q3) reduces to (QI V Q2) A (Ql V Q3)
For example, the following are not in conjunctive normal form: b. V and A are commutative operations
+Q), l(P V Q), l(P A Q), T(T), and -J(D). c. Q V r reduces to 7
2. No subformula of P whose principal connective is V contains A, T, d. Q V 0 reduces to Q
or Cl. That is, no A, 7, or q occurs within the scope of any V.
(The scope of a connective is the wff(s) to which it applies.) 3. We ensure no occurrences of r or 0 within the scope of any A by:

3. Neither r nor Cl occurs within the scope of any A. n a. Q A 7 reduces to Q

b. Q A 0 reduces to 0
For example, the following well-formed formula is in conjunctive
normal form: 4. To put our formula in reduced conjunctive form we eliminate
repetitions of the same proposition letter within the scope of any
V as follows: Assume P is in conjunctive normal form. Let Q
be a subformula of P of the form PI V P2. By condition 2 of
conjunctive normal form we know PI and P2 are built up by V’s
DEFINITION: P is in reduced conjunctive normal form if, in from Pi’s and ~Pi’s, where each Pi is a proposition letter. Since
addition to satisfying the three conditions above,
V is both commutative and associative, we can rearrange the Pi’s
4. NO proposition letter appears more than once within the scope of and lpi’s so that if Q contains two occurrences of some Pk we
any V. n can assume Q contains a subformula of one of the forms:

The reduced conjunctive normal form of the example given above a. Pk V Pk, which reduces t0 Pk
is: b. -Pk V Pk, which reduces2 to r
fjCV7B)A(CV7AVB) C. +k V lpk, which reduces to lpk
Every wff P can be reduced to an equivalent formula P’ such that
For example, we can use the reductions above to transform the wff
P’ is in reduced conjunctive normal form.

1. By repeated application of the following reduction rules we get +(A-(B+C))-+((A-+B)-+(A4)))

a formula satisfying criterion 1 of the definition of reduced con- ‘Note that the introduction of 7 may require that we again make use of the
junctive normal form. elimination rules 2c and 3a above.
26 Propositional Logic 2.4 Computation 27

truth and deduction, we can consider the symbols r and Cl as abbrevi- a. -Q reduces to Q
ations for A V TA and 1A A A, respectively. (It is convenient to assume b. $QI V Q2) reduces to l&r A 7Q2
that our wffs are expressed using only the connectives V, A, and 1.
c. l(Ql A Q2) reduces to l&l V 7Q2
Clearly this is possible since we have already shown (page 16) that
{ 7, A, V } is a sufficient set of connectives.) d. v reduces to •I
e. 10 reduces to 7
DEFINITION: A wff P is in conjunctive normal form, abbrevi-
ated cnf if all of the following are satisfied: 2. The following rules eliminate occurrences of A, 7, or Cl within the
scope of any V.
1. No subformula Q of P whose principal connective is 1 contains
another occurrence of 1, or an occurrence of V, A, 7, or Cl. a. &I V (Q2 A Q3) reduces to (QI V Q2) A (Ql V Q3)
For example, the following are not in conjunctive normal form: b. V and A are commutative operations
+Q), l(P V Q), l(P A Q), T(T), and -J(D). c. Q V r reduces to 7
2. No subformula of P whose principal connective is V contains A, T, d. Q V 0 reduces to Q
or Cl. That is, no A, 7, or q occurs within the scope of any V.
(The scope of a connective is the wff(s) to which it applies.) 3. We ensure no occurrences of r or 0 within the scope of any A by:

3. Neither r nor Cl occurs within the scope of any A. n a. Q A 7 reduces to Q

b. Q A 0 reduces to 0
For example, the following well-formed formula is in conjunctive
normal form: 4. To put our formula in reduced conjunctive form we eliminate
repetitions of the same proposition letter within the scope of any
V as follows: Assume P is in conjunctive normal form. Let Q
be a subformula of P of the form PI V P2. By condition 2 of
conjunctive normal form we know PI and P2 are built up by V’s
DEFINITION: P is in reduced conjunctive normal form if, in from Pi’s and ~Pi’s, where each Pi is a proposition letter. Since
addition to satisfying the three conditions above,
V is both commutative and associative, we can rearrange the Pi’s
4. NO proposition letter appears more than once within the scope of and lpi’s so that if Q contains two occurrences of some Pk we
any V. n can assume Q contains a subformula of one of the forms:

The reduced conjunctive normal form of the example given above a. Pk V Pk, which reduces t0 Pk
is: b. -Pk V Pk, which reduces2 to r
fjCV7B)A(CV7AVB) C. +k V lpk, which reduces to lpk
Every wff P can be reduced to an equivalent formula P’ such that
For example, we can use the reductions above to transform the wff
P’ is in reduced conjunctive normal form.

1. By repeated application of the following reduction rules we get +(A-(B+C))-+((A-+B)-+(A4)))

a formula satisfying criterion 1 of the definition of reduced con- ‘Note that the introduction of 7 may require that we again make use of the
junctive normal form. elimination rules 2c and 3a above.
28 Propositional Logic 2.4 Computation 29

to its reduced conjunctive normal form. First we eliminate the impli- The distinction will become more important in the next chapter. We
cations. abbreviate “ground” by gr.
Sometimes it is convenient in the following to think of a gr sentence
+(A-+(B-tC)) v ((A-tB)+(A-+C))) as a set of gr clauses, and a gr clause as a set of literals. This will cause
++A v (B+C)) v (+A-+B) V (A+(I))) no confusion since we know that a gr sentence is always a conjunction
$+A v (‘B v C)) v (+A V B) V (‘A V C))) (A) of its elements (gr clauses), and a gr clause is a disjunction (V) of
its elements (fliterals). In the following definitions we use the form
Then we move the negations inside the parentheses. of a context-free grammar, but indicate some alternatives using set
notation rather than the strictly syntactic juxtaposition of symbols (U
-+A v (‘B V C)) A +(-A V B) V (‘A V C)) indicates set union).

Last, we eliminate double negations and superfluous parentheses.

gr-clause ::= “0” ( &literal
(TA v TB v C) A (11(1A V B) A +A V C)) ( &literal U gr-clause
grsentence ::= gr-clause
(TAVTBVC)A((-AVB)A-AAlC)
) gr-clause U grsentence
(7A~7BVC)~(7AVB)AAA-C

To describe the process of CLE on formulas in reduced conjunctive Now we are prepared to define complementary literal elimination
form, we need to define a few more terms. The following style of syntax (CLE) on formulas in reduced conjunctive normal form.
specification is called a context-free grammar. Each line represents a
rule for transforming a nonterminal symbol into a string of symbols. DEFINITION: Pk and lpk (where Pk is a proposition letter) are
A terminal symbol is placed inside quotation marks; the symbol ::= complementary literah l
is to be read “is a”; the symbol ( is to be read “or” (representing
alternative reduction rules on the same line); and juxtaposition of IC DEFINITION: If A and B are gr clauses, and Pk E A and lpk E B,
and y indicates that z is followed by y. Thus the first clause below can then the ground resolvent of A and B with respect to Pk and TPk
be read “a plus-or-minus literal is a plus-literal or a minus-literal.” is:3

*literal ::= fliteral 1 -literal (A - {pk}) u (B - {+k}>

+literal ::= proposition-letter
-literal ::= “7” propositionletter That is, the resolvent clause is the union of the two original clauses,
ground-clause ::= “0” 1 *literal with the exception that the complementary literals have been elimi-
) &literal “V” ground-clause nated. Since we consider 0 to be the empty clause, the ground resol-
groundsentence ::= ground-clause vent of Pk and lpk is Cl.
1 ground-clause “A” ground-sentence Ground resolution, or CLE, is a more general application of MP,
which states that from A and A+B, we can deduce B. This could also
The empty clause is 0. A wff in reduced conjunctive normal form is be written: Prom A and (‘A V B) (which are in reduced conjunctive
a ground sentence. The adjective “ground” refers to the fact that the
clause or sentence it precedes contains no variables. Since there are no 3A - B indicates the set difference of A and I?. It consists of those elements of
variables in Propositional Logic, all clauses and sentences are ground. A that are not in B.

- ---- -- ..-__lll_ll_- .-.-. ._--

_-~---- _-._-_ -
28 Propositional Logic 2.4 Computation 29

to its reduced conjunctive normal form. First we eliminate the impli- The distinction will become more important in the next chapter. We
cations. abbreviate “ground” by gr.
Sometimes it is convenient in the following to think of a gr sentence
+(A-+(B-tC)) v ((A-tB)+(A-+C))) as a set of gr clauses, and a gr clause as a set of literals. This will cause
++A v (B+C)) v (+A-+B) V (A+(I))) no confusion since we know that a gr sentence is always a conjunction
$+A v (‘B v C)) v (+A V B) V (‘A V C))) (A) of its elements (gr clauses), and a gr clause is a disjunction (V) of
its elements (fliterals). In the following definitions we use the form
Then we move the negations inside the parentheses. of a context-free grammar, but indicate some alternatives using set
notation rather than the strictly syntactic juxtaposition of symbols (U
-+A v (‘B V C)) A +(-A V B) V (‘A V C)) indicates set union).

Last, we eliminate double negations and superfluous parentheses.

gr-clause ::= “0” ( &literal
(TA v TB v C) A (11(1A V B) A +A V C)) ( &literal U gr-clause
grsentence ::= gr-clause
(TAVTBVC)A((-AVB)A-AAlC)
) gr-clause U grsentence
(7A~7BVC)~(7AVB)AAA-C

To describe the process of CLE on formulas in reduced conjunctive Now we are prepared to define complementary literal elimination
form, we need to define a few more terms. The following style of syntax (CLE) on formulas in reduced conjunctive normal form.
specification is called a context-free grammar. Each line represents a
rule for transforming a nonterminal symbol into a string of symbols. DEFINITION: Pk and lpk (where Pk is a proposition letter) are
A terminal symbol is placed inside quotation marks; the symbol ::= complementary literah l
is to be read “is a”; the symbol ( is to be read “or” (representing
alternative reduction rules on the same line); and juxtaposition of IC DEFINITION: If A and B are gr clauses, and Pk E A and lpk E B,
and y indicates that z is followed by y. Thus the first clause below can then the ground resolvent of A and B with respect to Pk and TPk
be read “a plus-or-minus literal is a plus-literal or a minus-literal.” is:3

*literal ::= fliteral 1 -literal (A - {pk}) u (B - {+k}>

+literal ::= proposition-letter
-literal ::= “7” propositionletter That is, the resolvent clause is the union of the two original clauses,
ground-clause ::= “0” 1 *literal with the exception that the complementary literals have been elimi-
) &literal “V” ground-clause nated. Since we consider 0 to be the empty clause, the ground resol-
groundsentence ::= ground-clause vent of Pk and lpk is Cl.
1 ground-clause “A” ground-sentence Ground resolution, or CLE, is a more general application of MP,
which states that from A and A+B, we can deduce B. This could also
The empty clause is 0. A wff in reduced conjunctive normal form is be written: Prom A and (‘A V B) (which are in reduced conjunctive
a ground sentence. The adjective “ground” refers to the fact that the
clause or sentence it precedes contains no variables. Since there are no 3A - B indicates the set difference of A and I?. It consists of those elements of
variables in Propositional Logic, all clauses and sentences are ground. A that are not in B.

- ---- -- ..-__lll_ll_- .-.-. ._--

_-~---- _-._-_ -
Other documents randomly have
different content
 698 BREWSTER GENEALOGY Grant of East Windsor, born
there, ^December 12, 1797; died at South Windsor, Conn., February
21, 1886. They resided in the Grant homestead at East Windsor Hill,
Conn., the L of which was built in 1697, and the main part added in
1757, by Frederick Grant's ancestors. Children, horn at East Windsor
Hill: i Frederick William^, b. Aug. 29, 1843 ; d. June 12, 1856. ii
Anna Stoughton. b. Jan. 14, 1846; d. at So. Windsor, Aug. 31, 1871;
m. at E. Windsor, Nov. 22, 1870, Frederick Oliver Newbery, son of
Horace Newbery of S. Windsor, b. there, July 3, 1845. Ees. South
Windsor, Chn., h. there: 1 Anna" Newbery, b. Aug. 27, 1871; m. at
So. Windsor, Sept. 13, 1893, Louis E. Hart of New Britain, Conn., and
res. there, iii EoswELL, b. July 28, 1848; m. at Cambridge, Mass.,
Sept. 29, 1874, Annie Elizabeth, dau. of Captain Thomas Mickell of
Cambridge, b. at Hong Kong, China, Apr. 20, 1850. He atd. college at
Monson and Franklin, N. Y. Ees. E. Windsor Hill. He held various
town oflBces and was clerk of the society for more than twenty
years. Chn., &. at E. Windsor Hill: 1 Maud Mickell" Grant, b. July 10,
1875. 2 Frederick William Grant, b. Aug. 29, 1876. 3 Eoger Wolcott
Grant, b. Sept. 19, 1884. iv Lucy Elizabeth, twin of Eoswell ; d. Aug.,
1849. v Elizabeth, b, Aug,, 1850; d. Sept., 1851. 790 Honorable
JOHN" WETMOEE« STOUGHTON (Sara¥ Weimore, Sarah^ Brewster,
Elisha^, William*', William^, Love^, William^), married first, at East
Windsor, Conn., February 17, 1847, MAEY E., daughter of Stoddard
and Clarissa (Alden) ELLSWOETH of East Windsor. She died February
23. 1854, aged 32 years. Married second, June 4, 1856, his first
wife's sister, SAEAH BUCKLEY ELLSWOETH, who died at New Britain,
Conn., April 12, 1899. John W. Stoughton removed from East
Windsor to New Britain about 1879, where he resided in 1899, " He
entered Yale College in 1836, but, owing to ill health, left in his
junior year. About 1843 he was justice of the peace at South
Windsor, was selectman one term, judge of Probate of East Wind 
 EIGHTH GENERATION 699 sor District one term, and at one
time a trustee of the Connecticut Theological Institute at East
Windsor. In 1845 he was elected senator from the Second District in
General Assembly, was re-elected in 1860, when he was chairman of
Committee on Eailroads, and in 1873, when he was chairman of
Committee on Education." Children : i Judge John Alden^, b. June
28, 1848 ; m. July 19, 1876, Ellen, dau. of Ebenezer Pinney of South
Windsor, and widow of Henry L. Goodwin of Hartford. He grad. Yale
Law School, 1882. Has served as judge of Probate Court of East
Hartford, Conn. The author of Windsor Farms. Clin., h. at Hartford: 1
Ellen Katherine^° Stoughton, b. March 7, 1878. 2 Mary Theodosia
Stoughton, b. Oct. 1, 1879. 3 Elizabeth Clio Alden Stoughton, b. Jan.
31, 1882. ii Mary Brewster, b. Jan. 12, 1851; m. Oct. 27, 1875,
Frederick Wooster Giddings. Ees. New Britain. 791 OLIVEK
WETMORE« TREADWELL (Hannah'' Wetmore, Sarah^ Brewster,
Elisha'^, 'William*, William^, Love^, William^), married July 31,
1834, ANNA HELENA, daughter of Frederick and Mary (Eefew)
KRAMER of Baltimore, Md., born there, October 10, 1810; died at
Hamden, Conn., January 11, 1891. He died at Hamden, April 7,
1879. Oliver W. Treadwell entered Yale College, but went with
Professor Olmstead, of Yale, to the University of Carolina (N. C),
where he was graduated in 1826. " He was made an honorary
member of Yale Alumni in 1861. He remained in the South, having
charge of schools for boys at Mount Hope, Baltimore, and at
Rockville, Md. Just previous to the Civil War Mr. Treadwell removed
with his family to a suburb of New Haven, Conn., so that his sons
could be educated at Yale College. His life was spent in study and
teaching and he was honored and beloved for his usefulness and
scholarly habits." He was a member of the Board of Education and
deacon of Whitneyville (Conn.) Congregational Church. Children,
horn at Baltimore, Md.: 1300 i Anna Helena Dorothy^, b. Oct. 21,
1837. ii Oliver Ferdinand, b. June 25, 1841 ; d. Aug., 1898 : m.
Sarah Baracleugh. He grad. Yale Coll., 1862, and Yale
 700 BREWSTER GENEALOGY Med. School, 1865. He
succeeded his father as deacon of Whitneyville Church, s. p. 1301 iii
George Edwards, b. March 9, 1843. 1302 iv Lucy Whittlesey, b. Dec.
12, 1847. 792 JOHN GOODWIN« TEEADWELL {Hannah' Wetmore,
Sarah^ Brewster, Elisha^, William'^, William^, Love-, William^),
married at New York City, April 30, 1841, ELLEN TINKER, daughter
of Jacob and Hannah (Tinker) HOLMES of New London, Conn., born
there, November 24, 1815; died at Albany, N. Y., January 3, 1870.
He died at Monmouth, 111., June 25, 1900. John G. Treadwell was a
member of the firm of Treadwell and Perry, largely engaged in the
manufacture of stoves at Albany, N. Y. He served as U. S. assessor of
internal revenue for the counties of Albany and Schoharie, N. Y., and
was an officer of the Congregational Church for nearly fifty years.
Children, i-iv horn ai N. Y. City, v-viii Albany: 1 Hannah Wetmore®,
b. Jan. 20, 1842; d. at Cornwall, Vt., Sept. 5, 1864; m. at Albany,
May 19, 1863, Dr. Martin Luther Mead, son of Rufus Mead of
Cornwall, b. there, Jan. 23, 1834 ; d. at Highland Lake, Colo., Sept.
5, 1899. A physician at Albany and at Highland Lake. He atd.
Middlebury (Vt.) Coll., 1852-55, and Albany Med. Coll. Ch., h. at
Albany: 1 Hannah TreadwelP" Mead, b. June 18, 1864; m. at San
Francisco, Cal., Apr. 2, 1891, Samuel Watson Royston, son of Cyrus
A. Royston of Limestone, Tenn. Chn., b. at San Francisco: (1)
Clarence Edgar" Royston, b. June 15, 1893. (2) Howard Lester
Royston, b. Aug. 20, 1896. ii Ellen Holmes, b. Aug. 10, 1843; d. July
31, 1844. iii Sarah Wetmore, b. Aug. 15, 1845; m. at Albany, June
15, 1870, Rev. Amos Hammond Dean, son of Amos Dean of Albany,
b. at Bethlehem, N. Y., June 16, 1843. He grad. Hamilton Coll., 1864,
and Union Theol. Sem., 1869. A Presbyterian minister at Monmouth,
111. He is a trustee of Monmouth Coll. and president of Board of
Education of Monmouth. Served in Civil War as private, reg.
commissary sergeant, reg. quartermaster supt., and 2d lieutenant 3d
N. Y. Light Artillery. Chn., 1-2 b. at Albany, 3-6 Joliet, III.:
 EIGHTH GENERATION 701 1 Ellen Edwards^" Dean, b. May
16, 1871; m. at Monmouth, Oct. 38, 1897, Dr. Frank E. Wallace of
Monmouth. 2 Frederick Porter Dean, b. Feb. 16, 1873. 3 Edna Dean,
b. Dec. 21, 1874; m. at Monmouth, Dec. 30, 1897, Eev. John Samuel
Pollock of Wilsonville, Neb. 4 Mary Dean, b. Aug. 20, 1876. 5 Alice
Dean, b. July 30, 1878. 6 Ethel Dean, b. March 7, 1882. iv Mart
Treby, b. March 23, 1847; d. at Montclair, N. J., May 16, 1884; m. at
Albany, March 23, 1871, Samuel M. Porter, b. at Waterbury, Conn.,
May 17, 1835; d. at N. Y. City, March 19, 1876. A merchant in New
York City. Chn., 1 h. at Jersey City, 2 N. Y. City: 1 Bertha
TreadwelP<> Porter, b. Dec. 13, 1871; m. at Montclair, Nov. 12,
1896, Dr. Seth Cook" Comstock, son of Marco M. and Mary
Elizabeth^" (Cook)* Comstock of New London, Conn., b. there, Nov.
30, 1871. He grad. Coll. of Physicians and Surgeons, N. Y., 1894. A
physician in New York City. Ch.. h. there: (1) Dorothy Brewster"
Comstock, b. May 12, 1898. 2 Charles Willson Porter, b. Nov. 8,
1875. V John Goodwin, b. Nov. 23, 1848; d. in infancy. vi John
Pomeroy, b. Dee. 7, 1852 ; m. at Chicago, 111., Sept. 10, 1874,
Sadie C. Canfield. vii Thomas, b. Nov. 18, 1854; d. Nov. 21, 1854. viii
Edward Norton, b. Sept. 7, 1856 ; m. at Newark, N. J., Elma Fischer,
widow. A musician in Southern California, Ch.: 1 Goodwin Holmes^"
Treadwell, b. Aug. 17, 1891. 793 WILLIAM BEEWSTEE« TREADWELL
{Hannah' Wetmore, Sarah^ Brewster, Elisha^, William*, William^,
Love^, William^), married May 14, 1844, MARY ELIZA, daughter of
Roland and Lydia (Webster) ADAMS of Albany, N. Y., born there,
January 15, 1827, and died there, April 16, 1869. They resided at
Albany, where for many years William B. Tread* Mary Elizabethio
Cook. Elizabeths Smith. Harriets Prentis, StephenT Prentls. Annes
Starr, Samuels Starr, Jonathan* Starr, Hannahs Brewster, Jonathans,
Williami.
 703 BREWSTER GENEALOGY well was a member of the
firm of Treadwell and Perry, stove manufacturers. Children, born at
Albany: i Elizabeth®, b. Aug. 20, 1845; unm. Res. Albany, ii
Frederick K., b. May 27, 1847 ; d. Aug. 13, 1852. iii Franklin Adams,
b. July 20, 1849; m. at Albany, Sept. 7, 1876, Magdaline Amanda,
dau, of William Crounse of Albany, b. May 25, 1854. Ees. Albany.
Chn., b. there: 1 William Adams^" Treadwell, b. June 6, 1877; d. ge.
1 month. 2 William Adams Treadwell, b. May 9, 1878. 794 Doctor
SAMUEL EDWARDS^ TREADWELL {Hannah' Wetmore, Sarah^
Brewster, Elisha^, William^, William^, Love', \Yil~ Ham}), married
September 12, 1836, ANNA, daughter of Mordecai STAMP of Talbot
'County, Md. He died at New York City, April 30, 1860. Samuel E.
Treadwell was a practicing physician at Havre de Grace, Md., for
many years. Children, born at Havre de Grace: i Helena Cramer®, b.
Nov, 19, 1838; m. at H. de G., Sept. 28, 1863, William Wright, son of
William Wright of H. de G., b. there, March 18, 1826; d. at Santa
Clara, Cal., Apr. 22, 1890. Chn., b. at Santa Clara: 1 Dora TreadwelP"
Wright, b. Jan. 29, 1866; m. at Santa C, Oct. 1, 1889, Marvin Porter
Gibson, Gh., b. at San Francisco, Gal.: (1) Dolores" Gibson, b. Jan. 3,
1891. 2 William Tarleton Wright, b. Oct. 15, 1875; m. at Santa C,
Oct. 6, 1897, Mary Eleanor Steel. ii Elizabeth Alice, b. Jan. 16, 1840;
m. at Havre de Grace, May 30, 1860, Thomas Headley Allen, who d,
at H. de G,, Apr. 3, 1872. 2 chn., d, inf. iii Sarah Wetmore, b. June
12, 1842 ; d. at Havre de Grace, Jan. 15, 1865; m. there, Jan. 3,
1861, John M. Gilbert, son of Janett Gilbert of H. de G., b. there, Apr.
3, 1836. Chn., b. at H. de G.: 1 Annie Isabel" Gilbert, b. June 17,
1862; m. at H. de G., Apr. 23, 1891, James Amos Anderson. Ch.: (1)
Edna Graham" Anderson, b. July 22, 1899.
 EIGHTH GENEEATION 703 2 Janett Treadwell Gilbert, b.
Oct. 13, 1863; m. at H. de G., March 26, 1889, Irene Bradus Adams.
Chn.: (1) Wilton Greenway" Gilbert, b. June 1, 1890. (2) Annie
Isabel Gilbert, b. Dec. 9, 1892. (3) Treadwell Gilbert, b. Jan. 9, 1894.
iv Lucy Wetmoee, b. Feb. 8, 1844; m. at Havre de Grace, Dec. 20,
1867, as 2d wife, John M. Gilbert, whose 1st wife was her sister
Sarah. Ch., &. at H. de G.: 1 Harry Millard^" Gilbert, b. Sept. 12,
1867; m. at H. de G., March 15, 1894, Mary Emile Carroll. Chn.: (1)
Grace'^ Gilbert, b. Nov. 14, 1895; d. Feb. 4, 1896. (2) Lucy Marie
Gilbert, b. Aug. 25, 1897. (3) Harry Grillet Gilbert, b. Oct. 24, 1898.
V Martha Morris, b. May 9, 1846 ; m. at Santa Clara, Cal., Sept. 14,
1869, Abel Aldenson Withrow, son of Abel Withrow of Eichmond, Va.,
b. at Lawrenceburg, Ind., Nov. 2, 1832. Chn., h. at Santa Clara: 1
Elsie Briggs^" Withrow, b. Aug. 19, 1870; m. at Santa C, June 27,
1893, William J. Kirkpatrick. Ch.: (1) Marion Withrow" Kirkpatrick, b.
Aug. 6, 1894. 2 Ealph A^argrave Withrow, b. Feb. 23, 1872; m. at
Santa C, Dec. 4, 1895, Mabel E. Boynton. Ch.: (1) Earl Boynton"
Withrow, b. March 11, 1898. 795 EDWARD FEANCIS« TEEADWELL
(Hannah' Wetmore, 8arah^ Brewster, EUsha^, William*, William^,
Love^, William^), married at New York City, April 21, 1847, EOSINA,
daughter of Thomas and Elizabeth (Carter) HAMILL of Baltimore,
Md., born there, January 4. 1823; died at Flushing, N. Y., January 11,
1888. He died at Flushing, February 9, 1868. Edward F. Treadwell
was a prominent lawyer in New York City, and resided at Flushing.
Children, i-iv born at N. Y. City, v Flushing : i William Brewster®, b.
March 10, 1848 ; m. at Alameda, Cal., Sept. 19, 1870, Adelaide A.,
dau. of Eobert Kirk of Alameda, b. at N. Y. City, June 2, 1848. He
atd. Coll. of City of New York. A lawyer at San Francisco, Cal. Chn.,
h. in Cal.: 1 Eosina HamilP° Treadwell, b. Aug. 2, 1871. 2 Laura Olive
Treadwell, b. May 21, 1873.
 704 BREWSTER GENEALOGY 3 Edward Francis Treadwell,
b. May 19, 1875: m. at San F., March 31, 1900, Eulila Ayers. Ch., h.
at San Francisco : (1) Earl Francis^i Treadwell, b. June 2, 1901. 4
James Wetmore Treadwell, b. Nov. 32, 1877; m. at San F., June 14,
1899, Edith L. Day. 5 Oscar Sidney Treadwell, b. March 17, 1881. G
Albert E. Treadwell, b. Apr. 30, 1887. ii EosiNA Hamill, b. March 5,
1850; m. at Flushing, Jan. 7, 1874, Fredk. W. Dorr, son of Benjamin
Dorr of Medford, Mass., b. at New Hampton, N. H., March 22, 1847.
"Res. Medford. Ch., h. there: 1 Leila Osgood^'' Dorr, b. Jan. 7, 1875.
iii Cornelia La Tourette, b. Feb. 29, 1852. iv James Wetmore, b. Feb.
26, 1854; m. at Flushing, Nov. 13, 1879, Lillian Livingston, dau. of
Benjamin Holt of Flushing, b. at Baltimore, Md., Nov. 30, 1854. A
lawyer in New York- City. Chn., h. at Flushing: 1 Margaret
Livingston^" Treadwell, b. Nov. 6, 1880. 2 Chauncey Edwards
Treadwell, b. May 23, 1883. 3 Bertha Holt Treadwell, b. Apr. 6, 1885.
4 Edward Wetmore Treadwell, b. Oct. 12, 1889. 5 Charlotte
Burroughs Treadwell, b. Jan. 3, 1891. 6 Agnes Carter Treadwell, b.
Aug. 4, 1894. v Bertha Frances, b. Nov. 13, 1856 ; d. March 29,
1858. 796 EMILY GALE« GOODWIN (Clarissa' Wetmore, Sarah''
Brewster, Elisha^, WilUnm*, William^, Love^, William'^), married at
Hartford, Conn., November 27, 1830, Professor BENJAMIN
SPILLSBUEY BAECLAY, born at Worcester, England, in 1806. She died
at Athens, Ala., September 4, 1848. Benjamin S. Barclay was
graduated from Oxford University. Children, i horn at Hartford, ii-iv
Philadelphia, Pa., v Camden, N. J.: i Thomas S.^ b. Feb. 18, 1833 ;
d. July 10, 1834. 1303 ii Emily Spillsbury, b. Feb. 1. 1835. iii Ida
Williams, b. Nov. 29. 1836 ; d. at Fulton, Miss., Oct. 19, 1855 ; m.
Clement C. Moore of Fulton, iv Edwards Wetmore, b. Jan. 3, 1842. V
Clara Hannah, b. July 8, 1845 ; m. at Germantown,
 EIGHTH GENERATION 705 Pa., July 27, 1865, Alfred
Wallace Stoddard, b. at Marshfield, Mass., July 21, 1845. Ees.
Germantown. He served in Civil War in Co. C, 32d Eegt., Mass. Inft.
Chn., b. in Mass. : 1 Benjamin Barclay^" Stoddard, b. July 5, 1867. 2
Linda Dodge Stoddard, b. Apr. 13, 1870 ; m. Otis K.^"* Stewart
(Emily^ Barclay, Emily* Goodwin above). See his record. 3 Son, b.
and d. May 10, 1873. 4 Ida Palmer Stoddard, b. June 3, 1882. 797
HAREIET CLAEA« DODGE (Clarissa' Wetmore, Sarah^ Brewster,
Elislia^, William}, William^, Love^, William^), married at
Philadelphia, Pa., February 17, 1840, Lieutenant THOMAS
GEEEXLEAF CHASE, U. S. A., son of Thomas and Sarah (Greenleaf)
Chase of Boston, Mass., born at Bolton, Mass., March 3, 1793 ; died
at Philadelphia, March 2, 1871. They resided at Philadelphia, Pa.
Children, horn at Philadelphia: i Clara ANNE^ b. Nov. 12, 1840; m. at
Phila., Apr. 17, 1865, I. Addison Bush, b. at Unionville, 0., Apr. 1,
1835 ; d. at Phila., June 29, 1894. Ees. Phila. Chn., h. there: 1
Alleyne 0.^° Bush, b. June 13, 1867; unm. 1898. 2 Beryl Bush. b.
Jan. 10, 1870; d. Aug. 7, 1871. 3 George B. Bush, b. Oct. 11, 1872;
m. at Phila., June 14, 1894, Flora H. Widmer. ii Thomas, b. Jan. 11,
1843; m. at Phila., Oct. 13, 1865. Elizabeth Lawrence. Ees. Phila. He
served in U. S. Navy as asst. engineer, 1863-65. Ch., b. at Titusville,
Pa. : 1 Mabel Clara^" Chase, b. June 27, 1869 ; unm. 1898. iii
Alleyne Gardiner, b. March 26, 1849; d. June 19, 1868 ; unm. iv
George Emanuel, b. March 8, 1852; m. at Phila., Oct. 14, 1885, Ella
Peeves. Chn., b. at Phila.: 1 Stacy Eeeves^° Chase, b. Aug. 14,
1886. 2 Thomas Chase, b. Oct. 1, 1888 ; d. June 29, 1897. V
Emmeline, b. March 27, 1854; d. Aug. 15, 1856. 798 JOHN
DOBSON« SOUTHMAYD (Sophia' Wetmore, Sarah'' Brewster,
Elisha^. Willia)n*, William^, Love-, William'^), married at
 706 BREWSTER GENEALOGY Middletown, Conn., August
14, 1844, HAEEIET H. NOETH. He died at Middletown, October 11,
1847. Children, born at Middletown: 1 Frederick Giles^ b. Dec. 3,
1845; m. at Springfield, Mass., June 15, 1880, Jennie Whipple, dau.
of John H. Nutting of Springfield, b. at Amherst, Mass., Oct. 6, 1855.
Ees. Springfield. Chn., h. there: 1 Pearl Agnes^" Southmayd, b. May
20, 1881. 2 Leon Nutting Southmayd, b. Oct. 21, 1883. 3 Phillip
Helming Southmayd, b. March 8, 1890. 799 ELIZABETH*
SOUTHMAYD {Sophia' Wetmore, Sarah"" Brewster, Elisha^,
William^, William^, Love^, William^), married at Middletown,
Conn., May 5, 1853, SAMUEL BEOWN, son of Elisha and Lina (Cone)
Brown of Westchester, Conn., horn there, November 16, 1809, and
died there, June 1, 1884. Children, horn at Westchester: i Thomas
Southmayd®, b. June 20, 1854; m. East Hampton, Conn., Oct. 11,
1877, Elizabeth Cone, dau. of Horatio Chapman of E. Hampton, b.
July 29, 1856. Ees. East Hampton. He was a member of the Conn.
Legislature and sheriff of Middlesex County, Conn. Chn., h. at E.
Hampton : 1 SamueP" Brown, b. July 12, 1883 ; d. Apr. 16, 1885. 2
Thomas Wolcott Brown, b. July 14, 1886. 3 Anna Laura Brown, b.
Jan. 6, 1889. ii Lina Cone, b. May 13, 1857 ; m. at Westchester, June
7. 1877, Amatus Eobbins Bigelow, son of David Bigelow of
Westchester, b. there, Sept. 18, 1853. Ees. Westchester. Chn., h.
there: 1 Abby Elizabeth" Bigelow, b. March 2, 1879. 2 Leon Bigelow,
b. July 28, 1888; d. Oct. 11, 1888. 3 Euth Lina Bigelow, b. Aug. 9,
1891. 800 CHAELES* SOUTHMAYD (Sophia'' Wetmore, Sarah"
Brewster, Elisha^, William*, William^, Love^, William^), married at
East Windsor, Conn., June 17, 1845, MAETHA JANE«, daughter of
John and Sarah^ (Wetmore)* STOUGHTON, born at East Wind*
Sarah^ Wetmore, Sarah* Brewster above.
 EIGHTH GENERATION 707 sor, January 23, 1831 ; died July
15, 1876. He died at Grand Forks, No. Dak., October 31, 1885. They
resided at Middletown, Conn. Children, i horn at E. Windsor, ii-vii
Middletown: i Sarah Sophia^, b. May 8, 1846 ; d. Aug. 23, 1848. ii
Anna Stoughton, b. Jan. 8, 1850; d. at Southbury, Conn., July 15,
1881; m. at MiddletoMTi, Nov. 26, 1874, Charles Kelsey Osborne.
Ees. Southbury, Cli., h. there: 1 Anna Schultze^° Osborne, b. Nov.
10, 1875. iii John Edwards, b. Aug. 7, 1853 ; unm. 1898. iv Henry
Albert, b. Nov. 16, 1854; d. March 12, 1855. V Mary Wetmore, b.
Feb. 5, 1857; d. Aug. 30, 1884; unm. vi Charles Everett, b. Jan. 7,
1859. Ees. So. Dak., 1899, vii Egbert Winthrop, b. June 10, 1862 ;
m. Johnson of Middletown. Ees. So. Dak., 1899. 2 elm. 801
CHAUNCEY EDWAEDS« WETMOEE (Chaunceif Wetmore, Sarah^
Brewster, Elisha^, William*, William^, Love"^, William^), married at
Wethersfield, Conn., October 8, 1845, MAEY MIX, daughter of Dean
BUCK of Wethersfield, born there, in 1818; died at Clayton, Cal., July
10, 1894. He died at Clayton, September 29, 1894. Chauncey E.
Wetmore attended Middletown (Conn.) Academy, took a collegiate
course and attended lectures in New York City. He then taught
school in Herkimer County, N. Y., was principal of the public school at
Middletown and at Wethersfield, Conn,, then spent six months in the
Teachers' Academy at Andover, became a tutor in Mt. Hope College,
Baltimore, Md., and later opened and conducted a private school in
that city, but his health failing he sold the school and engaged in
mercantile pursuits. On Dec. 22, 1847, Mr. Wetmore, his wife and
brother Henry sailed for California, and after stopping at Valparaiso,
Callao, and other points of South America, and at Honolulu, S. I.,
they arrived at San Francisco, Aug. 27, 1848. Here he was a
merchant until April 3, 1849, when he and his family settled at
Benicia, Cal., where he formed a partnership with his brother-in-law,
Samuel C. Gray, who had brought a large consignment of goods
from Baltimore and New York, and Mr. Wetmore continued in Benicia
until Nov. 22, 1853. During his residence there he was active in
founding Benicia Female Seminary and building the first Presbyterian
Church, and was a trustee and superintendent of public schools.
From Nov., 1854, to
 708 BREWSTER GENEALOGY Aug., 1857, he was a
merchant at Vallejo, Cal., and was active in the piiblic interests of
that town. He was a delegate to the first Eepublican County
Convention and to the first State Convention of that party in
California, and made the only Eepublican speech at Vallejo during
the campaign. His politics having caused him serious business losses
he sold his interests and removed to Diablo Valley, where he began
farm life. In 1858 he was elected justice of the peace and associate
justice of the Court of Sessions; was a member of the Eepublican
State Convention of 1860, and Oct., 1862, was appointed assessor of
internal revenue for Contra Costa County, holding the position for
four years. On November 1, 1874, he sold his farm and removed
with his family to Clayton, Cal., and in 1875 to Oakland, Cal., where
he engaged in the real estate business. In 1882 Mr. Wetmore retired
and returned to his farm near Clayton, where he spent his declining
years. His most characteristic traits were his unswerving and
uncompromising devotion to principle, his high ideals and public
spirit. Children, i horn at San Francisco, ii Benicia, iii Vallejo, iv
Clayton: i Edith^ b. Sept. 24, 1848 ; d. at Clayton, Apr. 13, 1893 ;
m. there, Sept. 24, 1872, David Jesse Horswill, son of Frederick and
Mary (Lawyer) Horswill of Sydney, Australia, b. there, Apr. 6, 1848;
d. at Tacoma, Wash., July, 1886. She was the first child to receive
Protestant baptism in California. Ees. Clayton. Chn., 1 b. at Clayton,
2-3 OaHand, Cal.: 1 Alfred Dean^« Horswill, b. I^ov. 6, 1873 ; unm.
1900. 2 Lucy Horswill, b. Xov. 23, 1875; d. May 21, 1891. 3 Winifred
Grace Horswill, b. Apr. 1, 1877; unm. 1900. ii William Brewster, b.
Apr. 1, 1851; unm. 1900. A stock raiser near Barry, Wash, iii Henry
Clarence, b. Apr. 27, 1856 ; m. at Oakland, Nov. 25, 1882, Georgie
Alfreda, dau. of George W. Fountain of Oakland, b. at Brooklyn, N.
Y., Feb. 21, 1855. Ees. Clayton. Chn., h. there: 1 Mary EtheP"
Wetmore, b. Aug. 6, 1884. 2 Edna Wetmore, b. Oct. 1, 1885. 3 Ealph
Emerson Wetmore, b. Sept. 8, 1893. 4 Hazel Belle Wetmore, b. May
17, 1899. iv Frederick Chauncey, b. Oct. 24, 1860 ; d. May 5, 1875.
 EIGHTH GENERATION 709 802 LUCY« WETMORE
{Chauncey' Wetmore, Sarah^ Brewster, EJisha^, William'^,
William^, Love^, William^), married at Middletown, Conn.,
December 14, 1847, as first wife, SAMUEL COTTON GRAY, son of
Samuel and Joanna (Powers) Gray of Boston, Mass., born there,
October 11, 1816; died at Benicia, Cal.. June 19, 1892. She died at
Benicia, July 2, 1879. They resided at Benicia, " Mrs. Gray was a lady
of energy and character, and fine natural abilities, faithful and
conscientious in all her relations of life." Mr. Gray was a descendant
of Eev. John Cotton, pastor of the first church in Boston. Children,
horn at Benicia: i Dr. Edward", b. N"ov. 17, 1849 ; m. 1st, at London,
Eng.. Aug. 31, 1876, Gertrude, dau. of Rev. Henry M. and Lucy
(Tuttle) Colton of Middletown, b. at Avon, Conn., Nov. 15, 1856; d.
at Ft. Gaston, Cal., July 10, 1884; m. 2d at San Francisco, Cal., N"ov.
16, 1889, Maria M., dau. of Samuel Hopkins and Martha (Jeffers)
Willey of San Francisco, b. there, Dec. 7, 1856. He grad. Yale Coll..
1871, and Coll. of Physicians and Surgeons, New York. 1875 ; was at
the Nursery and Children's Hospital, N. Y.. 1874, Presbyterian
Hospital, 1874-75, General Hospital and Poliklinik, Vienna, Austria,
1875-77, and was government physician and surgeon at Ft. Gaston,
1881-84. A physician at Berkeley, or Eldridge, Cal. He is a member of
the San Francisco County Medical Society and of the Royal
Microscopical Society, London, Eng. Besides his professional
attainments he is an excellent Spanish scholar and is editor of the
revised edition of Velasquez Dictionary. Clin., 1, 6 &. at San
Francisco, 2, 5 Benicia. 3 Cherokee, 4 Ft. Gaston, all in Cal.: 1 Henry
Colton^" Gray, b. Apr. 16, 1878; unm. 1899. 2 Samuel Herbert Gray,
b. June 29, 1879; unm. 1899. 3 Theodore Gray, b. March 6, 1881; d.
March 6, 1882. 4 Percival Gray, b. March 7, 1883; d. Aug. 13, 1892.
5 Beulah Gray, b. Oct. 18, 1890 ; d. Jan. 7, 1899. 6 Girard Gray, b.
July 24, 1896. ii Franklin, b. Aug. 20, 1851 ; d. Jan. 18, 1852. iii
Samuel Arthur, b. Nov. 24, 1853; unm. 1899. A photographer at
Benicia. iv Theodore, b. Oct. 24, 1855; m. at San Francisco, March
 710 BREWSTER GENEALOGY 1, 1884, Anna Turner, dau. of
Epaphroditus Seymour and Mary Emma (Weaver) Farnsworth of San
Francisco, b. at Benicia, July 31, 1856. He atd. Un. of California,
Berkeley, 1873-77, and Hastings Law Coll., San Francisco, 1879-81.
Ees. San Francisco. Cli., h. there: 1 Harold Farnswortli^<> Gray, b.
June 29, 1885. V Lucy Gertrude, b. May 25, 1866; m. at San
Francisco, Apr. 25, 1894, David Marion Hart, son of Joseph and
Babette (Kullman) Hart of Centerville, Cal., b. at Lafayette, Ind., Dec.
13, 1855. She atd. Ladies Sem., Benicia. 1878-85, and Housatonic
Hall, Great Barrington, Mass., 1885-86. Ees. Los. Angeles, Cal. Chn.,
1-2 h. at Benicia, 3 Los Angeles: 1 Percival Grayi<> Hart, b. Sept.
19, 1895. 2 Richard Hubbard Hart, b. Sept. 22, 1898. 3 William
Brewster Hart, b. Oct. 31, 1904. 803 HAERIET^ WETMOEE
(CJiauncey' Wetmore, Sarah'' Brewster, Elisha^, William*, William^,
Love^, William^), married at Middletown, Conn., June 8, 1852,
JOHN WESLEY JONES, son of Elder Benjamin and Hannah or Mary
(Milliken) Jones of Lincolnville. Me., born there, August 31, 1821 ;
died at Benicia, Cal.. November 14, 1897. She died at Benicia, May
14, 1901. They resided at Benicia. John W. Jones was graduated
from Wesleyan University, Middletown, Conn., in 1849. Children,
horn at Benicia: i Walter Brewster^ b. July 2, 1856; m. at
Middletown. Sept. 12, 1894, Annie Katharine, dau. of Charles E. S.
Hall of Middletown, b. there, Jan. 4, 1863. He atd. Un. of California
(Coll. of Chemistry), 1874-78. Chn., h. at San Francisco: 1 Marion
Brewster" Jones, b. Feb. 18, 1896. 2 Katharine Hall Jones, b. July 4,
1897. 3 Walter Douglas Jones, b. Dec. 14, 1898. 4 Euth Gleason
Jones, b. Sept. 14, 1900. ii Bertha, b. June 16, 1859 ; d. Dec. 6,
1861. iii Daisy, b. May 15, 1863 ; m. 1st, at Benicia, June 8, 1883,
John Frederick Davis, b. at Plymouth, Eng. ; m. 2d, at San Francisco,
June 4, 1898, Horace Jefferson Dobbins. Ch., h. at Benicia: 1
Howard Wetmore" Davis, b. March 20, 1884.
 EIGHTH GENERATION , 711 804 HEI^EY GOODWIN"«
WETMOEE {Chaunceif Wetmore, Sarah^ Brewster, Elisha^,
William^, William^, Love^, William^), married in Suisun A^alley,
Cal., April 6, 1852, MAEY TOWNSEND, daughter of Captain Edward
and Eliza (Harvey) WINN of Benicia, Cal., born at Salem, Mass.,
November 16, 1835; died at Fairchild. Cal., November 13, 1871. He
died at Sonora, Cal., April, 1889. Henry G. Wetmore resided in
California from 1848 until his death, his last residence being at
Sonora. In early life he was a farmer and later a mining man. He
served as county recorder of Solano County, Cal., 1866-70. Children,
horn in Suisun Valley: i Franklin Henry^, b. Dec. 1, 1852 ; m. at
Cordelia, Cal., Dec, 1889, Mary, dau. of Justice J. B. Nichols of
Cordelia, b, at Benicia, Oct., 1867. Ees. Vallejo, Cal. Ch., h. at
Cordelia: 1 EtheP" Wetmore, b. Sept. 3, 1890. ii Charles Augustus, b.
Jan. 26, 1855; m. Oct. 8, 1895, Annie, dau. of Thomas Jenness of
Sonora, and widow of Carr, b. July 31, 1866. Ees, Sonora. Ch., h.
there : 1 Myron Wilfred^^ Wetmore, b. Dec. 14, 1897. iii Chauncey,
b. Dec. 29, 1856; m. Dolly Sieber. Ees. Sonora. Chn.: 1 Henry
Goodwin^" Wetmore, b. abt. 1894, 2 Benton Sieber Wetmore, twin
of Henry. iv Lucy, b. June 15, 1859 ; m. at San Francisco, Apr. 19,
1893, Myron Dennis Cooley. Ees. Washington, Cal. v Eebecca, b.
Feb, 18, 1862; m. at Sonora, June 6, 1883, Nathan Dean Nichols,
son of William Nichols of Benicia, b. there, July 25, 1855. Ees, San
Diego, Cal, Chn., h. at Benicia: 1 Dora Edna^° Nichols, b. Oct, 30,
1884, 2 Clarence Lee Nichols, b. Apr, 19, 1886. vi Sherman, b. Dec.
28, 1863; m, at Columbia, Cal., Sept. 24, 1887, Kate Elizabeth, dau.
of John Coffey of Columbia, b. there, Dec. 20, 1868. Ees. Columbia.
Chn., h. there: 1 Franklin Edward^" Wetmore, b. Sept. 11, 1888, 2
Gertrude Lucille Wetmore, b. Sept. 6, 1889. vii Dora Margary, b. Jan.
15, 1866 ; m. at Benicia, Oct. 12, 1887, Simon Clemens Bowman,
son of Moses Bowman
 712 BREWSTER GENEALOGY of Canada, b. there, March 1,
1853. Res. Colusa, Cal. Clin., 1 h. at Vallejo, 2 Colusa: 1 Addison"
Bowman, b. July 10, 1889; d. Dec. 5, 1897. 2 Dorothy Clemens
Bowman, b. June 21, 1900. viii Edward, b. Feb. 19, 1865; d. Feb. 21,
1865. ix Herbert, b. Aug. 1, 1868 ; d. Nov. 3, 1868. 805 MARY
ELLEN^ WETMORE (Chauncey' Wetmore, Sarah" Brewster, Elislia^,
William'^, William^, Love^, William'^), married at Flushing, Long
Island, K. Y., April 14, 1866, LEWIS GRAY "WELLS, son of Lewis
Wheeler and Affa Maria (Gray) Wells of Columbus, Ga., born there,
June 17, 1841. She died at Baltimore, Md., April 16, 1874. Children, i
born at Oregon, Wis., ii Grand Rapids, Mich., iii-v Baltimore: i Rev.
Hubert Wetmore'', b. Dec. 29, 1866; m. at New York City, Dec. 25,
1890, Lucy, dau, of Rev. Henry M. and Lucy (Tuttle) Colton of
Middletown, Conn., b. there, Aug. 31, 1865. He grad. Yale Un., 1889
; became assistant rector of Grace Episcopal Church, New York City,
then rector of the Episcopal Church, Waltham, Mass., and in 1900,
rector of St. Andrews Church, Wilmington, Del. Chn., 1 h. at
Westville, Conn., 2 N. Y. City, 3 Waltham: 1 Colton" Wells, b. Aug. 6,
1891 ; d. Apr. 27, 1894. 2 Margaret Colton Wells, b. Oct. 5, 1893. 3
John Colton Wells, b. Nov. 23, 1896. 4 Child, b. 1902. ii Prof. Philip
Patterson, b. Feb. 5, 1868 ; m. at New Haven, Conn., May 22, 1893,
Eleanor Duncan, dau. of Rev. Theodore Thornton and Elizabeth
(Duncan) Munger of New Haven, b. at Haverhill, Mass., Feb. 2, 1868.
He grad. Yale Tin., with deg. B. A., 1889. Lawyer — librarian and
instructor in Law School of Yale University. Ch., h. at New Haven: 1
Lewis Gray" Wells, b. June 9, 1896. iii Ernest Hubbard, b. June 12,
1870 ; unm. 1900. Res. Louisville, Ky. He grad. Yale Un., 1893. iv
Prof. Chauncey Wetmore, b. May 25, 1872 ; m. at Burlington, N. J.,
Sept. 8, 1897, Mary Prescott of New
 EIGHTH GENERATION 713 Haven, Conn. Ee grad. Yale Un.,
1896. Res. Berkeley, Cal. V Mary Eleanor, b. Jan. 5, 1874; d. June
28, 1874. 806 CHAELES FRANCIS^ MORRISON {John' Morrison,
HopestiW Brewster, EUsha^, William^, William^, Love^, WiUiam'^),
married at Willington, Conn., July 6, 1841, LOUISA, daughter of
Amos and Martha (Taylor) PRESTON" of Willington, born there,
February 13, 1819, and died there, April 6, 1896. He died at
Willington, September 13, 1882. Charles F. Morrison served as
selectman and as a member of the School Committee of Willington,
and was twice elected to the Connecticut Legislature. Children, horn
at WiJUngton : i Katheeine Louisa^, b. May 26, 1842 ; unm. 11
Sylvester Preston, b. Nov. 17, 1844; m. at Willington, Apr. 22, 1866,
Ellen, dau. of William Slingsby of Willington, b. in England, Oct. 6,
1845. Res. Willington. He served in Civil War as private in 22d Conn.
Inft. and 1st Conn. Heavy Artillery. Clui., h. at Willington : 1 Maude
Louisa^" Morrison, b. Apr. 6, 1867 ; m. at Willington, May 30, 1885,
Charles F. Macfarlane. 2 Nellie Persis Morrison, b. June 4, 1874;
unm. 1899. Hi Charles Frank, b. March 12, 1847; d. at Whitehall, N.
Y., July 25, 1893 ; m. 1st, Josephine Benchly ; m. 2d, Ada E. . He left
2 sons and 1 dau., who probably res, at Chicago, 111. Iv John
Henry, b. March 29, 1849 ; d. at Willington, June 20, 1896; m. at
Willimantic, Conn., Jan. 29, 1874, Emma Antoinette, dau. of John S.
Smith of Willimantic, b. there, Oct. 1, 1850. A mechanical engineer
at Pawtucket, E. I. Chn., h. at Norwich, Conn.: 1 Clifford Brewster^"
Morrison, b. Sept. 27, 1880. 2 Gertrude Louise Morrison, b. March
29, 1881. v Edgar, b. June 30, 1851; d. July 8, 1852. 807 EDWARD«
MORRISON (John' Morrison, HopestiW Brewster, Elisha^, William*,
William^, Love^, William^), married at Willington, Conn., August
25, 1841, SARAH JANE, daughter of Ralph and Sarah (Eider) HOLT
of Willington, born there, April
 714 BREWSTER GENEALOGY 11, 1821, and died there, May
23, 1854. He died at Willington, August 18, 1894. Edward Morrison
attended Wilbraham (Mass.) Academy. He was a wheelwright,
plowmaker, and later a farmer at Willington. He served as a member
of the Connecticut Legislature in 1862. Enlisted in the Civil War, Aug.
11, of same year, as private in Company D, 14th Eegiment,
Connecticut Volunteer Infantry; promoted sergeant, and was with his
regiment, " The Fighting Fourteenth," throughout the war. Children,
horn at 'Willi7igton: i EoMELiA Jane^, b. Nov. 5, 1842; d. March 26,
1849. ii Charles Edward, b. Apr. 10, 1844; d. Feb. 22, 1867; unm. iii
Walter Goodell, b. May 12, 1846 ; m. 1st, at Colchester, Conn., Oct.,
1868, Mary Lord Walton of Colchester, who d. March 2, 1869; m. 2d,
at E. Hartford, Conn., June 21, 1871, ISTellie Amelia, dau. of Eev.
Lorenzo W. and Elizabeth (Smith) Blood of Mason, N. H., b. at
Eockville. Conn., May 12, 1850; d. at Willimantic, Conn., March 2,
1898. A silk machinery manufacturer at Willimantic. Chn., h. there: 1
Alice Claire" Morrison, b. Sept. 24, 1879. 2 Huber Lorenzo Morrison,
b. Oct. 25, 1884. 1304 iv Emma Jane, b. July 19, 1848. v Sarah
Elizabeth, b. Jan. 19, 1852 ; d. Aug. 13, 1853. vi Clara Elizabeth, b.
May 3, 1853 ; m. at Willington, Sept. 18, 1876, Alfred Bennett
McCoy, son of Isaac McCoy of Marion, 111., b. there, July 19, 1851.
Ch7i., h. at Marion: 1 Hattie Carpenter" McCoy, b. July 25, 1877. 2
George Lewis McCoy, b. Feb. 18, 1879. 3 Walter Morrison McCoy, b.
Nov. 24, 1880. 4 John Edward McCoy, b. June 9, 1888. 808 Doctor
ALBEET^ MOEEISON (Jo/in^ Morrison, EopestiW Brewster, Elisha^,
William*, William^, Love-, William'^), married at Hartford, Conn.,
June 25, 1851, HAEEIET ELIZABETH, daughter of Eoswell and Sarah
(Stone) BAETHOLOMEW of Hartford, born there, July 15, 1824; died
at Windsor, Conn., May 12, 1864. He died at Windsor, July 18, 1873.
Albert Morrison attended Pittsfield (Mass.) Medical College and
 EIGHTH GENERATION 715 was graduated from College of
Physicians and Surgeons, New York City, in 1847. He was a
practicing physician at Windsor for many years. Children, torn at
Windsor: i Harriet^ b. Apr. 26, 1852; d. at Windsor, March 17, 1873;
unm. ii Albert Eoswell, b. Feb. 13, 1854; m. at Putney, Vt., May 17,
1877, Catherine S., dau. of Simon W. Houghton of Putney, b. there.
May 8, 1855. The superintendent of a silk mill at Kutztown, Pa. Chn.,
h. at WiUimantic, Conn. : 1 Helen Jeannette^° Morrison, b. July 15,
1883. 2 Harriet Houghton Morrison, twin of Helen. iii Mary Elizabeth,
b. Sept. 27, 1857; m. at Windsor, Oct. 25, 1881, William Johnson
Morgan, son of William Morgan of Brooklyn, N. Y., b. at New York
City, July 22, 1855. Ees. Eutherford, N. J. Chn., 1-2 h. at Doree, Del.,
3 Rutherford: 1 Clarissa^^ Morgan, b. Sept. 21, 1883; d. Oct. 15,
1883. 2 William Albert Morgan, b. Jan. 14, 1885. 3 Amey Morrison
Morgan, b. Jan. 18, 1896. iv Jeannette Bartholomew, b. Feb. 10,
1860 ; unm. 1900. V John Stone, b. March 28, 1862 ; unm. 1900.
809 FEEDEEICK« MQEEISON (John' Morrison, HopestiW Brewster,
Elisha^, William*, William^, Love-, William^), married at
Wethersfield, Conn., June 10, 1856, ELIZABETH DEMING, daughter
of Deacon George and Sarah (Whiton) STILLMAN of Weathersfield,
born there, January 9, 1834. They resided at Holyoke, Mass., and
Berlin, Conn. Children, i-ii, v horn at Berlin, iii-iv Westhoro, Mass.: i
Frederick^, b. Oct. 21, 1857 ; d. Dec. 3, 1857. ii Frederick Stillman,
b. Oct. 20, 1858; m. at Holyoke, Dec. 28, 1882, Lucy, dau. of Allen
Higginbottom of Holyoke, b. there, Dec. 22, 1858. He grad. Yale
Coll., 1880. An instructor of Greek and vice principal of High School
at Hartford, Conn., 1900. Chn., 1-2 h. at Wallingford, Conn., 3-5
Hartford:
 716 BREWSTER GENEALOGY 1 Keith^° Morrison, b. ^^ov.
7, 1883 ; d. Sept. 10, 1884. 2 Marguerite Morrison, b. March 31,
1885. 3 Helen Eathbnn Morrison, b. June 29, 1888. 4 Elisabeth
Rosina Morrison, b. March 9, 1800. 5 Frederick Brewster Morrison, b.
Dec. 4, 1893. iii James Whiton, b. Jan. 31, 1860; d. Apr. 20, 1861. iv
Mabel, b. Oct. 4, 1862 ; d. Nov. 15, 1869. V William Brewster, b. July
21, 1867; m. at Auburn, ]Sr. Y., June 20, 1894, Lillian Bradley, dau.
of Edmund B. Fay, b. at New York City, Aug. 16, 1871. He grad. Yale
Coll., 1889. A mechanical engineer at Auburn. Ch., h. there: 1
Katharine Stillman^" Morrison, b. June 25, 1897. 810 CATHERINE^
COOKE (Sara¥ Van Deursen, Lydia^ Brewster, Elislia^, William*,
William^, Love-, William^), married first, at Brooklyn, N. Y.,
February 15, 1837, JAMES M. SEYMOUE of New York City, who died
at Galveston, Texas, August 30, 1847 ; married second, February,
1854, E. D. NASH of Columbia, Texas, who died at Brooklyn,
October, 1860; married third, May 6, 1872, JOHN ADRIANCE, whose
first wife was her sister, Lydia Ann. She resided at Columbia, Texas,
in 1902. Children, i horn at BrooMyn, ii-v Galveston, vi Columbia, vii
Medina, N. Y.: i James» Seymour, b. Apr. 11, 1838; d. Jan. 12, 1839.
ii Catherine Seymour, b. July 12, 1840 ; m. 1st, at Columbia, Nov.,
1858, Dr. B. G. Salmon, who d. at Galveston, Aug., 1867; m. 2d, July
12, 1871, C. B. Maynard, who d. 1888. She res. at Bastrop, Texas.
Chn.: 1 Kate^*' Salmon, b. Nov., 1859 ; d. se. 2 yrs. 2 Mary Hulburt
Salmon, d. a?. 1 yr. 3 Lucy Gantier Salmon, d. se. 2 yrs. 4 Annie
Salmon, twin of Lucy; d. te. 1 yr. 5 Roberta Salmon, b. 1867; d.
1871. 6 Seymour Maynard, b. 1872 ; m. Oct., 1895, . 1 son. 7 Carrie
Maynard, b. Nov., 1873 ; m. July 12, 1897, C. Tittger. 8 Wanno (?)
Maynard, b. June, 1876; unm. 1902. 9 Bertha Maynard, b. Nov.,
1878; unm. 1902.
 EIGHTH GENERATION 717 iii William Seymour, b. Feb. 24,
1843; d. March 8, 1844. iv Charles Seymour, b. May 3, 1845; d. July
19, 1846. V James Madison Seymour, b. Sept. 3, 1847; m. at
Galveston, Xov., 1866, Caroline Lewis. Ees. New York City. Chn.: 1
Allen Lewis^° Seymour, b. Oct. 4, 1867 ; m. Miss H. B. Ayers. 4 chn.
2 Carolyn Seymour, twin of Allen; m. "Walter M. Keck. Ees. E.
Orange. 3 James Madison .Seymour, b. 1871 ; m. Etta Belle Heddon.
Have chn. 4 Alfred Seymour, b. Oct. 4, 1873 ; unm. vi Edwin
Adriance Nash, b. Feb., 1855.; d. 1864. vii William Eufus Nash, b.
Apr. 21, 1860; m. at Columbia, Dec. 10, 1884, Ina Young. Ees.
Columbia. Ch.: 1 Catherine^o Nash, b. May 6, 1886. 811 LYDIA
ANNS cOOKE (Sarah' Van Deursen, Lydia^ Brewster, ElisJia^,
William^, WiUiani^, Love-, WiUiam'^), married at Watertown, N. Y.,
September 24, 1846, as first wife, JOHN ADEIANCE, who was born
in New York State. She died at Columbia, Texas, January 30, 1871.
Th«y resided at Columbia, Texas. Children, i, iii-iv horn at Columbia,
ii Watertown: i Catherine% b. Oct. 13, 1848 ; d. Oct. 21, 1848. ii
Sarah Bush, b. Sept. 5, 1850; unm. Ees. Columbia, iii Hannah Dyer,
b. July 19, 1857; m. at Columbia, Feb. 6, 1878, G. C. Munson. Ees.
Angleton, Texas. Chn.: 1 Daughter^", b. Nov. 14, 1878; d. ge. 12
days. 2 Lydia Munson, b. Nov. 20, 1879 ; unm. 1902. 3 Sarah
Munson, b. Aug. 29, 1882 ; unm. 1902. 4 Henry W. Munson, b. Jan.
4, 1883. 5 Adriance Munson, b. Nov. 24, 1885. 6 Mordella Munson,
b. Feb. 22, 1893. 7 Euth Munson, b. Oct. 18, 1895. iv Duncan, b.
Feb. 16, 1867; d. in Texas, Feb., 1897; m. June 27, 1894, Miss M. B.
Webb. He grad. at A. and M. Coll. of Texas, and was a professor in
Chemistry Dept. of same college for ten years. Ch.: 1 Guy W.^«
Adriance, b. June 19, 1895.
 718 BREWSTER GENEALOGY 812 ELIZABETH PEARCE^
COOKE {Sara¥ Van Deursen, Lydia^ Brewster, Elisha^, William*',
William^, Love^, William^), married at Watertown, N. Y., September
23, 1850, Eeverend RUFUS DOANE STEARNS, son of Isaac and Lois
(Doane) Stearns of Pulaski, N. Y., born at Edmonston, N". Y.,
February 17, 1820; died at Omaha, Neb., September 10, 1900. Mrs.
Stearns resides at Omaha. Mr. Stearns was an Episcopal clergyman.
Children, i-iv horn at Medina, N. Y., v Chicago, III.: i William Rufus»,
b. June 13, 1853 ; d. Aug. 21, 1854. ii Emily Benedict, b. Dec. 24,
1854; d. July 10, 1870. iii Edwin Adriance, b. Nov. 23, 1857; m. at
Omaha, Jan. 1, 1900, Catherine Rood. Res. Kansas City, Mo. iv
Charles Wild, b. Oct. 9, 1861 ; d. at Elgin, 111. ; m. there, Aug. 20,
1880, Jessica M. Mallery. Res. Elgin. Chn. : 1 Richard Mallery^''
Stearns. 2 Martha Elizabeth Stearns. V Bessie B., b. Feb. 18, 1875 ;
unm. 1902. Res. Omaha. 813 Reverend RUSSELL DUDLEY^ VAN
DEURSEN (Peter' Van Deursen, Lydia^ Brewster, Elisha^, William'^,
William^, Love"^, William}), married first, at Richmond, Va., June
12, 1856, MARGARET BARKER, who died at Richmond, May, 1858 ;
married second, at Wellsville, Ohio, December 4, 1860, MARGARET
BRODIE, daughter of Samuel and Mary (Frazier) MACKENZIE of
Wellsville, born at Allegheny, Pa., September 25, 1837. He died at
Eureka Springs, Ark., January 21, 1887. Children, i horn at
Richmond, Va., ii-iv GallipoHs, Ohio: First ivife: i Eleanor Young®, b.
Sept. 3, 1857; m. at Paris, 111., Sept. 15, 1881, William Hamilton
Levings, son of Calvin W. Levings of P., b. at Rockville, Ind., Sept. 15,
1859. Res. Minneapolis, Minn. Chn., 1-2 h. at Paris, 3 Minneapolis: 1
Mary Read" Levings, b. June 26, 1882, 2 William Seymour Levings,
b. Jan. 26, 1885. 3 Margaret Esther Levings, b. Apr. 10, 1888.
 EIGHTH GENERATION 719 Second wife: ii Emma Baldwin,
b. Jan. 26, 1863 ; d. at San Antonio, Texas, Feb. 4, 1899; unm. iii
Lucy Eead, b. Nov. 1, 1864; m. at Paris, 111., June 10, 1884, as 2d
wife, George Edward Dilley, son of George M. Dilley of Dallas, Texas,
b. at Shelbyville, 111., Feb. 18, 1858. Ees. Palestine, Texas. Chn., h.
there: 1 Edna May^° Dilley, b. Jan. 16, 1886 ; unm. 2 Clarence Van
Deursen Dilley, b. Apr. 12, 1887. iv Dr. George Livesay, b. Jan. 19,
1867; m. at Lowell, Mass., Oct. 6, 1892, Elizabeth A. Tarbell of L. A
physician at Lowell, s. p. 814 WILLIAM PLUMMER* VAN DEUESEN
{Peter' Van Deursen, Lydia^ Brewster, Elislia^, William*', William^,
Love^, William^), married at Cincinnati, Ohio, December 20, 1859,
CHAELOTTE BOWES, daughter of Joseph B. and Martha (Pendery)
HUMPHEEYS of Cincinnati, born August 3, 1835; died at Wyoming,
Ohio, April 17, 1882. He died October 9, 1876. William P. Van
Deursen removed from Eichmond, Va., to Cincinnati in 1856, when
he became employed in the office of the Ohio and Mississippi
Eailroad Company. About 1870 he went to Texas as auditor of the
International and Great Northern Eailroad Company, with
headquarters at Palestine, Texas, and held that position until August,
1876, when he visited Cincinnati and while there was stricken by
sunstroke, from which he never recovered. Children, horn at
Cincinnati: i William Hammond^, b. Aug. 27, 1861 ; d. Aug. 2, 1869.
ii Frances Belle, b. Sept. 7, 1869; m. at Wyoming, June 15, 1892,
William Augustus Haven, son of James L. and Eebecca (Paddack)
Haven of Cincinnati, b. there, Dec. 16, 1866. The president of the
Haven Malleable Castings Company at Cincinnati. Chn., h. there: 1
Eebecca Van Deursen" Haven, b. May 30, 1893. 2 Charlotte Bowes
Haven, b. Feb. 21, 1895. 3 Irwin L. Haven, b. Apr. 1, 1898 ; d. Aug.
27, 1899. 815 MAEY EMILY* KIPPEN (George'' Kippen, Ruhif
Brewster, Elisha^, William*, William^, Love^, William^), married at
Bridgeport, Conn., July 17, 1834, EZEKIEL BIEDSEYE, son of
 720 BREWSTER GENEALOGY Nathan Gould and Polly (Piatt)
Birdseye of Huntington, Conn., born there, November 10, 1807; died
at Bridgeport, November 5, 1896. She died at Bridgeport, March 25,
1866. Ezekiel Birdseye was a leading drygoods merchant at
Bridgeport for many years. CJdldren, horn at Bridgeport: i George
Kippen^ b. March 15, 1837; m. at Bridgeport, Apr. 21, 1858, Martha,
dau. of Horace and Mary (Tyson) Hatch of Bridgeport, b. there, Sept.
25, 1839. A merchant at Bridgeport. Chn., h. there: 1 Frank Hatch^°
Birdseye, b. Sept. 8, 1859 ; m. at Bridgeport, Jan. 13, 1886, Ida,
dau. of Charles Coley of Bridgeport, b. there, Jan. 20, 1862. Pes.
Bridgeport. Ch., h. there: (1) Beatrice^i Birdseye, b. Sept. 28, 1891.
2 Herbert Birdseye, b. Dec. 4, 1860; m. at Bridgeport, May 15, 1890,
Jennie Elizabeth, dau. of Rev. William Garner of Bridgeport, b. at St.
Johns, N. B., Dec. 28, 1866. Pes. Bridgeport. Ch., h. there: (1)
Garner Kippen" Birdseye, b. Sept. 12, 1892. 3 Mary Gertrude
Birdseye, b. Apr. 6, 1868. ii Mary Emma, b. May 21, 1841; m. at
Bridgeport, Dec. 2, 1862, William Henry Pockwell, son of Jonathan
and Clarissa (Gregory) Pockwell of Pidgefield, Conn., b. there, Oct.
14^ 1834; d. at Bridgeport, Apr. 11, 1897. Pes. Bridgeport. Chn., h.
there: 1 Ada Birdseye^" Pockwell, b. Jan. 10, 1866; m. at
Bridgeport, Dec. 2, 1885, Edwin Glover Sanford, son of Edwin G.
Sanford of Bridgeport, b. there, Oct. 21, 1862, and d. there, March
1, 1891. Pes. Bridgeport. Ch., h. there: (1) Emily Judson" Sanford, b.
May 3, 1890. 2 William Gregory Pockwell, b. Apr. 15, 1872; unm.
1899. iii Elizabeth Anlien (?), b. Feb. 24, 1855; m. at Bridgeport,
Aug. 7, 1883, Franklin Curtis Clark, son of David W. Clark of Tidioute,
Pa., b. at Bridgeport, March 13, 1854. A druggist in No. Dak. Ch., b.
at Valley City, No. Dak.: 1 Herbert Birdseye^" Clark, b. Sept. 11,
1886.
 EIGHTH GENERATION 721 816 Captain WILLIAM
FEANCIS^ KIPPEN {George' Kippen, Rubtf Brewster, Elisha^,
William*, William^, Love'^, William}), married at Bridgeport, Conn.,
April 19, 1843, OLIVE CLOVER, daughter of Captain Elisha and
Susan (Gifford) DOANE of Cape Cod, Mass. She died July 29, 1898,
aged 81 years. He died at Oregon City, Ore., May 11, 1877. William
F. Kippen was a sea captain. Children, horn at Bridgeport, Conn.: 1
William Frank», b. Feb. 2, 1849 ; d. May 25, 1901 ; m. at Bridgeport,
June 22, 1870, Emma Sterling, dau. of George S. and Elizabeth
(Waterman) Lewis of Bridgeport. Pes. Bridgeport. Clin., h. there: 1
Pobert Waterman^" Kippen, b. Apr. 2, 1871. 2 William Francis
Kippen, b. July 5, 1878. it Helen Stewart, b. Jan. 19, 1852; m. at
Astoria, Ore., Apr. 25, 1874, William Lambert Worthington, b. in
Tenn., 1852; d. at Dallas, Ore., Apr. 26, 1883. Pes. Bridgeport. Chn.,
1, 3 h. at Astoria, 2 Oregon City, 4 Dallas: 1 Grace Olive^^
Worthington, b. Apr. 29, 1875. 2 Gertrude Lambert Worthington, b.
Sept. 13, 1876. 3 Herbert Spencer Worthington, b. Jan. 18, 1880, 4
Harold Williams Worthington, b. Nov. 1, 1882. 817 SUSAN"
MEEKER^ KIPPEN (George' Kippen, Ruhy^ Brewster, Elisha^,
William*, William^, Love^, William^), married at Bridgeport, Conn.,
September 20, 1842, HENRY WILLIAM CHATFIELD, born at Derby,
Conn., September 4, 1818. She died at Bridgeport, February 24,
1851. Children, horn at Bridgeport, Conn.: i Henry Whitney®, b.
Aug. 2, 1843; d. at Volusia, Fla., Feb. 5, 1865; unm. He served in
Civil War as 1st lieut. of 17th Regt. Conn. Vols, ii William Kippen, b,
Oct., 1846; d. se. 2 yrs. iii Daughter, d. in infancy. iv Charles Nichols,
b. May 16, 1850; d. at Brooklyn, N. Y., Aug. 3, 1884; m. at Brooklyn,
Dec. 3, 1873, Harriet Taylor of Brooklyn. Chn., h. there: 1 Henry
Whitney^" Chatfield, b. March 27, 1876. 2 Bertha Taylor Chatfield, b.
Nov. 15, 1879.
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!

ebookball.com